IL 1 L 1. 1 11 1 i CORNELL UNIVERSITY. I$a$t(ie(( THE THE GIFT OF ROSWELL P. FLOWER FOR THE USE OF THE N. Y. STATE VETERINARY COLLEGE 1897 Cornell University Library R 111.C76V Contributions to medical research, dedica 3 1924 000 273 486 Jr® Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924000273486 CONTRIBUTIONS TO MEDICAL RESEARCH ^ C-d-T^"^ ^1/ Ci^^c^c.^^^/iv'ii.ylyU .-^ CONTRIBUTIONS TO MEDICAL EESEAKCH DEDICATED TO VICTOK CLARENCE VAUGHAN BY COLLEAGUES AND FORMER STUDENTS OF THE DEPARTMENT OF MEDICINE AND SURGERY OF THE UNIVERSITY OF MICHIGAN TWENTY- FIFTH ANNIVERSARY OF HIS DOCTORATE ^ ANN ARBOR, MICHIGAN GEORGE WAHR, PUBLISHER 1903 /^ /,- J ■ 'BRAR No -S-vsU R. R. DONNELLEY & SONS COMPANY CHICAGO CONTENTS. I- — Operative Possibilities in Cases of Advanced Carcinoma of the Breast. Lewis Stephen Pilcher, A. M., M. D., LL. D. Pages 1-11. II. — Second Contribution to tlie Study of the Pathology of Early Human Embryos. Franklin P. Mall, A. M., M. P. 1 plate, 5 figures in text. Pages 12-28. III. — On a Common Form of Reduplication of the First Sound of the Heart Due to Extra-Cardiac Causes. Heney Sewall, Ph. D., M. D. Pages 29-33. IV. — The Lymphomatous Tumors of the Dog's Spleen. Herbert U. Williams, M. D., and Frederick C. Busch, M. D. 1 double plate. Pages 34-41. V. — The Aboriginal Physicians of Michigan. Edmund Andrews, M. D., LL. D. Pages 42-50. VI. — Observations upon the Cause of Shock, and the Effect upon It of Injections of Solutions of Sodium Carbonate. William H. Howell, Ph. D., M. D. Pages 51-62. VII. — On the Organic Peroxides. Paul C. Freer, Ph. D., M. D., and Frederick G. Nov^-, Sc. D., M. D. Pages 63-127. VIII. — Cancer of the Gail-Bladder and Biliary Passages. William J. Mayo, A. M., M. D. Pages 128-137. IX. — The Function of the Suprarenal Glands and the Chemical Nature of Their So-called Active Principle. John J. Abel, M. D. Pages 138-165. X. — A Consideration of Ovarian Fibromata, Based o^ a Study of Two Recent Cases and Eighty-two' Collected from the Literature. Reuben Peterson, A. B., M. D. 4 figures in text. Pages 166-187. XI. — On the Morphology of the Pyloric Glands of Vertebrates. Lydia M. DeWitt, B. S., M. D. 25 figures in text. Pages 188-203. XII. — The Mental State of Anarchists and of Others Who Kill or Attempt the Life of Rulers or Public Personages. Richard Dewey, A. M., M. D. Pages 204-215. XIII. — The Changes Produced in the Hemolymph Glands of the Sheep and Goat by Splenectomy. Aldred Scott Warthin, Ph. D., M. D. 2 double plates. Pages 216-236. XIV.— The Aeration of Milk. Charles E. Marshall, Ph. D. Pages 237-275. XV. — The Laws of Physics and Ballistics: The True Explanation of the Lodgment and Deflection of the Jlajorityof Modern Military Small-Arm Projectiles; Not the Ricochet Hypothesis. Charles B. Nancrede, M. D., LL. D. Pages 276-279. XVI. — The Tendon Action and Leverage of Two-Joint Muscles of the Hind Leg of the Frog, with Special Reference to the Spring Movement. Warren P. LoMBABD, A. B., M. D. 2 plates, 5 figures in text. Pages 280-301. XVII. — On the Action of Lithium. Clarence A. Good, M. D. Pages 302-313. VI CONTENTS. XVIII. — On the Secretion of the Urine and Saline Diuresis. Arthuk R. Cushny, A. M., M. D. 4 figures in text. Pages 314-342. XIX.— On the Diuretic Action of Alcohol. Charles W. Edmtjnds, M. D. Pages 343-350. XX. — On Nutmeg Poisoning. Gbokgb B. Wallace, M. D. Pages 351-364. XXI. — A Contribution on the Morphology of Sudoriparous and AUied Glands. G. Carl Huber, M. D., and Edward William Adamson. 16 figures in text. Pages 365-389. XXII. — On the Preparation and Use of Collodium Sacs. Charles S. Goksline, A. B., M. D. Pages 390-394. XXIII. — The Phylogeny of the Forearm Flexors. J. Playpair McMurrich, Ph. D. 13 figures in text. Pages 395-424. XXIV. — Clinical Observations on Congenital and Acquired Transposition of the Vis- cera. James Rae Arneill, A. B., M. D. 6 figures in text. Pages 425-439. XXV. — On the Superior Sphincter of the Rectum. Robert C. Bourland, A. B., M. D. 4 figures in text. Pages 440-445. XXVI. — Parotitis following Abdominal Section. William H. Morley, Ph. B., M. D. Pages 446-452. XXVII. — Report of Some Observations on the Blood of Pregnancy and the Puerperium. George R. Pray, M. D. Pages 453-461. XXVIII. — A Study of the Pulse Rate in Man as Modified by Muscular Work. Wilbur P. BowEN, M. S. 9 figures in text. Pages 462-494. XXIX. — On the Presence of Yellow Elastic Tissue and Reticular Tissue in Tumors of the Skin. Frederick A. Baldwin, A. B., M. D. Pages 495-504. XXX. — On Triphenylmethyl. Moses Gomberg, So. D. 2 figures in text. Pages 505-534. XXXI. — A Study of a Malignant Tumor in the Domestic Fowl, with an Attempt at Transplantation and with Special Reference to Amyloid-like Changes Occur- ring in the Liver. Frances Nichols Boynton. 2 figures in text. Pages 535-540. XXXII. — Idiopathic Myositis Involving the Extra-Ocular Muscles. John Elwin Gleason. 2 figures in text. Pages 541-548. XXXIII. — On the Cultivation of Trypanosoma Lewisi. Ward J. McNeal, A. B., and Frederick G. Novy, Sc. D., M. D. Pages 549-577. XXXIV.— Studies on Neuroglia Tissue (No. 2), Neuroglia Cells, and Neuroglia Fibers of Vertebrates. G. Carl Huber, M. D. 3 double plates, 1 figure in text. Pages 578-620. OPERATIVE POSSIBILITIES IN CASES OP ADVANCED CAECINOMA OF THE BREAST. BY LEWIS STEPHEN PILCHER, A.M., MX)., LL.D., Surgeon to the l\[ethod)st Episcopal and to the German Hospitals. Editor of "The Annals of Surgery." The history of adequate attempts to satisfy the indications presented b}- tlie pathology of Carcinoma of the Breast in operative procedures for its removal is brief and very recent; practically it is included within the surgery of the past twenty-five years. Up to the beginning of this period a condition of hopelessness filled in general the surgical mind when confronted by cancer of the breast, since the ablest surgeons found themselves compelled to acknowledge the absolute failure, in almost every instance, of their efforts to eradicate the disease by operation. In 1867 Charles H. Moore^ had read a paper before the Royal Medical and Chirurgical Society of London on "The Influ- ence of Inadequate Operations on the Theory of Cancer, " in which he urged the importance of removing, together with a diseased breast, any textiu-e ad- joining the breast which is even approached by the disease, especially skin, lymphatics, fat and pectoral muscle. This paper contained the germ of an advance in the surgery of the breast, the development of which became evident ten years later in the first paper published by Mr. Mitchell Banks^ of Liverpool (1877), in which was definitely formulated the teaching that, whenever there was found an appreciable disease in the breast, the axillary lymph nodes were to be regarded as already involved in the disease, and that, whether enlarged glands could be detected by palpation or not, the fatty and glandular contents of the axilla should be systematically cleared out in all cases as a part of the operation for the removal of a cancerous breast. Nearly simultaneous with this movement among English surgeons was a similar and even more general departure by their colleagues in Germany, marked especially by the writings of Volkmann' (1882) and of Kiister* ( 1883) , who, to the systematic clearing out of the axilla, added the practice of care- fully dissecting from the surface of the pectorsilis major muscle the fascia which covers it. In the United States, also, the younger Gross,^ particularly, by his papers and by his work on tumors of the mammary gland (1880) was emphasizing the importance of a wider extirpation of possibly infected neigh- boring tissues and lymph glands. A more important step was made in 1889, when Heidenhain" made a report to the German Surgical Congress in which he demonstrated the perva- 2 POSSIBILITIES IN CASES OF CARCINOMA OF THE BREAST. siveness of the carcinomatous process throughout the whole of a breast in which any part was affected, and the extension as a rule into the underlying fascia of any carcinomatous process in a breast, however limited it might seem to be. He further demonstrated the impossibility of removing this fascia without leaving behind disease-bearing fragments of it, however care- ful may have been the attempts to dissect it away from the subjacent muscle. Thus was explained the frequency of local recurrence, even after the most painstaking removal of this fascia. The logical inference was that the pectoralis major muscle, as well as the fascia covering it, should be removed together with the overlying breast and the axillary contents, in every case of carcinoma of the breast, in order to insure the highest possible certainty of complete removal of the disease in any given case. In 1894 was made the first publication by Halsted' of the unusually favora- ble results which he had been able to secure at the Johns Hopkins Hospital by a careful and thorough carrying out in his work of the precautions of his- predecessors already named; adding, perhaps, as the special feature of his own work, a more free removal of. the overlying skin, and a more frequent invasion of the regions above the clavicle. During the earlier years of the work, which the writer was privileged to do in Brooklyn, namely, in the period extending from 1872 to 1888, his work in connection with mammary cancer was conducted on the lines until then generally taught and practiced by surgeons, and which he now recognizes as having been incomplete. The record of his cases during this period is not complete enough for him to make use of them for statistical purposes. As a whole, however, they were a dreary record of recurrence, with quickly fol- lowing death. In two instances, however, the advantage gained by his opera- tions was so marked as to stimulate him to continued efforts, and to make him the more ready to adopt the more radical measures of the surgeons whose work has just been referred to, when they were brought to his attention. In one case, after the removal of a breast, followed later by a removal of perceptibly large axillary glands, the patient survived for six years, finally dying with symptoms indicating carcinoma of the liver. In a second case, which presented itself to him with recurrence in the scar of operation made by another surgeon, with well-marked involvement of axillary and supraclavicular glands, after a very free extirpation of all the diseased struc- tures, in the course of which the internal jugular and subclavian veins and the axillary artery were ligated, the patient remained free from local recur- rence and enjoyed six years of good health; but then evidences of intratho- racic disease manifested themselves, terminating in death two years later. When, by the opening of the Seney Methodist Episcopal Hospital in 1887, the writer was placed in a position to do more systematic work and LEWIS STEPHEN PILCHER. 3 to preser^•e the records of the same, he endeavored to take advantage of the opportunities presented, and the present study is based, with one exception, upon the cases which have come under his care in that institution, during the period from 1888 to 1900, inclusive. The operations recorded were in most instances made by himself personally ; in some cases they were conducted by his assistant, Dr. Warbasse. In reviewing any series of cases of carcinoma of the breast with reference to the value of the measures instituted for their relief, it is of the highest importance to classify the cases according to the nature and extent of the disease presented by them. All cases in which operations have already been done, and which after- ward present themselves with recurrence, more or less widespread, constitute a category entirely distinct from those cases in which the primary growth is met with as j^et undisturbed by any surgical procedure. In the mentioned class a long-standing existence of the disease is certain, together with evident widespread diffusion of the disease elements, a condi- tion carrying with it almost a certainty that multiple points of disease exist, not yet sufficiently developed to be grossly perceptible, in addition to the nodules which are already palpable. Efforts at a radical removal, therefore, are extremely likely to be incom- plete and to be followed speedily by the appearance of further points of disease. Numerous cases of such secondary operation have engaged the efforts of the writer during the period under consideration, but in no case has ulti- mate success been attained thereby. With this reference to them, and this admission, they are excluded from further consideration in the present study. The number of primary cases, in which it seemed that benefit from operation might be expected, which have presented themselves during the period in ques- tion, was exactly fifty. All of these were subjected to operation. In seven of them, however, the operation itself revealed that the disease had already extended beyond the possibilities of entire removal, either by reason of the degree to which the wall of the thorax was plainly involved, or by reason of the evident extension of the disease to the mediastinal glands. In one of these, in which a portion of a rib was removed in the course of the operation, a pneumonia developed which proved fatal on the third day. This was the only fatality attributable to the operation in the whole series. There remain, therefore, forty-three cases in which supposedly com- plete extirpation of a cancerous breast and adjacent disease was done. It is hardly necessary to premise that in all cases the operative attack that was made was guided by a desire to go so wide of the disease that no vestige of affected tissue should be left behind. StiU, in dealing with carcinoma in any region of the body the surgeon ever finds it difficult to hold a just balance between the natural reluctance 4 POSSIBILITIES IN CASES OF CARCINOMA OF THE BREAST. to sacrifice unnecessarily apparently healthy tissue, to inflict unnecessary deformity and disability, and to increase to an unnecessary degree the hazards to life that attend his work, and the demands for radical, wide-extending extirpation arising from the unquestioned fact that a wide margin of appar- ently healthy tissue of indefinite extent must always be regarded as already invaded by microscopic disease-elements, disseminated along the lymphatic paths that lead from the recognizable grossly affected foci. Too often the later history of his cases shows him that wha,t was intended to be judicious conservatism was really imperfect and useless surgery, that had tended greatly to lessen the original possibilities of ultimate cure. Be this as it may, it is my desire now to record the actual results attained, and to draw from them such lessons as may seem warranted thereby. In all the cases, the general procedure was conducted in accordance with the teaching that the incisions through the overlying skin should go wide of apparent disease, and that the breast and axillary lymphatics, with the connective tissue and fat in which they were imbedded, should be dissected out as an unbroken piece. In looking over the details of the technique employed in the carrying out of this general plan in the several cases, they naturally divide themselves into the following classes: I. Ablation complete to apex of axilla, without removal of any pectoral nmscle. Two cases. II. Ablation complete to apex of axilla, with removal of pectoralis major muscle only. Eleven cases. III. Ablation complete to apex of axilla, with removal of both pectoral muscles. Twelve cases. rV. Ablation complete to apex of axilla, with removal of one or both pectoral muscles and invasion of the supraclavicular region. Eighteen cases. Of the two cases in Class I, the first remained well for six years after ■operation, when symptoms of intrathoracic disease developed which jDroved fatal within the year. In the second case, a local recurrence in the border of the pectoralis major muscle had developed eighteen months later, for which a second operation was done, when the whole muscle was removed. The patient remained well for five years, when similar disease appeared in the other breast; this second breast was then removed, together with the contents of the axilla and the pectoraHs major muscle on that side; no further recurrence of exter- nally located cancer took place, but after three years symptoms of cancer of the liver developed, resulting in death ten years after the first operation for mammary disease. In Class II, the first case, at the end of ten years, still remains well. LEWIS STEPHEN PILCHER. 5 The second patient died of apoplexy one year after operation, having had no sign of recurrence up to the time of her decease. The third case, at the end of eight and one-half years, remains well. The fourth case remained well for six years; during the seventh year there became evident carcinoma in the ribs, below the site of the primary disease, from which death ensued after a further course of about two yeais. The fifth case, at the end of eight years, remains well. The sixth case, at the end of eight years, remains well. The seventh case died two years after operation from intrathoracic metas- tasis. The eighth case is the one mentioned in the preceding class, in which disease appeared in the second breast six and one-half years after the removal of the first breast for cancer, and in which case ultimate death resulted three and one-half j'ears after the last operation from supposed carcinoma of the liver, no local recurrence having taken place. In the ninth case, six months after operation, the supraclavicular glands were perceptibly enlarged ; the space above the clavicle was then cleaned out, when it became evident that already the disease had extended into the medias- tinum. Death followed within two years. In the tenth case, the patient was well one year after operation, since which time no report has been obtained. The eleventh case died two years after operation, with local and regional recurrence. In these two classes — Class I and Class II — of the thirteen patients con- tained in them, the later history of all but one being known, it appears that four have remained well to the present time, periods of from eight to ten years having elapsed since operation; that three more enjoyed a period of immunity lasting for six years, and then in each case developed renewed can- cerous disease; that in three cases evidences of recurrence in distant regions showed themselves within three years after the operation, and that in but two cases did local recurrence take place. It is pertinent and important to remark as to these cases, that they com- prise those in which, of all which presented themselves for treatment, the disease apparently had made the least advance; in which no muscular involvement was detected, and in which the involvement of the glands of the axilla was not so extensive as to make difficult the cleaning out of that space. In three of the cases, however, the result showed that the primary opera- tion was not complete: namely, the second case of Class I, in which the pectoralis major muscle was not removed, and in which muscle the disease developed within a few months after the primary operation; in the ninth case of Class II, in which within a few months after operation, the supra- clavicular glands became noticeably enlarged and were then attacked, but. 6 POSSIBILITIES IN CASES OF CARCIxNTOMA OF THE BREAST. as the operation revealed, not until after the disease had extended into the mediastinum; and in the eleventh case of the second class in which within a year such extensive local recurrence had developed as to make any further operative effort impracticable. That so, large a proportion of absolute recoveries, or of freedom from disease for many years, should have been secured by the operative measures employed, is full of encouragement as to the possibilities of successful attack in the earlier stages of breast carcinoma. On the other hand, one cannot but harbor the thought that if a wider excursion of the operative attack had been made at the first in the three last cases mentioned, such as the removal of a greater area of skin, of the pec- torahs major muscle and of more of the axillary connective tissue, and the extension of the incisions above the clavicle, the number of definite cures might have been yet larger. This experience serves to confirm the value of the more recently advo- cated methods of dealing with carcinoma, the results simply being in accord with those which have been secured by many other surgeons working on the same lines. With Class III come into consideration cases in which the disease had attained a more advanced stage so that, in order to facilitate the thorough removal of the axillary contents, the pectorahs minor muscle was removed as well as the pectoralis major. The character of the results obtained by operation presents a most marked change from those presented by the preceding classes. Twelve cases are included in this class. In one case the later history is unknown ; of the remain- ing eleven, one lived five and a quarter years after the operation free from recurrence, and then died from an acute pneumonia at the age of seventy years. Two others are well at the present time, three years and three years and four months, respectively, after the operation. One is living five and a half years after operation, but with slowly advancing recurrence in axilla and above the clavicle. One other is still living, three years after operation, without external recurrence but with evidence of carcinoma of liver. Six patients have died at periods varying from twelve months to five years and a half after operation. In all but one of these cases the development of supra- clavicular disease was among the earliest evidences that the primary operation had been incomplete. In so many instances did this demonstration of the extension of the disease to the lymph glands above the clavicle occur, notwith- standing that no such glandular involvement was perceptible to examination, that it has seemed to the writer to be reasonable to regard the supraclavicular lymphatic tissues as diseased in all cases in which the glands at the apex of the axilla were markedly affected, and that in all such cases the rational procedure for the surgeon to pursue is to open up the supraclavicular spaces LEWIS STEPHEN PILCHER. 7 and clean them out, as well as to deal in the same mannei' with the axilla. Accordingly in a very considerable number of cases, notwithstanding the absence of any evident supraclavicular disease, that region was opened and explored. The number of such cases was eight, and in nearly all of them were uncovered and removed small nodules distinctly cancerous, though too minute to be detected by palpation when covered liy the intact fascia and skin. The total number of cases in which the supraclavicular space was opened is eighteen, in ten of whom palpable supraclavicular nodes existed. Of the entire number of this class — Class IV — but two have remained well. These two cases ha^'e now passed respectively six and four and a half years since operation in good health, entirely free from any suggestion of cancerous disease. One other still lives, more than two and a half j^ears since operation, in good general health, but with a beginning enlargement of the costochon- dral articulations on the affected side which are indicative of recurrent disease in the ribs. The others are all dead, the majority of cases from intrathoracic metastasis. Of the two cases that may be pronounced as cured, one was a woman sixt3'-eight years of age with a diffuse infiltration of the right breast, and with perceptibly large axillary and supraclavicular nodes. For three years she had been aware of the presence of this disease. In the second case, like- wise for more than three years, the patient had been aware of the presence of a tumor in her breast; the whole gland had become manifoldly involved in the disease, and had become converted into a large ulcerating tumor. In this case the axillary nodes were enlarged but the supraclavicular nodes were not perceptible. The operative attempt to remove the disease was conducted in two stages; first was done an excision of the breast and the axillary con- tents and the pectoral muscles; two weeks later, primary union of the first wounds having been secured, the clavicle was exposed by incision and double division of the bone made so as to leave its middle third loose, and attached merely by the subclavius muscle. This bone muscle flap was then turned down so as to give complete access to the base of the neck, which was then carefully cleaned out. Some carcinomatous nodules were detected in the tissue removed frona the neck; the osteoplastic flap was then replaced in its proper relations to the rest of the bone ; rapid operative recovery from this second operation was secured. Four and a half years have now elapsed, and up to the present time she has remained perfectly well. The experience of these years has emphasized most stronglj^ to my con- sciousness the fact that nothing is more illusive than the apparent local ex- tent of a carcinomatous process. In many instances the epithelial invasion which constitutes the essential element of the process has been for a long time slowly, insidiously, painlessly, and imperceptibly progressing, without producing manifest tumor, and without attracting the attention of the person 8 POSSIBILITIES IN CASES OF CARCINOMA OF THE BREAST. affected, irntil by accident her attention is at last drawn to the alteration in the texture of the breast which has occurred; hence little of importance can attach to the subjective symptoms which are elicitable in the history of these cases, and nothing is more unreliable than the statements of patients as to the length of time the disease has been present. On the other hand, the size of the local growth and the rapidity with which its bulk has increased since its presence was detected, and the tendency to breaking down which it may exhibit when it comes to the surgeon's notice, are no positive index to the number and distance of the secondary outlying deposits which may have occurred along the outgoing lymphatic paths. It is true that there are certain gross evidences of advanced carcinoma which, when they are present, are unmistakable as to their character and meaning, such as fixation of the gland to the subjacent muscle, palpable enlarged lymph nodes in the axilla and above the clavicle, and nodules in the circumjacent skin. Cases presenting such conditions fall without dispute into the category of cases of advanced carcinoma of the breast. Different from these are some growths which, from the first, exhibit a tendency to local rapid increase in size and early necrosis, without corresponding tendency to the de^'elopment of metastases. They early attract attention and speedily come to extirpation, which, when done in the complete manner required by the pathological knowledge of the present day, is likely to result in perma- nent cure. While these latter acutely developing cases may also very properly be classed as cases of advanced carcinoma, the prognosis attending efforts at their removal is much more favorable than that which attaches to the cases of the more slowly diffused epithelial invasion. In the latter class of cases, the skin over and adjacent to the breast may be apparently healthy and still harbor multiple points of metastatic deposit, as yet microscopic in size. Upon the whole, one is almost driven to the conclusion that clinically the surgeon never sees carcinoma of the breast in any other than an advanced stage. Some cases when brought to his notice may certainly be further advanced than others, but, without dispute, every case when first brought to his atten- tion has behind it a long period of development, and has connected with it e^'ery probabihty of many and distant metastases. Hence those surgeons alone are rational and correct who insist that in every case that comes to opera- tion a far-reaching and wide-extending removal of overlying and adjacent tissue shall be made, together with the removal of the affected breast itself. Every tissue related to the affected breast by propinquity or by connecting absorbent ducts rests under suspicion; the less the apparent advancement of the primary disease, the greater, of course, the probability of the successful result of the surgeon's efforts, and hence the greater the importance of the most radical and far-reaching extension of his removal of possibly affected tissue in presumably early cases. A wide area denuded of skin may readily LEWIS STEPHEN PILCHER. 9 be covered again by plastic flaps or l)y grafts; removal of both' pectoral mus- cles entails surprisingly little ultimate disability, and most extensive wounds in the axilla and above the clavicle heal with certain promptness when matle under the precautions required in the surgery of the present day. Wlien the statistics of Billroth^ at Vienna wen- published in 1S7S, it appeared that of seventy-three women from \\hom he had removed the breast and axillary glands for cancer, t\\'enty-seven had died as the result of the opi-ration, more than one in e\'ery three operated upon! At the present day many lists of more than one hundred similar consecutiAC cas(>s ha\'e been published without a death. Among the fifty cases now reported liy the writer there was but one operative death. Such a remarkable absence of mortality attending the extensive and prolonged dissections now employed in operations for cancer of the breast is due to three causes, all characteristic of the perfected methods of wound treatment of the present day, viz.: the prevention of shock, careful haemostasis, and scrupulous antisepsis. Increasing appreciation of the patho- logical indications for radical operative measures and increasing perfection of operative technique have progressed mth equal steps. Upon the combina- tion of the two depends the great change for the better which has been effected in the operative possibiUties in cases of carcinoma of the breast. It ought to be unnecessary at the present day to call the attention of educated physicians to the high importance of immediate surgical interfer- ence in every case of even suspected carcinoma of the breast, unless there be circumstances attending the case which contraindicate any operation, and yet it is stiU the case that frequently patients are not presented to the surgeon imtil after they have been under the observation of physicians for many months, delay having been advised by the latter while they watched the progress of the gro^vth. It is not rare that a patient who has discovered something wi'ong in her breast is told by her physician not to worry, to come back again in six months, or "in the fall, " or to try some kind of treatment, and is thus led to postpone surgical relief until a period when the probability of its successful application is greatly lessened, if not absolutely destroyed. Nor is it the 3'oung and inexperienced or obscure practitioner that is always the greatest sinner in this respect. Of the fifty cases under consideration in the present report, thirty-seven had in this way postponed application for surgical relief after they had become aware of the existence of the disease for periods varjdng from six months to three years or more, the record being : for six months in twelve instances, one year in ten instances, two years in eight instances, and three years or more in seven instances ! It cannot be too strongly emphasized that practically every case of carcinoma of the breast, when it has reached that degree of development at ivhich a palpable tumor is formed, is already in an advanced stage, such an advanced stage that, as a rule, metastatic deposits have already begun to be formed, beginning in 10 POSSIBILITIES IN CASES OF CARCINOMA OF THE BREAST. the nearby lymphatic paths, and that onl}' by an immediate far-reaching removal of both the discernible disease and the adjacent tissue that may inclose metastatic points can even a moderate probability of permanent cure be secured. The differential diagnosis of neoplasms of the breast rarely presents any uncertainties to one who is familiar with them; the character- istics of the retention cysts, the adenomata and the inflammatory indurations, which constitute nearly all the non-malignant tumors of the breast, are usually well marked and readily made out. If in any case any doubt exists, it is far wiser to give the benefit of the doubt to malignancy and to proceed at once to its extirpation. Roger Williams analyzed 2,422 consecutive cases of pri- mary mammary neoplasms, and found of this number 1,974 that were ma- lignant — that is, over eighty-one per cent ! The operations required for the accomplishment of the wide reaching removal of tissue called for by the known pathological conditions present in mammary cancer are laborious and time consuming, and for their best and most successful performance require a high degree of technical skill and a full equipment of assistants and of material. The multiplication of hospi- tals and the increasing number of able men with operative training and ex- perience connected with them, however, place in most communities all the needed requisites for the more frequent performance of operations for cancer that comply with the demands of pathology. Present experience warrants the statement that surgery can promise a very large proportion of absolute cures to cases of cancer of the breast, if its resources are employed as soon as the presence of the disease is determined, even though it be acknowledged that the disease is then already in an ad- vanced stage. It is not to be wondered at that in the past, with its records of high opera- tive mortality and low ultimate immunity from recurrence, both patients and physicians have preferred to postpone efforts at extirpation until the burden of the local disease has become intolerable. The influence of this attitude of a past generation still lingers, and to it is due much of the hesitancy to seek at once surgical relief which we have been deploring. As the knowledge becomes more general as to what has been and can be done by surgery for the cure of cancer of the breast, less hesitancy will be dis- played by its victims in at once availing themselves of the help which is offered, and the proportion of permanent cures effected will be increased. LEWIS STEPHEN PILCHER. U REFERENCES. 'Moore — The Influence of Inadequate Operations on the Theory of Cancer. Medioo- Chirurgical Transactions, 2d series, Vol. XXXII, 1867, p. 278. ' Banks — On Free Removal of Mammary Cancer, with Extirpation of the Axillary Glands as a Necessary Accompaniment. British Medical Journal, December 9, 1882; Liverpool and Manchester Medical and Surgical Reports, 1877. ^ Sprengel— Archiv fiir Klinische Chirurgie. Bd. XXVII, 1882. *Kiister — Zur Behandluug der Brustkrebses. Verhandlungen der deutschen Gesell- sohaft fiir Chirurgie. 1883. * Gross — Tumors of the Mammary Gland. 8vo, 1880. I). Appleton & Co. "Heidenhain — Uber die Ursachen der lokalen Krebsrecidive nach Amputatio Mam- mse. Verhandlungen der deutschen Gesellschaft fiir Chirurgie. 1889. ■" Halsted — Results of Operations for the Cure of Cancer of the Breast. Johns Hop- kins Hospital Reports, ^■ol. IV, No. 6, 1894. ^ Wiuniwarter — Beitriige zur Statistik der Carcinome. Stuttgart, 1878. SECOND CONTEIBUTION TO THE STUDY OF THE PATHOLOGY OF EAELY HUMAN EMBEYOS. FRANKLIN P. MALT., Professor of Anatomy, Johns Hopkins University. Three years ago I published an account of fifty-three pathological ova, bringing together all of the data I could collect, with a provisional classification of the specimens* Since then twenty new pathological specimens have been added to my collection, and it is the object of this communication to give an account of them and to verify or to revise the classification made in the first communication. It appears that from a morphological standpoint, the first classification is not only confirmed but further strengthened by the study of the second group of pathological ova, to be described in this paper. The fifty-three specimens of the first communication were provisionally arranged into four groups, as follows : 1. Arrested development of the embryo, with continued growth of the ovum; 2. Degeneration of the embryo, leaving only the umbilical cord; 3. Ova normal in form, without embryos and uterine moles; and 4. Vesicular forms of pathological embryos. The new specimens fortunately do not fall parallel with the grouping of the old, but fill out in many respects stages which are missing. All of the stages together may now be considered as arrested development in some form, with or without destruction of some part of the embryo or its appendages. It is easily understood how groups two and three may successively arise from group one, after the second week, that is, after the foriiiation of the body of the embryo, the. amnion and the cord. With the new stages we have a series arising from group one, then passing to four, and finally to the younger ova of three. This series includes ova in which the changes affected specimens less than two weeks old, i. e., before the body of the embryo is outHned and, therefore, before the blood-vessels have reached the chorion through the ventral stalk. In our new group we have specimens of the third week which could easily be placed either in group one or four, thus giving the transition forms between these groups as the first communication did between one and two. When the embryos are entirely destroyed, all the * This paper is to be considered as a continuation of one that appeared in the "Welch Festschift, Johns Hopkins Hospital Reports, Vol. IX. "With the exception of No. 158 all specimens with numbers below 160 will be found described in that communication. 12 FRANKLIN P. MALL. 13 specimens fall into group three, and once in this group their characters are usually marked enough to enable one to determine through what course the specimens passed in the destruction of the embryos, and thereby at what time the pathological process began. This naturally brings us to the question of the cause of the pathological changes in the embryo, which changes are primary and which are secondary. If all of the data given in the two communications are brought together, it will be found when considering the wall of the chorion, the villi of the chorion, and the syncytium as coequal, there are, altogether, 154 distinct records. Of these sixty-eight are in the first group, eight in the second, thirty-nine in the third, and thirty-nine in the fourth. When taken together about one- half of the records state that the tissues of the membranes are normal and the other half that the}- are pathological. The pathological changes, however, are not equally distributed in the different groups. They are lowest in number in the second group (twenty-five per cent), a little higher in group four (thirty- two per cent), in group three forty-six per cent, and eighty-one per cent in group one. This indicates that the primary change is usually in the embryo in groups two and four and in the membranes in groups one and three. I have alwaj's made repeated efforts to obtain the history of the case from which the specimen was obtained, but have rarely gotten any data of importance. This is the more to be regretted, for at this point human em- bryological studies could be of considerable value to the practicing physician. In a few instances inflammation of the mucous membrane of the uterus was suspected from the exanunation of the chorion, which when treated, permitted women who were rarely pregnant, and who when pregnant aborted, to bear children. On the other hand, good histories of cases will enable the embry- ologist to recognize the cause of the change in the embryo in pathological cases. I need only to refer to specimens 110 and 141 from the same woman, showing the same kind of an invasion of the chorion by leucoc5rtes in a case of chronic endometritis. In both ova the embryos showed like changes, due to their strangulation. (See, also, No. 207.) If aU of my data are taken together, it is found that fourteen specimens in thirty-five show^ed leucocytic infiltration of the chorion, in group one ; seven in fifteen, in group three ; one in fifteen, in group four ; while in the six speci- mens in group two none are so infected. Omitting the older stages of moles (as given in Table III of the first communication), it is very apparent that the chorion is infiltrated with leucocytes in one-third of the specimens of group one. In the older moles, undoubtedly, the invasion of leucocytes is not pri- mary but secondary, long after the embryo is destroyed. It appears, then, that an endometritis frequently causes a strangulation of the embryo by infecting the chorion, as shown in the specimens of group one ; while in the other groups, when the embryo is rapidly destroyed, wholly or in part, the cause 14 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. lies within the embryo, the chorion renaaining normal for a considerable time, being only secondarily invaded by leucocytes if abortion does not follow. If I am fortunate enough to collect fifty or one hundred new pathological specimens during the coming five years, I hope to make a further study of all of the specimens at my disposal, considering especially the tissue changes in the embryo and its membranes when its nutrition is impaired, due to stran- gulation. These changes are of the most interesting kind, as specimen 201 shows, and may throw some light on the formation of monsters in general. In order to aid such studies it is urgently requested that physicians will send me well-preserved specimens accompanied by good histories. We are now at a point where we can hope for a knowledge of the pathology of the tissues of the human embryo. ARRESTED DEVELOPMENT OF THE EMBRYO WITH CONTINUED GROWTH OF THE OVUM. It is fortunate that the embryos to be considered under this heading do not duplicate to any marked extent, but rather supplement, those of the first communication. There are four specimens of the third week, while before there were but two, and only one of these. No. 115, can be considered with these four. The smallest specimen. No. 162, is in many respects the most remarkable, for the changes in it must have taken place very slowly. We have here the remnant of an embryo of fourteen or fifteen days, in an ovum of from eight to twelve weeks, considering its size and the menstrual his- tory of the mother. The process of distortion and degeneration must have extended during a period of at least six weeks, and still we can recognize the main structures of the embryo as Fig. 1 indicates. There are different kinds of tissues within the nodule: vascular, epithelial, mucoid, and a solid mass, possibly nervous. The mucoid tissue is arranged as a column ex- tending from the heart to the apex of the nodule and may represent a degenerated chorda. A stage a little more advanced, that is, the embryo was a little older when it began to degenerate, is found in specimen No. 166. It is also of long standing and represents a stage between No. 162 and No. 115. The amnion is also fully formed, but there are no blood-vessels m the chorion. In fact, the process is so far advanced that the embryo and am Fig. 1. Section through embryo of No. 162. X 15 times; ch, chorion; am, amnion; h, heart; umb, um- bilical vesicle; in, intestine; all, allantois or possibly liver. FRANKLIN P. MALL. 15 the amnion are detached entirely from the chorion (Fig. 2). Within the only the remnants of the central nervous ttflu Fig. 2. Section through the embryo of No. 166. X 15 times; ch, chorion; am, amnion; n, nervous system; x, heart or umbilical vesicle. cylinder-like embryonic mass system can be recognized with certainty. Specimen No. 196 is in- structive for it is from a case of tubal pregnancy, and the cause of the changes within the embryo can be definitely located on the outside of the chorion. It may be that the chorion was restricted in its growth by the wall of the tube or that its nutrition was cut off, due to conditions which in- terfered with the growth of the villi of the chorion. Fre- quently perfect embryos are obtained from tubal pregnan- cies even if the amniotic cavity is much smaller than normal and the chorion is greatly retarded in its growth. That something very severe has happened in the chorion is indicated by the condition of the villi, its syncytium, and by the amount of blood pigment taken up by the blood-cells of the embryo. The embryo itself did not continue to grow, but gradually solidified, as is so frequently the case, when malnutrition causes strangulation. Another most interesting specimen of the third week is No. 189. In the arrest of its development the central nervous system remained open throughout its entire length, the dorsal ganglia did' not form, but some of the ventral roots grew from the medul- lary plate into the body of the embryo (Fig. 3). The vascular system is also over- thrown pretty well, the large vessels of the body being irregu- lar and the heart hav- ing been reduced to a bag filled with blood ^ r, ■ , ,,,,,-, X' -.on ., o^ and barely attached to Fig. 3. Section through the head of embryo No. 189. X 30 i i i times. The medullary plate, to, is open throughout its ^'^^ DOCly. whole extent. It appears reason- 16 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. able to think that, if the process of destruction, as indicated by the changes seen in the embryos of the third week, had not been interrupted, the embryos would have been destroyed completely in the course of a few months, forming moles without any remnant of an embryo left in them. The specimen of the fourth week, No. 205, is about twenty-eight days old, judging by the development of the tissues. Its preservation is in a perfect state, so its history is given somewhat in full at the end of the article. There seems to be a tendency for the tissues to become hypertrophic, this being the case in the chorion, amnion and epidermis. In the course of time the heart stopped and all of the blood of the membranes accumulated in the embryo when the blood-cells migrated into the tissues, for they are gorged, while those of the chorion and the cord are empty. The four specimens of the fifth week, Nos. 177, 182, 174, and 200, all show practically the same kind of changes within their tissues. They are all more or less macerated and filled with migrating cells. In this respect they differ much, from Nos. 54, 69, and 132, which show some indication of growth after strangulation. While a pretty sharp line can be drawn between strangulated embryos in which the tissues cease to grow and are gorged with blood and those in which the strangulation is not so complete and is followed by a further and irregular growth of the tissues, the two forms are more or less blended. In the embryos of the fifth week, it may be pointed out that in No. 177 there is a peculiar advanced growth of some cartilages in the head; in No. 182 there is necrosis of the tissues in the front of the head; in No. 174, two peculiar horn-like processes arise from the olfactory pit; and in No. 200 there is extensive destruction of the front of the head. While a few tissues may grow to an abnormal degree in embryos which are strangulated and filled with migrating cells, it is also noticeable that, when the tissues disintegrate, the beginning is usually on top and in front of the head. The last three embryos of this group belong to the end of the sixth week and together are as instructive as those of the third week. Here we have the two varieties of changes, well marked when the nutrition of the embryo is impaired, due to a destruction of the chorion. Of all of the pathological embryos I have studied, thirty-five belong to the group in which the arrest of development is due to strangulation of the embryo, with or without continued growth of the chorion. In twenty-five of these, it appears as if the strangulation of the embryo were pretty complete from the beginning, causing a complete arrest of development of the embryo. No doubt the heart soon stopped, the central nervous system macerated, and the blood-cells of the embryo wandered out into the tissues, as is so frequently the case. But there is a second group of ten specimens in which the strangulation did not appear to be so complete, and as a result of the im- paired nutrition the growth of the embryo was not arrested but only delayed. FRANKLIN P. MALL. 17 In these embryos the main change is no longer a simple maceration of the tissues and a migration of blood-cells into thoni, but a gradual and distorted one, i. e., a pathological growth of the organs and the tissues. In general, these changes appear to be of the degenerative type and may be spoken of as such. There are four specimens (Nos. 162, 166, 189, and 201) of this second type described in this paper and six (Nos. 115, 135, 54, 69, 132, and 142) are given in the first communication. 1 have tabulated the specimens of arrested development of the embryo from the thirty-five specimens in my collection, which give some additional data, omitting Nos. 136 and 124. If these figures are arranged in two columns it will be found that in the cases in which the embryo AAas strangulated, the size of the ovum is on an average thirty per cent too large and its age fifty per cent too great. On the other hand, when the strangulation only caused delay of development of the embryo, the corresponding figures are respectively 250 and 300. In other words, when the strangulation causes death of the embryo, the abortion soon follows and the size of the chorion and its age are but a little in advance of what it would be in case the specimen were normal; while if the strangu- lation only delays the development of the embryo, the diameter of the ovum and its age will be from two to three times the normal for corre- sponding embryos. Possibly a better interpretation is that when the strang- ulation does not arrest but only delays the development of the embryo, the chorion continues to grow for a considerable time accompanied by ex- tensive pathological changes in the tissues and organs of the embryo. Specimens 188 and 207 give again the old story of arrest of development of the embryo with the consequent tissue changes. In No. 188, it appears as if the cause of the strangulation is to be sought in the activity of the syncy- tium, whether primary or secondary cannot be determined. In No. 207, on the other hand, we have a specimen in which the chorion is greatly affected and infiltrated by leucocytes. Within it there are two embryos which, due to the strangulation, show like changes, one a little more advanced than the other. (Fig. 1, Plate I.) The cords of both embryos are atrophic, both showing changes in them similar to those in No. 32. Extensive destructive processes have taken place in both embryos accompanied by an infiltration of migrating cells from the blood-vessels. Specimen No. 201 is one of the most remarkable cases of retarded develop- ment in my collection and, therefore, has been given a more extensive descrip- tion at the end of this article. The chorion is much diseased, being smooth and infiiltrated with leucocytes and syncytium. The external form of the embryo is given in the accompanying illustration. (Fig. 2, Plate I.) The central nervous system is broken up into several segments. The two eyes have shifted toward each other and are fully united (Figs. 4 and 5), forming a Cyclops of the embryo. There are two lenses, a double retina, a common 18 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. P'iG. 4. Section through the top of the double eye of embryo No. 201. X 30 times. The eyes are buried deep in the head, being covered with mesoderm and epidermis. optic nerve and a common pigment layer drawn together between the two halves like an hour-glass. The alimentary canal is partly obliterated, the mouth and anus being closed. The muscular system is largely destroyed. The changes within the embryo are accompanied by an extensive growth of connective tissue. DEGENERATION OF THE EMBRYO, LEAVING ONLY THE UMBILI- CAL CORD. There is but one new spec- imen (No. 198) of this group which falls between Nos. 32 and 25, and in general is like that series with the exception of some changes within the chorion which may be second- ary. It does not appear that *■ the primary cause is to be sought in the chorion in these cases, for the intermediate stages from strangulated em- bryos over to this variety are wanting entirely and because there is little or no change in the chorion. For the present it appears probable that the primary change was with the embryo, which caused its death and destruction, leaving the chorion to continue to grow. Such a specimen if not ex- pelled will, in the course of time, become the nucleus of a mole, and the changes in the chorion in No. 198 indicate this. No. 77 and probably Nos. 93 and 159 (for the amnion is present in these also) are only advanced stages which have undergone sec- ondary degeneration long after the embryos had died and disintegrated. OVA NORMAL IN FORM WITHOUT EMBRYOS. In the present communication I have five specimens to report in which apparently normal ova contained no remnants of the embryo whatever. I opened the mall most carefully, and, as no trace of an embryo could be seen. Fig. 5. Section through the optic nerve and double eye of embryo No. 201. X 30 times. FRANKLIN P. MALL. 19 I stained the pieces and cut them into serial sections. By this method tlie smallest remnant of an embryo would undoubtcnlly be found in every case if present. I wish in this connection to state that an ovum measuring 16x11x11 mm. (No. 191), which was obtained fi-om a criminal abortion, was ex- amined in the same way but no trace of an embryo could be found. The specimen had been torn a little and two nodules were found within the ovum which proved to be pieces of ^■illi. The specimen had been in alcohol a long time before it was gi-\-en to me, and it is possible that the embryo was lost in rough handling. I have excluded this specimen from the list, although no trace of an embryo or its attachment could be found. The five specimens given are to be compared with the first three (Nos. 71, 20, and 29) in the table of the first communication. No. 20 is figured in Plate IV, and in general it represents all eight specimens. Of these, the chorion appears normal in five specimens and cedematous (No. 181) or fibrous and infiltrated with syncytium (No. 29) or with leucocytes (No. 185) in the re- maining three specimens. This last specimen was obtained from a patient in the Johns Hopkins Hospital, but it was impossible to obtain any data from her. I have alwaj's been inclined to think that the specimen was one similar to No. 134, in which the woman punctured the ovum in producing the abor- tion. In the course of time the ovum became filled with pus through the open- ing and the chorion became infiltrated through the ccelom. No doubt ova without embryos may be retained in the uterus for a long time, and I am inclined to view moles with a cavity within them (the ccelom), which is not lined with an amnion as arising from such specimens. Specimens Nos. 55, 153, 70, and 82 all show this. The second group of moles in which the ccelom is Uned with an amnion (Nos. 93, 159, and 77) undoubtedly arose from specimens in which the embryo was destroyed after the amnion was formed, leaving only the cord. The table of ova without embryos and of uterine moles should now, I think, be divided into three sub-groups, to correspond with the above description. The changes then in the chorion of Nos. 181, 29, and 185 are to be viewed as secondary changes, which are more advanced in Nos. 55, 153, 70, and 82. VESICULAR FORMS OF PATHOLOGICAL EMBRYOS. It is now apparent that the two varieties which appeared most obscure have gradually become best understood through the study of some seventy patho- logical ova. The rapid destruction of the embryo leaving only the umbilical cord has been discussed sufficiently. It is mentioned in this connection for a similar destruction of the embryo before the amnion is well formed would probably give rise to the vesicular variety of pathological embryos. At present I have two specimens (Nos. 158 and 180) to add to the thirteen already described. These two specimens show nothing new and only confirm 20 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. what has aheady been said in the first communication. No. 158 is especially- noteworthy for it came from a case of tubal pregnancy, which indicates that the primary trouble is to be sought in the chorion. The viUi of the chorion are wanting altogether and its main walls are somewhat fibrous. The rem- nants of the embryo within are shriveled and necrotic, indicating its death before the abortion. The shriveled remnant is as a double sac with the walls poorly defined, to be compared with Nos. 13, 134, and 143, representing both amnion and umbilical vesicle. No. 180 is a clear case of entire destruction of the embryo and amnion leaving only the umbilical vesicle, which is barely connected with the chorion. It is to be compared with Nos. 123, 130, and 147. The chorion is fibrous, the villi normal, and the syncytium excessive. The vesicle is composed of one layer of cells, the mesoderm with its islands of blood-cells. There are, also, blood-vessels in the chorion, which go to show that at one time the embryo was present in order to allow the blood-vessels to grow from the umbilical vesicle to the chorion. In general, the appearance of the chorion in these specimens is normal, so much so that the pathological nature of their contents was usually not sus- pected before the ova were opened. Yet no conclusion as to the cause of this kind of pathological change can be drawn from this fact, for in one specimen (No. 134) with a normal chorion, the embryo had been infected, due to the puncture of the ovum, In another (No. 123) the normal chorion was im- bedded in a mass of pus; while in the case of tubal pregnancy (No. 158), all of the villi had been destroyed. CONCLUSIONS. From the study of seventy pathological ova can be drawn the following conclusions : 1. The embryo may be destroyed quite rapidly in the young ovum, and the cause probably lies within the embryo itself, for the chorion is usually normal. The changes which take place in the chorion in some of the cases may for the present be considered as secondary. Under this heading there are three main subdivisions : a. The embryo is not formed at all or is destroyed entirely in the young ovum. No. 71 is the best representation of this type. b. The embryo is destroyed during the end of the second week, leaving only the umbihcal vesicle (No. 180). c. The embryo is destroyed after the amnion is well formed, leav- ing only the umbihcal cord (No. 198). A specimen from & or c may be changed into an ovum without an embryo by a further degeneration of its remnants, or the chorion may undergo a variety of secondary changes to form a uterine mole. FRANKLIN P. MALL. 21 2. Primary changes in the chorion may cause strangulation of the em- bryo, which is followed by a variety of pathological changes. These may be brought together under two main heads : a. The strangulation may cause complete arrest of development of the embryo. In such cases the chorion continues to grow for a short time, the embryo becomes necrotic, ceases to grow, and its tissues fill with blood-cells, which wander out from the blood-vessels. b. The strangulation only delays the developm.ont of the embryo. In such cases the chorion continues to grow for a long time, and, the nutrition not being cut off entirely, the embryo continues to grow in an irregular fashion and a variety of pathological changes talces place in it. The organs become irregular, usually degenerate, and may wander. The tissues usually show fibrous atrophy, and not infrequently there are all sorts of irregular hypertrophy. TABLE I. Giving SuioiARY of Measurements of Twenty Pathological Ova. I. Arrested Development of the Embryo. TWO TO three weeks. No. 162 166 196 189 Dimensions of the OvTini. mm. 70x30x30 40x40x40 12x12 28x25x15 Length of the Embryo . mm. 1. 2.5 3. 4. Time between Last Period and Abortion. days 81 71 REMARKS. Atrophy of chorion. Tubal pregnancy. THREE TO FOUR WEEKS. 205 40x30x30 4. 28 (?) Leucocytio infiltration of chorion. FOUR TO FIVE WEEKS. 177 182 174 200 35x25x20 12. 13. 13. 15. 56 Leucocytio infiltration of chorion. FIVE TO SIX WEEKS. 207 188 201 70x45x45 45x40x40 80x60x50 16. 17. 20. 66 Leucocytic infiltration of chorion. Leucocytic infiltration of chorion. II. Degeneratic )n of the Embryo, Leaving only the Umbilical Cord 198 25x2.5x25 7. Chorion is fibrous. 22 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. III. Ova Normal in Form without Embryos. No. Dimensions of the Ovum. Length of the Embryo. Time between Last Period and Abortion. REMARKS. 204 181 190 195 185 mm. 14xl2x 8 18x18x10 25x22x12 30x30x30 40x25x15 mm. days. 35 Chorion is normal. Chorion is oedematous. Chorion is normal. Chorion is normal. Leucocytio invasion of chorion. IV. Vesicular Forms of Pathological Embryos. 158 180 12x 6 20x15x10 2. 1.5 Chorion is atrophic. Chorion is fibrous. Description of the Specimens. No. 158. Tubal pregnancy. Prof. W. T. Howard, Cle^•eIand, Ohio. The specimen came to me imbedded in celloidin and mounted on blocks ready to cut. From each block sections were cut, three of which proved to be through the chorion. In one of these sections there was the remnant of an embryo within the chorion; from this piece I removed the celloidin and reimbedded it in paraffin and cut it into serial sections 50 /* thick. The microscopical examination of the sections shows that the chorion is denuded entirely of its villi, being in apposition and apparently continuous with the wall of the uterine tube. Occasionally the line of separation is marked by a row of irregular cells, probably the remnants of the epithelial covering of the chorion. The mesodermic portion of the chorion is somewhat fibrous, being smooth on its coelom side and without an adhering amnion. The nodule within is shriveled and necrotic, only a few of its nuclei staining. It appears as a double sac, together measuring 2 mm. in diameter, with a clump of necrotic cells, appearing like those of the umbilical cord, between them. In none of the sections is the embryonic mass attached to the chorion. At one place, however, the cord-like structure runs into a long process toward the chorion with a blood-vessel .( ?) filled with blood in its center. My interpretation of the embryonic mass is that it is composed of amnion and um- bilical vesicle of about equal size, shri-s-eled and partly torn into pieces, but still held together by the remnants of the embryo and umbilical cord. No. 162. Large, solid uterine mole measuring 70x30x30 mm. Dr. Alfred Wanstall, Baltimore. The specimen came to me in formalin with the following note from Dr. Wanstall : " Last period from September 2d to 7th, that is her usual time, five days. The woman began bleeding November 9th, and passed the specimen on November 22d. She is the mother of five children and says that this is the only time she has aborted. There is not the slightest indication of uterine disease." Within the specimen there is a cavity measuring 35x12x12 mm., lined with a smooth wall and filled with a jelly-like substance, within which there is a very small embryo which was cut into serial sections 50 /* thick. The sections show a remarkable atrophy of the embryo and umbilical vesicle. The chorion is very thin and is composed of meso- derm only. The villi and epithelial cells are wanting, but in their place there is a thick layer of mother's blood. The entire chorion is lined with amnion and into its cavity the nodule-like embryo projects. Its tissues are not uniform, being thickened at some FRANKLIN P. MALL. 23 points, necrotic at others, and mucoid at others. Throughout the center of the nodule there are some capillaries filled with blood. At the point of juncture between the amnion and chorion there are three projections from the embryo into the ea4om — 1, the uml^ili- cal vesicle; 2, the allantois; and 3, the heart. That the second is the allantois is indicated by its cavity, which is multiple at points. The heart is within a poclu't of the ccelom and has an irregular lumen which is well filled with blood. At the base of the nodule there is a short tube which communicates \\'ith the allantois, the intestine. Xo. 166. Uterine mole measuring -10x40x40 mm. Dr. H. F. fVissett, Baltimore. Last period on October ISth. On December 29th there was a discharge of blood which continued until the 31st, when the mole was expelled. The mole is composed of \-erv thick, fleshy walls within which there is a cavity M'ith a smooth wall, measuring 30x20x20 mm. On one side there is a small atrophic embryo 2.5 mm. long. The sections of the chorion show that its A-illi are well formed and are imbedded in a ■ mass of blood from the mother. Possibly the syncytial layer of epithelium is increased. The coelom side of the chorion is smooth and is in contact with the amnion. Attached to the amnion there is the embryonic mass or remnant which does not reach to the chorion; there is no umbilical vesicle to be found. The amnion and embryo are completely sepa- rated from the chorion. There are no blood-vessels in the chorion. The embryo is cj'lindrical in form, being attached throughout half of its length to the amnion and passing through it. In the center of the embryo there is a solid column of cells quite sharply defined — the remnants of the central nervous system. At the tail end of the embryo there is a blind tube, the allantois. The coelom of the embryo, which is as a pocket on its ventral side, contains an irregular sac which may be either the heart or the umbilical vesicle ; probably the former. No. 174. Ovum 35x25x25 mm., embryo 13 mm. long. Dr. Gibbs, Baltimore. Last period January 11, 1900; bleeding five weeks later, which continued until the eighth week, when the abortion followed. The ovum is smooth, having but few villi, and is filled with a granular magma. Sections of the chorion show a marked degeneration of its walls, nearly all of its villi having been destroyed. Those few fragments of villi which remain are imbedded in blood and are riddled with the cells of the syncytial layer. The mesodermal layer of the chorion is no longer sharply defined and is more or less filled with cells with frag- mented nuclei, the origin of which caiinot be determined. The embryo is of the stage of five or six weeks with pretty sharply defined organs and tissues which are more or less macerated and infiltrated with wandering cells. Most of the epidermis has fallen off ; in the region of the olfactory pit (which is almost obliterated) the epidermis forms two marked horn-like elevations. The central nervous system is swollen and macerated more than the remaining tissues of the body, the change being greater in the brain than in the cord. The vascular system is gorged with blood which is beginning to invade the surrounding tissues. This increase is most marked in the umbilical cord, which appears cedematous. No. 177. Embryo with rounded head. V. B. 12, N. B. 11 mm. Dr. K. G. Harrison, Baltimore. The sections show well outlined all of the organs of an embryo of the end of the fifth week, but they are macerated and swollen. So extensive is the maceration in the head that the brain has become practically solid, the vesicles being nearly obliterated. The process is not so extensive in the spinal cord. Most of the epidermis has fallen off. The vascular system is again greatly distended with blood which is infiltrating the tissues, e.specially those surrounding the larger arteries and veins. In general the tissues 24 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. show the changes always seen in embryos which have been gradually strangulated before the abortion. In this specimen there' is one marked variation in the changes usually found. The precartilage outlines all of the vertebrse and ribs, but no true cartilage is as yet formed in them. Back of the eyes in the occipital region there are on either side of the head two cartilages well developed, much too advanced for embryos of this stage. A more ad- vanced stage of this cartilage will be found in embryo No. 135. The head is also beginning to become stumpy; the frontal process is necrotic and is beginning to fall off. No. 180. Ovum 20x15x10 mm., with a vesicle within, about 2 mm. in diameter. From Dr. C. W. Dodge, Rochester, N. Y., who sends the following history of the specimen : "I have in my possession a human embryo which, if I may judge from some of your papers which I have seen, is likely to be more valuable to you than to me, and for this reason I have kept it intact, instead of sectioning it as I have been sorely tempted to do. Its history is as follows: The woman from whom it came is a patient of Dr. Edward Mott Moore, Jr., of this city. On March 28th, last, her right ovary was removed. She left the hospital on April 15th, and coitus occurred May 13th. On June 19th menstrua- tion appeared and this ovum was expelled, which was brought to me in a pill box (the membrane being broken in handling), and put at once into four per cent formalin, in which solution it still remains. As the dates given above are vouched for by the patient and the physician, it seems to me that we have here an unusually accurate and perfect history of the embryo, and, while it is not so very young, its history may give it additional interest. " Sections of the chorion show that its mesoderm is of normal thickness but is fibrous and rich in nuclei. Throughout the main wall of the chorion, but not in its villi, there are numerous blood-vessels filled with blood, showing that at one time an embryo must have existed. The villi are normal in form with a very extensive syncytial layer of cells over them. At points the syncytium forms large islands, which can easily be seen with the naked eye. Immediately over the vesicle within, an island of this kind, a millimeter in diam- eter, arises from the main wall of the chorion and sends processes up between the villi already formed. The mesoderm immediately below this island is thinner than the rest, making it appear as if the violent growth of the syncytium took everything before it, but that in the attempt to produce new villi the fibrous mesoderm of the chorion would not follow. At many points between the villi there is a slimy mass of albumen quite well infiltrated with round cells and numerous small islands of syncytium, some of which can be followed back to their origin from the villi. The vesicle within is composed of but one layer of cells, those of the mesoderm with blood-islands imbedded within it. No trace of an entoderm can be made out, although the lumen of the vesicle extends into the pedicle, which as a single strand of cells attaches itself to the chorion. No. 181. 0^'um 18x18x10 mm., filled with reticular and granular magma. Dr. D. S. Lamb, Washington. No remnants of the embryo found, although every particle which might be the embryo with the adjoining chorion was cut into serial sections. The mesoderm of the chorion and villi is cedematous; the epithelial covering is poorly developed, often being composed of but one layer of cells. No. 182. Atrophic head and lower end of the body of an embryo about five weeks old. The viscera and ventral side of the embryo have fallen off. Dr. D. S. Lamb, Wash- ington. Sections of this embrvo show a most extreme degree of disintegration of the embryo. FRANKLIN P. MALL. 25 The brain has become con\-crted into a mass of cells filling the central canal entirely and extending into the surrounding tissues of the embryo, the line of demarcation being oblit- erated. The large ^-eius of the body are gorged with blood which also extends into the surrounding mesoderm. On the frontal side of the head there is a straw-colored necrotic mass with some migrating cells -within it. The cartilages alone are still well defined. No. 185. Ovum 40x25.xl5 mm. .\bortion se\'en weeks after the beginning of the last period. Dr. Sabin, Baltimore. The specimen was brought to me in formalin, and upon opening it I found that it was stuffed '\\'ith reticular and granular magma. No trace of an embryo could be found, although the entire ovum was cut into serial sections. The main \\all of the chorion is completely filled with leucocytes from the mother and show all stages of fragmentation of the nuclei. They form a fairly sharp border on the ccelom side, making the chorion appear as the wall of an abscess. The invasion with leucocytes must have been merely from the ccelom side, as the villi are not invaded to any extent. Some of the villi are oedematous, others atrophied, being covered with a normal amount of syncytial cells. No. 188. Ovum 45x40x40 mm., filled with granular magma. Embryo V. B. 17, N. B. 16 mm. From Dr. G. N. J. Sommer, Trenton, New Jersey, who has kindly given me the following history: "Last menstruation began January 6th; bleeding began March 19th, and ended in a few hours with the abortion. The unopened ovum was immedi- atel)' placed in ninety-five per cent alcohol." The chorion is very fibrous and the villi are mostly wanting. The tissue of the meso- derm is very rich in nuclei, none of which appear to belong to leucocytes from the mother. Three kinds can easily be recognized — 1, those which normally belong to the mesoderm; 2, blood-cells from the embryo ; and 3, an extensive invasion of the syncytial cells. This third group can be traced directly from large mounds of syncytium lying upon the chorion, from which they extend throughout the mesoderm, frequently entering the larger blood- vessels. Often large giant cells are seen, showing the usual characteristics of the syncy- tium after it has invaded the mesoderm of the chorion. The villi are afi'ected less than the main walls of the chorion. No cells from the syncytial layer of the chorion were found in any of the blood-vessels of either the embryo or the umbilical cord. The organs of this embryo are all normal in form and of the proper degree of develop- ment for an embryo of this size. The tissues are macerated somewhat, the most marked being the brain. The veins of the body are all gorged with blood, with but little migration of blood-ceUs into the surrounding tissues. No. 189. Ovum 28x25x15 mm., filled with granular and reticular magma. It con- tains a deformed embryo 4 mm. long, lying within a distended amnion 8mm. in diameter. Umbilical vesicle 0.5 mm. in diameter. From Dr. T. E. Oertel, Augusta, Georgia. The umbilical vesicle and amnion appear to be normal for an embryo of this size; the body, however, is greatly deformed, the central nervous system being open throughout its extent and encircled around the dwarfed embryo like a broad hoop around a ball. A number of the motor roots of the spinal nerves are developed, more in the region of the tail than elsewhere. There are no cranial nerves. The heart is a vesicle filled with blood, hanging into the ccelom and slightly attached to the body wall. Its vascular connection with the body is cut off entirety. The blood-vessels of the body are irregular in shape and entirely changed from the normal type. They are filled with blood which extends through their walls into the surrounding tissues. The branchial arches correspond to an embryo of this size. There are still traces of optic vesicles, chorda and possibly allantois pres- ent, the liver and stomach and intestine having degenerated. No. 190. Ovum 25x22x12 mm. Dr. C. M. Ellis, Elkton, Maryland. The ovum is filled with reticular magma within which no trace of an embryo could be found, although 26 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. the entire specimen was stained and cut into serial sections. The chorion and villi are apparently normal, with blood-\'esseIs from the embryo. No. 195. Ovum 30x30x30 mm. No embryo could be found although the entire ovum was cut. The specimen is well covered with villi and contains some reticular magma. Dr. D. S. Lamb, Washington. The mesoderm of the chorion and villi appears normal and is rich in blood-vessels filled with embryo's blood. No. 196. Tubal pregnancy, operation November 2, 1901. Kindly given me by Mr. Brodel of Baltimore. Hardened in formalin. The specimen contained two sus- picious bodies which were both cut into serial sections. One of these proved to be the embryo greatly deformed, representing a stage about three weeks old. The tissues of the embryo are quite homogeneous, only the central nervous system being recognizable. One eye and a large blood-vessel can still be faintly outlined. At points the amnion and umbilical vesicle are blended completely with the chorion. The outside of the chorion has attached to it a few long and thick villi which do not branch. The chorion and these villi are covered with a layer of syncytium of unequal thickness, which frequently invades the mesoderm. The whole chorion is imbedded in a large mass of mother's blood. The most remarkable part of this specimen is found within the blood-vessels of the chorion. They are gorged with nucleated blood corpuscles filled with a pigment of the same color as that of the surrounding mother's blood. It appears as if the syncytium, in destroying the mesoderm of the chorion and the mother's blood, at the same time made it possible for the blood of the embryo to take up the blood pigment thus liberated. At any rate, the blood of a human embryo three weeks old contains no pigment and the sec- tions of this specimen permit of this interpretation. There is also a considerable quantity of mother's blood within the ovum around the embryo, but as the specimen was opened before it was hardened and the corpuscles are all perfect, they need not be taken into con- sideration in the interpretation just given. , No. 198. Ovum 25x25x25mm. Dr. Larsen, Chicago. The specimen came to me har- dened in a mixture of bichromate of potash and formalin. The interior is filled with considerable reticular magma and large lumps of granular magma. Imbedded in this there is a large cylindrical pedicle 7 mm. long bent upon itself. Sections of this specimen show that pedicle to be the urribilical cord rounded off at its former juncture with the embryo. The mesoderm of the cord, chorion and villi is fibrous, having also an excess of spindle- shaped cells. The blood-vessels are all very large, all of those of the villi and most of those of the main wall being gorged with blood. The large blood-vessels of the cord are empty. . Within the cavity of the amnion scattered throughout the magma there are numerous flakes of tissue of the embryo and a great many free cells. No. 200. Ovum 35x25x20 mm. Embryo broken and macerated, with stumpy arms ; 14 mm. long. From Mr. Brodel, Baltimore. The central nervous system is macerated very extensively, the form of the brain and spinal cord being entirely lost. The organs are all deformed, the liver in addition being necrotic, as it does not stain at all. There is ulceration of the front of the head, but over the rest of it, in spite of the extensive internal change, the epidermis is intact. The walls of the umbilical vesicle are broken down entirely and its lumen is filled with a mass of necrotic cells. The amnion, chorion, and villi are more fibrous than normal. No. 201. Ovum 80x60x50 mm. without villi, filled with a fluid which hardened into a jelly in formalin. Embryo (20 mm. long) atrophic with a necrotic mass on top of its head. Dr. Schelly, Baltimore, through the kindness of Mr. Brodel. FRANKLIN P. MALL. 27 The fleshy chorion proved when sectioned to be a mixture of true chorion, villi, blood fibrin, decidua, blood sinuses, pus and syncytium. The layers are not at all in regular order and show what may be all stages of disintegration. The mesoderm of the villi is fibrous and trequenth' in\-aded l>y leucoc>'k\s and syncytium. At other points the syncytium in\ades the blood clot and l're(iueutly maternal blood sinuses are filled with pus and syncytium. Within the embryo most extensi\e changes luu-e taken place. The brain is greatly deformed and is se\-ered, through a growth of tissue, from the spinal cord in the region of the medulla back of the deformed ear. In fact, the -brain is included within the cap-like body on top of the head. The spinal cord begins quite abruptly in the upper cervical region and ends in the same way in the upper lumbar region. At its end there is a curious fibrous tumor measuring half the diameter of the cord. The cord, so far as it is developed, appears to be normal, but it is somewhat macerated. Below the upper lumbar region the spinal cord is wholly wanting, the spinal canal being filled up with mesodermal tissue rich in blood-A'essels. ^^'here the cord is missing most of the spinal nerves appear to remain and many dorsal ganglia can be made out. This all indicates that the changes in the central nervous S}-stem took place after the spinal nerves were developed from it. The two eyes are united into a single one with a double retina, two lenses, a single choroid, and a single optic nerve back to the optic commissure; back of this it is double again. It certainly appears as if the t\\'0 eyes have wandered together and ha\'e united in the middle line. The epidermis is quite complete, being broken through at the back of the head. This extensive ulcer is \-ery rich in blood-vessels, involves the walls of the brain, but does not reach into its ventricle. At the highest point of the head the epidermis has developed into a papilliforni body; below this there is a large necrotic area in which there is a great quantity of yellow pigment granules. The mouth is closed, although the alimentary canal from there through the stomach appears normal. The intestine is matted together, the cloaca and anus being obliterated. The epithelium of the upper portion of the intestine shows marked growths into this matted mass. The thoracic region, liver and vascular system have undergone practically no change. The extensive growth of mesodermal tissue throughout the embryo has caused an ex- tensive destruction and arrest of further development of the muscular system. No. 204. Ovum 14x12x8 mm. Dr. D. S. Lamb, Washington. The ovum was filled with a mass of granular magma. The whole specimen was cut and stained, but no trace of an embryo could be found. The chorion and villi appear normal. No. 205. Ovum 40x30x30 mm. Embryo V. B. 6, N. B. 6 mm. The specimen is about four weeks old and is from a woman who had been married three months. Syphilis is suspected in the case. Dr. D. S. Lamb, Washington. The chorion is partly encircled with the decidua, w^hich is more or less necrotic and well infiltrated with lencocytes, showing that an inflammatory process was present in the uterus. The chorion is fibrous at points and at others cedematous, with but few blood- vessels present. The villi are irregular and often very fibrous, being hypertrophied as well as atrophied. Their outlines are irregular and covered with a dense and very irregu- lar mass of syncytial cells. But few of the villi have blood-vessels within them and these are all empty. The amnion is completely adherent to the chorion throughout its extent, making these two membranes appear as one. On the amnion side there are numerous fibrous tuberosi- ties which appear much as small villi inverted. At other points the epithelial covering of the amnion builds by itself a double layer of cells, which often gives rise to papiUiform 28 STUDY OF THE PATHOLOGY OF EARLY HUMAN EMBRYOS. processes much like the syncytium on the outside. Sometimes this layer of epithelium is raised, forming a blister filled with a fibrin-like substance, possibly magma reticulare, throughout which are scattered transparent round cells with very small nuclei. The umbilical cord is quite fibrous, with large irregular openings scattered throughout it. These are filled with a mucoid substance in which a few nuclei are scattered. The blood-vessels are all obliterated with the exception of the point of the attachment of the cord to the embryo, where irregular vessels are present filled with blood. The external form of the embryo is well retained and is covered entirely with epidermis which is much thickened. The brain and spinal cord are swollen, the former being prac- tically solid in the region of the fore brain. The heart and large vessels are gorged with blood which extends from them into the surrounding tissues, obliterating them almost entirely. Within this mass of migrating cells can be seen the outlines of some of the organs of an embryo about four weeks old. The liver, stomach, and lungs are riddled, and but the faintest mark of an endocoelom can be seen. It appears as if all the blood of this specimen accumulated within the embryo, the cord and the chorion being free, the extensive epidermis preventing the migration of the blood-cells into the amniotic cavity. No. 207. 0\um 70x45x45 mm. From Dr. Horn, Baltimore, through the kindness of Mr. Brodel. The specimen came to me unopened and hardened in a strong solution of formalin. Its exterior is smooth with small villi at one of its poles. Within there are two embryos, both macerated, with atrophic heads. The larger embryo measures V. B. 16, N. B. 10 mm. The other is a little smaller, but as it is broken, an exact measurement could not be made. The cords of both embryos are atrophic. There is some granular magma within the amniotic cavity with several large clumps in the ccelom, where the two amnions meet. Sections of the membranes show that the chorion is denuded of m.ost of its villi, with the exception of the point over the attachment of the cord of the broken embryo. The entire chorion is covered with its decidua, which is rich in blood-sinuses and infiltrated with leucocytes. But few remnants of the syncytial layer of the chorion remain. The whole embryo is still co\'ered by epidermis excepting on top of the head, at the tail end of the body, and at the attachment of the umbilical cord. At these points there is a marked destruction of the tissues, which are beginning to disintegrate. The top of the head is ulcerated, its front necrotic and pigmented, as is frequently the case in other embryos. The nervous system shows the usual changes seen in strangulated embryos. The vascular system of the embryo is gorged with blood, but none is within the vessels of eitlier the cord or the chorion. Within the body there is quite an extensive migration of blood-cells into the tissues, obliterating them in part, but the process of destruction is not so far advanced as in No. 205. The majority of the organs can be still outlined. We have here a rapid infiltration with migrating cells of an embryo of forty days, with cellular destruction rather than irregular growth of the tissues. The changes in the broken embryo are practically the same as in the unbroken one, although they are more ad\'anced. Only the head extremities and cord remain entire, and in these the changes are more marked than in the corresponding parts of the unbroken embryo. In the former it is practically a mass of indi^'idual cells, while in the latter the brain is swollen and quite solid. Desckiption of Figures on Plate I. Fig. 1. Photograph of ovum No. 207, with twin embryos. X i^ times. One embryo is greatly macerated and its pieces lie beside the ragged end of the umbilical cord. Fig. 2. Embryo No. 201. X 2 times. . , Fig, I, Fig, 2. ON A COMMON FOEM OF EEDUPLIOATION OF THE FIEST SOUND OF THE HEAET DUE TO EXTEA-CAEDIAC CAUSES. HENRY SEWALL, Ph.D., M.D., Professm- of Physiology in the Denver and Gross College of Medicine. There is urgent need of a graphic method analogous to the musical score by which the \-arious phenomena attending the heart-beat may be objectively portrayed. It is exceedingly difficult to describe intelligently a cycle of sounds so that their characters and time relations may be distinctly represented to the mind of one not actually engaged in hearing them. Hence arises the considerable confusion existing in the use of terms descriptive of abnormal- ities in the heart sounds. There are few more interesting themes in the cardiac clinic than the so- called reduplications of the heart sounds. These reduplications may be attri- buted to doubling of either the first or second normal sounds of the heart. For example, most observers agree that the sigmoid valves at the base of the aorta and pulmonary artery, respectively, do not always close synchronously, but, probably on account of the varying ratio of pressure in these two vessels, the closure of the pulmonary valves becomes normally relatively more or less delayed at the end of inspiration so that the arterial valve sounds may be heard separately. This gives rise to a doubling of the second sound at the base of the heart and may be described as a true reduplication of the second sound, because there is a repetition of the phenomena generally acknowledged to be the cause of that sound. On the other hand, in certain pathological conditions, especially, at times, in cases of mitral stenosis, there may be heard immediately following the normal second sound a third abnormal sound, usu- ally resembling the true second sound in duration and quality and following it after a short interval. This abnormal sound is described in the text-books as "the reduplicated second sound." But it is characteristically heard at the apex rather than at the base of the heart, and since there is good reason to be- lieve its occurrence has nothing to do with closure of the sigmoid valves, Sansom' has well called it the " simulated " reduplication of the second sound. Like that just considered, most abnormal sounds are added to the cardiac rhythm during the diastole of the ventricles and accordingly fall in the pause between the normal second sound and the first. A very common phenomenon whose causation is not generally agreed upon, ' Diagnosis of Diseases of the Heart and Thoracic Aorta, 1892, page 207. 29 30 REDUPLICATION OF THE FIRST SOUND OP THE HEART. is what seems to be a true reduplication of the first sound of the heart. This consists in a division of the first sound into two elements as if there were a more or less complete splitting of the sound at its acme. The present paper is concerned with the explanation of an adventitious heart sound which is by no means uncommon. I have heard it in, perhaps, fifty cases. A study of the literature makes it evident that some able observers have designated this sound element as a reduplication of the first heart sound, and for a considerable time I labored under the same mistake. But further study threw doubt upon its origin, and recently in an essay devoted to a consideration of reduplication of the heart sounds, it was characterized and dismissed in the following terms: "A not very rare reduplication, occurring in appar- ently organically normal hearts, is one in which the first sound seems to be completely divided, the first element of the double sound coinciding with the beginning of the cardiac impulse as felt at the apex, and the second element following at an interval equal to nearly half that included between the first element and the normal second sound. I have not been able to satisfy myself whether this is a real or false reduplication, or whether due to intra-cardiac or extra-cardiac causes. It is often suggestive of that reduplication of the first sound heard in some cases of mitral stenosis; but in the latter condition the first element of the double sound is pre-systolic, and in the former systolic. ^" A case of reduplication described by me in another place,^ I now believe, belongs to the group which is the subject of this article and not to the class to which it was there relegated. The adventitious or added element of the cardiac rhythm now being dis- cussed lies nearly^midway between the normal first and second sounds, rather nearer the former than the latter, and, therefore, occurs during the contractile movement of the ventricles. In my notes it is, to avoid confusion, usually termed the echo of the first sound. In its duration and quality it is somewhat variable. In cases in which it has long persisted it is usually short, sharp, and clear, and may be the loudest element of the triple rhythm. In other cases it is more prolonged, muffled, and slightly frictional in character. The sound is usually most prominent over the area supposed to coincide with the lower half of the right ventricle, whence it diminishes in intensity and it is retained with pressure of the flexible stethoscope.^ Though the sound often persists throughout the respiratory rhythm, it is always more prominent in a definite phase of respi- 1 On the Clinical Relations of the Papillary Muscles of the Heart. — Phila. Monthly Med. Journ., September, 1899. 2 The Clinical Significance of Reduplication of the Heart Sounds.— Am. Journ. Med. Sci., June, 1898, page 13. 5 Cf. On the Value of Stethoscopic Pressure in Clinical Examination of the Heart. — N. Y. Med. Journ., December, 1897. HENRY SEW ALL. 31 ration and usually confined to it. It is in the expiratory phasi' of respiration and in suspended expiration that the sound is most often hcvird. It is often more prominent in the supine than in the erect position of the body, and it may become accentuated after exertion. The characters of this adventitious heart sound and the conditions of its occurrence are so constant that it forms one of the most sharply defined groups of abnormal rhythm known. The sound has no pathological significance as regards the condition of the heart itself; what, then, is its origin? After studying a large number of cases in which the triple cardiac rhythm under discussion was manifested, it was clear that there was a pathological condition demonstrably common to most of them; this consisted in a recent or remote inflammatory affection of the left pleura. It is easily conceivable that, just as inflammatory roughening of the surfaces lining the pericardial cavity gives rise to the well-kno\\ai friction murmur of pericarditis, so the approximate surfaces of the pleura and the pericardial sac may, if appro- priately roughened, give rise to an audible note when the movements of the heai't cause them to be rubbed over each other. Gerhardt' cites cases in wliich clinical exanrination disclosed reduplicate heart sounds over limited areas, and in which, examined -post mortem, there were found tendinous patches on the heart confined to the areas of reduplication, whose friction had caused the phenomenon. In those cases in which bands of adhesions connect the pleura with the pericardium, it is easy to see that under certain relations of lungs and heart the connecting band of tissue would suddenly be made taut by the energetic sj'stolic movement of the ventricles and produce a sharp report, following more or less closely upon the first sound of the heart. In a case of pulsating empyema of the left side recently seen in the person of a boy about nine years old, there could be heard at the distance of several feet from the patient, when his glottis was open, a sharp click attending the contractions of the ventricles the first sound of the heart itself being ill marked. The heart of the patient was crowded to the right by the excessive pleural effusion; after removal of the fluid the "click" completely disappeared. There can be little doubt that the explanation of the phenomenon proposed at the time was correct, namely, that displacement of the heart had so stretched some tissue adherent to the pericardium that the downward movement of the base of the heart in systole was able to take up the slack and produce the well-known snapping note of a band of fabric suddenly stretched tight. This simple explanation of the very striking "reduplication" of the first sound of the heart I was re- cently able to support by the anatomical conditions found at the autopsy on the body of a man dying with small red kidneys and general arterial fibrosis. The clinical examination of the heart several days before death revealed 'Lehrbuch d. Auscultation u. Percussion, 1890, S. 210. 32 REDUPLICATION OF THE FIRST SOUND OF THE HEART. over the lower half of the right ventricle, and especially during expiration, the reduplication of the first sound under discussion. The triple sound had no resemblance to the so-called "gallop rhythm. " Post-mortem examination showed the heart to be normal except for a considerable degree of hypertrophy and dilatation, and calcareous infiltration of the aortic valve. The lungs were free except for adhesions along the posterior border of the left and, what is of especial importance in this connection, there was a band of adhesion about one inch in length and three-quarters inch in width connecting the center of the anterior surface of the pericardium with the lower edge of the upper lobe of the left lung about an inch from its anterior margin. The relations of this band of adhesion made it apparent that in expiratory retraction of the lungs the band would have been made taut, so that sudden tension could have been produced in it by the known systolic movements of the heart. In persons manifesting this so-called reduplication of the first sound who have had no recent pleural inflammation, the reduplicated sound is usually very abrupt and may be intense ; it is easily distinguished by its quality from the true first sound, even when, as is seldom the case, it coincides with the end of the latter. In chronic cases I have observed this reduplication apparently imchanged for several years in succession. But passing pleural inflammation may only temporarily give rise to the conditions producing it, as is evidenced by the following notes on a case brought to the Arapahoe County Hospital, May 14th, with fully developed lobar pneumonia of the left lung. Nothing peculiar as to the heart was noted on entrance. The patient recovered. On "June 10th, in sitting position, the heart sounds are normal. In supine position, the first and second sounds are normal. About midway between the first and second sounds is heard a new sound not connected to them. It is sharp, short, not blowing or sawing, and has a distinct interval preceding and following it. It is heard all over the cardiac area, but is loudest an inch below and an inch inside the left nipple. It disappears on holding the breath after full inspiration, but is maintained during forcible expiration. It is maintained while firm pressure is made on the stethoscope. There is apparently no differ- ence in its intensity as heard in inspiration and expiration. It is not trans- mitted. " Five days later the patient was examined under the same conditions and the clinical notes record, "Only occasionally do we get the echo (redu- plication) of the first sound in the fifth space, parasternal line. " A little reflection will make it evident that, granting the extra-cardiac origin of such adventitious systoUc' heart sounds, the quality and duration of the sounds may be expected to vary widely according as they are produced by the tensions of adhesion bands, the friction of rough areas on adjacent serous surfaces, or the systohc pulling apart of the pleuro-pericardial mem- branes adherent by gummy exudate. I have just seen a feeble example of this reduphcation in a child afflicted with an acute attack of malarial fever. HENKY SEW ALL. 33 After a few days, and recovery under ordinary treatment, the reduplication had completely disappeared. The results of these observations may be briefly summarized as follows : Simulated reduplication of the first sound, following the latter after a distinct interval of silence, is not infrequent in the normal heart. 8uch re- duplications are due to some abnormal anatomical relation between the left pleura and the pericardium, and a pathological affection of the former tissue is usually demonstrable. Commonly, when the reduplication in question is heard, it points to some present or past inflammation of the left pleura. The majority of persons manifesting this false reduplication of the first sound of the heart are the subjects of pulmonary tuberculosis. THE LYMPHOMATOUS TUMORS OF THE DOG'S SPLEEN. HERBERT U. WILLIAMS, M.D., Professor of Pathology and Bacteriology, University of Buffalo, AND FREDERICK C. BUSCH, M.D., Professor of Physiology, University of Buffalo. The present paper consists of studies upon certain tumor-like nodules, having the structure of lymphoid tissue (or spleen-pulp or both of these) found frequently in the dog's spleen. Tumors of this character are rare in the human spleen. Our original purpose was to make attempts to transplant such tumors into the healthy dog's spleen. Owing to a number of adverse cir- cumstances, only a few experiments in this direction were carried out. None of them yielded positive results. We were, however, able to gather an abun- dance of material for histological and other studies.' Our observations are based upon autopsies made on healthy dogs. The animals had been recently killed by suffocation with illuminating gas at the city pound. The number of dogs examined was seven hundred and twenty. The number having tumors in the spleen was seventeen, or two and one-third per cent. None of the dogs in whom the tumors occurred gave any macro- scopic evidences of leukemia or pseudo-leukemia. No enlargement of lym- phatic glands or other lymphoid structures was seen; nor were lymphoid growths, visible to the naked eye, detected in other organs. In seven of the above seventeen cases where splenic tumors were found, the autopsies were made by the writers themselves. In most of those seven cases, culture-tubes were inoculated from the growths, portions of hver, kidney, and lung were taken for microscopic examination, cover-glass preparations of blood were secured, and the heart was examined for evidences of endo- carditis. The last is to be noted in connection with a possible relation of the splenic tumors to infarction. Endocarditis was, however, in no case ob- served. General Description of the Tumors. — The number of tumors seen in a single spleen varied from one to seven. The tumors usually occurred in otherwise normal spleens; rarely, there was also distinct hypertrophy of the whole spleen. Two or three enlarged Malpighian bodies were a few times detected in spleens containing the tumors, but this was not the rule. In one case only, moderate ' We are indebted for many specimens to our former colleague, Dr. A. T. Kerr, Assist- ant Professor of Anatomy, Cornell University, Ithaca, New York. 34 HERBERT U. WILLIAMS AND FREDERICK C. BUSCH. 35 enlargement of most of the Malpighian bodies coexisted with the splenic tumors in question. The tumors wcn-e seen in all parts of the spleens; when large they protruded just beneath the capsule, and often on both inner and outer surfaces. In shape they were roughly spherical, in diameter from five milli- meters to four centimeters. A capsule of connective tissue was present in one case only (Figs. 3-4). On section, they appeared to be composed of white translucent material, more or less in the form of small, soft masses, resembling Malpighian bodies, but lai-ger. They contained also a variable amount of a soft, dark red substance looking like spleen-pulp or blood-clot, in which the white translucent bodies were imbedded. Sometimes the dark red substance predominated and had only a few of the translucent bodies in it. The relations of nodules of this char- acter to hemorrhages of the spleen will be discussed further on. It may be remarked here, however, that it was not uncommon to find tumors of this char- acter and consisting of this dark red material in the same spleen with others consisting only of the translucent nodules. Histological Examinations. — The tissues were placed in Zenker's fluid or in formaldehyde for fixation. The former was used in most cases. The white translucent nodules proved to consist of lymphoid tissue. The soft, reddish substance appeared to correspond with the splenic pulp; it con- tained red blood-corpuscles in its meshes, and sometimes considerable masses of blood. With low magnification the structure of the tumors was seen to differ conspicuously from that of the spleen in three particulars (Fig. 6) : 1. In the absence of trabeculse, except at the edges. 2. In the large size of the nodules of lymphoid tissue as compared with the ]\Ialpighian bodies of normal spleen. 3. In the absence of a central artery in these nodules, as a rule, though not invariably. The same observation was made by Weichselbaum years ago, in a similar tumor found in the spleen of a woman. The margins of the tumors were not well defined. The neighboring splenic pulp was shghtly compressed. The growths sometimes appeared to encroach slightly upon the normal splenic tissue. Thrombosis of a large vein in the vicinit}' was noted once only. No other disease of the blood-vessels was ob- served. The dense lymphoid tissue of the nodules consisted of lymphoid cells, with a framework of connective tissue, plainly shown by Mallory's method (phospho-molybdic acid and aniline-blue). The connective tissue formed a delicate network among the lymphoid cells. Mitoses of these cells were some- times seen, but they were not numerous. Pale areas, composed of. cells larger than lymphocytes, were rarely present at the centers of the nodules. They appeared to correspond with the germinal centers of lymph-nodes. It is note- worthy that areas of necrosis or caseation were not encountered. Giant- 36 THE LYMPHOMATOUS TUMORS OF THE DOGS'S SPLEEN. cells also were absent. Cells with fragmented nuclei were few in number. Among the lymphoid cells, there were seen certain small, round bodies which stained deeply with hematoxylin and basic dyes. They were not numerous, and will be discussed in connection with the pulp-like substance in which they were more frequent. The dark pulpy substance has already been mentioned as constituting a part of the tumors, varying greatly in amount in different examples. This possessed a fine connective tissue reticulum, the distribution of which wsls quite uneven. Elastic fibres appeared irregularly and to an unimportant extent. Unstriped muscle-fibres were not seen. The blood-vessels were not numerous, and no definite plan of distribution for them could be discovered. This part of the tumors was rich in cells, which, furthermore, showed a great variety of form and structure. The most important of these were large cells, with large bodies of various shapes, their protoplasm staining with eosin. Their nuclei were good-sized, oval, pale, with an indistinct chromatin net- work. Cells of elongated or stellate form were also seen. Multinucleated cells were rare. Cells with fragmented nuclei, or with nuclei of unusual and much contorted forms, were often observed. Related with these, possibly, were the small, round bodies, staining deeply with hematoxylin, with basic dyes, and by Gram's method. These bodies appeared to be free. Pigmented and other phagocjrtic cells were not abundant. The amount of pigment was greater in tumors having a dark color. It was also greater in tumors containing many red blood-corpuscles. It was sometimes smaller in amount in the tumors than in the normal spleen. Lymphoid cells were numerous in the pulpy substance. Cells whose protoplasm was stained a pale violet with alkaline methylene-blue were frequently seen. These we are disposed to consider identical with the plasma-cells of Unna, which they resembled in all respects except in not showing the nodes of chromatin at the margin of the nucleus in a characteristic manner. They contained distinct nucleoli. We confess to not having reached positive conclusions in respect to the relations of the above-described forms of cells to one another. This is rendered difficult by the somewhat indefinite state of our knowledge of the normal cells of splenic pulp. Polynuclear leucocytes were not numerous in the pulpy substance. Eosino- phihc leucocjrtes were abundant in one case only, and in this case they were not increased in numbers in the circulating blood. Cells with basophilic granules (mast-cells) were not seen at all. The number of red blood-corpuscles varied exceedingly in different tumors. In some the number was insignificant ; in others considerable masses of blood could be seen, even to the extent of constituting hemorrhages. Nucleated red corpuscles were abundant in one specimen. Other Organs.— In six cases microscopic examinations were made of the HERBERT U. WILLIAMS AND FREDERICK C. BUSCH, 37 liver, kidney, and lung for deposits of lymphoid tissue. They were not found with the exception of a few small collections of lymphoid cells between the tubules of the kidney, in two cas?s. Blood Examinations. — In four of the dogs presenting splenic tumors, smears were made of the heart's blood. The smears were fixed by heat and stained with the Ehrlich Triple Stain. The ratio of white corpuscles to red corpuscles was normal except in one case, where the white corpuscles were slightly in- creased in number, and among these there was a decided increase in the pro- portion of eosinophilic leucocytes. The red corpuscles were of normal size and shape in all cases. Bacteriological Examinations. — In six cases, sections of the splenic tumors were examined for bacteria by a variety of staining methods, including the stain for the tubercle bacillus. No bacteria were found. The deeply staining round bodies, previously mentioned, may be recalled in this connection, though the -wTiters think it more likely that they are the result of some degenerative or other change of nuclei. Culture-tubes were inoculated from the splenic tumors in seven cases. Large pieces of tissue were rubbed over the surfaces of the media. A variety of culture-media was used. Among these, in every case, at least two tubes containing a mixtru-e of dog's ascitic fluid with agar were inoculated. One of these tubes and some of those containing other media were kept in anae- robic conditions according to Buchner's method. All tubes were placed in the incubator. As a rule, no growth of bacteria developed. In the few instances where gro'wths appeared, they proved to be different in different tubes. A large bacillus, resembling the hay bacillus, was the only organism seen more than once. All of these were undoubtedly contaminations. In view of the fact that the autopsies were made in a cellar under a stable, it is not surprising that some contamination of the tubes took place. Transplantation Experiments on Animals. — Owing to adverse circumstances, such as inability to secure normal dogs, death of the animal within a day or two after the operation, etc., only one case can be considered. Under strictly aseptic precautions two pieces of one of the splenic tumors, each about three millimeters in diameter, were pushed through a sterile glass cannrda into the spleen of a healthy dog. The animal made a perfect recovery. The spleen was examined eleven and one-half weeks later. The points of inoculation were represented by small scars. No tumors had developed. Laceration and Hemorrhage of the Spleen. — The well-known text-books of Kitt and of Friedburger and Frohner inform us that injuries to the spleens of animals, leading to hemorrhage, frequently occur. Under the name of hemor- rhagia intrapulposa subcapsularis Kitt describes masses of blood clot forming protruding tumors, the cause assigned being contusion or embolism. Cicatri- zation is the result imless absorption is delayed, when a hematoma remains. 38 THE LYMPHOMATOUS TUMORS OF THE DOG'S SPLEEN. While collecting the tumors which we have described above, we also en- countered others which belong in the category of hemorrhages. Superficially, these may resemble closely those tumors of a lymphadenomatous variety which had a dark red color on account of containing much pulp (Fig. 5). In such cases the true nature of the nodule could be settled only by micro- scopical examination. We have sometimes found it difficult to decide whether we were dealing with a blood clot in process of resolution and being invaded by granulation tissue, or with a new growth of atypical spleen-pulp and lym- phoid tissue. One such specimen which we have not included in the list of tumors above given, consisted of a dark red mass two centimeters in diameter and dotted with small white bodies. The dark portion consisted of red blood-corpuscles, more or less well preserved, with some long and large cells with pale oval nuclei. Trabeculse were absent. The white bodies proved to be lymphoid nodules. From their margins, groups of lymphoid cells could be seen extend- ing into the blood clot about them. Among these were many plasma-cells. Nuclei in process of division were fairly numerous. Certain features which the hemorrhagic areas and the tumors of the spleen present in common, seem to point to a possible relation between them. These are: 1. Their similarity in form, size, and position. 2. The absence of trabeculse in both, and the lack of sharp definition from the surrounding spleen-pulp. 3. The lymphadenomatous tumors often contain large masses of blood. 4. The hemorrhagic areas may be invaded by multiplying Ijmaphoid cells. 5. It is well known that splenic tissue possesses the power of regeneration in a high degree. On the other hand, aside from the improbability of a new growth arising from a mere laceration, the following facts argue against any such relation : 1. The lymphadenomatous tumors were not, as a rule, conspicuously pig- mented. 2. The frequency with which the lymphadenomatous tumors are seen makes it incredible that hemorrhage could be the cause of most of them; for in any case such a result after hemorrhage could only be exceptional. 3. The absence in the lymphadenomatous tumors of evidence of injury to the blood-vessels. There seemed, however, sufficient encouragement to warrant our observing the progress of experimental lacerations. The results, it will be seen, were always negative. In three normal dogs, slight lacerations were made of the spleen substance beneath the capsule. In every instance the laceration was represented, at autopsy, by scar tissue only. HERBERT U. WILLIAMS AND FREDERICK C. BUSCH. 39 In a further series of experiments, an attempt was made to imitate the formation of these tumors, by producing more extensive lacerations. In four normal dogs anesthetized with ether, the spleen was well lacerated with a tenaculum, the spleen-pulp being separated from the capsule over a considerable area. More or less well-marked blood-tumors occurred within a few minutes, macroscopically closely resembling the genuine tumors of dark red color which we have described. The spleens so lacerated were again examined in, respectively, fourteeij, eighteen, nineteen, and twenty-four days. In all cases the sites of injury were represented by small scars only, the original blood-tumors having entirely disappeared. In another dog in which the lower end of the spleen was thus lacerated, the vessels entering and leaving the upper end were tied. The spleen opposite the ligatured vessels grew dark and slightly swollen. Sixteen days later, at autopsy, no infarct was discovered at the site of ligature and no tumor at the point of laceration. In another dog a laceration was made in a portion of the spleen where the vessels had been tied. At the autopsy, twenty-four days later, no tumor or infarct was foimd. In still another dog the veins were tied at several points along the inner surface of the spleen and lacerations made opposite these points. Charac- teristic bloody swellings occurred in a few minutes. After seven weeks the spleen was examined. The points that had been lacerated were the seat of slight atrophy, pigmentation, and fibrous tissue formation. No other change had occurred. Historical References. Many text-books of pathology make no allusion to the occurrence of pri- mary tumors of splenic tissue. They were, nevertheless, described many years ago. Virchow' speaks of the occurrence of nodes of scrofulous origin, and also of hyperplastic lymphadenomata, particularly in animals. He describes their presence in the dog's spleen. Rokitansky^ describes a tumor of the human spleen, consisting, apparently, of spleen-pulp. Lancereaux^ gives a figure of a similar case and mentions one seen in the dog. Friedreich^ gives an account of an autopsy where multiple ' nodes, consisting of splenic pulp, were observed. He mentions one, previously described by Griesinger. Weichselbaum^ describes two tumors of the human spleen, which appear to be similar to these tumors found in the dog, and which he caUs lymphomata. 'Die krankhaften Geschwiilste, Bd. II, p. 661. ^Lehrbuch der pathologischen Anatomie, 3 Aufl., 1861, Bd. Ill, p. 303. 'Traite d' Anatomie Pathologique, Tome II, p. 596. * Archiv f. pathol. Anat., Bd. XXXIII, p. 84. ' Archiv f. pathol. Anat., Bd. LXXXV, p. .565. 40 THE LYMPHOMATOUS TUMORS OF THE DOG'S SPLEEN. In the well-known text-books of Orth^ and Birch-Hirschfeld^ this subject is briefly discussed. In the works of Kitt' it is more fully considered. According to Kitt, nodules, in the spleen, consisting of lymphoid tissue (excluding those which occur in leukemia and pseudo-leukemia) may be of two sorts: 1. A universal h3rpertrophy of the Malpighian bodies, with much enlarge- ment of the spleen (Hyperplasia folUcularis splenis, Splenoma, Splenadenoma, etc.). 2. Circumscribed white tumors of lymphoid tissue, which may be single or multiple and may reach the size of a potato (Lymphoma). The names malignant lymphoma or lymphosarcoma, are used for the second variety also, apparently when the tumors are multiple and produce great enlargement of the spleen. Both kinds are common in the dog, hog, cow, and horse. Kitt alludes to the possibility of an infectious origin. The tumors which we have described appear to belong to the second variety mentioned by Kitt, with the possible exception of one case to which we have already alluded, where general enlargement of the Malpighian bodies was pres- ent. However, there seems to be no sharp dividing line between the first variety (Hyperplasia foUicularis splenis) and the second (Lymphoma), since Kitt says of the latter that the tumors arise from, the follicles ("die Follikel- inseln des Organs den Mutterboden abgeben") and that hundreds of tumors may be present in a single spleen. Kitt does not state that the tumors classi- fied as lymphomata (or lymphosarcomata) give rise to metastases. In our cases, no metastases were seen, unless the small collections of lymphocytes ob- served on two occasions in the kidney be placed in this category. We were fortunate in having at our disposal, for purposes of comparison, a case of undoubted pseudo-leukemia in the dog*. There was universal hyper- trophy of the Malpighian bodies of the spleen without any tumor-like forma- tions: lymphoid cells in process of division were numerous. ' Summary. 1. The spleens of seven himdred and twenty normal dogs were examined. In seventeen cases, or two and one-third per cent, the spleens contained tumor- 1 Pathologische Anatomie, Bd. I, p. 112; Anatomische Diagnostik, 5 Aufl., pp. 322, 329. ^ Pathologische Anatomie, Bd. II, p. 217. ^Lehrbuch der pathologisch-anatomischen Diagnostik fiir Thierarzte, etc., Bd. II; also, Pathologische Anatomie der Hausthiere, 3 Aufl., Bd. II, pp. 398, 404. •* There was great enlargement of aU the lymphatic glands of the body in this case. Lymphoid deposits in the liver and kidney were conspicuous. The blood showed no in- crease of leucocytes and no eosinophilic leucocytes; poikilocytosis was marked; nucleated red blood-corpuscles were present in small numbers. WILLIAMS AND BUSCH. Plate L Fig. 1. J'iff. 3. Mj;. A. ififir. 5. jFig. «. Explanation op Figures on Plate I. Fig. 1. Dog's spleen with lymphomatous nodule of moderate size. Fig. 2. Dog's spleen with numerous smaU lymphomatous nodules. Fig. 3. Dog's spleen with large lymphomatous nodule. Fig. 4. Same as Fig. 3, shown in section. Diameter of the tumor, 4 cm. ■ Fig. 5. Dog's spleen much enlarged and showing a tumor-like mass consisting of blood clot. Fig. 6. Low power sketch of one of the lymphomatous nodules stained by Van Gieson's method. The new growth is shown above and at the right; the normal splenic tissue appears at the left and below, with a normal Malpighian body and artery, and with numerous trabeculae which appear dark on account of the acid fuchsin stain. (See page 35.) HERBERT U. WILLIAMS AND FREDERICK C. BUSCH. 41 like nodes consisting of lymphoid tissue, and to a less (>xt('nt of tissue resena- bling splenic pulp. 2. The number of tumors found in a singk> spleen varied from one to seven. In diameter they were from five millinxeters to four centimeters. 3. They differed from the normal spleen in having no trabecuhT, except at the edges. The lymphoid collections of which they were chiefly composed were larger than the Malpighian bodies of the spleen, and were not usually found about the small arteries. 4. Certain features in their structure suggested a possible origin in injuries of the spleen in some cases. No relation to injury could be demonstrated experimentally. 5. A single attempt to transplant one of the tumors to a healthy dog's spleen gave no result. 6. Bacteriological examinations gave negative results. THE ABOEIGINAL PHYSICIANS OF MICHiaAN. EDMUND ANDREWS, M.D., LL.D., Late Professor of Surgery, Northwestern University. The Indian tribes of Michigan were of mixed origin. They were mainly TUgonkins of the Ojibwa group, but they had lived for ages alongside several outlying Iroquois tribes until their traditions, their medical usages, and their magical rites became somewhat mingled and cannot be separately treated. Even their great chief, Hiawatha, was originally an Iroquois hero in central New York, and is supposed by Iroquois writers to have been an actual chief who lived and flourished there about the year 1600. However, by the chances of war, fragments of the Canadian Iroquois were driven westward and settled among the Chippewas on Lake Superior, and Schoolcraft, who married into an Indian family, seems to have found Hiawatha's name and fame along the southern shores of that great lake. Finally the poet Longfellow, following Schoolcraft's lead, has so embalmed the old hero's memory in the resinous breath of the northern pines and firs, that it will remain there as long as the Pictured Rocks endure. Even in Ann Arbor, Iroquois echoes exist in the name of River Huron. I shall be pardoned, therefore, if in my Statements I intermingle somewhat the folk-lore of these two groups of aborigines. The healing art among the Indians consisted in what may be called the four M's, viz. : Mythology, Medicine, Magic, and Massage. The very first beginning of their traditions is remarkable because it relates the first obstetric case ever recorded on this continent. I find four somewhat contradictory versions of this myth. The substance of the story is this: A long time ago, for , judging by David Cusick's Tuscarora chronology, it must have been over three thousand years, the only comfortable place in the universe was the world above the clouds. In this region the clouds themselves were the floor or soil; the atmosphere was always warm and bright, and deer, bears, and every kind of desirable game trooped over it. The people there are not described, but they seem to have been gods, mani- tos, and other superior beings. One of them, a woman, who was pregnant with twins, wished to go hunting for a bear. She accordingly set forth, accompanied by two dogs. By acci- dent, she, the dogs, and the bear all fell through a rift in the clouds. At that time the world below the clouds was one great, cold, dismal swamp, which far down at its bottom presented a boundless, half-submerged marsh whose 42 EDMUND ANDREWS. 43 only inhabitants were turtles, snakes, frogs, beavers, and other aquatic crea- tures known in the Tuscarora tradition as monsters. These so-called monsters, however, were of a kindly disposition. When they saw the goddess descending from the sky, they were much concerned for her safety. A great turtle volunteered to receive her on his back, and immediately began to expand to a broader and broader area. The beaver dived into the water and brought up a quantity of soft mud, which he spread on the turtle's back to soften it. The turtle continued to expand, and the mud became wider, thicker, and drier until it developed into a great area of soft, rich, and fertile country. Bushes and trees grew v;p suddenly upon it, and by the time the goddess alighted she was received without injury on a very comfortable couch. The turtle and soil continued to expand, and the trees to grow and spread, and were ultimately converted into the present great continent of North America, mth hills, valleys, streams, forests, and all kinds of vegetation. As the time for the delivery of the goddess drew near, the two young god- lings in her womb began to discuss the best method of accouchement. One of them, the more intelligent of the two, named Enigo-rio, the Good Mind, decided to be born in the natural way; but the other, who was always going contrarj' to good rules, and whose name was Enigon-ha-het-gea, the Bad Mind, determined to make his way into the world by tearing a hole through his mother's side under one arm. They each executed the plan proposed, but in consequence of the violent method adopted by Bad Mind, the goddess shortly died. Thus there were two powerful manitos ushered into existence, of w^hom one was constantly desirous of doing good and the other always trying to do evil. The Good Mind, seeing the coldness and dimness of the new earth, took the head from the body of the goddess, and raising it in his hands, passed it upward into the sky, where it became the sun, the greatest of all spirits. He then took the rest of the body and formed the moon and sent that into the sky to illuminate the night. From this time, Good Mind seems to have been laiown by the name of Ni-yoh, and is in the myths frequently spoken of as the Creator. Having finished the earth, the forests, the lakes, and the streams, he now placed in it an abundance of fishes, birds, and game animals of all kinds, and also in- troduced men. But the Bad Mind tried to resist and destroy all the useful works of his brother. He created dangerous rocks and precipices in the hills, endeavored to destroy all the deer and other game, and created noxious insects and poi- sonous reptiles. This brought on a battle between the two brothers, concerning the results of which different versions of the myth contradict each other. One ver- sion states that Good Mind proved the stronger. One of his blows on the 44 THE ABORIGINAL PHYSICIANS OF MICHIGAN. face of Bad Mind drove the left corner of his mouth upward toward the eye and left it permanently in that distorted position, but he could not ab- solutely expel the Bad Mind from the country. However, he seems to have compelled him to a compromise, and to agree to desist from injuring the people and to devote himself to the relief and cure of the sick. The myth is a little obscure, but as I understand it, the Bad Mind seems to have become a reformed god or spirit. He is transformed into an Indian ^Esculapius, or an Iroquois god of medicine and of magic. A secret society still exists among the central New York Iroquois to com- memorate the powers and virtues of this divinity. It consists of initiated "Medicine Men" of two or more grades, and they are known as the Society of the " False Faces, " in allusion to their custom of wearing at their meetings masks of wood carved to represent the distorted oblique mouth, which the Good Mind inflicted on their tutelary divinity. I have not heard that the society has in later years discontinued the mask, but they may have done so. They are still called by their old name. False Faces. Part of their annual meeting services consist, or at least used to do so, in visiting all the nearby sick, and in trying to relieve them by medicine and by magical arts. The Ojibwa and other northern tribes also have their secret societies of initiated "Medicine Men." Bishop Baraga's Ojibwa dictionary gives them the EngUsh name of "The Grand Medicine." The Indian name of the society is Mide-wiwin (pronounced mee-day-wee- win). If I understand correctly, there are at least two grades of members, the Mide and the Jossakeeds. The latter are both medicine men and prophets. Medicine men not initiated into the society are known as the Wabinos, and there are a large number of them. In fact, wild Indian tribes support a much larger proportion of "doctors" as compared with their whole population, than any civilized communities do. As a consequence many of the medicine men obtain their Uving mostly by the same industrial means as do the re- mainder of the tribe. The Iroquois do not all agree that the god of the False Face society gave them all their medical lore. They say the nation was once devastated by two inamense warriors called the " Stonish Giants, " who were clothed in some form of stony armor. The Iroquois captured one of them in battle and compelled him to reveal to them the medicinal plants and, roots. (David Cusick's Traditions of the Six Nations, published in 1825. Cusick was a Tuscarora.) The materia medica of the Iroquois, the Algonkins, and the Dakotas em- braced one hundred and fifty or more trees, plants, and roots, of which the following are specimens: 1. The pines, firs, and terebinthinates generally were used much as by white physicians, for coughs, for gonorrhea, and various kidney troubles. The EDMUND ANDREWS. 45 turpentine of the balsam fir, tamarack, etc., was applied as a dressing to any raw surface. 2. Oak bark, hemlock bark, and other products containing tannin were their astringents and antiseptics. They were used in infusions for diarrhaias, and made into astringent washes for wounds and ulcers. 3. The yellow birch, which like gaultheria contains methyl salicylate in abundance, was given as a tea for its well-known effect on the kidneys. 4. Culver's physic and several other plants were used as cathartics. 5. The Dakotas were well acquainted with several emetics. 6. Lobelia syphilitica was employed to cure syphilis and some other dis- eases. The mercurials and iodides were not known to them. 7. The Senega snakeroot was a favorite expectorant, and our own use of it for the same purpose was derived from the Iroquois. 8. Sarracenia purpurea, or the Adam's cup, was used by the Canadian Indians for smallpox. A British army surgeon stationed in Canada became convinced by his observation among the natives and introduced it enthu- siastically into general practice. American physicians tried it extensively, but shortly discontinued its use as it seemed to possess no value. 9. Spearmint and various other aromatics were employed as infusions for headaches and other pains, much as in our domestic practice. 10. The do^vn enveloping the seed of the various poplars was gathered as an absorbent dressing for wounds and discharging ulcers. 11. The natives of Arizona chew a small, round plant known as the "mes- cal button, " for its intoxicating effect and for its magical influence. It is a very powerful intoxicant, exalts and deranges the mental powers, and finally induces a prolonged sleep. By analysis it contains three strong alkaloids. Its effect has been likened by some to a combination of cocaine and cannabis indica. It has been used by army surgeons to abate pain. 12. From Mexico, Central America, and South America, the Indians have furnished a great number of valuable and well-known drugs, such as cinchona, jalap, condurango, etc. The surgical knowledge of the Medicine Men was very primitive. They knew how to bleed ages ago, by using sharp splinters of flint. They dressed fractured bones by binding on concave strips of bark freshly peeled from trees, or at other times stitched straight twigs of trees or small reeds between pieces of deer skin and wrapped that around the limb. According to one Chippewa legend, the first case of skin grafting ever per- formed on this continent was executed by a young hero south of Lake Superior. I found the myth in a collection of similar stories gathered by an old Catholic priest, diiring a lifetime of service among the Chippewas of Mackinaw and Lake Superior. I do not think he ever published them, but they were found among his papers after his death. The skin-grafting myth is briefly as follows : 46 THE ABORIGINAL PHYSICIANS OF MICHIGAN. A brave young chief traveled westward in search of adventures. He stopped at the lodge of an old chief who was in great affliction; for his enemies had surprised him, scalped him in great haste, and fled carrying off the scalp, without stopping to kill him. The old man was inconsolable, and the young chief determined to recover the scalp of the old chieftain. Being not only bold, but a most potent magician and medicine man, he transformed him- self into a strong bird and flew to the top of the lodge built by the enemy for a great scalp-dance, where they had tied the scalp to the lodge-pole. Descend- ing through the opening in the roof, he aUghted on the scalp, untied with his beak the string, clutched his claws into the scalp-lock, rose with it through the roof opening and flew away to the lodge of the old scalped chief, where he settled down and carefully fitted the scalp to the head of the old chief. The scalp grew to the head and the young chief won by it great glory. Moreover, the operation was antiseptic, for the creasote in the lodge smoke must have destroyed all the pyogenic microbes. If the operation was ever repeated by later magicians I am not aware of the fact. At any rate, the old missionary's collection of myths does not record any more cases of this nature. The Medicine Men made extensive use of the vapor bath, in what is called the " Sweating Lodge. " A small wigwam was built, perhaps four feet high, and well covered with blankets or skins. The patient then seated himself in the middle and a quantity of hot stones were laid near him. Water was then thrown upon the hot stones and the opening instantly covered with a blanket or skin. In a short time the patient was in a dripping sweat. He then came out and plunged into a lake or river, if one existed there, or if not, his friends threw cold water over him. This is said to cure many minor diseases, but it is alleged by white men living among them that measles and other eruptive fevers treated in this way almost always proved fatal. I think the mental condition of the average medicine man is peculiar. He is well aware that his cathartics, emetics, etc., have real virtue, and a patient gorged with venison and corn mush is greatly relieved by them, but to impress the minds of his tribe he must also resort to tricks and willful deceptions; but this is not all. He is born and bred in a community completely steeped in amazing superstition. He profoundly beUeves in the value of the magical ceremonies, which have been transmitted to him through ages by his Mide- wiwin secret society, and he is unable to rise above his surroundings. He believes in his magic more than he does in his medicines and his willful deceptions both combinied. There is a streak of honesty in his belief in magic. The aborigines generally are the most superstitious of men. In their belief, every stick, stone, stump, or log may be, or may contain, a dangerous manito. Every animal, plant, utensil, or weapon, in short every object from the earth EDMUND ANDREWS. 47 up to the clouds, sun, moon, and stars, has its ghost, or else is itself a mighty manito whose spiritual powers may be dangerous. A gentleman- who spent most of his life among the Dakotas said that, though reasonably brave by day, yet when night set in they became the most abject cowards, from their fear of ghosts. If compelled to go out at night they would do so onlj^ in companies, carrying torches and keeping up a shouting. So prevalent was this superstition that he believed, even in time of war, a per- son could safely tra\'el all over their country by the simple device of lying hid by day and traveling only by night, and then avoiding any large col- lection of shouting men with torches. The Medicine Men are full of supersti- tion, like the rest of them. A dispute has arisen whether the phrase Kije-Manito, Great Spirit, expresses an idea original to the red men, or was invented by the missionaries to con- vey the idea of God. To a considerable extent it is a discussion about words. The early French'missionaries in Canada incorporated the term in their teach- ings a long time ago, and taught their converts to use it in the Christian sense, but I have not seen anything to describe whether they found the phrase already applied in their mythology or not. About the same time John Eliot of Mas- sachusetts translated the Bible into Mohican, and not finding an Indian word suitable for his purpose, introduced the English word God and taught its meaning. If by Great Spirit is meant the infinite, eternal, and omnipotent God, in the Christian sense, then it is true that the natives had no phrase or idea with that high significance. They were eminently polytheists, having many gods, some great and some small, but none of them infinite nor reaching up to the height of the monotheistic Supreme Being. They may, however, for aught we Imow, have often applied the phrase " KijS-Manito " to the sun, or some other of their higher divinities. They did not look upon their many deities as standing all on a democratic equality. They filled the universe with gods, small and large, and most of the tribes conceived of the sun as holding a higher rank than the others, and in aU probability could hardly avoid speaking of him at times as the Great Spirit. Major Powell, in the Seventeenth Annual Report of the United States Bureau of Ethnology, says of that great group of nations on the western plains called the Kiowas, that they all look upon the sun as the greatest spirit, and hence their worship of him in the great " Sun Dances. " Mr. Riggs, who has a very broad knowledge both of the Dakotas and of other nations, says that the sun is the highest deity known to them. The dispute is of little consequence. They almost all held the sun as the greatest spirit, but none of them sup- posed he was infinite. They had no settled theologic system and their notions were to the last degree obscure, confused, and contradictory. It is impossible 48 THE ABORIGINAL PHYSICIANS OF MICHIGAN. now to ascertain whether they used before the white men came any phrase similar to "The Great Spirit" as apphed to any one single deity. Possibly the mental attitude, called by Max Miiller " Henotheism, " may exist among some of the Indians. That is to say, particular individuals or villages may devote their worship so exclusively to one favorite divinity as to pay little or no attention to others; and while they do not consciously deny the existence of others, or even of greater spirits, yet practically they may worship only one, and their cult becomes a form of monotheism. The Medicine Men claim to be able either to kill or to cure by magic and some relate traditions of raising the dead. One of the manuscripts of the old missionary before mentioned, recites this tradition: A valiant chief lived with his tribe south of Lake Superior. They were attacked by their enemies and the chief and the whole tribe slain with the exception of his infant sdn and its mother, who escaped to the shelter of a friendly tribe. When the boy became of age, he was given his father's bow, and his medicine bag containing three magic arrows. Taking these he went to the battlefield where the bones of his father's tribe still lay thick under the trees. Taking out the three magic arrows he shot them upward toward the sky, and as they turned to fall, he shouted, " All you that lie on the earth get up or you will be hurt. " Instantly all the skeletons rose up, clothed with flesh and skin, and became living men. He then became their chief and ruled them prosperously for many years. The Iroquois have a similar tradition of a number of skeletons, both of men and animals, being raised by shooting three magic arrows above them and calling out to them, "Get up, or you will be injured." No battle, however, is mentioned by this version. The Medicine Men, though curing the sick and raising the dead, seem to leave midwifery mostly to the women, who practice it as a mechanic art. Robert Kennicot of Chicago, a noted scientist, who traveled far north in Brit- ish America, gives this account of the procedure. When the labor commences, the patient with a few female friends retires to the woods, if the season per- mits. They bind a smooth pole horizontally across two trees, about eighteen inches above the ground. The patient kneels facing it, and flexes her body over the pole in such a way that the latter presses strongly against the top of the uterus, and thus helps to force the foetus down. Kennicot states that the period of labor was believed to be materially shortened by this method. When the aboriginal doctor has a case in which he deems it important to impress the tribe with the great power and influence which he receives from the spirit world, he erects near his patient a "Medicine Lodge," which is a small but strong wigwam in which he holds interviews with unseen beings. This wigwam is built of poles, bark, etc., braced and stiffened with EDMUND ANDREWS. 49 cords drawn firmly around it and through it, in various directions, so as to bear, without coming to pieces, the groat shalting which the spirits are oxiiocted to give it. This shaldng is one of the chief evidences of supernatural power, and, therefore, very important to the doctor's reputation. The lodge being completed he solemnly enters and tightly closes the entrance with a blanket or skin so that no peering eyes can see within. After a period of silence the wigwam begins to tremble and jerk and finally to shake with the violence of the spiritual influences, and the awe and astonishment of the people without is great. When the Medicine Man has thus become filled with magical healing power, he comes forth to minister to the waiting patient. The shaking of the lodge may not have been purely a voluntary action of the doctor. If on entering he sat down and took firm hold of the tight cords or poles, with a mind under tense expectation, his muscles would soon begin to tremble and jerk, and shake the hut without any direct consciousness of willing the mo- tions. Thousands of our old experiments with amateur clairvoyant mediums have made this fact familiar to us. A Chippewa Medicine Man of western Michigan, many years ago, renounced his magic and became an officer in a mission church. A friend afterward asked him, "What made the medicine lodge shake when you went into it?" He rephed seriously, "I do not know." The iledicine ]\Ien were profoundly superstitious, as well as willfully deceitful. It is probable that some of their muscular actions were involuntary, as in certain analogous experiments of white men, and that they really believed in part of their spiritual manifestations. The Indian doctor taught that many pains and swellings of the human body were caused by witches blowing splinters of bone and other small ob- jects into the flesh, and one of their frequent tricks consisted in pretending to extract the object, and showing it to the patient and his friends. They also believed that thej' could blow with greater strength than the witches, and thus blow away insanity and many other diseases. Longfellow in his poem of Hiawatha has faithfully delineated these aborigi- nal ceremonies in the famous passage about the cure of Hiawatha's melancholy. The Medicine Men, having been in the spirit lodge in the hero's behalf, come to cure him and one of them commences his incantation. " Mystic songs like these he chanted. I myself, myself, behold me! 'Tis the Great Gray Eagle talking. The loud speaking thunder helps me ; All the unseen spirits help me. I can blow you strong, miy brother ; I can heal you, Hiawatha. 50 THE ABORIGINAL PHYSICIANS OF MICHIGAN. I myself, myself, the prophet ! When I speak the wigwam trembles, Shakes the sacred lodge with terror, Hands unseen begin to shake it; I can blow you strong, my brother, I can heal you, Hiawatha. Like a man from dreams awakened He was healed of all his madness ; Straightway from his heart departed AH his sorrow and affliction. " These old Mides (the doctors) and Jossakeeds (the prophets), no doubt, were often useful. They cured some patients and comforted many hearts with hope. Their surgery was simple, but it was soothing and safe. No one died of the shock of their operations nor of hemorrhage from their incisions. Some seventy years ago, Dr. Zina Pitcher of Detroit joined with others and founded a State Medical Society, and supposed it to be the first one ever founded on Michigan soil. They were mistaken. Ages before that time the smoke of the Midewiwin fires curled up through the elms that shaded the banks of Detroit River. This new society died. About 1854, I proposed to Dr. Pitcher and others that we renew the effort, and that we establish the Peninsular Medical Journal to strengthen and cooperate with the society. Dr. Pitcher, Dr. Morse Stewart, Dr. Brodie, all the faculty of the department of medicine and surgery of the University of Michigan, with many others, united in the effort. We supposed this to be the second State society, but we were again mistaken. Back of both, in a dim antiquity, smoked the fires of the ancient Midewiwn Society, surrounded by red aboriginal physicians, some of whom, though less learned than ourselves, may have been men of great natural sagacity, and our superiors in mental power. Peace to their ashes ! They had their faults. They were partly impostors and partly self-deceived, but they were also partly true and partly honest, and they helped many a suffering patient into hope and convalescence. They were better men than we have been accustomed to suppose. OBSERVATIONS UPON THE CAUSE OF SHOCK, AND THE EFFECT UPON IT OF INJECTIONS OF SOLUTIONS OF SODIUM CARBONATE. WILLIAM H. HOWELL, Ph.D., M.D., Professor of Physiology, Johns Hopkins University. In the course of experiments upon mammals, it happens not infrequently that toward the end of a long series of observations the blood-pressure falls to an abnormally low level, indicating more or less complete vaso-motor paralysis, and the animal exhibits in other respects the symptoms of the con- dition usually designated by the term shock. On one such occasion in my experience, some of the solution of sodium carbonate used in the experiment was allowed by accident to enter the circulation, with the result that the blood- pressure was brought back for a certain period nearly to a normal level. As alkaline solutions of a certain strength are Icnown to favor the tone of plain muscle tissue, the observation suggested the desirability of investigating the action of alkaline injections in conditions of shock, particularly as the thera- peutic methods used to combat this condition have proved of uncertain value. The object of this paper is to report briefly a series of experiments made upon dogs, in which the animal was first thrown into a condition of shock by one means or another, and while in this condition was treated by vascular or rectal injections of a solution of sodium carbonate. The condition of surgical shock is defined differently in the various surgi- cal text-books, and usually in quite general terms as an exhaustion of the cen- tral nervous system or of certain centers in it. Most authorities lay stress upon the vascular conditions as the most significant, basing their observations upon the well-known experiments of Goltz made upon frogs. Although the vascular conditions vary somewhat, it may be said that experimental results, as well as observations upon the human being, indicate that in shock there is a very low arterial pressure, such as may be produced experimentally by de- struction of the medullary vaso-motor center, or by severance of its connections with the blood-vessels, combined with a rapid and feeble heart-beat. These vascular conditions alone, if prolonged, are sufficient to account for a fatal termination and may, therefore, be regarded as the most important symptom of shock. Methods Used to Produce Shock. — A number of different methods were em- ployed to throw the dogs into a state of shock as determined by the condition 51 52 OBSERVATIONS UPON THE CAUSE OP^ SHOCK. of the arterial pressure and pulse-rate. The pressure was measured in all cases in the carotid, and expressed, therefore, mean aortic tension. The animals were first brought into a condition of anesthesia by means of a subcutaneous injection of one to two grains of morphia, followed by the administration of ether through a tracheal cannula. While in this state shock was produced by one of the followdng methods : 1. Exposure and handling of the abdominal viscera. 2. Long-continued stimulation of the cutaneous nerves, effected usually by the application of hot-water bags to the skin. 3. By operations upon the brain involving removal of the skull and dura mater, and, in some cases, portions of the cerebrum. 4. By hemorrhage. The last- mentioned method is undesirable for experimental purposes. For, although a sufficient hemorrhage will always bring about a condition of shock, the main object of the experiments was to produce the same set of symptoms by opera- tive violence, with so little loss of blood that one could feel certain that the hemorrhage alone was not the responsible cause. The use of these different methods brought out one fact very clearly, namely, that operations which were sufficient to produce shock in one animal might in another have little or no effect of this kind. The difference depends, no doubt, upon the condition of the animal, but the experiments were not sufficiently numerous to throw light upon the precise nature of these conditions. In some animals, free ex- posure to the air of all of the abdominal viscera for hours had no marked effect upon pressure or pulse, that is, both remained within what might be called normal limits, and cardiac and vascular reflexes were easily obtained. A similar operation in other cases might bring on a condition of extreme shock, especially if the animal had been submitted previously to severe stimulation. In one case, for instance, hot-water bags with a temperature of 80° to 90° had been applied to the skin of the abdomen for an hour or more without any distinct effect upon the anesthetized animal, although the arterial pressure had fallen somewhat and the pulse was more rapid, both facts indicating a tendency toward shock. Subsequently the abdominal cavity was opened. The skin was much congested and oedematous, and the abdominal viscera also showed intense congestion. Shortly after the exposure of the intestines the animal fell into a condition of extreme shock, which was manifested first in the pulse- rate. The pulse before the abdomen was opened was ninety-six per minute, and had an amplitude as measured by the mercury manometer of 2.5 mm. Shortly after the exposure of the intestines the pulse-rate increased to 192 per minute with an amplitude of less than one millimeter. The blood-press- ure at this time was 100 mm. as contrasted with a pressure of 134 mm. just before the abdomen was opened. Within a short time, however, and without any further stimulation than simple exposure, the arterial pressure also began to fall and gradually sank to zero. The most certain by far of the several methods used to produce shock was WILLIAM H. HOWELL. 53 by operating on the brain. In a number of cases, simple removal of the skull- cap was followed soon by a condition of extrcine shock. In other cases, a similar operation had little or no effect, but if subsequently portions of the cerebrum were removed, a condition of shock followed in almost all cases. Under these last conditions the resulting hemorrhage, undoubtedly, aided in bringing about a state of shock, although, according to my experience, this factor alone was not sufficient to account for the permanent effects observed on ]ire,ssure and pulse. Where extensi^'c removal of the skull-cap alone produced sev(>re shock, the effect nuist be attributed mainly to stimulation of the dura mater. This membrane, contrary, perhaps, to general belief, is extremely sensitive, as shown by its reactions upon the heart, blood-vessels, and respirations. If the ether anesthesia is A'cry deep, irritation of the membrane may give no obvious physiological reaction; but if the ether is removed in part, then even slight mechanical pressure upon the inner surface of the dura may call forth very marked reflex effects. The violence of the respiratory movements in such cases and the marked disturbance of the circulation, usually in the direction of a depressed blood-pressure and accelerated heart -beat, indicate that in such an operation as removing the skull-cap and dura the medullary centers are exposed to a severe stimulation . The Physiological Conditions of Cardiac Shock. — A consideration of the vari- ous records obtained in these experiments shows that the two most striking S3rmptoms of shock, namely, the rapid, feeble heart-beat and the paralysis of the vaso-constrictor center, may occur independently of each other to a cer- tain extent and justifies the distinction sometimes made between cardiac shock and vascular shock. In cardiac shock the result is practically similar to that obtained by section of both vagi with the exception that the heart -beats are usuaUy more feeble. We may assume that in cardiac shock the most im- portant factor is a partial or total suspension of activity of the cardio-inhibi- tory center, but together with this factor thei'e may be an associated depres- sion of activity of the augmentor center, and, furthermore, the marked fall in arterial pressure which usually forms a part of the condition of shock will co-operate also in making the heart-beats more feeble. As to the cause of this paralysis of the cardio-inhibitory center it is natural to assume, when we bear in mind that it occurs as a result of violent sensorj^ stimulation, that the paralysis results from an over-stimulation of the center, which leads to exhaustion. If this simple explanation were entirely correct, however, we should expect first a marked slowing of the heart rate — so-called vagal beats — as a result of the initial stimulation of the cardio-inhibitory center, and that this effect should give way gradually to an acceleration as the excessive stimulation paralyzed the center. But such an order of events is not observed, at least not usually. The cardiac shock manifests itself, under the conditions of these experiments at least, as a sudden or progressive 54 INJECTIONS OF SOLUTIONS OF SODIUM CARBONATE. increase in rate and diminution in amplitude of the heart-beat. The effect cannot be attributed to a mere ischa^mia of the heart or of the medullary- centers, since not infrequently it may be obtained while the arterial pressure remains practically unchanged. The following experimental data may be used to illustrate this point : Experiment of January 16th. — Removal of portion of cerebrum and mechanical stimulation of the dura mater. Before the operation : After the operation : Arterial pressure=83 mm. Arterial pressure=82 mm. Pulse-rate=66 per minute. Pulse-rate =120 per minute. Amplitude of heart-beat =3. 5 mm. Amplitude of heart-beat=2 mm. Experiment op January 29th. — Destruction and removal of portions of the cerebrum. Before the operation : After the operation : Arterial pressure=150 mm. Arterial pressure = 104 mm. Pulse-rate=87 per minute. Pulse-rate =162 per minute. Amplitude of heart-beat=7 mm. Amplitude of heart-beat =1.5 mm. Experiment of February 14th. — Removal of skull-cap. Before the operation : After the operation : Arterial pressure=124 mm. Arterial pressure=99 mm. Pulse-rate=70 per minute. Pulse-rate=171 per minute. Amplitude of heart-beat =4. 5 mm. Amplitude of heart-beat=l mm. Experiment of March 5th. — Exposure of abdominal viscera after long-continued stimulation of the skin of the abdomen by a hot-water bag with a temperature of 80°-90° C. Before the operation : After the operation : Arterial pressure=134 mm. Arterial pressure=100 mm. Pulse-rate=96 per minute. Pulse-rate =192 per minute. Amplitude of heart-beat=2 mm. Amplitude of heart-beat= — 1 mm. Bearing in mind the physiological mechanisms involved, a study of these and similar records has convinced me that the paralysis or loss of function of the cardio-inhibitory center in shock is caused by over-inhibition, if one may use such a phrase, rather than by exhaustion from over-stimulation. With regard to the cardio-inhibitory center, we have an abundance of physiolog- ical evidence to prove that its activity may be increased or inhibited by nor- mal physiological reflexes. Indeed, the marked variations in heart-rate that occur under the changing conditions of life may be explained satisfactorily as due to varying degrees of reflex inhibition of the tonic activity of this center, rather than to a reflex stimulation of the accelerator center. There is nothing unphysiological in assuming that a long-continued and powerful inhibition may have a long after-effect, and my records, so far as they can be interpreted, certainly favor the view that cardiac shock is due essentially to this cause. This condition of cardiac shock, as I have said, may occur more or less inde- WILLIAM H. HOWELL. 55 pendently of serious vascular shock. In spite of the rapid, feeble heart-beat, the arterial pressure may remain at a level not greatly below the normal, indicating a. considerable degree of vascular tone. My experimental evidence indicates that in such cases, so long as the vaso-motor center retains an ap- proximately normal condition of tone, the condition of the animal is not hope- less. At least, throughout the short experimental periods of several hours, the arterial pressure shows no tendency to sink, and it would seem probable that proper protection of the animal would insure final recovery. The Physiological Condition of Vascular Shock.— By vascular shock is meant a more or less complete loss of arterial tone, which expresses itself in a corres- ponding fall in general arterial pressure. All the physiological evidence indi- cates that this condition is owing to a loss of activity in the vaso-constrictor center. In extreme vascular shock the arterial pressure falls as low as 20 to 40 mm. of mercury, the same level as is reached when the vaso-constrictor center is destroyed or the cervical spinal cord is severed. In the experiments that I have made, I ha^-e never obtained this condition of complete vascular shock independently of cardiac shock. An extreme and permanent fall in arterial pressm-e following upon operative violence was always accompanied or preceded by the appearance of cardiac shock, the result differing in this respect from the transitory fall in pressure obtained by stimulation of depressor nerve fibers. jMy experiments, like those of many other observers, indicate that this loss of activity in the vaso-constrictor center is the most serious phase of shock. Where the loss is extreme or complete, no therapeutical means has been discovered by which the tonic activity of this center can be restored, and under experimental conditions at least the animal never recovers. If left en- tirely to itself the heart -beats become still weaker, the pressure falls toward zero, and the respirations gradually cease. Under some circumstances this condition comes on rather suddenly, and the general explanation that has been offered is that it is due to exhaustion from over-stimulation. But as in the case of cardiac shock, the occurrences preceding its appearance do not tend to support this view. Here, again, if the theory were correct, we should ex- pect the loss of tone to be preceded by a marked vaso-constriction that would give way to a dilatation as the exhaustion became evident. On the contrary, the pressor effects are usually not conspicuous, the vaso-dilatation may come on at once and pass slowly or quickly to the stage of complete loss of tone. In fact, the experiments of Crile, as well as my own, indicate that stimulation of those sensory regions such as the testes or dura mater, which usually give a depressor effect, lead most often to a condition of shock. Crile has noted that in stimulating sensory nerves the primary pressor effect usually observed from stimulation of a mixed nerve-trunk gives place to a primary depressor effect after repeated stimulation. He draws the justifiable conclusion that the pressor fibers or their action on the vaso-constrictor center are more quickly 56 INJECTIONS OF SOLUTIONS OF SODIUM CARBONATE. exhausted than the depressor iibers. Seemingly when the animal is in such a condition that irritation of a mixed nerve-trunk will no longer give a rise of arterial pressure, this condition alone does not suffice to induce shock, further stimulation with a resulting depressor effect precedes the actual appearance of a long-lasting loss of tone. Theoretically, a reflex depressor action on blood- pressure may be brought about either by stimulation of the vaso-dilator centers, ca^ising peripheral inhibition, or by central inhibition of the vaso- constrictor center. The physiological evidence seems to indicate that generally the reflex fall of pressure obtained by stimulation of sensory surfaces is to be attributed to the latter cause, and it seems permissible in this case, as with cardiac shock, to assume that the condition of vascular shock results from an over-inhibition, probably of the vaso-constrictor center, which leads to a permanent or long-continued cessation of tonic activity. Injections of Solutions of Sodium Carbonate. — After the animal had been brought into a condition of partial or complete shock, the effect of injecting the alkaline solutions of sodium carbonate into the veins or the rectum was tested. Two solutions of sodium carbonate were used, one a 0.5 per cent solution which was made up in a 0.5 per cent solution of sodium chloride and was injected in large quantities, and one a 5 per cent solution which was injected in smaller quantities, usually 10 cc. at a time. With the latter solu- tion it was found that the initial beneficial effects to be described below might be succeeded by a distinctly injurious action upon the heart and respirations when the amount injected exceeded 2^ to 3 grammes of sodium carbonate, for dogs weighing from six to eight kilos. Practically, the more dilute solutions proved to be safer and the effects, so far as they could be observed, equally beneficial. Effect of the Alkaline Injections on Cardiac Shock. — la moderate as well as in severe conditions of shock, the cardiac feature of a rapid, feeble pulse is usually conspicuous. Injections of the solutions of sodium carbonate had but little distinct effect upon the pulse-rate. In none of the experiments made was it possible to bring about a complete return to the normal rate; we may, therefore, infer that the alkaline solutions are not capable of restoring the cardio-inhibitory center at once to its normal tonic activity. In some of the experiments the pulse-rate was practically not affected by the injection; in other cases it was diminished, and in still others it was increased. While the effect of the alkaline injections on the rate of heart-beat was uncertain, the effect on the force of the individual heart-beats was unmistakable; in all cases the amplitude of the pulse-waves was increased to a marked extent. The mercury manometer is, of course, a very inadequate means of measuring the amplitude of the pulse-wave, so far as absolute values are concerned, but it gives a sufficiently reliable indication of any increase or decrease in the am- plitude, especially if the pulse-rate remains about the same. As measured AVILLIAM H. HOWELL. 57 by the mercury manometer, the amphtude of the pulsi-wave in the state of shock was frequently- less than one millin;eter, and the injections of sodium carbonate might cause an increase to a firm, strong beat of se\'eral millimeters. The following figures may be quotetl to illustrate these points: ExPEnmENT OP January 29th. — Shock produced by operation on the cerebral hemi- spheres. Normal pulse-rate=S7 ; amplitude=7 mm. Pulse-rate dviring shock=U)2; amplitHde = l mm. After injecting 10 cc. of solution of sodium carbonate 5 p?r c('nt=168; amplitude— 3 mm. Experiment of February 7th. — Shock produced bij removal of the skull-cap. Normal pulse-rate = 78; amplitude=5 mm. Pulse-rate during ,shock=108; amplitude=l mm. After injecting 20 cc. of solution of sod. carb. 5 per cent = 126; amplitude=3 mm. Experiment op .January 21st. — Shock produced by removal of skull-cap. Normal pulse-rate =75; amplitude=7 mm. Pulse-rate during shocks 150; amplitvide= — 1 mm. After injecting 40 cc. of solution of sod. carb. 5 per cent=108; amplitude=2-|- mm. Experiment op January 31st. — Shock produced by operations on cerebrum. Normal pulse-rate =80; amplitude=6 mm. Pulse-rate during shock=177; amplitude=1.5 mm. After injecting 10. cc. of solution of sod. carb. 5 per cent=129; amplitude=3 mm. Experiment of February' 14th. — Shock produced by removal of skull-cap and opera- tion on cerebrum. Normal pulse-rate =70; amplitude=4.5 mm. Pulse-rate during shock=189; amplitude=l mm. After injecting 175 cc. of a solution of sod. carb. 0.5 per cent=114; amplitude=4 mm. In the same experiment more severe shock was produced by removing portions of the cerebrum after above injection: Pulse-rate during this phase of shock=168; amplitude= — 1 mm. After injecting 50 cc. of a 0.5 per cent solution of sod. carb. = 180; amplitude=3 mm. Experiment of January 16th. — Shock produced by operations on cerebrum. Normal pulse-rate =66; amplitude=3.5 mm. Pulse-rate during shock=102; amplitude = — 1 mm. After injecting 50 cc. of a solution of sod. carb. 5 per cent=117; amplitude=4 mm. Experiment op February 23d. — Shock produced by removal of skull-cap. Normal pulse-rate=135; amplitude=7 mm. Pulse-rate during shock=180; amplitude=2.5 mm. After injecting 400 cc. of solution of sod. carb. 0.5 per cent into rectum = 201 ; amplitude = 2.54- mm. This favorable action of the solution of sodium carbonate on the force of the heart-beat cannot be attributed to the mere bulk of the injection, since it was apparent in some cases when only 10 cubic centimeters of the strong solution were injected. The effect was long-continued and might be renewed, when it wore off, by a new injection. Together with the action on the heart -beat, there was also a corresponding increase in blood-pressure. Since the two effects 58 INJECTIONS OF SOLUTIONS OF SODIUM CARBONATE. were parallel, it seems probable that the rise of arterial pressure was due en- tirely to the stronger heart -beat, and that there was no actual increase in ar- terial tone. I have had only one opportunity to test the efficacy of these alkaline solutions upon the human being. The case was a very unfavorable one. A small boy was brought into the hospital in a condition of extreme shock following on appendicitis. He was in a comatose condition; the radial pulse was imperceptible and the heart -beat could not be distinguished with the stethoscope. As death was seemingly inevitable, an infusion was made of a 0.5 per cent solution of sodium carbonate in normal saline. The effect upon the heart -beat was marked ; it could be heard easily with the stethoscope and the radial pulse also became perceptible. In this condition an operation was begun, but death followed shortly after the abdomen was opened. Effect of the Alkaline Injections upon Vascular Shock. — The influence of the injections of the solutions of sodium carbonate was always in the direction of increased arterial pressure, and this effect usuall}'' lasted for a considerable period. My original idea, however, that the increased alkalinity might bring about a restoration of vascular tone by direct action upon the peripheral vessels was not reahzed in cases of extreme shock. When the condition of vascular shock was only moderate, that is, when the general arterial pressure had not fallen lower than 60 to 70 mm., the effect of the alkaline injections was most favorable. Under these circumstances the arterial pressure could be increased to as much as 100 mm., and this favorable pressure, together with the more vigorous hearts beat, lasted throughout the period of experimentation. There can be little doubt that in such cases the effect of the alkaline injections was to prevent any further tendency toward shock, and to restore the vascular tension to an approx- imately normal level beyond the danger of collapse. When, however, the condi- tion of vascular shock was extreme, that is, when aortic pressure had fallen to 20-40 mm., which may be considered as an indication of complete loss of arterial tone, then the injection of sodium carbonate gave less gratifying results. In such cases the pressm-e could be raised to 40, 60, or even 70 mm., but never to an approximately normal level. The increase of pressure thus obtained was long continued, and it is possible that if the period of experimentation could have been extended for 12 to 24 hours, with repetition of the injections at appropriate intervals, the shocked centers might eventually have regained their tone. For the short experimental period, however, of two to four hours it was noticed that if the animal was left to itself the arterial pressure grad- ually declined to zero. It is impossible from the experiments made to state positively whether or not the alkaline injections had any influence upon the tone of the peripheral arteries, but the impression that I have obtained from a study of the records is that they acted solely as a stimulant to the heart, and that the increased arterial pressure was due chiefly, if not entirely, to a more vigorous heart-beat. WILLIAM H. HOWELL. 59 While the results of the alkaline injections in exti'enio shock were somewhat disappointing, they were far more hopeful than those obtained by the use of other substances which have been suggested as having a possible value under these circumstances. Adrenal extracts, hypophysis extracts, strychnin, alcohol, and large infusions of normal saline all proved entirely negative, or, at the best, gave only a temporary and unimportant rise in arterial pressure when used upon animals in a condition of extreme shock. To show some of the actual results obtained, the following figures are quoted from the protocols of the experiments : Experiment of January 5th. — Shock produced by removal of skull-cap and operations on the cerebrum. Arterial pressure before operation = 110 mm. Pressure after production of shock;=60 to 75 mm. Final effect of intravenous injection of 35 cc. of a 5 per cent solution of sodium carbonate in three successive doses=95 to 100 mm. Experiment op January 16th. — Shock produced by stimulation of dura mater and removal of part of cerebrum. Arterial pressure before operation=100 mm. Pressure after production of shock=22 mm. Final effect of intravenous injection of 40 cc. of a 5 per cent solution of sodium carbonate in four successive doses=46 mm. Experiment of January 21st. — Shock produced by removing skull-cap. Arterial pressure before operation=114 mm. Pressure after production of shock=20 mm. Final effect of intravenous injection of 40 cc. of a 5 per cent solution of sodium carbonate in four successive doses=40 mm. Experiment of February 7th. — Shock produced by removing skull-cap. Arterial pressure before operation=100 mm. Pressure after production of sliock=22 mm. Final effect of intravenous injection of 20 cc. of a 5 per cent solution of sodium carbonate in two successive doses=56 mm. Experiment op February 14th. — Part I. — Shock produced by removing skull-cap and part of one cerebral hemisphere. Arterial pressure before operation = 124 mm. Pressure after production of shock;=82 mm. Final effect of intravenous injection of 175 cc. of a 0.5 per cent solution of sodium carbonate, slow continuous injection=102 mm. Part II. — Further shock by removing a part of second cerebral hemisphere. Pressure at this phase of shock=34 mm. Final effect of intravenous injection of 50 cc. of a 0.5 per cent solution of sodium carbonate in two doses=70 mm. Experiment op February 23d. — Shock produced by removal of skull-cap. Arterial pressure before operation=114 mm. Pressure after production of shock=60 mm. Final effect of rectal injection of 400 cc. of a 0.5 per cent solution of sodium carbonate = 101 mm. 60 IXJEGTIONS OF SOLUTIONS OF SODIUM CARBONATE. Experiment of March 6th. — Shock produced by several long-continued operations in the abdomen; temporary ligation of superior mesenteric arteries, etc. Arterial pressure before operation=126 mm. Pressure after production of shock=30 mm. Final effect of intravenous injection of 140 cc. of a 0.5 per cent solution of sodium carbonate in two parts (100 cc. and 40 oc.)=62 mm. Effect of Cutaneous Stimulation upon Animals in Shock. — When the condi- tion of shock was severe but not extreme, it was thought that possibly the vaso- motor center might be stimulated to greater activity by strong irritation of sensory nerves. For this purpose electrical stimuli were applied to the sensory nerve-trunks or sensory membranes, or in some cases hot--Svater bags were applied to the skin. It was found that in all such cases, in accordance with the observations of Crile previously quoted, a depressor effect was obtained, the vaso-constrictor center was inhibited instead of stimulated, and the end- result was that the condition of shock was augmented rather than lessened. One may draw from these experiments the practical conclusion that in the case of human patients in a condition of shock it is unwise to attempt to whip the vaso-motor center into activity by sensory stimulations, since the tendency, if any, toward a recovery of normal vascular tone will thereby be retarded or completely destroyed. The proper procedure in such cases would seem to be to give the patient as complete protection as possible from all external stimulation. The Respirations During Shock and After Injection of the Alkaline Solutions. — The respiratory movements were recorded in these experiments by means of a tambour applied usually to the side of the thorax, and connected with a recording tambour which wrote upon the kymograph. While this method was satisfactory, as far as the rate of the respiratory movements was concerned, it left much to be desired in regard to the relative amplitude of the movements. A slight variation in the pressure of the receiving tambour upon the chest- wall in consequence of a movement of the animal or from any other cause, inter- fered at once with anything like an accurate comparison of the records at different periods in the experiment. While the respiratory records proved, therefore, somewhat unsatisfactory in many details, one or two general facts seemed to be indicated with sufficient certainty. The production of a condi- tion of shock was accompanied in most cases, after the operations leading to the shock had lost their temporary effect, by a distinct diminution in the amplitude of the respirations. The rate also showed a tendency to decrease, and in time, if the animal was left to itself, the respirations were much slower and more feeble than in the normal animal. While in this condition injections of the solutions of sodium carbonate exerted an influence which varied with the amount of sodium carbonate and the condition of the animal. In general, the injection tended to increase the amplitude of the respirations, but fre- WILLIAM H. HOWELL. 61 quently the initial effect was a temporary inhibition of the rate, which soon passed off and was followed by stronger respirations that might or might not vary in rate from the condition before the injection. When the injection of sodium carbonate was carried beyond the limits spoken of in the beginning of this paper, the respiratory center was obviously inhibited, and might be brought to a standstill or be thrown into a condition of rapid feeble movem(>nt that was obviously abnormal. Effect of Injections of Blood from an Animal in a Condition of Shock. — ^The appearance of shock in some of the animals experimented upon was so sud- den and complete that the possibility was suggested that it might be due to a toxic change of some kind in the blood. As a crucial experiment, blood was taken from an animal before shock was produced, and again from the same animal after it had been throwm into a condition of extreme shock. Both specimens were defibrinated and were injected into the veins of a normal animal. Several experiments of this character were made, but so far as could be deternained, the results were entirely negative; the shock blood, in the quantities used, had no distinct effect upon blood-pressure or pulse, and gave no indication of possessing a distinct toxic action. Summary. The conclusions reached in these experiments may be restated briefly, as follows : 1. Shock is characterized by a long-continued, low arterial pressure (vas- cular shock) due to partial or complete loss of activity of the vaso-constrictor center, and by a rapid feeble heart-beat (cardiac shock) due in part, at least, to a partial or complete loss of activity of the cardio-inhibitory center. 2. Cardiac shock may occur more or less independently of vascular shock, but vascular shock is always preceded or accompanied by cardiac shock. The respirations in shock are diminished in amplitude and usually in rate. 3. Shock may be produced experimentally by severe operations of various kinds, but most often by extensive operations on the brain. 4. The physiological evidence in experimental shock indicates that the condition is due fundamentally to a strong inhibition of the medullary centers (vaso-constrictor, cardio-inhibitory), leading to a long-continued suspension of activity, partial or complete. 5. Injections of alkaline solutions of sodium carbonate, intravenously or into the rectum during shock, increase markedly the amplitude of the heart- beat and bring about a rise of arterial pressure. When the shock is moderate (aortic tension of 60-70 mm. Hg.), the injections may restore arterial press- ure to an approximately normal level. When the shock is severe (aortic tension of 20-40 mm. Hg.), the injections may increase arterial pressure by about 100 per cent for long intervals, and the effect when it wears off may 62 INJECTIONS OF SOLUTIONS OF SODIUM CARBONATE. be restored by repeating the injections. The effect of the injections is due chiefly or entirely to a direct action on the heart. 6. Stimulation of sensory nerve-trunks or sensory surfaces in an aninaal in a condition of shock leads to a further fall of pressure, and to this extent augments the condition of shock. 7. The blood of animals in a condition of shock has no toxic action when injected into the circulation of a normal animal. ON THE OEGANIC PEROXIDES. PAUL C. FREER, Ph.D., M.D., Professor of General Chemistry, AND FREDERICK G. NOVY, Sc.D., M.D., Professor of Bacteriology. {From the Laboratories of Genera! Chemistry and of Hygiene, University of Miehigan.) I. The Formation axd Decomposition of the Peroxides. Hoitsema^ has shown that the amount of hydrogen dissolved in pallad- ium is, -ft-ithin certain limits, directly proportional to the pressures exerted, but if these are very small, then the quantity is proportional to the square root. The conclusion to be reached is, therefore, that, under these condi- tions, the hydrogen is present in the form of individual atoms. The forma- tion of water from hydrogen and oxygen, under ordinary conditions, takes place with unmeasurable slowness, but is increased to measurable rapidity by the presence of certain substances, so that with finely divided metals, such as platinum or palladium sponge, the rate of change is often so rapid that explosion may occur. Unquestionably, in these cases, the metal has a specific action, but the intensity of this is increased by increasing the surface. The gases in these cases are, perhaps, best considered as being in solution in the solids, and the dissociating action of solvents would produce the mona- tomic condition made apparent by Hoitsema. Mond, Ramsay, and Shields^ have shown that palladium foil, sponge, or black take up equal quantities of hydrogen, so that the solubility of the gas is independent of the surface. The hydrogen in these solutions has the characteristics of a metal, and ob- serving the analogy between this and fluid solutions of one metal in another, it also seems probable that the hydrogen is in the form of individual atoms. The increased activity can be explained not only by the greater concentra- tion of hydrogen in unit space, but also by the dissociation which has taken place. The resistance to the oxidation of gaseous hydrogen may rest upon the slow rate at which the molecules are split into atoms, and this disso- ciation takes place more rapidly in metallic solution, thus accounting for the increased rate of oxidation. The same must be true, to a greater or less extent, of all surfaces, although not all have the same solvent or dissociating power for hydrogen. ' Zeitschr. physik. Chem., XVII, p. 1. ^ Proc. Roy. Soc, LXII, p. 290. • 63 64 ON THE ORGANIC PEROXIDES. In the case of oxygen there is no such marked absorption by metals, so that instances in which oxygen is undoubtedly rendered active by the presence of catalyzers are not frequent. The change from sulphur dioxide to sulphur trioxide by the action of platinized asbestos or platinum is un- questionably an example.^ The oxidation of mixtures of benzaldehyde and acetic anhydride, when spread on sand and exposed to the air, produces, according to the conditions, either benzoyl peroxide^ or benzoyl acetyl peroxide.^ These reactions were, in our opinion, also cases in which the catalytic action of the solid surface brought about the oxidation, so that, while undertaking some bacteriological studies on the action of organic peroxides, we also studied the influence of catalytic agents on the forma- tion of this class of bodies. Nef has shown that the formation of benzoyl acetyl peroxide (accom- panied by that of benzoyl acetyl oxide and benzoic acid) is much more rapid when the mixture of benzaldehyde and acetic anhydride is spread upon sand than when it is exposed to the air in glass vessels; this acceleration, in his paper, is tacitly ascribed simply to mechanical causes, owing to the greater ease of contact between the air and the mixture. In order to determine whether the contact of a sufficient volume of air is alone necessary for rapid peroxide formation, or whether the surface action of a solid is necessary, we performed the following experiments : Fifty grams of benzaldehyde and 50 grams of refractioned acetic anhy- dride were placed in an ordinary gas washing-bottle, the entrance tube of which was drawn out to a number of capillaries, so as to insure a very fine division of the air bubbles. This wash-bottle was placed in a freezing mix- ture, and air, previously cooled by passing through a worm and drying bottle with concentrated sulphuric acid, both packed in a freezing mixture, was passed through in a rapid stream for seven hours, at the end of which time the contents of the bottle suddenly solidified. The crystals were separated and recrystallized from petroleum ether, when they proved them- selves to be pure benzylidene diacetate. The yield was only a little less than quantitative, and only minimal quantities of either benzoyl acetyl peroxide or benzoic acid were formed. Repetition of the experiment, at the same temperature, produced the same result. At higher temperatures (13° to 14°), even after a current of air had freely passed for twenty-four hours, only a trace of peroxide was produced, but at the same time no benzylidene diacetate, or at least no quantities sufficient for isolation from the mixture in the flask, could be observed. The benzaldehyde and acetic anhydride were almost completely unchanged. The rapid formation of benzylidene ' For an extended review of tliis subject see Bodlander in Ahrens' Vortrage, III, p. 385. 2 H. Erlenmeyer, Jr. : Ber. d. chem. Ges., XXVII, p. 1959. « Nef: Ann! Chem. (Liebig), CCXCVIII, p. 280. PAUL C. FREER AND FREDERICK G. NOVY. 65 diacetate, therefore, seemed to be dependent upon the condition of low tem- perature. The result shows that, when the mixture of benzaldehyde and acetic anhydride is kept agitated by a rapid current of nearly dry air, the formation of benzylidene diacetate is much more rapid than when the mix- ture is left standing alone, for Nef found only 2 grams of benzylidene di- acetate formed from 4S grams of benzaldehyde, even after five days' standing. In Nef 's experiments a large quantity of acetic acid (24 grams) was added, whereas in the above, the acetic anhydride was twice refractioned, and so could contain only a trace of acetic acid. A large quantity of acetic acid, therefore, has no influence in increasing the rate of action, but on the con- trary it seems to diminish it. If Nef's assumption that the formation of benzyhdene diacetate is due to a previous addition of acetic acid, separation of the latter to form nascent benzaldehyde, and subsequent addition of acetic anhydride, is correct, j;C< 9 > >C-(); H^ ^0C0CH3 H^ i I H^ I I \cOCH3 H^ "^OCOCH, then, the increase in the concentration of the acetic acid should accelerate rather than retard the rate of reaction. In order to test whether substances other than acetic acid would also affect the rate of formation of benzylidene diacetate, 10 grams of aldehyde and 10 grams of refractioned acetic anhydride were placed in a tube with a strip of platinum foil, 1 cm. wide and 5 cm. long, and dry air was run through for twenty hours. The current was very slow, so as to secure merely an agitation of the liquid. At the end of this time 16.5 grams of benzylidene diacetate had been formed; calculated 19.5 grams. By far the greater portion of this diacetate was produced in the first hour, and the tube at the beginning became quite warm, showing that under these conditions the addition of acetic anhydride to benzaldehyde is much accelerated, although only a trace of acetic acid could be present. .On repeating the above experi- ment, aU conditions being the same except that the tube was cooled to 10°, the formation of the diacetate was retarded so that the change was not com- plete for three days. The same mixture, sealed in a tube and placed in the shaking machine, solidified completely, forming benzylidene diacetate after seven hours, so that agitation by means of a current of air did not accelerate the reaction. Other metals have even a more marked effect. Zinc and tin, when placed in tubes containing the mixture of aldehyde and anhydride, 1 Ann. Chem. (Liebig), CCXCVIII, p. 278. 66 ON THE ORGANIC PEROXIDES. effect the change in about half an hour, with a slow current of air passing through, and with iron the reaction-products become heated at once, and the addition is complete after a few minutes. Various metals, therefore, exert a marked catalytic action on the forma- tion of benzyhdene diacetate, independently of the amount of acetic acid which may be present. Acetic acid, however, may act as a catalytic agent, increasing the velocity of the reaction, but not altering the final equilibrium. Indeed, in accordance with Nef 's results, we have found that acetic anhy- dride, which had been freed from acetic acid by standing over sodium and then distilling, did not form benzylidene diacetate in appreciable quantities, with iron and a slow current of air, until thirty-six hours had elapsed. And probably, as in Nef's case, mixtures of benzaldehyde and acetic anhydride, purified in this manner, if simply standing in a closed flask, would react with such extreme slowness that amounts of benzylidene diacetate sufficient for isolation would not be produced for months. We have not completed our work on the action of catalysis in the formation of benzylidene diacetate, as very slight variations in the purity of the materials or of the catalyzers pro- duce such marked variations in the rate of formation that concordant results are difficult to obtain. Neither have we determined whether a mixture of benzaldehyde and acetic anhydride, purified by standing over sodium, will eventually produce benzylidene diacetate, but the work has gone far enough to demonstrate that various substances have different effects on the rate of addition, and that even traces of acetic acid are just as effective as the large amounts used by Nef. The conclusion, therefore, is that the action of acetic anhydride on benz- aldehyde is purely one of addition, and that the course of the reaction is simpler than the one proposed by Nef. It is nevertheless probable, in view of the work of Baeyer and Villiger' on benzoperacid and on the oxidation of benzaldehyde, as well as the other additive reactions in which benzaldehyde takes part, that a methylene derivative, takes part in the reaction. Many changes which aldehydes undergo can best be explained by the existence of an equilibrium between the two isomers R^ R I. >C = and II. ^C^ . H^ H— 0^ ^ Only a trace of the enol form (II) need be present to start reactions. As soon as form II is removed, the equihbrium is restored by a renewed change ^ Ber. d. ohem. Ges., XXXIII, p. 1582. PAUL C. FREER AND FREDERICK G. NOVY. 67 from I (keto form) to II (enol form), and so on until reaction is complete. The ease with which the change can be brought about would determine the rate at which a given aldehyde would react. The existence of an equili- brium between keto and cnol form in acetoacetic ether is now extremely probable, in view of the work of Schiff,' as well as the large mass of modern literature on desmotropic forms of substances, which seems to show that an equilibrium between the two forms exists in many liquid aldehydes and ketones. This view need only be extended to ordinary aldehydes of the form ^C = to have a rational explanation of the various reactions of this class of bodies. Catalyzers, as in the above case of the formation of benzylidene diacetate, would then simply affect the rate of change of keto to enol form, on exactly the same principle as we suppose that catalyzers can increase the rate of dissociation of the hydrogen molecules into atoms in the formation of water. ^ The relative inactivity of the carbonyl oxygen atom in the carboxyl group R, .0 R. ,0H ^C^ iS > J^C versus R. ^0 R >C^ S > would thus be explained by the mere fact that two distinct forms, one of which is the intensely reactive unsaturated form, and the consequent equili- brium, are not possible, and hence carbonyl in the carboxyl group cannot behave like carbonjd in aldehydes and ketones. As the rapid formation of benzylidene diacetate, provided acetic anhy- dride which has merely been refractioned is used, causes the contents of the tubes through which air is passed to solidify before appreciable quan- tities of peroxide are formed, we were compelled, in all experiments relating to the influence of surface on the rate of oxidation, to use acetic anhydride which had stood over sodium for some time, and which was then poured off and refractioned. By this means the formation of benzylidene diacetate is so greatly retarded that, in most cases, we were able to run air through the mixture of benzaldehyde and acetic anhydride long enough to obtain quantities of benzoyl acetyl peroxide which could be quantitatively deter- mined. The problem of ascertaining the amount of peroxide which is present in a mixture of benzaldehyde, acetic anhydride, acetic acid, benzylidene > Ber. d. chem. Ges., XXXI, p. 205, and following. 2 Bodlander: Ueber langsame Verbrennung, Ahrens' VortrSge, III, p. 439. 68 ON THE ORGANIC PEROXIDES. diacetate, and benzoyl acetyl oxide, all of which bodies must be present in the reacting liquid after air has been run in for a sufficient length of time, is not an easy one, and we soon discovered that such means as decolorizing indigo or titration with potassium iodide could not give accurate results. We finally adopted the method of heating a measured quantity of the mix- ture, resulting from oxidizing benzaldehyde and acetic anhydride by means of a current of air, and measuring the evolved gas. That this gives a reason- ably accurate means of quantitatively determining benzoyl acetyl peroxide in the presence of all of the other substances occurring during the reaction is shown by the following experiments : A weighed quantity of benzoyl acetyl peroxide was dissolved in about 5 c.c. of benzaldehyde and the solution placed in a small, round-bottomed flask, which was connected with a carbon dioxide apparatus on the one side and a Schiff 's azotometer on the other. After all air had been expelled the flask was heated with a free flame until evolution of gas began, then the flame was removed, and only brought under the flask from time to time, as the reaction required. The decomposition must take place as rapidly as possible without causing explosion. Calculated gas volume for 1 molecule gas per Gram peroxide used. 2 molecules peroxide. Found. 0.2445 15.1 14.3 0.1199 7.4 7 3 0.2084 12.9 12 7 0.2018 12.5 12.3 0.0941 5.8 6.0 0.3306 20.5 21.2 As will be seen from the above, 1 molecule of gas is developed for every 2 molecules of peroxide, and the method is sufficiently accurate for all pur- poses of measurement. No gas whatever is obtained on heating pure benz- aldehyde or a mixture of benzaldehyde and acetic anhydride, or any other mixture not containing peroxide under the above conditions, as we have proved to our satisfaction. In the presence of benzaldehyde the yield is always quantitative. The gas collected over potash, much to our surprise, proved to be methane. The gas collected from 5 grams of benzoyl acetyl peroxide was allowed to stand over strong potash solution for twenty-four hours and was then analyzed by combustion. Calculated for CH4. Found. C 75.0 75.4 H 25.0 24.6 A gas analysis, kindly carried out by Mr. White, of this laboratory, by the usual methods, confirmed the combustion and proved the gas to be methane. The gas does not reduce a solution of palladium chloride, hence no hydrogen is present. PAUL C. FREER AND FREDERICK G. NOVY. 69 If the gas given off by heating benzoyl acetyl peroxide is collected over mercury, then the volume is equal to that calculated for 4 molecules of gas for every 2 molecules of peroxide decomposed, and of these four molecules three are carbon dioxide. I. 1.111 grams substance in 40 c.c. of xylene were decomposed by heat- ing to 105° in an oil-bath, the temperature being gradually increased to 155°, and the total gas collected over mercury. The gas was then passed through a potash bulb by means of a current of air, free from carbon dioxide, and the carbon dioxide weighed. Found 0.379 gram CO,. II. 1.238 grams substance in 20 c.c. of xylene were decomposed as above, care being taken not to heat above 110°. Found 0.431 gram CO,. Calculated for 3 mole- cules COa to 2 mole- Found. cules peroxide. I. II, COa 36.8 34.1 34.8 The percentage of gas obtained is slightly low, but the decomposition is almost quantitative. A very minute quantity of benzoyl acetyl peroxide, therefore, decomposes differently. The volume of methane obtained also varies slightly from the calculated. The volume of gas not absorbable by potash varies somewhat if the peroxide is very slowly decomposed in xylene, but is fairly constant, as shown above, if the substance is heated over a free flame. Two molecules of benzoyl acetyl peroxide, therefore, decompose into three molecules of carbon dioxide, one of methane, and a liquid residue. The empirical equation would be : 2C,H,0. = SCO, + CH, + C„H„0,. The high-boiUng residue was examined as follows: The xylene was distilled under diminished pressure from the residue obtained by heating five grams. The remainder was then taken up in ether and extracted with sodium carbonate. On acidifying the carbonate, 0.11 gram of an acid was isolated. This acid has not been identified. The ether was then distilled off and the non-acid remainder boiled for six hours on a water-bath with two grams of alcoholic potash. The alcohol was then distilled off and the whole taken up in water. The non-saponifiable portion was carefully extracted with ether. The alkaline solution, when acidified, gave 0.7 gram of benzoic acid. The non-saponifiable portion was fractioned at 18 mm. pressure. Two grams went over between 160° and 185°, and about an equal weight of high-boiling residue, which would not distil up to to 265°, remained. The distillate is colorless and boils between 260° and 285° at ordinary pressure. It cannot be crystallized. No constant-boiling product 70 ON THE ORGANIC PEROXIDES. could be obtained from it. An analysis showed this remainder to be a mix- ture of high-boiling hydrocarbons. Weiffht of substance 0.1332. ° Found. ' Gram. Per cent. H 0.0114 8.5 C 0.1193 89.6 Total, 0.1307 98.1 The hydrocarbons have not, as yet, been identified. The two grams of high-boihng residue finally passed over, when heated with a free flame to 335°, leaving practically no residue. This high-boihng portion is not a hydrocarbon. The result shows that, while the gaseous products are related to the original body in a very simple manner, the liquid remainder is by no means as simple. A large portion consists of hydrocarbons of high boihng-point, so that the result of the decomposition has been the couphng of the residues. The saponifiable portion, which yields the benzoic acid, is probably phenyl benzoate. Hitherto the amount of substance at our disposal has been too small to follow out this interesting decomposition in such a way as to identify the liquid portion, but we intend to return to this question in the near future, and clear up the mechanism of this change as well as the decom- position of other peroxides by heat, for the reaction may be made one of great synthetic importance. The simplest change, cnno.oo.cocH, c^h.co.oio.coch, C.H^CO.OO.COCH, ■' C.H,CO.OiO.COCH3 C,H,|CO OjO C0;CH3 c,Hjco"biO"coicH, by which diphenyl and ethane would be produced, does not take place, and it is a remarkable fact that the methane, which must unquestionably be pro- duced from one of the methyl groups by reduction of a second methyl group, should result at all, in preference to ethane. The fact that this reaction is quantitative in the presence of benzaldehyde would seem to indicate that this substance takes part in the reaction. Its hydrogen may go to form temporarily acetic acid, which then may break up into COj and CH^. The cleavage of COj in this way, especially by bacteria, is known to occur. The further study of the decomposition products will probably do much to clear up the constitution of organic acid peroxides. The above results having fully established the feasibility of determining the quantity of benzoyl acetyl peroxide which can be produced by passing air through a mixture of benzaldehyde and acetic anhydride, we first made the following blank test : PAUL C. FREER AND FREDERICK G. NOVY. 71 Twelve test-tubes were each filled with a mixtures of five grams of benz- aldehyde and five grams of acetic anhydride, both dry, and a current of thoroughly dried air was slowly passed through all the tubes at once, placed in series. The first six tubes were opened at intervals of one-half hour, and the contents tested for peroxide. The amount scarcely increased after the first half-hour, when 1 c.c. of gas only could be collected. One of the tubes was then opened and a small piece of filter-paper, from which the fat had been extracted, placed in it, with the result that the peroxide production increased by 200 per cent in two hours. The tubes which had run for about eighteen hours finally solidified to benzyhdene diacetate. This test shows that, unless surfaces other than those of the test-tube and delivery-tube are present, the production of peroxide is very slow, despite the large quantity of air which is run through in six hours. An equilibrium is finally reached in which peroxide is destro3'ed as fast as it is formed. The paper surface instantly produces a marked increase. The same series of tubes were now cleaned thoroughly, dried, and each filled again with, the same mixture. Pieces of pure metal, in strips, were now dropped in, care being taken to have the surface approximately the same in all, and the strips so placed that the entering air would come in contact with them. The current was admitted during fifteen hours, temperature 12°, when the tubes were taken out, the air stopped, and the peroxide present in each esti- mated in turn. By this method the tube opened last stood twelve hours, without air being admitted. Tube. Contents. 1 -Aluminum 2 Extracted filter-paper 3 Blank 4 Magnesium 5 Iron * 6 Copper 7 Tin 8 Platinum The accelerating effect of the surfaces is very apparent in the first portion of the series, but the gradual decrease of peroxide present the longer the tubes were allowed to stand before determining the gas given off, makes it evident that decomposition of the peroxide takes place, and becomes evident as soon as the current of air ceases. As a consequence we repeated the above series, placing the tubes in the current of air successively after a given interval of time, then allowing each tube to be subjected to the action of air for the same length of time, twenty-two hours and forty minutes, removing each in the order in which it was placed in series, and determining the gas at once. Each tube, therefore, stood the same length of time as all of the others. Absolute quantity Gas delivered. of peroxide. c.c. Gram. 12.79 0.2055 10.43 0.1691 4.14 0.0670 9.76 0.1580 7.83 0.1267 5.76 0.0732 1.56 0.0256 0.31 0.0050 72 ON THE ORGANIC PEROXIDES. ^be. Contents. 1 Aluminum 2 Extracted filter-paper 3 Blank 4 Magnesium 5 Iron 6 Copper 7 Tin 8 Platinum Gas delivered. CO. Absolute quantity of peroxide. Gram. 7.2 0.11656 13.6 0.2201 6.4 0.1036 63.2 1.0232 40.5 0.6556 1.7 0.0275 13.4 0.2169 15.2 0.2460 During this reaction the iron (2.728 grams) lost 0.0064 gram, the aluminum 0.0006 gram from 1.485 grams, and the copper 0.064 gram from 2.366 grams. The copper was decidedly attacked therefore, whereas all other metals were only altered, if at all, to a sHght extent. In order to test whether the fact that the metal was attacked would, in the case of copper, reduce the yield, we car- ried out a series of experiments with copper alone. Pieces of copper of equal size, previously completely reduced in a current of hydrogen, were placed in the tubes and the following results obtained : Time of determination after first admitting air. C.c. of gas Hours. obtained. 2 2.7 4 6.7 6 4.1 8 3.7 10 3.0 12 2.8 14 2.8 24 2.5 It is evident from this that in the case of copper the formation of peroxide rapidly increases to a maximum, the rate of decomposition then exceeds that of formation until a practical equilibrium is reached after twenty-four hours. The metal is attacked in all cases, but the cause of the decomposition is not to be found in the copper acetate produced, because a tube containing copper acetate only gave 11.9 c.c. gas after fifteen hours. This peculiar decom- position of copper will be the subject of future investigations. The above results show, beyond a doubt, that the rate of formation of per- oxide is dependent upon the surface action of substances coming in contact with air and the mixture of benzaldehyde and acetic anhydride at the same time, for the blank tubes, in each case, produce no greater quantity of peroxide than would naturally be expected from the surface of glass exposed. Of all substances used, magnesium seemed to have the greatest effect, so that this metal was used in an experiment carried out with the purpose of determining whether absolutely dry air can form benzoyl acetyl peroxide with a mixture of absolutely dry benzaldehyde and acetic anhydride. The point may be raised that if the acetic anhydride is not perfectly dry it may react with water and that the resulting acid acting upon the metal PAUL C. FREER AND FREDERICK G. NOVY. 73 will yield hydrogen, and thus hydrogen peroxide, which in turn reacting with acetic anhydride would yield acetoperoxide. This view is not supported by the following experiment : The apparatus consisted of a drying train consisting of a worm tube with sulphuric acid, a tube three feet long filled with glass wool dusted over with phosphorus pentoxide, and a tube four feet long filled with freshly glowed soda-lime, a reaction tube with the entrance tube for air sealed in and sur- rounded by a spiral of pure magnesium ribbon, and the exit tube sealed to the side and connected to the aspirator by means of a tube three feet long, filled with glass wool and phosphorus pentoxide. The exit tube of the reaction flask was connected by means of a T, with two distilling flasks which were so adjusted that the first and last portions of the distillate could be run off through side stop-cocks, while any quantity of liquid necessary could be introduced into the reaction flask. All parts of the apparatus were sealed together so as to form a continuous piece of glass. The acetic anhydride used was allowed to stand over sodium for several weeks ; the benzaldehyde was distilled several times. The apparatus was heated for several hours at 150°, with a current of air passing through; the acetic anhydride and benzaldehyde were then placed in the distilling flasks and distilled over into the reaction tube, the first portion of the distillate being rejected in each case, and the tube connecting with the apparatus was then fused off. A slow current of air was then passed through for twelve hours. Ten c.c. of mixture gave 12 c.c. of gas, corresponding to 0.1952 gram of per- oxide. The presence of moisture is, therefore, not necessary in the formation of benzoyl acetyl peroxide. In this experiment 1.234 grams of magnesium were used, which lost only 0.0041 gram during the time of the reaction, and this loss is probably entirely due to the formation of oxide, during the heating and drying, which was removed during the process of cleaning. The magnesium is, therefore, not acted on during the twelve hours of the experiment, although acetic acid must be produced according to any view which we can take of the formation of benzoyl acetyl peroxide from mixtures of acetic anhydride and benzaldehyde. The same result was obtained in other cases when air was run through the mixture for a sufficient length of time to produce acetic acid enough to more than dissolve all the magnesium present. Nef's view' of the formation of benzoyl acetyl peroxide supposes that, in the presence of water, benzaldehyde adds the latter to form dioxytoluene : C„H,CHO + HOH = >C<^ which again separates as follows: > Ann. Chem. (Liebig), CCXCVIII, p. 280. 74 ON THE ORGANIC PEROXIDES. "•='•><"'' = '■"■>< +H.0. The latter substance then again acts on water to form orthobenzoic acid and hydrogen : >C< + 2H0H = ..... >C< , + 2H. The hydrogen then forms hydrogen peroxide with molecular oxygen, while the benzoic acid forms benzoyl acetyl oxide with acetic anhydride. Finally, hydrogen peroxide is supposed to oxidize benzoyl acetyl oxide to benzoyl acetyl peroxide. In A'iew of what has been shown above as regards the ease with which benzaldehyde adds acetic anhydride to form benzylidene diacetate, it seems scarcely credible that, even in the presence of a trace of water, the benzalde- hyde should add the latter substance in preference to the former. Nef 's view of the mechanism of the reaction must, however, be definitely abandoned after the above demonstration that water is not a necessary factor in the formation of benzoyl acetyl peroxide. The intermediate formation of benzoyl acetyl oxide also does not necessarily accompany the action of air on the mixture of aldehyde and anhydride because, as will be shown below, we have succeeded in so conducting the oxidation as to obtain a crude peroxide, which crystallizes completely with only a minimal formation of benzoic acid. Were benzoyl acetyl oxide present in any quantity, it would either manifest itself by being present as an oil mixed with the neutral crystals of peroxide, or, if it were hydrolyzed, its decomposition would show itself by the presence of benzoic acid. The most reasonable and simple way in which to look upon this oxidation is to assume the existence of an equilibrium between the keto and enol forms of benzaldehyde. The latter would then manifestly add oxygen to form benzoperacid. HO^ "- H-0 0^ which, in turn, would react with acetic anhydride to form benzoyl acetyl peroxide.^ This view of' the reaction would not affect the question as to the structure of benzoperacid, because formula II, which is the more probable one, can easily result from I by a rearrangement similar to those so frequently observed in organic chemistry. In any event, the structural formulas of ' Clover and Richmond have shown that benzoperacid may be tested for quantita- tively by yielding benzoyl acetyl peroxide upon the addition of acetic anhydride. " With solutions of the latter peroxide it yields benzoperoxide. (Am. Chem. Journ., March, 1903.) PAUL C. FREER AND FREDERICK G. NOVY. 75 hydroperoxide and the organic peroxides are as yet undetermined, although the weight of evidence points toward a symmetrical structure, so that this ques- tion must be a subject for future investigation. The above results are in accord with those obtained by Baeyer and Villiger' on the oxidation of benzaldehyde. The views of Erlemheyer, Jr.,^ Jorissen,' and Engler and Wild* are all based on the supposition that benzoyl peroxide is the result of the oxidation of a mixture of benzaldehyde and acetic anhydride, and are consequently to be rejected because benzoyl peroxide is only a secondary product of the reaction, for we have convinced ourselves that acetic anhydride cannot convert benzoyl peroxide into benzoyl acetyl peroxide. The supposition of Erlenmeyer, Jr.,^ that possibly an addition-product of benzaldehyde (benzylidene diacetate) is first formed, which latter substance is then oxidized to a substance which is converted by the mixture into benzoyl peroxide and acetic acid is also not correct, for benzylidene diacetate cannot be oxidized to a peroxide, either by air or by hydrogen peroxide. The same view in regard to the mechanism of the reaction was advanced by Bodlander^ and by Michael,^ both of whom assumed the existence of benzoperacid, C„H,CO.O.O.H, before its discovery by Baeyer. The above experiments also prove conclusively that the formation of perox- ides is remarkably accelerated when certain surfaces are present as catalyzers, and that this influence varies with the nature of the surface. In order to prove that the extent of liquid surface exposed is not of influence, a mixture of equal quantities of benzaldehyde and acetic anhydride was placed in a Fernbach flask so that the bottom was covered to a depth of about 3 mm. The flask was allowed to stand for twenty-four hours, at the end of which time only a very small quan- tity of peroxide had been forrned. The same mixture formed in a large, flat watch-glass, and then drained off, so that a very thin layer was left, so thin that direct contact of the air with the glass surface was possible, was com- pletely oxidized to a mixture of benzoic acid and benzojd acetyl peroxide within flve minutes. These results prove conclusively that the oxidation of mixtures of benz- aldehyde and acetic anhydride cannot be referred in any great measure to a partial dissociation of the ordinary oxygen of the air, as is supposed by van 't 1 Ber. d. chem. Ges., XXXIII, p. 1582. ' Ibid., XX.VI1, -p. 1961. ' Ztschr. physik. Chem., XXII, p. 58. * Ber. d. chem. Ges., XXX, p. 1677. 5 Ibid., 1961. ' Akrens' Vortrage, III, p. 471. ' J. prakt. Chem., LX, p. 75. 76 ON THE ORGANIC PEROXIDES. Hoff.' As this acceleration of the rate of oxidation is brought about by a variety of surfaces, such as sand, paper, cloth, cotton, and various metals, it seems probable that the oxygen of the air, when occluded by any surface, is rendered chemically more active, and this increased activity is due to a partial dissociation of the occluded oxygen molecules, = = 0-0. 1 I This view is of great importance from a biological standpoint, as, for example, the ease with which hemoglobin is changed to oxyhemoglobin in the lungs can be referred to the large surface of tissue exposed. If the assumption of an equilibrium between the keto and enol forms is correct, then the difference between the activity of various surfaces can be explained by the assumption that the substances introduced into the mixtures of aldehyde and anhydride have different effects on the rate of change from the keto to the enol form, whereas the surface action on the oxygen remains alike in all. The above experiments clearly indicated to us the best means of procuring any desired quantity of benzoyl acetyl peroxide. A mixture of equal weights of benzaldehyde and acetic anhydride is pre- pared, the mixture placed upon strips of filter-paper or muslin, hung in a large glass Jar with a cover, and dry air gradually run through. Care must be taken not to place too many strips in the jar at one time, as contact combined with the rapid formation of peroxide will in a very few minutes bring about local heating to such an extent that the paper chars, and may even take fire. After the odor of benzaldehyde has disappeared, the paper or cloth is extracted with low-boiling petroleum ether, the solution washed with a little soda solution until the washings just become alkaline,^ then the petroleum ether is partly distilled off from a bath of luke-warm water (the bath must not exceed 80° in temperature, otherwise explosion may result), and the solution is then placed in a freezing-mixture. Crystals of benzoyl acetyl peroxide soon separate, and can be filtered off and recrystallized from the same solvent. On concentrating the petroleum ether further crops of crystals, somewhat less pure, are obtained. The yield often reaches sixty-eight per cent of the theoretical. It varies with the temperature, rate of admission of air, thoroughness of drying, and amount of sunlight. Benzoyl acetyl peroxide thus obtained is identical with that described by Nef . When pure it can be kep't apparently indefinitely, provided no moisture comes in contact with it. In the presence of impurities, such as moisture, traces of alcohol, ether, or acids, it gradually decomposes, becoming • Ztschr. physik. Cheiu., XVI, p. 411. 2 But a very small amount of benzoic acid is produced. This proves that under these conditions but little benzoyl acetyl oxide is formed, for this substance is hydrolyzed by moist air, and the resulting benzoic acid would appear in the alkaline solution. PAUL C. FREER AND FREDERICK G. NOVY. 77 liquid and then slowly depositing eiystals of dibenzoyl peroxide. The lique- faction is due to hydrol\'tic changes which will be presently considered in detail. Tubes sealed up, the contents of which had liquefied as above, showed no pressure, so that oxygen is not given off. These changes will be the subject of future investigation. The melting-point of pure benzoyl acetyl peroxide is 40° to 41°. Pure benzoyl acetyl peroxide is not an oxidizing substance, nor has it any odor, a fact which was also observed by Baeyer.' The following experiments will make this clear. Twenty grams of benzoyl acetyl peroxide were dissolved in 200 c.c. of abso- lute ether, and then 10 grams of anihne were added, and the well-stoppered flask allowed to stand in ice-water for twelve hours. At the end of this time 3.5 grams of a solid, which proved to be acetanilide, had crystallized out. The ether was filtered off and evaporated to one-third its bulk, in the cold, and 250 c.c. of ligroin were then added. There separated 10.5 grams of a red solid. This was extracted with sodium carbonate solution, which was then acidified, giving a crystalline precipitate of benzoic acid. The remainder was boiled with water until no more dissolved and the water decanted from a dark red, insoluble oil. The substance crystallizing from the water was acetaniUde. The dark red oil, when recrystallized from ligroin, melted at 34° to 35° and was azoxybenzene. The ligroin-ether solution was evaporated in vacuo in the cold, and the remainder extracted with low-boiling petroleum ether (40°), and the ether evaporated. The remainder had the penetrating odor of benzoyl hydrogen peroxide. It was taken up in ice-cold caustic soda, leaving a residue of ben- zoyl peroxide, which was filtered. The cold caustic soda solution was then precipitated with barium chloride, yielding a small quantity of a precipitate which decomposed on standing over night at 0°. It was probably benzoper- acid. All of the crude acetanilide which separated gave a sharp odor of phenyl- isonitrile on warming with alkalies pointing to the intermediate formation of phenylhydroxylamine, but not enough of this substance was present to be isolated. In aU there were obtained in a pure state 9.3 grams acetanilide 4.7 grams benzoic acid, 2.5 grams azoxybenzene, 0.5 gram benzoyl peroxide, and traces of benzoperacid. In another experiment in which we used one-half as much peroxide, we were able to isolate three grams of aniline benzoate also, with the other products the same, except that no benzoperacid could be detected. The first change in this reaction, which is very gradual, is always the separation of acetanilide, and no oxidation takes place until this change has begun. This fact proves conclusively that the benzoyl acetyl peroxide is not in itself capable of oxidizing aniline, but that it must be separated into ben- ' Ber. d. chem. Ges., XXXIII, p. 1574. 78 ON THE ORGANIC PEROXIDES. zoperacid and acetanilide before its oxidizing action can become apparent. The benzoperacid then changes the aniline to azoxybenzene, probably with the intermediate formation of phenylhydroxylamine, while it is itself converted to benzoic acid. If an excess of peroxide is present, then some benzoperacid can be isolated. At the end of the reaction, if an excess of aniline is used, then aniline benzoate appears. Pure benzoyl acetyl peroxide does not produce any marked effect on hydro- quinone in benzene for forty-eight hours, if water is excluded; after this time oxidation gradually sets in with the production of quinhydrone. That a change finally sets in is probably due to the fact that a minute quantity of water, which cannot be excluded in seahng up a flask, gradually causes hydroly- sis with the formation of benzoperacid, while the latter oxidizes the hydro- quinone to quinhydrone and water, so that the reaction, once begun, would go on with increasing velocity. Although pure benzoyl acetyl peroxide is not capable of oxidizing action, the aqueous solutions are, for they promptly liberate iodine from potassium iodide. This difference between the normal solid and the solution is due to hydrolysis. This change was followed out quantitatively in the following manner : Ten flasks, with necks drawn out for sealing, were filled with two grams each of benzoyl acetyl peroxide and 100 c.c. of water. The containers were then filled and placed in a shaker kept at 26° in a thermostat. The flasks were opened in the order given in the following table, the contents filtered, and the active oxygen in solution determined by means of acid potassium iodide and titration with N/20 sodium thiosulphate. In a certain number of flasks the solid residue was weighed to determine the point at which equihbrium, if any, was reached : Benzoyl Active oxy- acetyl per- Active oxy- Total Thiosul- gen in oxide repre- Solid gen in active Hours. phate. solution. sented. left. solid. oxygen. 18 44.098 0.0353 0.3967 24 50.125 0.041 0.4611 1.49 30 71.335 0,057 0.642 48 84.04 0,067 0.756 1,31 72 135.0 0.108 1.22 0,91 96 147.34 0,118 1.326 120 149.41 0,1195 1.345 144 156.25 0.125 1,407 0.701P 0,0464 0.1714 216 155.6 0.1245 1,326 0.7166^ 0.0469 0,1715 404 138.9' 0.1112 1,259 0.7467^ 0.0494 0.1606 Calculated total active oxygen in two grams of benzoyl acetyl peroxide, 0.1778. The solid, after twenty-four hours, melted for the most part at 39° to 40°, only a small portion having a higher melting-point; the amount of benzoyl ' Slight pressure in tube on opening. ^ Pure benzoyl peroxide. PAUL C. FREER AND FREDERICK G. NOVY. 79 peroxide gradually increased until, after one hundred and forty-four hours, it melted at 106° to 107° sharp, and was pure benzoyl poroxido. Before this time the residual solid was a mixture of benzoyl peroxide and unchanged benzoyl acetyl peroxide, so that, as the relative proportions were unknown, it was impossible to calculate the amount of active oxygen remaining. After all had been converted to benzoyl peroxide, as was the case in the last three experiments, it was possible to calculate the active oxygen remaining in the soUd and add it to that found in titrating the solution, and the sum of these two quantities determines whether active oxygen is lost during the hydrolysis. As will be seen from the last column, no such loss took place until four hundred and eight hours A\-ere reached, when the tube showed slight pressure on open- ing. The total quantity of active oxygen present in the benzoyl acetyl per- oxide therefore remains after the hydrolysis. Benzojd acetyl peroxide, at room temperatures, is slowly attacked by nor- mal caustic soda or caustic potash, giving off oxygen and depositing benzoyl peroxide, which latter is quite stable in the cold, a fact which was also ob- served by Nef.^ Baeyer demonstrated that the acid sodium salt of benzoper- acid also rapidly gives off oxygen and leaves sodium benzoate and benzoyl peroxide,- and to explain this change he assumes that three molecules of benzoperacid form one of benzoyl peroxide and one of an oxygenated ben- zoperacid, which latter then decomposes into oxygen and benzoic acid : aaco.o.oH iCeH,CO.O:.OH = CeHsCO.O.O.COCeHj + CeH^CO.O.O.O.H + H^O. :CeH,CO.O:.OH If this same reaction were to take place in aqueous solution, then, as there is no loss of oxygen in the above experiments, we would have to assume that the oxygenated benzoperacid is stable, except in the presence of alkalies, and that the primary result is the hydrolysis of benzoyl acetyl peroxide to benzoperacid and acetic acid. (1) CeH,C0.00.C0CH3 + HOH = C.H.CO.OO.H + CH3CO.OH. That the hydrolysis does not take place in the other direction, i.e., with the formation of acetoperacid and benzoic acid is proven by the following facts : 1. In the action of aniline or benzoyl acetyl peroxide, acetanilid but no benzanilid is obtained. 2. No benzoic acid can be isolated from the early products of hydrolysis of benzoyl acetyl peroxide by water. However, in the secondary hydrolysis of the peracid, benzoic acid results. ' Ann. Chem. (Liebig), CCXCVIII, p. 285. ' Ber. d. chem. Ges., XXXIII, p. 1570. 80 ON THE ORGANIC PEROXIDES. 3. Sodium ethylate, acting on benzoyl acetyl peroxide, gives the sodium salt of benzoperacid, but no sodium salt of acetoperacid can be detected. Positive proof of the absence of the latter, however, is wanting. The acetyl group is, therefore, separated as acetic acid by hydrolysis, the total active oxygen remaining with the benzoyl group. This conclusion, it will be seen, has been confirmed by the more recent studies of Clover and Richmond,^ who demonstrated the presence of benzoperacid in the hydrolyzed solution by its reaction with acetic anhydride, with which it yields benzoyl acetyl peroxide according to the equation : C„H,CO.OO.H + (CH3C0),0 = C.H^CO.OO.COCH, + CH^CO.OH. These workers were able to isolate only a very small amount of benzoic acid, which they ascribed to the secondary hydrolysis of benzoperacid. Thus, (3) C„H,CO.OO.H + HOH = C,H,CO.OH + HOOH. A small amount of acetoperacid is present in the hydrolyzed solution, but this they do not consider as resulting from a primary hydrolysis of benzoyl acetyl peroxide, but rather as the result of the reaction of benzoperacid with the unchanged peroxide, in which case benzoperoxide and acetoperacid form. This change may be represented by the equation : (2) C„H,CO.OO.H + C.H.CO.OO.COCH, = C.H^CO.OO.COC.H, + CH3.CO.OO.H. The formation of the insoluble precipitate of benzoperoxide in a solution of benzoyl acetyl peroxide is thus readily accounted for. In the presence of sulphuric acid, or even acetic acid, according to Clover and Richmond, this formation of benzoperoxide is prevented, and because of these facts, they incline to the belief that the peracid owes its reactivity to its ion rather than to the undissociated molecule. Summing up our present knowledge regarding the hydrolysis of benzoyl acetyl peroxide we may say that the first change consists in the formation of acetic acid and benzoperacid (equation 1). Some of the latter then reacts with the unchanged peroxide to form the insoluble precipitate of benzoper- oxide (equation 2). On prolonged standing the benzoperacid undergoes hydrolysis, yielding benzoic acid and hydrogen peroxide (equation 3). In other words, benzoyl acetyl peroxide on contact with water yields an insoluble precipitate of benzoperoxide, while the solution contains, chiefly as the active constituent, benzoperacid with an amount of acetoperacid, corresponding to that of the benzoperoxide formed, some acetic acid and when the solution is several days old, benzoic acid and hydrogen peroxide. The use of solutions of benzoyl acetyl peroxide as germicides rendered a ' Am. Chem. Journ., March, 1903. PAUL C. FREER AND FREDERICK 0. NOVY. 81 determination of the solubility in water necessary. The best method of arriv- ing at accurate results was to decompose the substance with N/20 potassium hydroxide, and to titrate the excess of alkali with N/20 hydrochloric acid. If a weighed quantity of benzoyl acetyl peroxide is treated with an excess of N/20 potassium hydroxide, and at once warmed on a water-bath until all benzoyl peroxide has disappeared, the results invariably show the neutrali- zation of more than the calculated quantity of caustic potash by one or two per cent. If, on the other hand, the peroxide is allowed to stand in the cold with the alkali for twenty-four hours, until the benzoyl acetyl peroxide is completely converted to benzoj'l peroxide, and the solution is then warmed on the water-bath, the results are very accurate and correspond exactly to those calculated for the production of two molecules of acid. 0.2263 gram substance treated with 70 c.c. N/20 potassium hydroxide required 10.9 c.c. N/20 hydrochloric acid to neutrahze. Calculated, 10.9 c.c. 0.2518 gram substance treated mth 70 c.c. N/20 potassium hydroxide required 14.3 c.c. N/20 hydrocliloric acid to neutrahze. Calculated, 14.1 c.c. Having estabhshed the accuracy of the method, we carried out the following experiments : The peroxide, purified by several recrystallizations, wag finely pulverized and then shaken with water at 25°. The amount of peroxide used did not affect the results, but the time of shaking did. It was necessary to shake for two or three minutes until a maximum solubility was reached. The increase in the amount in solution with length of shaking is due to the changes shown in the table given above, and as indicated, the rise is a slow one. Changes of temperature, of course, materially affect the result. Three-tenths gram substance was shaken with 110 c.c. water, at 25°, filtered quickly through asbestos with suction, and 50 c.c. removed and treated with 10 c.c. N/20 caustic potash. After warming on the water-bath for several hours the excess of caustic soda was titrated. N/20 tliiosulphate Time. Caustic potash used to determine Minutes. required. active oxygen. 1 5 7.2 7.0 2 10 7.3 7.3 3 15 7.4 7.4 4 20 7.5 7.5 5 30 7.7 7.7 6 40 8.2 8.2 The above table shows that for one molecule of peroxide in solution there is one atom of active oxygen, for the number of cubic centimeters of N/20 caustic potash and N /20 thiosulphate are the same, and two molecules of caustic potash in forming two molecules of salt from one of peroxide are equivalent to two molecules of thiosulphate in reducing iodine. This fact confirms the 82 ON THE ORGANIC PEROXIDES. view that the first change is hydrolysis with the production of benzoperacid,' as indicated above in equation 1. The sokibihty, after five minutes, is therefore the equivalent of 7.1 c.c. caustic potash ± one per cent, which corresponds to 0.639 gram per liter. This would represent 0.0568 gram active oxygen per liter. A one per cent hydrogen peroxide solution would contain 4.7 grams active oxygen per hter, or eighty times as much. Nevertheless, as will be shown below, these dilute solutions of benzoyl peroxide have a much greater germicidal action than even moderately concentrated ones of hydrogen peroxide. The reason for this is, a priori, not apparent, for in the one case we have a substance of the formula H.O.O.H and in the other CeH^CO.O.O.H. The germicidal action is, therefore, not due exclusively to the absolute quantity of active oxygen in unit space, so that some other cause, of which we cannot at present furnish an ex- planation, must be at the bottom of this remarkable fact. Hydroperoxide, when pure, is practically odorless, benzoperacid and acetoperacid, both of which have , the same intense germicidal action, have most penetrating odors, resembling hypochlorous acid. This latter fact suggests a study of hypochlorous acid from the standpoint of a peracid. Thus RO.OH the general formula for the acid peroxides would be represented by CI. OH, in which CI corresponds to RO. The acid can be looked upon as the hydrol}rtic product of chlor monoxide. All of these facts point to a difference between hydroperoxide and the organic acid peroxides which are germicides, which must be a subject of future investigation. The decomposition of benzoyl acetyl peroxide with sodium ethylate gives benzoperacid, and apparently no acetoperacid. Seven and a half grams of pure benzoyl acetyl peroxide were dissolved in 200 c.c. absolute ether and cooled to —12°. 0.94 gram of sodium dissolved in alcohol and cooled to — 12° was then slowly added, care being taken to allow no rise in temperature. The solution instantly forms a precipitate which almost solidifies the entire mass. The precipitate was allowed to stand for ten minutes, during which time it gave off some oxygen. Two hundred grams of ice were then added and shaken until all the solid went into solution. The ether was then separated and the remainder, after adding frozen dilute sulphuric acid, was extracted five times with ethyl chloride. On evaporating the ethyl chloride, a solid remained, having the intense odor of benzoperacid. This, when treated with sulphuric acid and alcohol, yielded only benzoic ester, and no acetic ester. The solid is, in all probability, a mixture of benzoperacid and benzoic acid, although the above experiment does not fully exclude the possibility of acetoperacid being present. A full confirmation of this view, however, has been supplied by the subsequent studies on hydrolysis. ' Electric conductivity measurements of solutions of benzoyl acetyl peroxide showed that rapid hydrolysis takes place. With 20,000 ohms' resistance the bridge readings fell from 850 to 340 in twelve minutes. PAUL C. FREER AND FREDERICK G. NOVY. 83 Acetyl Peroxide. The acet}--! peroxide used in our experiments was prepared as follows : Com- mercial barium dioxide was covered with water, cooled in a freezing-mixture, and then acetic anliydride and dilute hydrochloric acid were added alter- nately in small quantities, with frequent shaking, until all barium dioxide was used up. Care was taken not to have the liquid alkaline during the reaction. The mixture was finally acidified with hydrochloric acid, which facilitates the separation of acetyl peroxide ; the crystals of the latter body were filtered off, dried, and recrystallized from ligroin. Clover and Richmond,' emplo^-ing the following method, obtained nearly a theoretical yield of acetyl peroxide : Twenty-t\vo and a half grams of comparatively good barium dioxide (94 per cent) were covered with 30 c.c. of water and shaken for a few minutes with cooling. Twenty grams of acetic anhydride were added to 100 c.c. of ether and cooled to 0°. The suspen- sion of barium dioxide was then gradually added, with shaking, during the course of five minutes. The mixture was kept cooled to 0° and shaken for about fifteen minutes, when the dioxide had about disappeared. The ethereal layer was then removed by decantation, dried for several hours over calcium chloride, and the solvent then removed in a vacuum. The peroxide begins to crystallize from the cold, concentrated solution, and finally remains as a pure crystalline residue. It may be recrystallized from ligroin by cooling, in a freez- ing-mixture, a solution saturated at room temperature. Owing to the danger of spatter- ing, during the removal of the last portion of ether, it is advisable, on account of the ex- treme explosiveness of the substance, to dissolve the concentrated ethereal solution in ligroin, and allow the peroxide to crystallize in a freezing-mixture. The yield was 8.8 grams. As the peroxide is probably formed according to the equation 2(CH3CO),0 + BaO, = BaCCsHsOa)^ + C4HeO,, this amount represents 76 per cent of the total possible yield. It will be shown later that acetoperoxide on contact with water undergoes hydrolysis with the formation of acetoperacid and acetic acid. A secondary hydrolysis follows resulting in the cleavage of acetoperacid into acetic acid and hydrogen peroxide. Inasmuch as the acetoperacid represented the active germicidal constituent of the acetoperoxide solution, an attempt was made to isolate this substance, but unsuccessfully. For this purpose seven and one- tenth grams acetyl peroxide were dissolved in 50 c.c. of absolute ether and then 1.38 grams of sodium dissolved in the smallest amount of alcohol slowly added, the whole being well shaken and cooled to —20°. The pasty mass was then poured into two volumes of cold petroleum ether and filtered on a Briihl filter. When allowed to stand in vacuo, a white salt remained which exploded vio- lently when brought into a flame. The salt dissolved in water, but instantly began to give off oxygen. On acidifying with frozen dilute sulphuric acid, an oil was produced with a most intense odor of oxide of chlorine, which is char- acteristic of the peracids. The oil was extracted with ethyl chloride and the ' Am. Chem. Joum., March, 1903. 84 ON THE ORGANIC PEROXIDES. ■ latter evaporated. A small quantity of oil, with an indescribably intense odor, remained, but it was so volatile that in a few minutes all was gone. This substance was undoubtedly acetoperacid, but it decomposes so rapidly that its isolation in a pure state will be very difficult. Several attempts to obtain larger quantities gave the same result. The explosive sodium salt, mentioned above, soon loses this property, even when placed in a desiccator. Acetyl peroxide, on standing with water, gradually gives off oxygen, if the solution is concentrated. Weak solutions, hke those of benzoyl acetyl per- oxide, do not give off oxygen at the ordinary temperature. Five grams in 50 c.c. of water evolved oxygen during one month, and at the end of that period still contained active oxygen. A solution of acetyl peroxide, 1.2 grams in 50 c.c. of water, gave the theoretical amount (0.166 gram) of active oxygen on titration, even after standing but a few minutes, so that hydrolysis was practically complete. After seventeen days, however, this active oxygen had fallen to 0.03 grams. II. The Germicidal Action of the Peboxides. It is a well-recognized fact that ozone and hydrogen peroxide possess dis- tinct germicidal properties. Moreover, various metal peroxides, especially sodium, potassium, and calcium, have been studied with reference to their bactericidal value and have been shown to be fairly useful in that regard. On the other hand, the organic peroxides, which during the past few years have received considerable attention from the chemical standpoint, can be said to have been subjected to scarcely any study by the bacteriologist. And yet, on a ■priori grounds, it would seem as if. we might expect to find among these organic peroxides bodies which would possess germicidal properties greatly superior to those of hydrogen peroxide and of other well known dis- infectants. The considerations which led us to take this view were based upon the germicidal action of two agents, namely, metals and sunlight. The studies of Miller^ on mouth bacteria showed that copper amalgam possessed a strong antiseptic action. More unexpected was the observation that metallic gold exhibited a similar behavior. He was able to show that gold foil of different manufacture behaved in this respect differently. Thus, while one specimen showed a marked antiseptic action, others would show less, or none at all. Moreover, he found that an active preparation could be rendered inert by heating. This interesting observation by Miller was con- firmed by Behring,^ who found that not only dentist's gold foil, but also gold coins exerted at times this peculiar action when placed on gelatin plates which were inoculated with different pathogenic bacteria. Given an active gold ' Die Mikro-organismen der Mundhohle, 2te Aufl., 1892, pp. 277-281. ^ Zeitschr. f. Hygiene, IX, p. 432, 1890. PAUL C. FREER AND FREDERICK G. NOW. 85 preparation, the distance through which it would act — in other words, its zone of action — would vary with the species of bactcn-ia. Thus, with diphtheria bacilh, the width of the zone would be 3.5 cm., with anthrax bacilli 1.5 cm., with B. pyoeyaneus, 1 cm., with cholera, 0.4 cm., while glanders and typhoid bacilli were wholly uninfluoncod and would grow right up to the edge of the gold. Furthermore, while tin, lead, and iron wore inert, silver, mercury, cop- per, nickel, and zinc showed similar though slight antiseptic action. Moreover, while the sulphide of mercury was inert, the oxide and subchloride behaved the same as metallic mercury. Behring further showed that the action of these substances was not merely inhibitive, but that the organisms were actually destroyed. The action of metals, and more especially of copper, upon spirogyra was studied hx Naegeli.' His results, which were published after his death, were confirmed by Cramer,^ and also by Israel and Klingman,^ who demonstrated the germicidal action of " copper water" on several species of bacteria. Meade Bolton* hkewise made an extended study of the action of metals upon bacteria, and more recently Thiele and Wolffs have again subjected the question to an experimental study and arrived at the conclusion that silver, mercury, and copper were dissolved by the nutrient media, and that the solution was hastened under electrohiiic conditions. The explanation of this action of insoluble substances has not seemed satisfactory. Jliller looked upon the metallic surface as a means of effecting oxygen condensation. Behring also considered the possibility of gases being condensed upon the surface of the metal, but abandoned the idea, strange to say, because a given gold coin continued to exert the same antiseptic action when repeatedly transferred. Inasmuch as the gelatin beneath a piece of gold was not merely sterilized, but was so changed that after removal of the metal, if streaked with a fresh culture no growth resulted, Behring concluded that this was due to antiseptic products which were derived from the metal, and which persisted in the gelatin. In other words, through their waste products, minimal quantities of gold are dissolved and exert the peculiar antiseptic action observed. It is difficult, however, to see how bacteria will accomplish a solution of the gold before they have begun to multiply. Behring himself observed the total absence of even microscopic colonies in the sterile zone of action. It is, indeed, difficult to conceive how traces of a metal so minute as to escape quaUtative recognition should exert so marked a germicidal effect. • Neue Denkschriften d. aUg. schweizer. Ges. f. d. ges. Naturwissenschaften. Bd. XXXIII, Abth. I, 1893. 2 Ibid., p. 44. 2 Arch. f. pathol. Anat. CXLVII, p. 293, 1897. * Trans. Am. Assoc. Phys. IX, p. 174; XII, p. 488. 5 Arch. f. Hyg. XXXIV, p. 43, 1899. 86 ON THE ORGANIC PEROXIDES. That traces of metal are dissolved by the gelatin or by the bacterial products is quite possible, and according to Behring such traces are indicated by the darkening of the gelatin immediately around the metal. The change in color, however, was very marked with iron and lead, which, as already stated, exert no antiseptic action. It seemed to us that the germicidal action of metals could be explained in a different and more logical way. Unquestionably, metals do pass into solution, but it was quite possible that in addition the surface action of the metals led to the formation of hyperoxides, and if so, since hydrogen peroxide is itself a relatively feeble germicide, it was probable that more energetic organic peroxides were formed. Our studies upon the surface action of metals' demonstrate that hyperoxides are formed as a result of such action, and that the rate of formation varies within wide limits with the kind of metal or sur- face employed. These investigations have led to the preparation of a number of organic peroxides, some of which, we have been enabled to show, possess remarkable antiseptic properties. The demonstration of the formation of these bodies by the action of metals on nutrient media is wanting at present, but it is our intention to investigate this point at an early date. With the hypothesis that peroxides are formed by the action of metals, it becomes possible to ex- plain certain pecuharities observed by Behring. The peroxides, as is well known, are unstable bodies, and their decomposition can be brought about by some organic substances, as enzymes, and by alkalis, hydrogen sulphide, etc. If, for example, an organism produces an appreciable amount of either of the latter it is in a position to offset the action of the metal, and is, therefore, enabled to grow close to if not up to the very edge of such metal. The for- mation of other oxidizable bacterial products, as well as the reaction of the medium itself, must be taken undoubtedly into consideration. Again, the inertness of mercuric sulphide as compared with the oxide becomes easy of interpretation. The germicidal action of sunlight seemed to offer additional reason for the behef in the existence of very active organic peroxides. The interesting fact that solar rays in the presence of oxygen are capable of arresting putrefaction, and hence the development of bacteria, was first pointed out by Downes and Blunt.^ For some time this observation was apparently lost sight of, but eventually it received full confirmation at the hands of Janowski, Buchner, Ward, Esmarch, Kruse, and others. From the earhest investigation it was evident that oxygen was a necessary factor in this destructive process. More- over, it was clearly shown that the germicidal action of sunlight was due to rays of high refrangibility, and was not due to the increased heat consequent upon insolation. Obviously then the chemical rays of the sun exert a marked ' Am. Chem. Joum., 1902; see preceding part. ' Proc. Roy. Soc, XXVI, p. 488, 1877; XXVIII, p. 199, 1878. PAUL C. FREER AND FREDERICK G. NOVY. 87 germicidal effect, but a closer insight into this process was not gained until Richardson' showed that hydrogen peroxide was formed under the influence of sunlight. Experimenting with urine, as Downes and Blunt had done previous- ly, he showed that hydrogen peroxide was formed under precisely the same conditions as those which brought about the sterilization of the liquid, and consequently the inference seemed justifiable that this product of insolation was itself the active germicidal constituent. Dieudonn^^ promptly confirmed the fact that hydrogen peroxide was formed in the presence of sunlight and air. Agar plates exposed to intense sunlight gave the iodine reaction in a few minutes, while several hours were necessary in diffuse light to produce the same result. Gelatin required considerably more time than agar to develop the reac- tion, while tap water, especially in presence of ether, gave an intense and prompt reaction. When oxygen was removed either by hydrogen or by absorption, no peroxide was formed, and under such conditions considerably more time was necessary to destroy colon bacilh and anthrax spores than was required when air was present. Kruse,' while recognizing the formation of hydrogen peroxide by insolation, is not inclined to consider this substance as the germicidal agent, on the ground that it is formed more abundantly in water than in bouillon, and yet such insolated water is much less germicidal than an equally insolated beef tea. While the reaction of the medium seems to be without any special influence, it would appear that the presence of pepton or of other nitrogenous constitu- ents was necessary to the formation of the germicidal constituent. While both Richardson and Dieudonne state that the hydrogen peroxide formed by sunlight soon disappears when the objects are kept in the dark, Kruse points out that the germicidal substances produced by light are not destroyed by heating for two hours at 100°. It is evident that these observations will have to be extended before hydrogen peroxide can be credited as the bacteri- cidal substance formed during insolation. Accurate data as to the amount of hydrogen peroxide formed in the urine or other liquids under these conditions are necessary, and such work is now being carried on by us. The small amount of hydrogen peroxide present in insolated urine does not satisfactorily explain the fact that such urine can destroy colon bacilli within one minute, as we have been able to demonstrate. The formation of nitrites by the action of sunlight, arc-light, and by the induction spark is also due to the ultra-violet rays. Obviously the presence of nitrous acid complicates the test for peroxides, and it is more than likely that some observers have ascribed results as due to hydrogen peroxide when the reaction was really due to a nitrite. Unquestionably, hydrogen peroxide may by itself exert a distinct germi- 1 Joum. Chem. Soc, LXIII, p. 1109, 1893. ^ Arbeiten a. d. k. Gesundheitsamtes, IX, p. 537, 1894. = Zeitschr. f. Hygiene, XIX, p. 313, 1895. 88 ON THE ORGANIC PEROXIDES. cidal action, but it has seemed to us probable that more active organic peroxides or persulphates were formed by the action of sunlight in the presence of air. An observation of Berthelot^ may be pertinent in this connection. By passing dry ozonized oxygen through anhydrous ether he obtained an ethyl peroxide, which was found to be violently explosive on heating. It is a well-known fact that old ether contains hydrogen peroxide, and this is especially true if water is present. Nevertheless, as Richardson^ pointed out, hydrogen per- oxide will form in anhydrous ether when exposed to sunlight and air. The explosive properties of ether during evaporation, as sometimes observed, find a satisfactory explanation in the formation of peroxides. In the following pages the results of the bacteriological study of six peroxides are given. Two of these bodies, namely, acetone and dibenzoyl peroxides, as will be seen, are practically indifferent, whereas the diacetyl, benzoyl hydrogen, phthalmonoperacid and benzoyl acetyl peroxides possess remark- ably interesting germicidal properties. In order to avoid repetition, it may be well to state that the method of testing, unless otherwise indicated, consisted in dissolving the peroxide under examination in sterile distilled water. Frequent agitation for an hour was resorted to in order to effect saturation or solution. The solution was then filtered through sterile uncharred paper, and portions of 5 c.c. were transferred by means of sterile pipettes to sterile test tubes. Agar cultures of the organ- isms were employed and were usually developed at 37° for about twenty- four hours. A loopful of the culture was then transferred to the tube con- taining the peroxide, and thoroughly whipped for about three-quarters of a minute. In order to remove aggregations or masses of bacteria that might persist in the suspension, the latter was now filtered through a sterile glass wool filter and the filtrate was received into a sterile test tube or Esmarch dish. Owing to the ease with which charred paper or cotton is oxidized, it is not advisable to filter ^ peroxide solutions through such material. At definite intervals a large loopful (2 mm. loop) of the filtrate was transferred to a bouiUon tube and the entire set was eventually incubated for several days at 37°. At first the inoculations were made into tubes, each containing 10 c.c. of bouillon in order to minimize the effect of any peroxide carried over. When it was found, however, that the peroxide solution was destroyed or rendered inactive by bouillon, and that, therefore, there was no danger of inhi- bition, as with mercury salts and other antiseptics, the amount of bouillon in the tubes was reduced to about 5 c.c. Control tubes, it may be added, were made in all cases, and served for comparison of growths. SteriUzation or 1 Annal. de Chim. et da Phys., (5), XXVII, p. 229, 1882. Comp. Rend. Acad.- Sc, CVIII, p. 543, 1889. ' Joum. Chem. Soc, LIX. p. 51, 1891 ; LXIX, p. 1352, 1896. PAUL C. FREER AND FREDERICK G. NOVY. 89 failure to grow is indicated in the tables by — ; whereas typical growth is marked +. Acetone Peroxide. Two peroxides of acetone arc known. One of those, possessing a melting- point of 97° (90-94°) was prepared by AVolffenstein' by allowing acetone to act upon a concentrated solution of hj^drogen pen-oxide. Baeyer and Villiger^ prepared a different peroxide with a melting-point of 132° to 133° by the action of Caro's reagent on acetone. According to these investigators, the peroxide melting at 97° has the formula (CjH.OJ,, while that molting at 132° to 133° is (CjHjOj)^. The structure is indicated by the formula: CH,^ 0-0^ nn. The latter peroxide was employed for the following experiment : The pure white crystals of diacetone diperoxide, melting at 132°, were agi- tated for one hour with distilled water. 10 c.c. of the filtrate on evaporation gave a residue of only 0.0003 gram, which would correspond to a solubility of 1 : 33,000. The fact that the solution was tasteless and gave no reaction •with potassium permanganate or with potassium iodide and starch indicated a probable absence of germicidal power. This conclusion was confirmed by experiment, as seen from the following table, in which, as already stated, + indicates growth. Table I. Minutes. 1. 3. 5. 10. B. coli . . + + + + + + + + + + + + + + + + B. typhosus . . + B. icteroides + B. pestis + V. cholerse Asiat. . . + Fresh solutions of this acetone peroxide possess therefore no germicidal action. It is quite possible that on standing, as a result of hydrolysis, some such action may be manifested. Benzoperoxide. Dibenzoyl, or benzoperoxide, CjHjCO.O.O.CO.CjH^, was first prepared by Brodie.^ It can be easily prepared according to the method of Pechmann and Vanino,* by shaking hydrogen peroxide with benzoyl chloride in the presence of sodium hydrate. The odorless white crystals melt at 106° to 108° and are extremely insoluble in water. The water solution, such as it is, has no oxidizing action on potassium iodide, nor does it affect permanganate. ' Ber. d. Chem. Ges., XXVIII, p. 2265, 1895. 2 Ber. d. Chem. Ges., XXXIII, p. 124, 858, 1900. ^ Liebig's Annal. Suppl., 1863, p. 316. * Ber. d. chem. Ges., XXVII, p. 1510, 1894. 90 ON THE ORGANIC PEROXIDES. These facts point to its inertness with respect to bacteria. It will be presently shown that solutions of benzoyl acetyl peroxide deposit benzoperoxide as a result of the interaction of the products of hydrolysis. The germicidal properties of this substance have already been tested twice by other investigators, and as will be seen, our own results agree with those heretofore obtained. Frey and Vanino,* after inoculating agar tubes with different organisms, sprinkled the surface of the agar with crystals of peroxide. The growth was unrestricted. Silk threads or bits of filter-paper soaked in bouillon cultures were exposed to the action of aqueous suspensions of the peroxide, with the result that diphtheria bacilli and those of green pus were killed in fifteen minutes, while the colon bacilli and streptococci required fifty and ninety minutes, respectively. Nencki and Zaleski^ tested benzoperoxide with special reference to its use as an intestinal antiseptic. They found that dogs could tolerate relatively large doses (5 to 10 g.) and that while about a half would pass the intestines unchanged, the remainder was hydrolyzed by the pancreatic enzymes, and appeared in the urine as hippuric acid. Inasmuch as the conjugate sulphates, before and after administration of benzoperoxide were unchanged, they con- cluded that this substance was of no, value. The same conclusion was reached with reference to phthalyl superoxide. In our hands the aqueous solutions of dibenzoyl peroxide showed themselves to be as inert as those of acetone peroxide. In one experiment on mouth dis- infection the number of organisms present in the mouth was practically the same before and after washing with this peroxide. The result in no wise differed from that obtained in control tubes with sterile water. In another experiment crystals of dibenzoyl peroxide were allowed to remain in excess in the aqueous solution. Every day a portion of the liquid was filtered and tested, and the result, even at the end of seven days, was wholly negative. Table II shows the result of the final trial with a solution of benzoperoxide which has stood in contact with crystals for one week. The addition of 0.5 per cent acetic acid in no wise increases the efficiency of the peroxide, for while cholera and diphtheria bacilli are destroyed, the same result is accom- plished in control tests with acetic acid of like concentration. Table II. Minutes. 3. 5. 10. 15. B. anthracis + + + + B. diphtherise + + + + B. typhosus + -I- + + B. icteroides + + -|- -i- V. cholerae Asiat + + + + ' Pharm. CentralhaUe, XL, p. 209, 1899. 2 Zeitschr. f. physiol. Chem., XXVII, p. 487, 1899. PAUL C. FREER AND FREDERICK G. NOVY. 91 It will be seen from the above results that the germicidal value of benzo- peroxide is much less than what Fre^nd Vanino found. In fact, this peroxide should be considered as wholly inei^partly because of its extreme insolubility, and especially because of its structure. As will be seen later, an organic per- oxide is germicidal when it contains one hydroxyl group, or when it readily hydrolyzes into such a compound. The high molecular weight is undoubtedly a factor in preventing the hydrolysis of benzoperoxide. ACETOPEROXIDE. Acetoperoxide or diacetyl peroxide, CH3CO.O.O.CO.CH3, was first prepared according to the method of Nef.' The neutral peroxide was dissolved to saturation in distilled water. The water on prolonged contact with excess of acetoperoxide became strongly acid, as a result of the hydrolysis of the latter. This change is indicated by the fact that on titration with decinormal permanganate, while the fresh solution (10 c.c.) reduced but a fraction of a cubic centimeter of the reagent, eventually, in the course of several weeks, it reduced more than twice its volume. An understanding of the hydrolytic change of diacetyl peroxide is necessary in order to explain the germicidal action of this body. This substance, like benzo peroxide or benzoyl acetyl peroxide, itself possesses no oxidizing action. In other words, it does not react, or but very slowly, with potassium iodide, ■R-ith anilin, or with indigo. The behavior of acetoperoxide with potassium iodide has been studied by Clover and Richmond,^ who reach the conclusion that it possesses an oxidizing action even before hydrolysis takes place. In the absence of marked action, it follows that acetoperoxide, as well as the two above mentioned, are not likely to be active germicides. Solutions of these peroxides, excepting benzoperoxide, on standing, promptly undergo hydroly- sis, which, however, is not completed for some hours, and this change is at once indicated by the liberation of iodine from potassium iodide. The hydrolysis of acetoperoxide may be understood from the equation: CH3CO.O.O.COCH3 + HOH = CH^.CO.O.O^H + CH^.CO.OH. In other words, acetic acid and acetyl hydrogen peroxide are formed, and it is to the latter that are due the oxiiiizing and germicidal properties of so-called solutions of diacetyl peroxide. As already indicated, we have endeavored to prepare acetyl hydrogen peroxide, but our efforts in that direction were not successful. Clover and Richmond,^ however, have conclusively shown that acetoper- 1 Ann. d. Chem., CCXCVIII, p. 287, 1897. ' Locus citus. " Am. Chem. Joura., March, 1903. 92 ON THE ORGANIC PEROXIDES. acid is formed during the hydrolysis of the peroxide. They found that the hydrolyzed solution, which contained no unchanged acetoperoxide, on treat- ment with acetic anhydride, liberated less iodine from potassium iodide than before such treatment, indicating the formation of acetoperoxide, which, as already stated, does not itself possess an oxidizing action. Moreover, they succeeded in isolating in crystalline form the acetoperoxide thus formed. They express the reaction thus : CH3.CO.OO.H + (CH3C0),0 = CH3CO.OO.COCH3 + CH3CO.OH. Additional evidence of the presence of acetoperacid was obtained by treating the hydrolyzed solution with benzoyl chloride, whereby, apparently, benzoyl acetyl peroxide and benzoperoxide were obtained. The formation of the latter may be due to the action of hydrogen peroxide on benzoyl chloride. The study of the germicidal action of hydrolyzed solutions of acetoperoxide showed that the active constituent (acetoperacid) disappeared on standing. As in the case of Caro's acid, it was possible that the acetoperacid hydrolyzed into acetic acid and hydrogen peroxide. Clover and Richmond have since shown that this change does take place and that at the end of a month nearly the entire active oxygen content of the solution is represented by hydro- peroxide. The hydrolysis of acetoperacid is shown thus: CH3CO.OO.H + HOH = CH3CO.OH + H.OO.H. The first experiment with acetoperoxide was made with the above-mentioned solution, after it had stood two days in the ice-box. A preliminary trial with purely vegetating forms of bacteria, such as V. cholerse, B. typhosus, B. icteroides, B. anthracis, M. tetragenus, and S. pyogenes aureus, showed com- plete destruction of these organisms within one minute. This result is all the more remarkable when it is compared with the total inefficiency of the acetone, and dibenzoyl peroxides, as shown in tables I and II. From this experiment it was evident that the germicidal value of acetoperoxide was very great, and it was accordingly given a more severe test, as shown in table III. In this ex- periment the suspensions were made in the usual way, with rich, spore-bearing material. For comparison, tests were made at the same time and in like manner with a five per cent carbolic acid solution. The acetoperoxide solu- tion, it should be stated, had stood four days over an excess of the peroxide. 10 c.c. of the solution reduced 2.45 c.c. of N/10 KMnO,. This would indicate a solubility of about 1-690 mth approximately 0.02 per cent of active oxygen. Clover and Richmond have since shown that the solubility of acetoperacid is about 54.2 g. per liter. PAUL C. FREER AND FREDERICK G. NOVY. 93 Table III. Acetoperoxide. 5 per cent. Phenol. Minutes. 1. 3. 5. 10. 15, 30. 1. 3. 5. 10. 16, 30. Spores, B. anthrac " B. mes. viilg. , . . " B. siibtilis " B. megateriuni . , - - - - - - + + + + + + + + + + + + + + + + + + + + + + + + It will be seen from this table that acetoperoxide easily ranks as one of the most powerful germicides, destroying, as it does, the most resistant spores within one minute. Carbohc acid, on the other hand, in a five per cent solu- tion is incapable of accompUshing the same result within thirty minutes. At the end of four months, when all the excess of acetoperoxide had hydrolyzed, and the resulting acetoperacid had largely decomposed, the solution was still active, although not to the same extent as when fresh. A second specimen of acetoperoxide was prepared according to the method described on page 83. This time the peroxide was obtained in the pure crys- talline form. A saturated solution was prepared by shaking the white per- oxide crystals with 250 c.c. of sterile water for two hours, after which the solu- tion was filtered through sterile filter-paper. Portions of 5 c.c. of the filtrate were then distributed to several sterile tubes, which were then inoculated and treated according to the directions already given. The fresh peroxide solution possessed a well-marked odor, and liberated iodine instantly from potassium iodide. The following table gives the results obtained with this freshly prepared solution of acetoperoxide. The amount of acetoperoxide (hydrolyzed) in solution was determined by means of potassium iodide, and the liberated iodine was estimated by means of N/10 Na^S^Og. Thus, 23.5 c.c. of the solution required 6.623 c.c. of N/10 Na.S^Oj, which represents 0.0226 per cent of active oxygen, or 0.166 per cent of acetoperoxide. According to this the solubility of acetoperoxide is 1-602. Table IV. Acetoperoxide (1-600), 0.0226 per cent active oxygen. Minutes. 1. 3. 5. 10. IS. 30. 60. B. pyocyaneus B. coH B. tjrphosus B. diphtherias + + + - - - - - : Vib. cholerae S. pyogenes aureus Strept. pyogenes Spores, Anthrax B. ..... - Spores, Hay B Spores, Potato B 94 ON THE ORGANIC PEROXIDES. Control tubes inoculated with the organisms gave typical growths. It has already been stated that the solution before testing gave a prompt reaction with potassium iodide. In each instance above, after the sixty minutes' transplantation was made, the suspension was tested for peroxide with potas- sium iodide. The iodine reaction was positive in all. It is customary to ascribe the germicidal action of ozone and of hydrogen peroxide to the active oxygen which these bodies contain. A comparison of acetoperoxide with hydrogen peroxide will reveal the interesting fact that the germicidal action of the former is to a large extent independent of active oxygen, inasmuch as a hydrogen peroxide solution of like contents in active oxygen is enormously inferior as a disinfectant. This fact will be apparent on comparison of table IV with table V. Table V. Hydrogen peroxide (1-1000), 0.05 per cent active oxygen. Minutes. B. pvocyaneus B. coli B. typhosus B. diphtheriae Vib. cholerae S. pyo. aureus Strept. pyogenes . . Spores, Anthrax B . Spores, Hay B. . . . Spores, Potato B. . . 1. 3. 5. 10. IS. 30. -1- + + + + — + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + — + + + + + + + + + — — — + + + + + + + + + The suspensions were tested after the last transplantation with potassium iodide and ferrous sulphate. A good reaction indicating the presence of HjOj was sustained in all but two. The suspensions containing aureus and diphtheria germs did not react, showing that all the hydrogen peroxide had been decomposed. The hydrogen peroxide solution used above represents 0.1 per cent HjO„ and although it contains twice as much active oxygen as the 0.16 per cent solution of acetoperoxide, it is still greatly inferior as a germicide. As a matter of fact, a 0.5 per cent solution of HjOj will show no difference up to ten minutes' exposure from the results tabulated above. Indeed, a one to two per cent solution of HjOj is necessary to destroy the vegetating germs within one to three minutes, while a solution double this strength would be necessary to destroy the spore material in the same time. In other words, a two per cent solution of H2O2 which can hardly do the same work as the 0.16 per cent solu- tion of acetoperoxide, contains nevertheless more than forty times as much active oxygen. Obviously a different explanation of the germicidal action of acetoperoxide is necessary than such as is afforded by nascent oxygen. This subject will be again discussed under benzoyl acetyl peroxide. PAUL C. FREER AND FREDERICK G. NOVY. 95 In table VI are given the results obtained with nieroui'io chloride and car- bolic acid. These solutions were tested in exactly the same way as were those of the peroxides. Control tubes in all cases gave tlu^ usual typical growths. Table VL Mercuric chloride, 0.1 per cent. Plienol, 5 per cent. Minutes. 1. 5. 15. 30. 60. 1. 5. 15. 30. 60. B. pyocyaneus - - - - - + + + + + + + + + + + + B. typhosus Spores, Anthrax B. + Spores, Hav B. + Spores, Potato B. + In view of the method employed, it is not unlikely that the results with mercuric chloride are too favorable to this agent. No attempt was made to render inert the mercury that was carried over in the operation of transplanting and hence an antiseptic action is not excluded. From the above experiments with acetoperoxide, we are justified in draw- ing the conclusion that this substance possesses remarkable germicidal prop- erties. Measured bj' the amount of active oxygen given off, it is many times more active than hydrogen peroxide. In aqueous solution and acting on aque- ous suspensions, it may be said to be at least as active as 1-1000 mercuric chloride, and it is certainly more active than a five per cent solution of carbolic acid. There is httle hope, however, of utilizing this agent in the pure crystal- hne form as a practical germicide. Its instability and extreme explosiveness render it too dangerous. Recognizing these facts, we have endeavored to obtain an equally active, stable, and less or non-explosive peroxide. Such a body might indeed prove itself to be a valuable addition to the list of known germicides. Benzoperacid. The existence of benzoyl hydrogen peroxide was indicated on purely theo- retical grounds by Bodlsender,' but it is to Baeyer and ViUiger^ that we owe not only the method for its preparation, but also the fundamental facts re- garding its formation. According to the nomenclature proposed by these investigators, benzoyl hydrogen peroxide, C5H5CO.O.O.H, is designated as benzoperacid. It is hydrogen peroxide, HO. OH, in which one hydrogen atom is replaced by the benzoyl group. This peroxide possesses extraordinary oxidiz- ing power, as seen in the fact that iodine is instantly liberated from potassium iodide, and that aniUn is oxidized to nitrobenzol. In this respect, then, it differs enormously from dibenzoyl peroxide, which is wholly inert. When the latter does react, if at all, it is because of hydrolysis with the formation of ' Ahrens' Vortrage, III, p. 471. 2 Ber. d. chem. Ges., XXXIII, p. 858, 1569 (1900). 96 ON THE ORGANIC PEROXIDES. benzoperacid. The latter possesses an odor not unlike that of chlorinated Ume. As an oxidizing agent and a germicide it may be compared with the latter, whereas, the permanent dibenzoyl peroxide is odorless, and as shown heretofore, is wholly indifferent. The benzoperacid necessary for the bacteriological examinations was prepared by hydrolyzing benzoyl acetyl peroxide by means of sodium ethylate. It may also be prepared, according to Qover and Richmond,^ by treating benzoic anhydride in the presence of potassium hydrate, with hydrogen per- oxide. They found that, like other peracids, its solution slowly hydrolyzes to benzoic acid and hydrogen peroxide. The yellowish white crystals, possessing a strong odor of chlorinated Hme, were added to 200 c.c. of sterile water. After shaking the suspension for one hour the undissolved crystals were removed by filtration through sterile paper. The filtrate was then tested in the usual way. As in the case of acetoperoxide, it gave a strong iodine reaction, even after remaining in contact with the bacterial suspensions for one hour. The strength of this solution was not determined, but its germicidal action was tested on B. pyocyaneus, B. typhosus, S. p. aureus, V. cholera, and spores of anthrax. Suspensions of these organisms were sterihzed in less than one minute of contact. Benzo- peracid, therefore, possesses powerful germicidal properties, such as the in- tense oxidizing action would presuppose. For the purpose of comparison a fresh solution of benzoperacid was prepared of such strength that it contained 0.0046 per cent of active oxygen. This cor- responds to 0.0397 per cent solution of benzoperacid (1: 2500). The results obtained with this solution are shown in table VII. Table VII. Benzoperacid (1 : 2500), 0.0046 per cent active oxygen. Minutes. 1. 3. 5. 10. 15. 30. 60. B pyocyaneus. + + + + - - - - _ B. coli , . B. typhosus ; , B. diphtherias . . .__ Vib. cholera . . . ."" - Strept. pyogenes Spores, Anthrax B - Spores, Hay B Spores. Potato B. - Control tubes inoculated with the organisms gave typical growths. The iodine reaction was given promptly by all of the suspensions at the end of sixty minutes, showing that the reagent was still present. ' Locus citus. PAUL C. FREER AND FREDERICK G. NOVY. 97 On comparing the results shown in table VII with those in table IV, it will be seen that the benzoperacid solution (1 to 2500) containing 0.0046 per cent active oxygen is as active as the acetoperoxide solution (1 to 600) contain- ing five times as much active oxygen. A direct comparison with solutions of equal strength was not made, and it is highly probable that if carried out the two solutions would be found to be equally active. The disparity between the germicidal action and actiA'e oxygen contents will be seen on comparison with hydrogen peroxide (1 to 1000, see table \') containing more than ten times as much active oxygen, and hence a corresponding greater number of molecules. While this latter solution shows a very feeble germicidal action, that of benzo- peracid is extremely active. In order to secure approximately the same ger- micidal action \Adth hydrogen peroxide it would be necessary to employ at least a one per cent solution. Such a solution would contain 0.47 per cent active oxygen, or 100 times as much as that contained in the benzoperacid solution used above. Obviously, if we assume that the active oxygen in both instances is the same, in so far as its potential energy is concerned, then it follows that this element cannot be considered as the germicidal factor in either hydrogen peroxide or benzoperacid. Phthal Monopeeacid. PhthaUc acid, inasmuch as it is a dicarboxyl compound, may give rise to four peroxides. Thus the simplest member of the group would be phthalyl monohydrogen peroxide, or, following Baeyer's nomenclature, phthalmono- /CO.O.OH peracid. It possesses the formula C.H, ( and corresponds to benzo- \C0.0H peracid. Just as the latter unites with another benzoyl group to form the dibenzoyl peroxide, so the former may unite with the phthalyl radical and vield peroxidphthalic acid. ^CO. 0. O. CO. ^CO.OH CO.OH-^ Again, both carboxyl groups may receive active oxygen, in which case /CO.O.OH terephthalic diperacid, C,H, < , results. The union of the latter with XCO.O.OH the phthalyl group would yield phthal diperoxide, ^CO.O.O.CO' " ^CO.O.O.CO^ The latter is probably identical with the phthalyl superoxide of Pechmann and Vanino,^ and presumably it is insoluble and as inactive as benzoperoxide. 1 Ber. d. chem. Ges., XXVII, p. 1510, 1894. 98 ON THE ORGANIC PEROXIDES. The phthalmonoperacid was prepared by Mr. Richmond according to the method of Baeyer and Vilhger.^ No attempt was made to obtain it absolutely pure, and the product as used, therefore, contained about twenty-eight per cent 'of phthalic acid. The high melting-point of this peroxide, about 110°, made it very desirable to test its germicidal value. For this purpose 1.1 g. of the peroxide was added to 250 c.c. of sterile water. After shaking vigorously for one hour 50 c.c. of the liquid were filtered off. The remainder of the liquid in contact with undissolved peroxide was set aside in the dark. At the end of twenty-four hours, two, four, and six weeks, respec- tively, portions of 50 c.c. were filtered off and tested. The solution on treatment with potassium iodide instantly hberates iodine. In portions of 5 c.c. of the respective filtrates, the active oxygen was thus determined. The iodine set free was titrated with N/10 sodium hyposul- phite. The results obtained showed that the active oxygen was gradually disappearing under these conditions. Thus : 5 c.c. of the 1 hour solution required 5 c.c. " 1 day 5 c.c. " 13 " 5 c.c. " 28 " 5 CO. " 42 " At the same time that the active oxygen was determined as above, disin- fection experiments were made in each case with aqueous filtered suspensions of bacteria. In the following table only the results obtained with the one- hour and the six-weeks solution are given, since the intermediate trials with one, thirteen, and twenty-eight days agree exactly with one another and with those tabulated. It may be added that the crystals when exposed to diffuse, and even to direct, sunlight for three weeks (October) yielded solutions which gave results the same as those with the freshly prepared peroxide. Table VIII. 1.3 c.c. of the N/10 sol 1.1 c.c. 1.0 c.c. ti tt 0.9 c.c. cc a 0.65 c.c. ti l( 1 hour solution (1-422); 5 c.c. = 1.3 c.c. N/10 Na,S2O3=0.021% active oxygen. 6 weeks solution (1-844); 5 c.c. = 0.65 c.c. N/10 Na2SsO3 = 0.01% active oxygen. Minutes. 1. 3. 5. 10. 15. 30. 60. 1. 3. 5. 10. 15. 30. 60. B. coli Vib. cholera . . Spores, Anthrax Spores, Hay B . Spores, Potato. + + + + + + + + + + + + + + + + + + + +' + + + + + + + + Controls were made in all cases and gave typical growths. It will be seen from the above that solutions containing 0.01 per cent active oxygen will kill colon and cholera germs and spores of anthrax, but will not destroy the spores of hay and potato bacilli. The latter are not even destroyed by solutions of twice this strength. 1 Ber. d. chem. Ges., XXXIV, p. 763, 1901. PAUL C. FREER AND FREDERICK G. NOVY, 99 A determination of the thermal death-point of the spores employed in the above experiments was made. The suspensions were drawn up in capillary- tubes which were then sealed and immersed in boiling water. An exposure of one minute sufficed to destroy the anthrax and hay spores, while five minutes' heating did not kill the potato baciUi. It is evident from the above that phthalmonoperacid is an active germicide, since it promptly destroys vegetating bacteria and weak spores. On com- parison with acetoperoxide (table IV) of about the same oxygen strength, it will be seen to be less effective. This is also true when compared with the much weaker benzoperacid solution (Table VII) or with the benzoyl acetyl peroxide solutions (IX, X). On the other hand, it is more effective than hydrogen peroxide solutions of greater oxygen contents (Table V). Benzoyl Acetyl Peroxide. This compound was first obtained by Nef by exposing a mixture of benz- aJdehyde and acetic anhydride, soaked up in sand, to the action of air. The change in this case we have shown to be one of surface action, and by replacing the sand with paper or muslin cloth we have been able to prepare this peroxide in large quantities (page 76). The change which takes place is one of auto- oxidation, and occurs, according to Baeyer and Villiger,' in two phases, repre- sented by the equations : (1) CeH^CHO + 0, = CeHjCO.O.OH. Benzoperacid. (2) CeHjCO.O.OH + 0(COCH3)2 = CeH^CO.O.O.COCHs + CH^CO.OH. Benzoyl Acetyl Peroxide. If no acetic anhydride is present, or if present and unfavorable tempera- ture conditions exist, the benzoperacid may react with benzaldehyde, in which case it yields benzoic acid. Thus, CeH^CO.O.OH + CeH^COH = 2C,H5CO.OH. It wiU be seen from the above that benzoyl acetyl peroxide is hydrogen peroxide (H.O.O.H) in which the hydrogen atoms are replaced by acetyl and benzoyl groups. In the purest condition it crystallizes in snowy white plates, which melt at 40° to 41°. The melting-point according to Nef is 37° to 38°, while Baeyer and Villiger place it at 38° to 39°. It is noteworthy that benzoyl acetyl peroxide itself possesses no oxidizing action, and as a matter of fact, is a comparatively inert substance. This is seen in the fact that it does not liberate iodine from potassium iodide. This chemical inertness is paralleled by the absence of any germicidal action. While benzoyl acetyl peroxide as such possesses no oxidizing or germicidal power, its solutions in water very promptly develop such properties. On shaking > Ber. d. chem. Ges., XXXIII, p. 1582, 1900. 100 ON THE ORGANIC PEROXIDES. with water the crystals may be brought into solution, though the amount dissolved at first is extremely small. The perfectly fresh solution is indifferent, or nearly so, to potassium iodide and to bacteria, but in a few minutes it acquires the power to liberate iodine and to destroy bacteria. This conversion of an inactive into an active substance is due to a hydro- Ijrtic change whereby the benzoyl acetyl peroxide is changed into more soluble and more energetic products. The hydrolysis, which takes place as already pointed out (page 79), may b&indicated by the equation : (1) CeHsCO.O.O.COCHg + HOH = CeHsCO.O.OH + CH3CO.OH. Benizopbhacid. Acetic Acid. This hydrolyzed solution is not permanent, since a secondary reaction gradually sets in between the benzoperacid and the unchanged benzoyl acetyl peroxide, resulting in the precipitation of benzoperoxide, which change may be thus expressed : (2) CeHsCO.OO.H + CeHsCO.OO.COCH, = CeHsCO.O.O.COCeHs Benzoperoxide. + CH3CO.OO.H. The fully hydrolyzed solution would, therefore, contain, as the chief active constituent, benzoperacid with some acetoperacid, the amount of the latter corresponding to the amount of benzoperoxide formed. The relative amounts of these two peracids will necessarily vary with the conditions under which solution takes place. Thus, in the presence of an excess of b'enzoyl acetyl peroxide it is conceivable that the benzoperacid is used up, as fast as it is formed, to produce benzoperoxide (Equation 2), and as a result, acetoperacid wovild predominate. However, the production of benzoperoxide is not instan- taneous, and hence the original peroxide, to a large extent, dissolves and hydrolyzes to benzoperacid. As pointed out by Clover and Richmond a solution of benzoperacid on prolonged standing undergoes further hydrolj^sis into benzoic acid and hydro- gen peroxide. Thus, (3) CeHsCO.OO.H + HOH = CeHsCO.OH + HOOH. From this it will be seen that very old solutions of benzoyl acetyl peroxide may lose their germicidal power to a very considerable extent. It is obvious, therefore that benzoyl acetyl peroxide solutions owe their activity to the extremely energetic benzoperacid (and acetoperacid) which results by hydrolysis. Such hydrolyzed solutions, as will be presently shown, are in every respect as germicidal as the similar solutions of diacetyl peroxide and the pure solutions of benzoperacid, the behavior of which has already been described. Inasmuch as the preparation of the pure acetoperacid, or benzo- peracid, is difficult, and the compounds themselves are relatively unstable, it is a distinct advantage, so far as the practical use in medicine is concerned, to PAUL C. FREER AND FREDERICK G. NOVY. 101 be able to handle a definite chemical substance, such as benzoyl acetyl perox- ide, which, although itself inert, nevertheless readily by hydrolysis yields the active peracids. As already indicated, benzoyl acetyl peroxide melts at, or slightly above, the temperature of the body. Like many other peroxides, it can be decom- posed with explosive violence, which, however, is by no means as energetic as that of diacetyl peroxide. This property is easily demonstrated by striking a minute crystal on an anvil with a hammer, or by rubbing it with a pestle in a mortar. When added to water which has been heated to 80°, and above, an explosive decomposition is likely to result. In view of these facts it is de- sirable that the benzoyl acetyl peroxide be kept in a cool place, in well- stoppered vials, and that solutions should be prepared with ordinary or slightly tepid water. It has been mentioned that the solubility of benzoyl acetyl peroxide itself is very slight. The direct determination of the dissolved substance by evapora- tion in vacuo of a definite volume of the filtered liquid gave at first results which indicated a solubility of about one in 30,000. This method, however, was soon found to be wholly unreliable, owing to the volatilization of the hydrolytic products. Two other methods were used to determine the solubility, and the results thus obtained were perfectly satisfactory. One procedure consisted in decomposing the peroxide with excess of standard alkali solution and then determining by titration the amount of uncombined alkali. The other pro- cedure, which is equally good, and more convenient, is to determine the amount of iodine hberated on addition of potassium iodide. This can be done by means of a decinormal solution of sodium thiosulphate. The result obviously repre- sents the active oxygen in the hydrolyzed product rather than the amount of benzoyl acetyl peroxide proper. From what has been said it is evident that when an excess of crystals of benzoyl acetyl peroxide is shaken vigorously with water, a very small amount dissolves, forming a saturated solution. The dissolved portion promptly hydrolyzes, yielding the easily soluble products mentioned above. This change enables a fresh portion of the peroxide to dissolve, and this in turn un- dergoes hydrolysis. The solution of the benzoyl acetyl peroxide and subse- quent hydrolysis is repeated until all of the original material undergoes this change. The amount of active oxygen in solution, representing acetyl and benzoyl hydrogen peroxide, increases slowly until the maximum is reached on about the sixth or seventh day, after which it declines. The rate of change wiU obviously be greatly influenced by the temperature of the water used to effect solution. Thus, in some experiments- we found that after shaking an excess of crys- tals with water at 25° for five minutes, the clear filtrate contained 0.0576 g. of active oxygen per liter, which would correspond to 0.648 g. of benzoy 102 ON THE ORGANIC PEROXIDES. acetyl peroxide unhydrolyzed, or one part to 1,543 parts of water. In another experiment, water of lower temperature being employed, the solution at the end of an hour contained 0.03 g. active oxygen per liter, representing 0.34 g. of the peroxide, a solubility of one to 2,941. In a series of experiments extending at intervals through seventeen days, the amount of active oxygen increased from 0.353 g. per liter at the end of eighteen hours to 1.25 g. at the end of the sixth day. These amounts corres- pond to 3.967 g. and 14.07 g., respectively, of benzoyl acetyl peroxide, and represent in the former case a solubility of 1 to 252 and in the latter 1 to 71 ; more correctly stated, that much of the peroxide has undergone hydrolysis. The perfectly clear filtered solution of the peroxide on standing gives a slight cloudiness and eventually a deposit of minute crystals of dibenzoyl peroxide. The decline mentioned above in the amount of active oxygen pres- ent in solution is due to the formation of this insoluble peroxide. When an excess of benzoyl acetyl peroxide is Icept in contact with water it gradually dissolves and hydrolyzes, as explained above. At the same time benzoper- oxide gradually deposits from the solution, and as a result a crystalline residue is always present, and at no time do we have, as might at first be expected, a complete solution. If the melting-point of the insoluble residue is taken, day after day, it will be found that while at first the crystals melt at 40°, the melting-point of benzoyl acetyl peroxide, after a few days they melt at a higher temperature, and finally, at the end of the sixth day, they melt sharp at 106° to 107°, the melting-point of pure benzoperoxide. When pure crystals of benzoyl acetyl peroxide are exposed to the air, they take up moisture and hydrolyze, and the contents, as a result, become liquid. Under these condi- tions benzoperoxide does not form, probably owing to the large amount of acid present (see page 80). Eventually the secondary hydrolysis gives rise to large crystals of benzoic acid. It will, therefore, be seen that a series of changes result when benzoyl acetyl peroxide is dissolved in water. These changes cannot be satisfactorily followed out in detail, but this much is easily demonstrable; the inactive peroxide dissolves, and undergoing hydrolysis, yields the extremely germicidal and chemically active peracids, and the inactive, insoluble benzoperoxide or dibenzoyl peroxide. The hydrolyzed solutions of benzoyl acetyl peroxide are, therefore, virtually solutions of benzoyl and acetyl hydrogen peroxides, and consequently they possess the same chemical and germicidal properties as do these peracids. This fact will be recognized on inspection of the disin- fection experiments recorded in tables IX and X, and comparing these with tables IV and VII. Action on Pathogenic Bacteria. It has been pointed out in connection with benzoperacid that solutions of this substance are considerably more germicidal than a solution of hydrogen PAUL C. FREER AND FREDERICK G. NOVY. 103 peroxide, which contains ten times as much active oxygen; and that in order to secure equal germicidal action it would be necessary to increase the con- centration of the hydrogen peroxide solution until it would contain at least 100 times as much active oxygen as is present in the solution of benzoperacid. These facts are equally true for the hydrolyzed solutions of benzoyl acetyl peroxide, and will be evident on comparing the following experiments with that recorded in table ^'II. Tabi,e IX. Hydrolyzed benzoyl acetyl peroxide (1-177) containing 0.05 per cent active oxygen. Minutes. 1. 3. 5. 10. 15. 30. 60. B. pvocyaneus . B. coli B. typhosus . . , B. diphtheria . . ^'ib. cholera . . , - — — - - - - S. p. aureus Strep, pyogenes. . Spores, Anthrax B. Spores, Hay B. . . . Spores, Potato B . 1. 3. 5. 10. 15. 30. 60 The above solution was prepared by shaking an excess of the peroxide with ^\•ater for about thirty hours. It contains the same amount of oxygen as, and is therefore equi-molecular with, the solution of hydrogen peroxide employed in table ^'. While the latter solution exerted only a very feeble germicidal action, the benzoyl acetyl peroxide solution destroyed all bacteria tested, including spores, in less than one minute's time. In order to make the com- parison more striking, a fresh solution of benzoyl acetyl peroxide was prepared by shaking the crystals for one hour with water. The results obtained with this relatively weaker solution are given in table X. Table X. Hydrolyzed benzoyl acetyl peroxide (1-2940) containing 0.003 per cent of active oxygen. Minutes. 1. 3. 5. 10. 15. 30. 60. B. pyocyaneus B. coli B. typhosus + + + + + -t- + + -f- + + + + - Vib. cholera S. pyogenes aureus Strept. pyogenes Spores, Anthrax B - " Hav B " Potato B, + It will be seen on comparing the above results with those in Table V that the freshly prepared benzoyl acetyl solution is more active that a hydrogen peroxide solution, which contains sixteen times as much active oxygen. To secure the same germicidal effect as that attained above, a one per cent, or even stronger, solution of hydrogen peroxide would be necessary. Since a one per cent hydrogen peroxide solution contains 0.47 per cent of active oxygen, 104 ON THE ORGANIC PEROXIDES. it follows that such a liquid contains more than 150 times as much active oxygen as the equally germicidal benzoyl acetyl solution. The germicidal action of hydrogen peroxide is usually ascribed to the prone- ness of the latter to give off nascent oxygen. While hydrogen peroxide instantly decomposes and liberates oxygen on contact with blood and other fluids, no such evolution of gas is observed with benzoyl acetyl peroxide, or, at all events, it is not appreciable. These facts render it highly probable that the germicidal action of phthal- monoperacid, benzoyl acetyl, diacetyl and hence of benzoyl hydrogen and acetyl hydrogen peroxides is in no wise dependent upon the amount of active oxygen. This is certainly true if we assume that the active oxygen in hydrogen peroxide possesses the same potential energy. Strictly speaking, however, no nascent oxygen is given off by these organic compounds. If the active oxygen in the organic peroxides is not directly responsible for the germicidal effects, it follows, in all likelihood, that in hydrogen peroxide itself the nascent oxygen does not play the role usually ascribed to it. The most rational explanation for the difference in the action of hydrogen peroxide and the organic peroxides seems to us to be afforded by the theory of dissociation into ions. Baeyer and Villiger have shown that in its chemical behavior benzojd hydrogen peroxide is intermediate between the weak hydro- gen peroxide and the extremely energetic Caro's acid, the peroxide of sulphuric acid. This difference they ascribe to the anions, and it is equally probable that the benzoyl or acetyl anion is more active than the hydroxyl anion of hydrogen peroxide. The dissociation of these compounds can be understood from the following : S- OH Ij - O .OH ** - CeH, = o - O OH c - CH; = o - o OH HO OH Card's Acid. HsHOs. Bbnzoperacid. CeHsCO.OjH. ACETOPEBACID. CH3C0.0,H. Hydropbboxidb. H202. It may be of value to note in this connection that Caro's acid, which is by far the most active oxidizing agent known, results, by the hydrolysis of per- sulphuric acid, in the same way that benzoperacid and acetoperacid form from benzoyl acetyl and diacetyl peroxides. The persulphuric acid, according to Baeyer and Villiger, oxidizes potassium iodide even more slowly than does hydrogen peroxide. And as has already been pointed out, this is also a characteristic of the benzoyl acetyl and diacetyl peroxides. s - O - O - OH - O O - - |H0 - O - sc — CeHi = o - o CHs O O CG - CH; = O - o CHa - O = O - Persulphuric Acid. Benzoyl Acetyl Peroxide. Diacetyl Peroxide. PAUL C. FREER AND FREDERICK. G. NOVY. 105 It will be seen from the above formulas, to which can be added those of dibenzoyl peroxide and hydrogen peroxide, that these relatively feeble com- pounds possess a symmetrical structure (or nearly so), whereas the energetic hydrolytic products are not so constituted. The reaction of diacetyl and benzoyl acetyl peroxides may be looked upon as similar to the decomposition of persulphuric acid which is expressed by the equations : H,S,08 + H2O = H2SO5 + H.8U.,. Caro's acid H.SOs + H,0 = H2O2 + H2SO4. Why dibenzoyl peroxide does not hydrolyze in the same way is difficult to understand. This may be due in part to the extremely feeble solubility of the compound, since hydrolysis may be induced, as Baeyer and Vilhger have shown, by means of sodium ethylate. In the case of organic substances it is possible for the dissociation to pro- ceed along other lines than ionization. Thus, alcohol may be considered as dissociating into ethj'l and hydroxyl groups, or into ethylene and water, or into aldehyde and hydrogen. Ipatiew* has shown that the last two decom- positions do take place at high temperatures. The possibility of the organic peroxides dissociating into other groups than the anions must be granted, though as yet no definite conclusion can be drawn as to its actual occurrence. The germicidal action of solutions of benzoyl acetyl peroxide obtained by shaking an excess of the crystals for one hour, as seen from Table X, is such as to promptly destroy the vegetating bacteria usually in less than one minute. The same result is obtained when the one hour solution is diluted with one or even with two volumes of water. The spores of anthrax and of potato bacilli are not readily destroyed under those conditions. When, however, the active oxygen in the solution is increased either by dissolving the peroxide in warm water or by allowing the liquid to remain in contact with the crystals for sev- eral hours, the germicidal power is increased so that even spores are promptly destroyed within one minute's time. Numerous similar experiments were made, and inasmuch as the results always corresponded to those given in Tables IX and X, it is unnecessary to reproduce these in detail. The fact should be borne in mind that the activity of a solution of benzoyl acetyl peroxide depends considerably upon the degree of hydrolysis; that is to say, upon the amount of active oxygen and more especially on the amount of the peracids which are formed. The hydrolytic process, as already mentioned, can be facilitated by the use of warm water or by allowing the water to act on the crystals for some hours. By the former process it is possible to secure a very active solution within five minutes. > Ber. d. chem. Ges., XXXIV, p. 3579, 1901. 106 ON THE ORGANIC PEROXIDES. Owing to the variable amount of active agent in solution it is obviously necessary in all exact experimental work to determine the actual quantity. This can be readily done by titrating the iodine which is set free upon the addi- tion of potassium iodide, by means of decinormal sodium thiosulphate. The germicidal action of benzoyl acetyl peroxide can be shown in a way other than the direct use of solutions. Inasmuch as the substance is slightly volatile its action can be exerted, as it were, at a distance. Thus, if a few crystals of the peroxide are placed in a small sterile watch glass or fragment of a glass bulb, and this is then set in the center of an agar or gelatin plate, it will be found that the plate will remain sterile or else the colonies of the organ- ism planted will develop only in the periphery, leaving a sterile central area, which will vary in diameter, according to the germ used, and the temperature of the experiment. In the following table are given some of the results which have been obtained in this way. It will be noted that benzoyl acetyl peroxide under these condi- tions exerts a marked action, whereas, iodoform and chloroform, under like conditions, are inert. The reagent used was placed, as mentioned above, in small watch glasses on agar Petri plates, which were then incubated at 35°. The diameter of the Petri dishes was 9 cm. The diameter of the sterile zone is indicated in the table. The development of colonies at the edge of the dish is largely due to the deeper layer of the medium at that point. When very flat dishes are used or just enough agar to cover the bottom is employed complete sterility may be obtained. This result was repeatedly obtained with the diphtheria bacillus. If a few crystals of the peroxide are placed directly on the agar in the center of the dish the result is about the same as when the watch glass is used. This experiment is of interest inasmuch as it shows how a small amount of the peroxide when volatilized may affect surrounding organisms. In one instance the loss in weight was only 5-6 mg. It will be shown later that organic matter interferes with the germicidal action of the peroxide and this is true for the peripheral zone in this experiment. Nevertheless, the bacteria at or near the surface of an organic medium, as agar or gelatin, can be destroyed. Control plates, it may be added, were made in all cases, and showed numerous colonies throughout the medium. Table XL Benzoyl acetyl peroxide in watch glass, on agar plates; diameter of sterile zone in cm. V. Deneke V. Metsohnikovi B. anthracis B. violaceus . . . . B. hog cholera . . B. icteroides . . . . V. cholera B. Friedlander . , 8.2 8.5 8.2 8.0 7.5 7.5 7.0 7.5 M. tetragenus . . . S. pyog. aureus. . . B. coli B. typhosus B. pyocyaneus . . . B. prodigiosus . . . Strept. pyogenes Monilia 7.5 6.3 5.5 5.3 5.0 4.3 4.0 4.2 PAUL C. FREER AND FREDERICK G. NOVY. 107 Action on Water Bacteria. The germicidal action on typhoid bacilli and cholera vibrios is especially important with reference to the practical disinfection of waters containing these organisms, and is, moreover, suggestive of the possible value of this sub- stance in the early stages of cholera and of typhoid fever. The following experiment was made with a bouillon suspension of typhoid bacilli: A cylindrical glass wool filter was prepared and a plug of benzoyl acetyl peroxide crystals, about 3 cm. deep, was placed on top of the glass wool. The suspension when passed but once through this filter gave a sterile filtrate, while control tubes gave the usual abundant growth. In another series of experiments the attempt was made to destroy cholera and typhoid germs which had been added to sterile tap water. In each case a liter of tap water was sterilized in an autoclave, and when cool a suspension of the organisms, previously filtered through glass wool, was added. Ten and twenty mg., respectively, of the peroxide were added and, after thorough shaking, portions of 1 c.c. of the liquid were taken out and planted in bouillon and in agar which was plated. Similar controls were made before the addi- tion of the peroxide and gave abundant growths, the plates yielding 600,000 to 800,000 colonies. The results are indicated in the table. Table XII. 10 mg. peroxide. 1 20 mg. peroxide. Min. Bouillon. Agar. Bouillon. Agar. Cholera sus- pension 5 — — — — 15 — — — — 30 — — — — 60 - - - - Typhoid sus- pension 5 + + + + 15 — — — — 30 — — — — 60 — — — — The amount of peroxide added in the above tests, it will be noticed, was extremely small, so small indeed that the test for active oxygen with potas- sium iodide and ferrous sixlphate was negative, or but faint, during the first half hour. The addition of 10 mg. represents one part in 100,000 and the active oxygen in this amoimt is only 0.9 mg., which corresponds to 0.00009 per cent. Nevertheless, this small amount was sufficient to sterilize the cholera germ within five minutes, and the typhoid bacilli within fifteen minutes. The above experiment repeated with unsterilized tap water gave similar results. That is to say, the cholera and typhoid bacilli were destroyed as before, but the waters were not rendered wholly sterile, owing to the presence 108 ON THE ORGANIC PEROXIDES. of relatively resistant fluorescing bacteria. A double control experiment with the same unsterilized water, to which the peroxide was added in amounts as above, but no cholera or typhoid germs, gave an enormous decrease in the number of the water bacteria. With 10 mg. peroxide the number of bacteria dropped from 900,000 and 600,000 to 7 and 10, respectively, at the end of one hour. With 20 mg. peroxide the decrease was from 1,600,000 and 500,000 to 5 and 8, respectively. To further test the possibility of effecting the complete sterilization of diverse spring and well waters, nineteen samples were collected from different localities. Twenty mg. of the peroxide were added to each liter of water, and cultures were made at intervals as above. Two of the waters, as shown by control tubes, were sterile. Five were rendered sterile in from thirty to sixty minutes. In the remaining twelve the number of bacteria at the end of sixty minutes Varied from 2 to 45, the average being 11^ colonies per sample. Con- trol counts showed that the waters usually contained several hundred, and in some instances many thousands of bacteria perc.c, before treatment. The colonies which developed on the sixty-minute plates of the test series were transplanted and studied further. The organisms were found to be bacilli, usually spore-bearing. The presence of resistant spore forms in these cultures was also indicated by the resistance to fresh solutions of peroxide, and even to boihng water. When suspensions of these germs were inclosed in capillary tubes, with one exception they all resisted 100° for one minute, and a number even survived after two minutes' exposure. The failure to completely sterilize natural waters by the addition of 20 mg. of benzoyl acetyl peroxide is, therefore, due in part to the small amount of the reagent used, and especially is due to the presence of resistant forms or spores. The important fact, however, stands out in these experiments that the pathogenic organisms, such as those of cholera and typhoid fever, as well as other vegetating bacteria, are speedily destroyed by extremely small amounts of this peroxide. The practical applicability of this agent for the purification of contaminated waters, especially in connection with military operations, is a subject which it is desirable to investigate further. As wiU be seen from the following experiments with sewage, even enormously contaminated waters can be rendered almost sterile by employing relatively small amounts of the peroxide. The further fact that the reagent soon disappears renders its use preferable to the ordinary chemicals. Sewage. — The germicidal action of benzoyl acetyl peroxide upon aqueous suspensions of bacteria is even more striking when sewage water is employed. Complete sterility is only exceptionally obtained, owing to the presence of very resisting forms. Nevertheless, the number of bacteria, as will be seen from the subjoined table, is enormously decreased within a few minutes. In the following experiment sewage from Ann Arbor and from Detroit PAUL C. FREER AND FREDERICK G. NOW. 109 was used. It was collected in sterile bottles, and before using was filtered through sterile glass wool to remove gross solid particles. To a liter of the filtered sewage 100 mg. of the peroxide was added (1-10000). The liquid kept at 40° was frequently shaken to facilitate solution of the reagent. At intervals, portions of 1 c.c. were removed, and agar Petri tlishes were made which were then incubated at 35°. Control plates were made with the filtered, but untreated, sewage. For these only 0.1 c.c. of water was used, but the results given in the table are expressed per c.c. The counting of these colonies was made under the microscope with the aid of Ehrlich stops in the eye-piece. Experiments were also made using 200 and even 300 mg. of the peroxide per liter of the sewage. The results obtained with the larger quantities of the reagent did not materially differ from those given in the table. That is to say, the great decrease in the number of bacteria occurred in the first fifteen minutes, after which a few resistant organisms persisted, even for sixty minutes. Table XIII. Sewage plus benzoyl acetyl peroxide (1-10000). Colonies per c.c. Control. 15 Min. 30 Min. 45 Min. 60 Min. Samnle Al 162,410 78,188 54,945 45,574 826,430 1,450,200 1,106,570 79,720 1,520,260 594,050 22,843 75 28 23 31 13 44 22 9 4 17,572 65 38 10 10 29 25 11 9 4,295 50 16 7 6 4 11 12 5 35,144 ' A2 50 ' A3 34 ' A4 ' Dl ' D2 ' D3 7 7 3 3 ' D4 . 12 ' D5 . 3 ' D6 1 Action on Saliva. The energetic action which benzoyl acetyl peroxide exerts on water bacteria is also seen in the destruction of mouth organisms. For the following experi- ment, four persons (A, B, C, D) were selected, and about 10 c.c. of saliva were gathered from each one. In order to secure as many bacteria as possible, these persons were requested to roll or rub their tongue for three minutes over the buccal mucous membrane and over the gums, and at the end of that time the sahva was received in sterile test glasses. A portion of 5 c.c. of each sahva was then transferred to a sterile test tube and an equal volume of the peroxide solution, obtained by saturation for one hour, was then added. The liquid was mixed thoroughly, and at definite intervals transplantations were made as follows: (1) a large loopful (2' mm. loop) was transplanted to 10 c.c. of bouillon; (2) 1 c.c. was transferred to melted agar, kept at 50°, and after mixing thoroughly Petri plates were made. In each case a control Petri plate was made with i c.c. of the saliva. The plates were developed for thirty- no ON THE ORGANIC PEROXIDES. six to forty-eight hours in the incubator after which the colonies were counted. Owing to the large number of colonies on the control plates it was necessary to count these under the microscope, using Ehrlich stops in the eye-piece. The results of this experiment are given in the following table : Table XIV. Saliva and benzoyl acetyl peroxide. A. B. C. D. Agar plate. Bouil- lon. Agar plate. Bouil- lon- Agar plate. Bouil- lon. Agar plate. Bouil- lon. Control 1 min 3 " 5 " 10 " 15 " 1,900,000 301 115 181 53 33 + + + + + 1,200,000 - 650,000 4 + -1- -1- 1,600,000 '63 5 - The strength of the peroxide solution used in this experiment was not deter- mined, but assuming that it was about the same as employed in Table X, it would represent approximately a solution of 1-6000, with an active oxygen contents of 0.0015 per cent. The fact that in two cases (B and C) the steri- lization was complete within one minute, and in another (D) within five minutes, indicates the remarkable destructive action of the solution. In case A the number of bacteria was greatly reduced, but complete sterilization was not effected in fifteen minutes. This may have been due, as in the experiments with water bacteria, to the presence of spores, or perhaps to the presence of small masses of bacteria, in which case it is of course conceivable that a few organisms in the center of such minute clusters might be protected against the action of the germicide. In the following experiment equi-molecular solutions of benzoyl acetyl peroxide, diacetyl peroxide, and hydrogen peroxide were used. The strength of these solutions was such that 1 c.c. required 0.14 c.c. of N/10 NajSjOg. This represents 0.0112 per cent of active oxygen. The saliva of nine students was subjected to tests with these solutions. The procedure was the same as that given above, and inasmuch as equal volumes of saliva and reagent were employed, it follows that the per cent of active oxygen was only 0.0056. The experiment indicates that benzoyl acetyl and diacetyl peroxides are equally efficient, and moreover, as heretofore pointed out, are very much more active than a solution of hydrogen peroxide containing a like amount of active oxygen. It will be noted that the benzoyl acetyl and diacetyl peroxides, even in the dilution employed, are capable of steriHzing a homogeneous saliva within one minute of exposure, excepting one instance where this result was obtained in three minutes. In this experiment bouillon tubes and agar plates were made each time. PAUL C. FREER AND FREDERICK G. NOVY. Ill o CD •O o o d so > X H < 3 C c ce > ^ t>: ■* 00 CO CD CO Tf a to 00 C CD .9 9S o ■* IN o 00 CQ !>. o -^ 00 lO CO t^ !>. lO »< »— 1 to- s" "3 g' oo" 00- -_^ cd- i^- co" OS_ oo" °i o" o^ CO i T-H ffi «e 00 t^ CO to CO CD o lO IN OS o 00 CO CO Oi c t^ ,._^ °i. iO Tt< OS 00 I-H co^ s if e-. x ^'" in" o- co- (N" o" »o 00 OS 00 co o 00 05 1— ( t^ lO T-H Os_ CO lO to IN o 00 o CO -^ 00 c:^ OS CO 00 OS T-H (N (N o c^ Tfl OS os__ lO OS CO ^^ 1 lO to" 00~ CO- IN o- oT CO*" o 00 ■^ IN lO t^ OS 00 CO IN^ T-( OS CO I>^ 00_ o o^ i-H T-H I"* *"• 00 IN « d 00 S s o ■ 1-H feo >1 1 lO lO -* o IN Q CO CO »o ojC!' I> cq -tf o to IN o o o lO m lO °i ^^ lO t^ co_^ lO fi in" oT o" I-H -a 'S d o i ^ -.00 >>co |2 *i lO o o t^ (^ CO o zo o M Tt< c^< 00 t^ o (M CO "s. o ■o_ CO os__ to '^« co_ o^ "^^ o ^ d 5 T*" -*^ i-T CO CO of t-- «d" 1— 1 G lO CO ^ o cq lO tH t- o cq to CO IN Tfl -^ 1— 1 c^ CO Tfi lO CO t^ o6 05 cS ,^ ^ ^ ^ ^ ^ ^ ^ ^ (S m 112 ON THE ORGANIC PEROXIDES. The former are not recorded, because they agreed exactly with the plates. That is to say, when the agar plates developed colonies the bouillon tubes showed a growth; when the plates were negative, the tubes were Ukewise. In the case of benzoyl acetyl and diacetyl peroxides transplantations were made as in the case of hydrogen peroxide, at 1, 3, 5, 10, and 15 minutes, but inasmuch as these were all negative, only the first is recorded. In another experiment with the above four persons 10 c.c. of the peroxide solution were placed in the mouth and vigorously rolled about for fifteen seconds. The Hquidwas then transferred to sterile test glasses and at same intervals as above a large loopful was transferred to bouillon. These mouth washings in three cases were sterile within one minute, and in the fourth person this resiilt was obtained within five minutes. The latter result was probably due to the presence of small masses of bacteria in the washings. The complete destruction of all the bacteria in the mouth is obviously difficult to attain. This is especially true when the edge of the gums are lined with decaying food particles which, as is well known, are extremely rich in bacteria. A minute particle of this material dislodged from between the teeth or from the edge of the gums will liberate a large number of organisms. Nevertheless, even without the previous use of a brush, the number of bacteria in the saliva may be enormously decreased by several successive rinsings with the peroxide. This decrease is much greater than that which foUows the similar use of sterile water, and justifies the belief that practically complete disinfection is possible. It will be seen from the experiments related that benzoyl acetyl peroxide may prove a valuable remedy in bacterial diseases of the mouth. Where the strong solution can be readily applied to a surface growth, the result will be prompt. The trials which have been made thus far in tonsillitis, pyorrhoea, as well as in other affections, have been satisfactory. The direct application of the crystals, as a rule, is undesirable, owing to the marked irritating action which the sohd substance exerts upon the mucous membrane, or even on the skin. In several instances reported to us by practitioners the crystals have been used with good results. In one case the crystals were placed within a suppurating gland and in another they were applied to a very intractable trichophytosis. In gonorrhoea a single irrigation will frequently stop the dis- charge, which, however, is Ukely to recur. At times, when older and hence stronger solutions are used in this disease, they cause considerable smarting. Action on Intestinal Bacteria. Nencki and Zaleski^ in 1899 attempted the disinfection of the aUmentary canal by means of dibenzoyl peroxide. Starting out with the conception that the bacterial decompositions were anaerobic in character, they endeavored ' Zeitschr. f. physiol. Ohemie., XXVII, p. 487. PAUL C. FREER AND FREDERICK G. NOVY. ' 113 to introduce, or rather generate, oxygen within the intestine by the use of this peroxide. On digesting benzoperoxide with pancreatic juice and bile they obtained a gas, nearly nine-tenths of which was carbonic acid and nitro- gen, the remainder being oxygen. They, therefore, concluded that oxygen was hberated in the intestines, but in amounts so small as to have no effect on the existing fermentations. Actual feeding experiments on dogs showed no variation or effect upon the amounts of conjugate sulphates or aromatic bodies, although the benzoperoxide was largely decomposed in the intestine and excreted in the urine as hippuric acid. About a half, however, was ehmi- nated with the feces. With phthahc acid peroxide they obtained equally negative results. The reason for this is perfectly obvious, since, as has already been pointed out, the organic peroxides, unlike hydrogen peroxide, do not readily Uberate free oxygen ; and furthermore, the peroxides employed by them are germicidally wholly inert. In view of the fact that the hydrolyzed solution of benzoyl acetyl peroxide is a powerful disinfectant, it was a matter of very great importance to test its action upon the organisms in the alimentary canal. For this purpose the feces of eleven persons were collected in sterile dishes. A portion was taken and whipped up with sterile water till an opaque suspen- sion resulted; this was then filtered through glass wool and an equal volume of a benzoyl acetyl solution (1-1000) was added. Companion tests were made with mercuric chloride (1-1000) and phenol (1-20). While the control plates showed the presence of large numbers of bacteria, as in the case of sew- age, those made after an exposure of 1, 3, 5, 10, and 15 minutes were invaria- bly sterile when the peroxide was used. The mercuric chloride and phenol solution gave almost the same results. This experiment, as well as those -with sewage and saliva, indicate that benzoyl acetyl peroxide is extremely germicidal to the organisms, as found in the ahmentary tract. The question which naturally presents itself is. Is it possible, by means of this agent, to suppress, or even reduce, the growth of bacteria in the intestines? The ingested peroxide hydrolyzes as under ordinary conditions. Eventually the products are absorbed and the benzoyl group is eliminated almost quan- titatively in the urine as hippuric acid. Thus, one dog which received 6.5 grams of the peroxide in the course of seven days excreted in the urine 6.3 grams of hippuric acid. When about two grams of the peroxide were fed to a dog a trace of the substance apparently appeared in the feces. At all events, this was indicated by the liberation of iodine from potassium iodide. But since this reaction is also given by nitrites, it follows that the elimination in the feces is open to doubt . The urine of a dog which was fed one gram daily showed no marked variation in the absolute amount of conjugate sulphates eliminated in the urine, and it is altogether probable that this amount of peroxide is 114 ON THE ORGANIC PEROXIDES. chiefly destroyed in the small intestine, and therefore exerts little or no effect upon the decompositions in the lower part of the canal. The administration of water suspensions of the peroxide, by means of a flexible catheter, to rabbits, does result in the practical sterilization of the con- tents of the stomach. Moreover, in several experiments with these animals the intestinal tract, apart from the coecal pouch, was found to be sterile. Neither bouillon tubes nor agar plates showed growths, although the con- trols gave abundant cultures. These experiments, though few in number, would seem to indicate that the continued administration of the peroxide in sufficient quantity may result in a more or less complete sterilization. The animal experiments having shown that the peroxide can be adminis- tered with safety it was decided to test the reagent as to its capability of influ- encing intestinal fermentations in man. ■ Six students volunteered to subject themselves to this experiment. They were placed upon an exclusively milk diet for six days. A few, as indicated in the table., in addition drank a water solution of the peroxide (1-1000). The milk used was sterihzsd in an auto- clave at 120° for half an hour. The subjects, moreover, frequently rinsed their mouths and brushed their teeth with a peroxide solution. During the first two days sterile milk only was given in order to establish an equilibrium for this diet. One person (Johnson) was reserved as a control and was given sterile milk only throughout the six days. After the second day one gram of benzoyl acetyl peroxide was added to each liter of sterile milk, and the other five men were requested to drink this freely, which they very willingly did. In the following four days Clift consumed 10.75 g., while Crane in the same time disposed of 13 g. of the reagent. The other three men (Hale, Hodges, and Myers) took the peroxide milk on three days only, and on the last, or sixth day, reverted to plain sterile milk. This was done in order to check the results of the previous days, and also those of the two men who continued to drink the peroxide milk. The feces were collected each day in sterile vessels. They formed, as a rule, hard, hght-colored masses, which were, occasionally, blood-streaked. The normal indol odor, as might be expected with a milk diet, diminished after the Burst few days. A microscopical examination of each specimen showed the presence of bacteria. At the same time aerobic and anaerobic plates were made, and in order to obtain fairly comparative results, a standard 2 mm. loop was used for transplanting the material. In spite of the care thus taken, and for reasons which can be readily understood, the colony counts showed considerable variation, which could not, altogether, be ascribed to the use of the peroxide. Taken as a whole, however, the number of colonies seemed to be less after the peroxide was taken than on the first two days, when sterile milk only was used. The control (Johnson) who received no peroxide gave a much larger number of colonies, but even in this case there PAUL C. FREER AND FREDERICK G. NOVY. 115 was a marked decrease in the number on the last two or three days. It is clear that a prolonged milk diet brings about a decrease in the number of organ- isms, and for that reason no d(>finite conclusion from the mere colony counts can be drawn. More satisfactory results were expected from the chemical examination of the urine. Each day the urine was collected in sterile bottles and was sub- jected at once to analysis. The laborious and painstaldng analyses were made by Mr. C. L. Bhss, without whose conscientious assistance it would have been impossible to do the work. The sulphuric acid estimations were made in the usual way. Indoxyl was determined after oxidation with ferric chloride, colorimetrically by comparison with a solution of indigo-blue of known strength. The phenol was estimated, after distillation, by means of iodine and deci- normal sodium thiosulphate. Duplicate analyses were made in all cases. Owing to incomplete results the urine examination of one of the subjects is not given in the table. The results of these analyses are given in Table XV. The conjugate sulphates, as is weU known, may be taken as a measure of the extent of intes- tinal putrefaction. This is particularly true of the ratio of simple to conjugate sulphates. An inspection of this column will show a marked decrease in the ratio "of these two substances on the days when the peroxide was taken, as compared with those of the control person, and with those days when the reagent was not used. In part this decrease may be attributed to the milk diet, but, on the other hand, it is sufficiently pronounced to justify the belief that the peroxide when used freely will reduce bacterial activity in the intestines. The results, perhaps, might be more conclusive, if instead of milk some other food was used which would not combine with the peroxide, and thus lessen its action. Obviously the complete suppression of intestinal bacteria can hardly be expected. When the peroxide solution is placed in a loop of freshly washed intestine it graduallj- disappears so that at the end of an hour scarcely any peroxide can be detected. In other words, in the absence of organic matter in solution or in suspension the peroxide combines with the proteids of the intestinal wall. Diverse Applications. The usefulness of benzoyl acetyl peroxide in mouth disinfection has already been pointed out, and in this connection it may be well to refer to spraying experiments made especially upon the guinea-pig. The animal was placed for this purpose in a glass cylinder of about two liters capacity and was exposed to a continuous spray of a saturated peroxide solution. Spraying may be said to be without any injurious action upon the animal. Two guinea-pigs were exposed to a spray daily for four hours for thirty days and throughout all this time presented no deviations from the normal, either in temperature or 116 ON THE ORGANIC PEROXIDES. weight. A large number were, moreover, sprayed for variable lengths of time as controls for diphtheria experiments, and likewise showed no bad effects. Table XY. Analysis of urine after administration of benzoyl acetyl peroxide in man. CI., Dec. 13. 14.. 15. 16. 17. 18. Milk and peroxide taken. 1500cc.;No. P . 3000 cc; No. P . 1500 cc; 1.5 g. P 3000 cc; 3 g. P.. 3000cc.;3g. P. . 2000 cc. ; 2 g. P. . + 1-} L. water = li g. P Vol. Urine. 783 508 490 1,605 1,590 1,650 Specific gravity at 15° 1.0233 1.0300 1.0298 1.0128 1.0167 1.0118 Total SO,. g- 1.418 1.702 2.365 2.896 2.3.36 Simple so,. g. 1.346 1.662 '2.3i4 2.903 2.280 Conj. SO,. g. 0.103 0.079 0.673 0.063 0.069 Ratio S:C. 13:1 21:1 31.7 :i 46:1 33:1 Total Phenol. g. 0.060 0.047 0.036 0.043 0.039 0.036 Total Ind'x'l. g. Trace. Trace 0.0019 Trace 0.0044 Practically no loss of urine. Trace of albumin and few hyaline and granular casts present each day; Fehling solution reduced on fourth, fifth, and sixth days. Total peroxide taken in four days= 10.75 g. Cr., Dec. 13. 14. 15. 16. 17. 18. 3000 cc; No. P. . . 2500 cc; No. P. . . 3000cc.;3g. P. . . 3000 CO. ; 3 g. P. . . 2500 CO.; 2.5 g. P. + 1000CC water = lg. P 2000 cc; 2 g. P. + 1500 cc. water = 1.5 g. P 959 1.100 763 1,288 1.0154 1.0137 1.0164 1.0125 1.260 1.565 1.756 1.169 1.542 1.715 0.101 0.070 0.659 11.5:1 22:1 ' '29:i 0.048 0.026 0.021 0.038 1,830 1.0107 2.254 2.214 0.091 24.3:1 0.038 1,200 1.0131 1.658 1.591 0.100 15.9:1 0.048 0.0090 0.0089 0.0201 Trace 0.0197 0.0319 Considerable and indefinite loss of urine at stool. No albumin or casts. on fourth, fifth, and sixth days. Total peroxide taken in four days= 13 g. Fehling solution reduced He. Dec. 13. . . 2250 cc. ; No. P. ... 855 1.0248 1.906 1.769 0.146 12.2:1 0.073 0.0172 14... 2750 cc; No. P. ... 1,540 1.0166 2.648 2.707 0.098 27.6:1 0.047 0.0145 15... 2000cc;2. g. P.. .. 762 1.0264 0.046 0.0151 16... 3000 cc. ; 3 g. P. ... 1,250 1.0192 2.482 2.435 0.092 26.4:1 0.037 0.0252 17... 2000oc.;2g. P. ... 1,055 1.0208 2.362 2.325 0.065 35.7:1 0.029 0.0112 18... 1700 cc; No. P. ... 1,077 1.0250 2.968 2.901 0.103 28.1:1 0.041 0.0142 Loss of urine on 14tli about 50 cc. ; on ISth, about 150 cc. ; otherwise none. No albumin or casts. Slight reduction of Fehling's solution on the last two days. Total peroxide taken in three days =7 g. Hs., Dec. 13... 1500 cc. ; No. P. ... 570 1.0251 1.369 1.293 0.082 15.7:1 0.043 0.0226 14... 3500 cc; No. P. ... 1,510 1.0160 2.431 2.497 0.087 28.7:1 0.047 0.0201 15... 2000 cc ; 2. g. P 920 1.0244 0.058 0.0086 16... 2000cc.;2g. P. ... 348 1.0220 0.804 0.801 0.025 32:1 0.047 0.0038 17... 3500 oc; 3.5 g. P. .. 1,800 1.0120 2.739 0.046 58.5:1 0.106 0.0120 18... 1700ec.;No. P ... 1,248 1.0198 2.710 2.645 0.092 28.7:1 0.04 0.0134 Loss of urme on the fourth day, about 200; on the fifth day, about 800 cc Albumin and easts negative. Reduction of Fehling's solution on the fifth day, marked. Total peroxide taken in three days = 6.5 g. Control. Jn., Dec. 13. . . 4500 CO.; No. P. ... 1,068 1.0182 1.567 1.458 0.132 11:1 0.056 0.0185 14... 2000 cc; No. P. ... 647 1.0196 1.343 1.318 0.085 16.5:1 0.034 0.0173 15... 2000 CO.; No. P. ... 519 1.0270 0.042 0.0210 16... 4000 CO.; No. P. ... 1,595 1.0156 2.721 2.628 0.133 19.7:1 0.070 0.0170 17... 2500 00. ; No. P. ... 2,490 1.0082 2.574 2.514 0.084 29.9:1 0.0201 18... 1,165 1.0166 2.034 1.964 0.118 16.6:1 0.068 0.0392 i ■c< i°^^,°^ urine on fourth day, about 400 cc. No albumin; few casts on three days. No reduction of Fehlmg s solution. This person received no peroxide, and hence served as a control. In order to test the value of the spray in laryngeal diphtheria, a series of twenty-eight guinea-pigs were inoculated intratracheally with a culture of the PAUL C. FREER AND FREDERICK G. NOVY. 117 diphtheria bacillus. One-half of these animals were sprayed every alternate four hours for five periods. The other half were not sprayed and were reserved as controls. Of the fourteen spra^'od animals three died from the effects of the operation and three from subcutaneous infection, while the remaining eight survived. Of the fourteen controls only four recovered, the others died of diphtheria infection, as was shown by post-mortem examination. On the other hand, the animals that died in the sprayed set, as a rule, gave a perfectly sterile trachea, showing, therefore, that benzoyl acetyl peroxide, when used as a spray, is able to destroy the bacteria in the air passages. Intraperitoneal injections of the peroxides are readily tolerated by guinea- pigs, rabbits, and white rats. In the former a dose of 5 c.c. can be administered daily for more than a month A^ithout any untoward effect. A series of experi- ments have been made upon animals infected with various pathogenic bacteria and with trypanosomes, and in many instances it was possible to retard or prevent a fatal termination. Unhke hydrogen peroxide, benzoyl acetyl peroxide may be injected intra- venously into rabbits without injury. Thus one animal received into the ear vein three separate injections, each of 5 c.c, in the course of the day. The following day a like number of injections, each of 10 c.c, were made. The weight and the temperature of the animal were unaffected. The absence of any recognizable deleterious action is due in the first place to the fact that the peroxide does not liberate oxygen on contact with blood, as does hydrogen peroxide. Moreover, as will be presently seen, the peroxide is destroyed by, or combines T^ith, organic matter. Intravenous injections of large quantities, five liters at a time, on two successive days, into horses do not affect the trypanosome of surra. The fact that heat destroys benzoyl acetyl peroxide has been incidentally referred to. This effect is brought about by exposing a saturated solution of the substance to a temperature of 75°. On the other hand, a temperature of 55°, acting for one hour, has no appreciable action, inasmuch as the solution is about as germicidal as in the beginning. Benzoyl acetyl peroxide, as may be readily inferred from its chemical nature, is easily destroyed by a variety of substances. Thus, the addition of a hydrogen sulphide solution, urine, milk, serum, or bouillon to an equal volume of the peroxide solution produces an immediate change which is seen in the fact that the solution no longer liberates iodine from potassium iodide. That the action of the peroxide is done away with is seen from the following experiment, in which a freshly prepared solution of peroxide, one hour old, was employed. The control tests were made by diluting this solution with an equal volume of sterile water. The other trials were made with equal volumes of reagent and the fluid to be tested. The sterile water and fluids were inoc- ulated with the trial organisms. 118 ON THE ORGANIC PEROXIDES. Table XVI. Efiect of medium on peroxide action. Control, semi-saturated. Urine. 1-1. Bouillon, 1-1. Minutes. 1. 3. 5. 10. 1. 3. 5. 10. 1. 3. S. 10. B. coli. - - - - + + + + + + + + + + + + + + + + + + + + + + + + + + + + B. typhosus . . . + B. icteroides + B. pestis + Vib. cholera B. coli + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + B. typhosus + B. icteroides . 4- B. pestis Vib. cholera + Milk, 1-1. 1 H jS-water, 1- L. 10% Serum, 1-1. It will be seen from the above that while the semi-saturated peroxide solution destroyed the test bacteria within one minute, the milk, serum, and hydrogen sulphides mixtures failed to exert any effect within the time of the experiment. This fact and the uniform absence of the iodine reaction showed that the peroxide was destroyed or neutralized. The action of hydrogen sulphide may be correctly interpreted as a destruc- tion of the peroxide, since the sulphide undergoes oxidation. On the other hand, milk, serum, and the proteid solutions seem to inhibit the action of the peroxide by giving rise to proteid compounds of the peroxide, which are salt- hke in character and are inert. This reaction will be studied more in detail. The urine and the bouillon mixtures still gave a faint iodine reaction at the end of the ten minutes, showing that the peroxide was not completely destroyed. The amount present, however, was so small as not to affect any of the organisms tested, except the very sensitive cholera vibrio. In this connection, it may be stated that a putrid sputum, rich in hydrogen sulphide, is promptly deodorized by the peroxide, and with a sufhcient amount of the latter is rendered sterile. It is obvious from the above that given an excess of peroxide, in other words, more than what is required to act upon the easily oxidizable constitu- ents, the mixtures will exhibit the usual germicidal power. In the case of urine it can be shown that urea, creatin, hippuric acid, and the phosphates do not act upon the peroxide. On the other hand, when uric acid is shaken up with a peroxide solution, the iodine reaction is promptly weakened and disappears altogether in about two minutes. In view of the above facts, that labile substances are easily destroyed, and in turn destroy the peroxide, it seemed desirable to test the action of the per- oxide upon bacterial toxins and upon enzymes. The experiments which have been made thus far along this hue indicate that toxins and enzymes are affected PAUL C. FREER AND FREDERICK G. NOVY. 119 by this reagent. Thus, it was found that 1 c.c. of the peroxide solution will destroy in one hour ten minimum fatal doses of diphtheria toxin. The same, and oven more marked, result is obtained with the tetanus toxin. The action of the reagent upon ptyalin, rennin, and pepsin is apparently not so rapid or so pronounced, but, nevertheless, is very distinct. These results are in accord with those obtained by Sicber,^ who studied the action of calcium per- oxide, hydrogen peroxide, and of several oxidases upon abrin, diphtheria, and tetanus toxins. At the temperature of the incubator these poisons were readily destroyed, and moreo\-cr, this result was obtained when the oxidase and toxin were injected either simultaneously, as a mixture, or separately into animals. In other words an antitoxic action, similar to that reported in con- nection •^^•ith extracts of the thymus gland, was demonstrated. The experiments heretofore described were, excepting Table XVI, made with aqueous suspensions of bacteria. In practical work, the conditions are by no means as favorable, and it was, therefore, deemed desirable to test the action of the peroxide upon cultures previously dried down upon such surfaces as cover-glasses, silk threads, muslin, and rubber sqviares, and on various instniments. For this purpose agar cultures were employed. By means of a bulb pipette sterile water was introduced into the freshly developed culture tube. A por- tion of the gro-ni;h was thoroughly whipped up with the water and the resulting thick suspension was transferred to a glass-wool filter. The fairly homogeneous filtrate was then smeared over the surface of oblong pieces of dentist's rubber or of cover-glasses, or over instruments. The silk threads and muslin squares were soaked in the suspensions for about an hour. After the surfaces were covered in this way by the organisms, the specimens were placed to dry in an incubator at 39°, usually for three to four Jiours. When thoroughly dry they were exposed to the action of the disinfectant. In the above experiment the cultures and three of the reagents were the same as those employed in Tables V, IX, and X. The results obtained with silk threads and cover-glass specimens are essentially the same as those recorded in the tables just mentioned. The comparison of the weak peroxide solution with the hydrogen peroxide and carbolic acid solutions is distinctly favorable to the former. In connection with this experiment a similar set of specimens were exposed to the action of a 1-1000 mercuric chloride solution. All of the specimens failed to develop, although they were washed in sterile water and planted in large amounts of bouillon (10 c.c). Assuming that the germs were killed within one minute's time it follows that the strong peroxide solution is at least as effective as mercuric chloride, 1-1000. In several similar disinfection experiments, rubber and muslin squares, employing the colon, typhoid, glanders, and green pus bacilli and the Staphyl- 1 Zeitschr. f. physiol. Chem., XXXII, p. 573, 1901. 120 ON THE ORGANIC PEROXIDES. a o3 ho 3 T3 GJ i-i -S X r O I— J K CD d to + + 1 1 1 ■f + 4- 1 1 d + + + 1 1 + + + 1 4- i d ^ to + + + 1 + + + 1 + 1 m g- ii U5 + + + + + + + + 1 ^f^ 1 o 1.° ii 0,^ ^ + + + + + + + + + + -s d + 1 1 + + + 1 ■si Eh O. O CO + 1 1 + + 1 « >, m w ffi ^ lo + + + + + + + ^-^ O w + + + + + + - + + + + + + + + d to 1 1 1 1 + + + 1 d CO 1 1 1 1 + + + 1 4 • d ^ to + + + 1 4 q d iO + + + + -f - 2g a) X ^^ K "3 -1 -p o Z o > -g ^ i D. c a: P4 « cc pq PQ PAUL C. FREER AND FREDERICK G. NOVY. 121 ococcus pyogenes aureus, were exposed for fifteen minutes to a freshly prepared, one-hour peroxide solution, after which they were transferred to bouillon, but failed to give any growth. Sterilization was, therefore, obtained in less than fifteen minutes. In another experiment twelve sterile instruments (scalpels, scissors, forceps) were dipped in a thin suspension of Staphylococcus p. aureus, after which they were dried in the incubator. They were then immersed in a solution of ben- zoyl acetyl peroxide. At the end of five minutes two of the instruments were removed and transferred to tubes of bouillon. The same was done at the end of ten, fifteen, thirty, and sixty minutes. All these specimens failed to develop, although two controls gave abundant growths. A similar set of instruments were covered with a thick suspension of B. pyocyaneus, but otherwise were treated in the way just mentioned. The five, ten, and fifteen minute speci- mens gave growths, but the thirty and sixty minutes did not. Obviously the conditions in the foregoing experiments are hardly to be expected in dental and surgical work. Nevertheless, it is evident that with fairh' active solutions of benzoyl acetyl peroxide, it is possible to sterilize such instruments. Inasnuich as this is accomplished without any injury to the articles, it is quite probable that the peroxide solution may prove useful, especiallj' with delicate instruments which will not bear ordinary methods of sterilization. Summary of Part II. This part comprises a study of the germicidal action of six, or rather seven, organic peroxides. Of these, the acetone and dibenzoyl peroxides possess a syxometrical structure, and are wholly inert, and remain so even on prolonged contact with water. The diacetyl and benzoyl acetyl peroxides are likewise symmetrical and inert, but in the presence of water they undergo hydrolysis, and yield the active peracids. Acetyl and benzoyl hydrogen peroxides are extremely germicidal, and easily rank with the most active disinfectants. The phthalic monoperacid is also active, though to less extent. Hydrogen peroxide is considerably weaker than these organic peroxides. The activity of the peracids and of hydrogen peroxide is not due to active oxygen, but is probably due to the acid ions. Their solutions do not give off oxygen on contact with organic matter, at least not readily. The peroxides, however, may be estimated by determining the amount of active oxygen which they contain. Benzoyl acetyl peroxide, on account of its relative stability in the crys- talline form, and its rapid hydrolysis in the presence of water, with the con- sequent development of intense germicidal power, may prove to be of value in the treatment of certain bacterial diseases. 122 ON THE ORGANIC PEROXIDES. The hydrolyzed solution of the latter, in the absence of oxidizable sub- stances, will promptly destroy even resistant spores inside of a minute. The concentration required for this result is about 1 to 1,000. Considerably weaker solutions will destroy the vegetating forms of bac- teria in the same length of time, provided they are in water suspension. Organ- ic solutions, such as bouillon, milk, blood serum, urine, are not disinfected as readily. The hydrolyzed solutions of this peroxide can be used to advantage for the sterilization of dental and other dehcate surgical instruments, for which purpose they are considerably more effective than a five per cent carbohc acid solution. Such solutions are not poisonous, and may be used for the disinfection of the mouth and for spraying the air passages, and possibly for the suppression of the growth of bacteria in the stomach and intestines. In typhoid fever, dysentery, and especially in cholera, this peroxide may be expected to yield useful service. The results indicate that it may be used for the purification of waters infected with typhoid and cholera germs. The hydrolyzed solution is not only destructive to bacteria, but also to toxins and to enzymes. III. Practical Application of Benzoyl Acetyl Peroxide. Of the several peroxides discussed in the preceding part it will be seen that there are some which are wholly inert, while others are extremely active. Among the latter acetoperoxide occupies a prominent place not only on ac- count of the readiness with which it hydrolyzes into the energetic germicide, acetoperacid, but also on account of the ease of preparation, and the relatively low cost of production. These facts might lead to the introduction of this peroxide into medical use were it not for the very great danger of explosion. Indeed, few substances can rival acetoperoxide, either in the ease with which they explode or in destructiveness. On the other hand, benzoyl acetyl peroxide can be prepared with relative safety. However, the pure crystalline peroxide cannot be dispensed, ordi- narily, for several reasons. In the first place, it possesses a low melting-point, which is only slightly above the temperature of the body, and consequently in a warm room, or in a warm climate, it would be in the melted condition. While in this shape it leaks readily through glass-stoppered bottles in spite of utmost care, and it has repeatedly happened that a shipment of the pure peroxide has reached its destination in the shape of empty bottles. Again, the peroxide readily takes up moisture from the air and undergoes hydrolysis. The crystalline contents become semi-liquid, strongly acid, and large crystals of benzoic acid may separate out. In either case the liquid contents fix the PAUL C. FREER AND FREDERICK G. NOVY. 123 glass stopper in place so that it can scarcely be removed without violence. Now, benzoyl acetyl peroxide is by no means inactive. It may explode when given a sharp blow, or when ground or scratched, or when heated, and several unfortunate accidents have occurred in this way. The objections to the use of the pure crystalline peroxide, referred to above, can be wholly obviated by making a mixture of equal parts of the reagent and infusorial earth. The fine powder thus resulting shows no deterioration by heat or moisture, and the danger of explosion is minimized. A high tem- perature, approximating that of boiling water, should however, be avoided. The dosage as given in the following pages refers to the pure peroxide, and where the fifty per cent powder is used, obviously the quantity taken must be doubled in order to secure like concentration. During the past two years benzoyl acetyl peroxide has been given extended trials in a varietj^ of diseases. It has been used as the pure crystal, as the fifty per cent powder, and in aqueous solution. The use of the crystalline peroxide was resorted to in purely experimental studies. In general medical work, for the reason already given, it is absolutely necessary to do away with the handling of the pure peroxide. The crystals or the diluted powder may be administered in the form of capsules in intestinal affections. The former have been given in three-grain capsules, every one-half or three-quarters of an hour, to normal individ- uals. In some instances the solution of the capsule in the stomach and the resulting concentrated local action led to considerable irritation, and to occa- sional vomiting. Under the direction of one of us (Dr. Freer^) double gelatin capsules containing 0.3 gram of benzoyl acetyl peroxide each were given in the beginning of the cholera epidemic in Manila. These were administered every four hours on an empty stomach, as when given on a full one they were Ukely to produce vomiting. Later on it was found expedient to substitute a smaller dose of 0.25 gram, to be given more frequently; the best results being finally obtained by the use of the latter, after coating with two layers of celloid- in. Similarly coated capsules containing 0.166 gram of the peroxide (2.5 grains) were also used. The object of coating with celloidin was to avoid solution of the capsules in the stomach in the expectation that this would occur in the intestines, in which case the germicidal action upon the cholera germs would be more marked. The capsules, however, were found to pass the small intestines undissolved, and were, in some instances, recovered from the feces, having lost most of the drug during their passage, and thus by osmosis the peroxide was gradually distributed along the length of the intestinal canal. The direct application of the peroxide crystals or powder is, as a rule, ' Preliminary Notes on the Preparation of Benzoyl Acetyl Peroxide and its Use as an Intestinal Antiseptic in Cholera and Dysentery. Report to the Secretary of the Interior, from the Bureau of Government Laboratories, Manila, 1902. 124 ON THE ORGANIC PEROXIDES. undesirable, owing to the severe local irritation which it will induce. Mois- ture, it will be understood, is essential to the hydrolytic cleavage of the per- oxide, and until this cleavage has taken place the germicidal action must necessarily be weak. The further tendency of the peroxide to combine with proteid matter, even before hydrolysis, explains why the direct application of the unhydrolyzed drug often fails to inhibit beginning diphtheria, local abscesses, and the like. One part of the fifty per cent powder with one hundred parts of boric acid gives a dusting powder, which has been found to give good results in nose and throat work, especially while the surfaces are still moist after a previous cleansing with Dobell's solution. In the form of an aqueous solution, the peroxide will be found to yield the best results. It will be remembered that on contact with water hydrolysis takes place with the production of the real active agents — benzoperacid and acetoperacid. Ordinarily, this change is slow, and may require several days. It may, however, be hastened by frequent shaking and by the use of warm, but not hot, water. The solubility of the peroxide, as such and unchanged, is about 1 in 1,543 parts of distilled water at 25° C. At 15°, and lower, the solubility is only 1 in about 3,000 parts; and on the other hand, when warm water, of 40° C. for example, is used, the solubility is materially increased. As a result of hydrolysis, the difficultly soluble original peroxide is changed into the very soluble peracids, and consequently, if the warm water is frequently shaken with an excess of the peroxide it is possible to obtain very strong hydrolyzed solutions corresponding to 1-500, or even 1 to 200 of the original substance. Freshly prepared solutions yield a deposit of inert benzoperoxide which, as well as the diluent powder, will eventually settle. If a perfectly clear solution is wanted the liquid may be decanted, after standing for a day or longer, or it may be filtered. The activity of the solution, however, is de- creased by passage through filter-paper. When solutions are prepared by thorough shaking with warm water (40°- 45° C.) the hydrolytic cleavage will be sufficiently rapid to allow the use of the solution within a few minutes. If cold water is used it is advisable to let the solution stand for some hours, or a day, with frequent shaking. The solution once prepared must not be expected to keep indefinitely. When kept cool the activity of such a solution may persist for weeks, but on the other hand, weak solutions, especially if kept warm, deteriorate after a few days or a week. This is due to the secondary hydrolysis of the peracids into hydrogen per- oxide, benzoic, and acetic acids. The strength of the solution to be used will obviously depend upon the work to be done. Thus, for spraying purposes, it is not advisable to use a solution weaker than 1 to 1,000. A 1 to 500, or even stronger solution, is pref- erable. On the other hand, an extremely sensitive urethra may call for a PAUL C. FREER AND FREDERICK G. NOVY. 125 1 to 3,000, or even weaker solution. In general, it may be said that the solu- tion should be used as concentrated as the patient can tolerate, and the appli- cation should be repeated as frequently as possible. Thus, spraying, gargling, swabbing, and in case of the ear, syringing, can be repeated hourly, whereas, rectal, urethral, and vaginal irrigations should be made at least several times a day. It is important to bear in mind that organic matter, in solution, or in suspension, tends to unite with the peroxide, and as a result the solution of the latter loses its activity. Hence, when irrigating the urethra, the bladder or rectum, it is advisable to first wash out the contents with a sterile 0.75 per cent salt solution, after which the peroxide solution should be introduced in as large an amount as possible, in order to secure utmost distension of the mucous surfaces. Subcutaneous injections of the peroxide in animals show no ill effects. They have also been used in man in special cases, with apparently very satis- factory results. Thus, Chisholm^ employed it with success in a case of malig- nant edema. According to Freer,^ a mixture of equal parts (half a liter each) of normal salt solution and of a peroxide solution (1 to 1,000) injected sub- cutaneously into cholera patients when in a state of collapse produced stimu- lating effects which were more direct and more lasting than when a salt solu- tion was used alone. The harmlessness of intravenous injections in animals has been already referred to (page 117). No such injections to our knowledge have been made in man, and it is probable that if made the germicidal action would not mani- fest itself, as in the case of surra, owing to the fixation of the peroxide by the proteids of the blood. The pronounced action of benzoyl acetyl peroxide upon the bacteria of the mouth has been discussed under the head of saliva (page 109), and in this con- nection it may be well to emphasize that the drug, when properly used, will undoubtedly yield good results in tonsillar infections. Frequent application of strong solutions, either as such or in the form of a spray, are indicated. Insufflation of the peroxide in the form of 1 part of the powder to 100 parts of boric acid has been satisfactory. The same considerations hold true for the nasal cavity, pharynx, and trachea. It may be noted in passing that good results have followed the use of the peroxide in pyorrhea, in otitis media, and in corneal ulcerations. On a 'priori grounds the 'administration of the peroxides in gastric fermen- tations could be expected to be beneficial, but we have no special data bearing on this point. In a few instances a marked relief has been obtained. In gonorrheal infections, especially in perfectly fresh cases, irrigation ' Therapeutic Gazette, Aug., 1902. ^ Locus citus, I, p. 12. 126 ON THE ORGANIC PEROXIDES. with large quantities of the solution has given satisfactory results. Owing to the sensitiveness of the urethra it is advisable to begin with a weak solution (1 to 5,000), and to gradually increase the concentration as the irritation subsides, till a solution of about 1 to 1,000 is given. Distension of the urethra by injections is likewise advisable. One case of gonorrheal ophthalmia was treated successfully. It is quite likely that in intestinal diseases, such as cholera, dysentery, and typhoid fever, benzoyl acetyl peroxide will prove of great use. For the reasons already set forth (page 115) it will be evident that a complete suppres- sion of the bacteria in the intestinal tract cannot be expected. A partial in- hibition, however, is indicated in the laboratory experiments, and the extended trials in cholera, and to a less degree in dysentery and typhoid fever, support this view. An opportunity to test the usefulness of the peroxide in cholera was afford- ed during the epidemic in Manila in 1902, and the results obtained may be briefly summarized.' The treatment consisted in administering the peroxide in two and one-half to five grain celloidin-coated capsules, in giving a 1 to 1,000 solution freely by the mouth, and in giving frequent rectal injections of large quantities of a 1 to 1,000 solution, either alone or diluted with an equal vol- ume of normal salt solution. In the state of collapse, subcutaneous injections of the latter mixture were used. The treatment was first employed at the San Lazaro Hospital, where the total number of patients was 120. Benzoyl acetyl peroxide was used exclu- sively after the expiration of the first five days. The total mortality in this hospital, including those not treated and those who died within six hours after admission, was 70 per cent, whereas, at that time, the cholera mor- tality in the city was above 78 per cent. In the Santa Mesa Hospital 186 patients were admitted, of which 152 died, or 81.7 per cent. Of these patients 93 received the benzoyl acetyl treat- ment and 26 recovered; whereas, of 29 who did not receive a peroxide treat- ment all died. The mortality with the peroxide treatment was 72 per cent as against 100 per cent with other treatment. Many of the latter, it should be added, were practically moribund at the time of admission to the hospital. The results at the Santiago Hospital were more definite. At the latter, 408 cases were treated either with benzoyl acetyl peroxide, or with benzoyl acetyl peroxide guiacol carbonate and calomel, and of these 169 recovered or 41 per cent. Five hundred and ninety-three cases received other treat- ment, and of these only 106 recovered, or 18 per cent. In another series, 136 cases were treated by the native physicians with their own methods and 49 recovered, or 36.02 per cent; whereas, 138 cases ' These data are taken from the report by Dr. Freer, who, as superintendent of the Government Laboratories, has been in Manila since September, 1901. PAUL C. FREER AND FREDERICK G. NOVY. 127 treated with peroxide by American physicians gave 61 recoveries, or 44.20 per cent. In Hospital No. 4, 24 cases were treated with peroxide, and of these 17 recovered, or 70.9 per cent. These cases, however, occurred at a time when the epidemic was less rigorous than at the beginning, but the percentage of recoveries is in excess of those in the other hospitals. In amebic dysentery, according to Dr. Strong, tlie results obtained with the peroxide treatment have been very encouraging. The number of cases, how- ever, was small. Out of eleven, two died, while the remaining nine were re- ported as showang no symptoms of the disease. No opportunity was afforded to test the, treatment on baciUary dysentery. We have had no personal experience with the treatment of typhoid fever by means of the peroxide. The reports which have been made, however, by Dr. Wasdin and others, indicate a marked degree of usefulness. It is to be hoped that extended and thorough trials will be made with the drug in this disease. CANCEE OF THE GALL-BLADDEE AND BILIAEY PASSAGES. WILLIAM J. MAYO, A.M., M.D., Surgeon to St. Mary's Hospital, Rochester, Minnesota. From June 24, 1891, to March 1, 1903, we had performed, for all causes, 501 operations upon the gall-bladder and biliary passages, and of this number, 22, or nearly 5 per cent, were for malignant disease. While this does not make it a common malady, it is sufficiently frequent to merit careful attention. The true proportion would, perhaps, be somewhat higher, as cancer of the gall-bladder, like malignant disease of most of the internal organs, is slow to be recognized and is often, if not usually, so far advanced as to render even an exploration unnecessary, the character of the trouble being only too mani- fest on physical examination. It is a question whether the relative number of cases in which cancer of the gall-bladder and bile passages is found and not subjected to opera- tion, would be more than five per cent of the number of gall-stone patients in which operation is refused, or the true nature of the condition is not recog- nized. Schroeder, after investigation of the subject, says that fourteen per cent of gaU-stone patients suffer at some time from cancer of the biliary apparatus. This at once brings up the query. Is there an etiological relationship be- tween gall-stone disease and cancer of the gall-bladder? Courvoisier found that seventy-fom: out of eighty-fo\u- cases of malignant disease of the gall- bladder had gall-stones. Siegert states that in ninety-five per cent of all cases of primary cancer of the gall-bladder, gall-stones are present, and adds this most significant fact, that calculi are found in only fifteen per cent of the cases of secondary malignant disease of this organ. Beadles, in the London cancer hospital, found that in fomr cases of primary cancer of the liver, gall-stones were present in all, and in thirty-six secondary liver cancers, gall-stones were not detected in a single instance. Musser gives the percentage of gall-stones in the cases of primary cancer of the gall- bladder which he found recorded as sixty-nine per cent, but says that in many of the instances in which no stones were discovered, it was probable that the calculi had passed. In all of the cases of cancer of the gall-bladder in our own series which were explored, gall-stones were present. Kelynack, in an ex- haustive examination of the subject of cancer of the gall-bladder and bile ducts, placed the proportion of primary cancer of the gall-bladder as seventy- 128 WILLIAM J. MAYO. 129 five per cent of the whole, giving twenty-five per cent as the number of pri- mary cancers of the bile ducts. Kelynack called attention to another note- worthy fact, that cancer of the gall-bladder was at least three times as frequent in women as in men, and this proportion is also true as to gall-stone disease. Siegert found seventy-nine females to fourteen males. Cancer of the gall- bladder is most common between the ages of fifty and sixty, which is the period of greatest frequency of gall-stones. Sutton states that columnar epithelium from the mucous membrane is the type of the carcinomatous process, but says that in most instances it could not be determined whether its origin was in the mucous glands or in the epithelium. It is to be noted that the cholesterin which forms the chief constituent of gall-stones is also a product of the mucous membrane. Butlin says that the cancerous ulcer is the most common form. Sutton has found general car- cinomatous infiltration of the gall-bladder, forming a hard, thick-walled tumor ^^•ith a small caA-ity containing stones, to be the more common variety, and with this latter group could be placed the cases which we have recorded. It will be observed that these two varieties which compose the majority of the reported specimens not only contained stones, but the primary lesions would suggest at once this probable source of irritation. Other gross forms of malig- nant disease of the gall-bladder are also found, particularly that type in which a tumor projects into the cystic cavity. The site of origin of cancer in the gall-bladder is usuallj^ near the fundus, although a considerable number have been discovered near the opening of the cystic duct. Perforation into the peritoneal cavity has rarely ensued, and general peritoneal infection taken place. The viscera may be found studded with small cancerous growths, causing ascites, or fistulous openings between the gall-bladder and the hollow viscera may occur. The most common method of extension is to the adjacent liver, either directly or through the blood-vessels, and to the lymphatic glands lying in the hepatic fissure. So far as we could find, there has been no recorded case of operation for primary sarcoma of the gall-bladder, although a few post-mortem specimens have been exhibited. Musser found three such cases recorded in the literature. The following indisputable facts attract attention: First, gall-stones are almost constantly present in primary malignant disease of the gall-bladder, and rarely in secondary; second, the relative proportion of gall-stone and malignant disease of the gall-bladder in women and men is practically identical; third, the pathological lesions found are best explainable on this theory; and, fourth, the similarity in age frequency. Certainly we are warranted in con- cluding that gall-stones are the most important etiological factor in malignant disease of the gall-bladder. As the proportion of cancerous disease involving the gall-bladder and bile tract to simple gall-stone disease was as one to 130 CANCER OF THE GALL-BLADDER AND BILIARY PASSAGES. twenty in 501 operations which we have performed upon the gall-bladder and bile passages, the question assumes a practical importance ; and while I woiild not wish to say that for this reason alone gall-stones should be removed, it certainly aids in deciding that early removal of active gall-stones, other things being equal, is sovmd svirgery, particularly so as nearly all the mortality-giving complications are the result of delay. In over 250 uncomplicated gall-stone operations, the mortality was less that one per cent. The diagnosis of primary cancer of the gall-bladder may be easy. As a rule, a hard tumor is to be detected in the region of the gall-bladder which is not very tender to the touch, and unless there is a peritoneal involvement, rigidity of the overlying muscles is not marked. There is progressive loss of flesh and later a cachexia is developed. A nodular tumor becomes apparent as the liver is involved, and jaundice from extension to the common and hepatic ducts in the later stages may occur. A previous history of gall-stone disease can usually be obtained. Naunyn says that at least one-half of the cases of jaundice diagnosticated as due to gall-stones are caused by or complicated with cancer. In our experience, this is too high a percentage. The jaundice, when it exists, is persistent and unchanging, and if a tumor is present, which is accompanied by a loss of flesh which appeared before the jaundice, there can be but little doubt as to the malignant nature of the trouble. Courvoisier long ago pointed out that stone in the common duct was accom- panied by a contracted gall-bladder in eighty-four per cent of cases, and in the latter disease, the jaundice in the early stages is intermittent and often accompanied by occasional chills and fever, due to an infective cholangitis. We have seen a cystic gall-bladder of long standing, due to stone impacted in the cystic duct, give rise to a hard tumor resembling malignant disease, and occasionally such a tumor may result from an adherent gall-bladder shrunken down upon a mass of stones. The local tenderness and history of sudden appearance with the general condition of the patient will usually be sufficient to differentiate. It sometimes happens that even on the operating table it is impossible to say whether a thick-walled gall-bladder is malignant or not. Robson reports a case in which no fluid was found on puncture ; the hardness of the enlarged gall-bladder and the bad general condition of the patient led him to the conclusion that it was malignant disease. Drainage was established and permanent recovery followed. Abb6 reports an almost identical case. The opposite mistake may also be made. The following case occurred in St. Mary's Hospital: Case I. — Cholecystotomy for gall-stone disease. Unsuspected malig- nancy. Mrs. J. R., age 64, admitted to the hospital September 9, 1901. History. — For some years patient had suffered from typical gall-stone colics. In September, 1900, had a very severe attack during which her physician, Dr. Slocumb of WILLIAM J. MAYO. 131 Pl'ainview, Minnesota, could distinctly feel the enlarged gall-bladder. This was suddenly relieved and the tumor disappeared. She has lost a great deal in weight and is in constant pain. Physical Examination. — Patient cachectic, slightly jaundiced, liver dullness indefinite, perhaps slightly enlarged. No tumor could be felt on palpation. Nothing else of impor- tance was noted. From the history a diagnosis of gall-stone disease was made. On Sep- tember 10, 1901, through the usual incision, an exceedingly adherent and very thick-walled gall-bladder was exposed. In the small central cavity a few gall-stones were found. One stone was impacted in the cystic duct. The patient was in bad general condition and drainage of the gall-bladder was instituted. Cholecystectomy would otherwise have been done. No suspicion existed that the gall-bladder was cancerous. The condition was supposed to be due to inflammatory thickening. Patient discharged in three weeks. The relief obtained from the operation was but temporary. The pain re- tm-ned and a tumor soon afterward appeared and rapidly enlarged, followed later by jaundice and death about four months after the primary operation. Cholecystectomy is gradually displacing drainage in the class of cases ac- companied by marked changes in the wall of the gall-bladder, and it is probable that manjr early cases of cancer will in this way be discovered and removed. Practically all of the unoperated cases die within the year, although a few have been reported in which life was prolonged ton early two years. Musser says that only a few months of life are to be expected, and that this has not been prolonged by operation in the majority of the reported cases. The first removal of the gall-bladder for malignant disease took place in 1887. The number of cases reported is small. Robson had four personal cases of which one died as a result of the operation. Butlin collected sixteen cases operated upon with twelve recoveries, and of these, nine died from, recurrence on inefficient removal within four months. The following two cases are of interest in this connection : Case II. — Cholecystectomy for cancer of the gall-bladder. Recurrence after one j^ear. Mrs. M. H., age 63, American, housewife. Widow. Admitted to St. Mary's Hospital, June 12, 1901. History. — For some years has had occasional attacks of cramps in the stomach lasting from a few^ minutes to several hours and necessitating the exhibition of morphia for relief. For three months there has been a steady boring pain in the region of the gall-bladder extending through to the back. There has been considerable loss of flesh. In the early history some transient attacks of jaundice followed the spells of pain, although the icterus was never marked. Personal and family history otherwise negative. Physical examina- tion revealed nothing beyond a deep-seated resistance in the region of the gall-bladder. Diagnosis, gall-stone disease. Operation, June 15, 1902. A thick-walled gall-bladder, densely adherent and containing a few smaU gall-stones in its nearly obliterated cavity. The 132 CANCER OF THE GALL-BLADDER AND BILIARY PASSAGES. appearance was suspicious and the gall-bladder was dissected out and drain- age established. Discharged in three weeks. Unfortunately the speciraen was lost and it was not until the patient presented herself with the subse- quent history that the diagnosis was established. She had remained well for over a year. July 2, 1902, began to have a return of the pain and a jaundice developed. An irregular, deep-seated tumor at the site of former operation could be discovered on deep palpation. A manifest recurrence. Case III. — Cholecystectomy for cancer of the gall-bladder. Recovery. Mrs. T. MoD., Irish, married, housewife, age 56. Admitted to St. Mary's Hospital, June 19, 1902. History. — For several years has been subject to attacks of cramping pain in the stomach lasting from a few minutes to several hours. For the past year these attacks have in- creased in severity and for several months the pain has been more or less continuous, ac- companied by loss of flesh and some jaundice. The latter symptom has been variable but never entirely absent. Otherwise, history negative. Examination. — Patient shows loss of flesh and strength, somewhat jaundiced; pulse 110, temperature 97J. Trace of albumin and considerable bile in urine. Liver dullness increased, no tumor in region of gall-bladder, tender to pressure in the epigastrium and right hypochondriac region. Diagnosis. — Gall-stones with probable stone in common duct. Operation, June 21, 1902. Straight incision through right rectus muscle, gall-bladder buried in a mass of adhesions. It is of small size with great thickness of the walls. Stone obstructing cystic duct at its juncture with the common duct, apparently compressing the latter. Appearance of gall-bladder suspicious. Cholecystectomy with hepatic drainage through cystic duct. Discharged in twenty-four days. Specimen shows carcinomatous infiltra- tion. In a number of instances the adjacent liver has been involved and yet the diseased process was sufficiently localized to permit of resection of the infil- trated parts. The surgical treatment of such tumors has been thoroughly reviewed by Keen, Ferrier, and others and personal cases reported. When complicated with jaundice, the mortality is very high. The following case illustrates the condition as found in a more advanced stage of the disease. Case IV. — Cholecystectomy and partial hepatectomy for cancer of the gall-bladder involving the liver. Recovery, followed by recurrence. Mrs. E. R., American, age 65. Admitted to St. Mary's Hospital of Rochester, Minnesota, April 18, 1900. History. — She has been in her usual health until within last six months. During this time she has suffered from a boring pain in right side, which of late has become almost constant. Stomach symptoms have been of moderate severity. There has been some loss of appetite and constipation, with a decrease of fifteen pounds in weight. No jaundice, nor history of colics. Examination reveals a somewhat movable tumor in the right hypo- chondriac region evidently connected with the liver. The mass has a nodular feel. WILIJAM J. MAVO. 133 Exploratory incision, April 21, 1900. A carcinomatous gall-bladder in- volved the adjacent portion of the hver and the cy.stic duct. There was some infiltration along the common duct and extending to the duodenum, upon which at one place was a considerable area of adhesions. A few glands in the angle between the cystic and hepatic ducts w(>r(> infected. The disease was so definitely circumscribed with such slight glandular involvement that its re- moval was decided on. The excision was begun at the common duct, two inches of which were removed with one-half inch of the hepatic duct. The vessels were caught and tied as divided ; an area of adherent duodenum the size of a silver half-dollar was included in the excision. Tlie opening in the intestine was closed by circular purse-string suture. The lower end being thus free, the gall-bladder -ndth the attached liver was removed with the Pacquelin cautery knife. The larger vessels were grasped with forceps. The free venous oozing from the liver substance was not controlled by the cautery, although easily checked by slight pressure, the blood current being of httle force. A piece of sterile gauze the size of the wrist was placed in the cavity, and a continuous suture of fine catgut was run through the liver substance on each side of and around the gauze, compressing the bleeding liver margins against it, and controlling the hemorrhage efficiently. The portal vein was exposed to a considerable extent in the bottom of the cavity. Adequate drainage was afforded, the bile being conducted to the surface. Recovery was uneventful. The gall-bladder contained a single stone three-fourths inch in diameter. Patient died four months after operation from recurrence. So far as I have been able to find, aU of the cases in which the liver was in- volved secondary to primary cancer of the gall-bladder have had early recur- rence. This is also true of the cases in which the lymphatic glands in the fissure of the liver were infected. Jaundice occurs in about one-half of the cases, and as it is a late symptom, operation is contraindicated. The increas- ing frequency of operation for gall-stone disease will lead to the accidental discovery of cancer in an early stage and the results of operative interference wiU vastly improve. Cancer of the Common Bile Duct. Primary carcinoma of the common duct is rare. Kelynack in 4,578 autop- sies found eight cases of primary cancer of the gall-bladder and but two having origin in the duct. Musser collected one hundred cases of carcinoma of the gall-bladder and eighteen of the bile ducts. The site of the neoplasm in the common duct is usually either at the junction of the hepatic and cystic ducts or near its duodenal termination. The eighteen cases collected by Musser showed three in the hepatic duct and foin-teen in the common duct, and of these latter, nine were at or near the papilla. Rolleston in seventeen cases of cancer of the ducts found fifteen in the common duct and ten of these at or near the papilla. 134 CANCER OF THE GALL-BLADDER AND BILIARY PASSAGES. As to the etiology of carcinoma of the common duct, there is some question. It must be conceded that gall-stones are the most common cause of cancer of the gall-bladder. In Musser's one hundred cases of gall-bladder carcinoma, sixty-nine contained gall-stones and good evidence that calculi had at one time been present in the majority of the remainder. Primary cancer of the gaU-bladder and gall-stone disease are more common in females than in males, and in about the same proportion. This is not true of duct carcinoma. Mahgnant disease of the common duct is equally frequent in the male and female, which does not favor the theory that gall-stones are the cause of the malignant process in the duct. Edes in twenty-two collected cases of cancer at or near the papilla of the common duct found gall-stones in but four, and three of these in the gall-bladder. RoUeston in thirty-six cases of cancer of the common duct found gall-stones in less than half. The extensive experience of Mayo Robson, however, entitles his opinion to great weight, and he states his belief that gall-stones are the most common cause of malignant neoplasms in the biliary passages, although the calculi have not remained in situ. The histological variety of carcinoma of the ducts is always of the columnar cell type, although Robson says that secondary degeneration of papillomata occurs. Systemic infection is rare; the growth usually progresses by con-, tiguity and sooner or later the lymph glands of the gastro-hepatic omentum are involved. In some of the cases reported, the growth was very small at autopsy, notably the case reported by Edes, in which it was not larger than a bean, even after a year or more of marked symptoms. Death usually occurs from debility, the result of the jaundice and infection of the biliary ducts. The symptoms are not distinctive and the diagnosis cannot often be made. The chronic jaundice and cachexia are not dissimilar to malignant growths of the head of the pancreas, and the occurrence of glycosuria and fatty stools is not sufficiently common in the latter disease to aid differentiation. In primary carcinoma of the common duct, pain is not usually severe, in this respect differing from stones in the same situation, but as so many cases have at one time had stones, the diagnostic importance of pain as a symptom is not great. There is usually no tumor present, although not infrequently the distended gall-bladder can be palpated below the margin of the liver, but not the distinct hard tumor of cancer of the gall-bladder, a sign commonly existing in cancer of the gall-bladder. The question of jai;ndice is an impor- tant one. In the beginning it is often intermittent, very like stone in the common duct. Rarely this will lead to an infection with the fever and chills of cholangitis. In the later stages, jaundice is complete. Maury reports a recent case with the typical symptoms of Hanot and Rendu, early inter- mittent jaundice, insidious onset, and diarrhoea with later complete ob- struction of the common duct. An exploration of the ducts in doubtful cases is the only way a positive diagnosis can be established. WILLIAM J. MAYO. 135 McBurney first called attention to the eiusc with whicli the duodenum could be opened for the purpose of renioving stones impacted in the diver- ticulum of Vater, and first performed the operation. Wv have several times successfully opened the duodenum for this purpose, and a large number of such operations are now on record. Carle in an address before the Italian Surgi- cal Congress strongly urges incision of the duodenum for removal of stones or growths from the duodenal end of the common duct, and cites cases in which stones were formed in the duct and later might give rise to carcinoma. Cancer higher up in the duct would necessitate union between the remaining fragment of the duct and the duodenum, as Halsted succeeded in doing in his case. As a palliation, cholecystenterostomy is the indicated procedure. The anastomosis may be made either between the gall-bladder and duodenum, or, if the latter is involved, with the transverse colon or jejunum. We have four times joined the gall-bladder to the transverse colon for inoperable obstruction of the common duct or clironic pancreatitis, and these cases did fully as well in every respect as six cases in which the duodenum was used as a receptacle for the biliary discharge. One case, supposed to be malignant, proved not to be so by living in good health six years after uniting the gall-bladder and colon. As a result of our own experience, I see no reason why the transverse colon may not serve as well as the duodenum, provided the latter, more favorable situation is not practicable. The proximity of the large bowel and the nature of its coats render anastomosis with the gall-bladder easy, and in palliation of malignant disease it is, perhaps, almost as good for the purpose as the duodenum. In the original work of Winiwarter, it was the chosen method. A few cases of enormous distention of the common duct have been reported. Robson details an instance in which he had been able to suture such a cystic formation to the surface of the body. Summers, in a most interesting case, united the common duct to the duodenum with a successful result. In these cases the obstruction was, however, non-malignant. The following case came under our observation : Case V. — Carcinoma ampulla of Vater. Excision. Recovery, followed by recurrence and cholecyst-duodenostomy. M. K., female, age 59, German. Admitted to St. Mary's Hospital, Rochester, Minne- sota, November 1, 1900. History. — For many years patient has suffered from sudden attacks of pain arising in the epigastric and extending to the right hypochondriac region. The suffering has been very severe, lasting from two to six hours and ending with an attack of vomiting accom- panied by prostration. At times has been somewhat jaundiced after these attacks. About one year ago appetite began to fail, distress in the stomach became more constant but less severe, and there has been a progressive loss of weight, over forty pounds in all. Family and personal history good. Examination. — Patient somewhat emaciated; there is a marked cachexia with a moderate jaundice. Pulse, temperature, and respiration normal. Liver can be outhned 136 CANCER OF THE GALL-BLADDER AND BILIARY PASSAGES. just below the free margin of ribs, gall-bladder cannot be felt. There is a tenderness and some rigidity of the muscles in this region, otherwise examination negative. Urine has a trace of albumin and much bile. Test meal developed free hydrochloric acid, and on distention with air the outlines of the stomach were found to be normal. Stools contained traces of bile, but were light colored. The history was clear as to the presence of gall- stones, but the patient had a distinctly cachetic look. Diagnosis. — Either gall-stones in the common duct or malignant disease. November 3d. Incision through the right rectus muscle. Liver somewhat larger than normal; gall-bladder enlarged, containing bile mixed with ropy mucous and a single, non-faceted dark-colored stone, the size and shape of a small pea. The cystic and common ducts were moderately dilated, but no stone nor other obstruction could be detected on roost careful exploration. Gall-bladder drained to the surface after attaching to the parietal peritoneum. The findings at the operation were unsatisfactory and did not account for the condition of the patient. For forty-eight hours drainage of bile was free but gradually increased in quantity up to two or more pints a day, the skin became greatly irritated from the discharge, and examination showed that a large part, if not all, of the pancreatic secretion was being discharged to the surface with all of the bile. Stools now contained no bile. A Jacob's self-retaining female catheter was inserted into the gall-bladder through the fistulous opening, and in this way the drainage was directed into a receptacle without contact with the skin. It was evident that there was an obstruction which had been overlooked at the duodenal extremity of the duct. Patient was in a very feeble condition and on November 30th was allowed to return home. Even with the continuous drainage she improved somewhat. The jaundice disappeared and on January 29, 1901, was readmitted to the hospital. On January 31, 1901, an incision four inches in length was made to the inner side of the fistula. The adhesions were separated and the com- mon duct and duodenum thoroughly exposed. At the extreme end of the common duct a hard body could be felt through the wall of the duodenum, the size of a filbert, and was supposed to be a stone lodged in the ampulla of Vater. An incisioa was made two inches in length in the anterior wall of the duodenum, exposing a grayish-white mass which was strictly localized to the site of the papilla of the common duct. Its size did not exceed the end phalanx of the forefinger. About one-third of its length projected into the free lumen of the duodenum, and two-thirds posterior to the intestinal wall. The tumor was excised, exposing the free end of the common duct. The re- moval was made partly with a knife and partly with the Pacquelin cautery, and finally the whole raw surface was seared with the cautery. The common duct was otherwise free of obstruction. The incision in the duodenum was sutured. No enlarged lymphatics could be detected. A small drainage wick was inserted and the wound closed. The attachment of the gall-blad- der to the skin was left undisturbed. The discharge from the fistula rapidly WILLIAM ,1. MAYO. 137 diminished and in three weeks had ctjmpletely ceasctl. Stools became normal in color, and the gain in the weight and general app(>a,rance was most rapid. For nearly a year and a half the patient remained well. About the middle of June, 1902, she began to notice sonie pain of a boring character in the epi- gastrium which was soon followed by jaundice which slowly progressed, and on July 7th the abdomen was opened in the region of the former incisions. Gall-bladder found distended. On opening the duodenum it was discovered that the growth had returned at the site of former operation. Enlarged lymphatics were present a.id also deep attachments to the pancreas. Chole- cyst-duodenostomy was performed with the Murphy button, with recovery. The specimen was examined by Dr. LeCount, Assistant Professor of Pathology, Rush Medical College of Chicago, and his report is as follows: "The tissue is from the duodenal wall and some sections show portions of Brimner's glands. The glands of Lieberkiihn may be traced to lower depths than normal through a very inflammatory mucosa that contains a few small lymph nodes and small areas of hemorrhage. Certain of these glands are directly continuous with groups of epithelial cells that lie deeply within the mucosa and the muscular coats. The epithelium in these invasions is altered as foUows: The cells lose their columnar shape, become possessed of larger and more deeply stained nuclei, possess karyokinetic nuclei in many instances, and do not retain their characteristic grouping, being, instead, arranged in disorderly clumps and bunches that vary in size ; these deeply lying collections of epithelial cells always possess an irregular cavity that simulates a gland of the simple tubular type or a gland duct. One must conclude that this is the tissue from a cylindro-cellular carcinoma. " So far as the writer can ascertain, the only case in which a carcinoma of the common duct has been excised previously to the one herewith reported was that of Wm. S. Halsted of Baltimore, published in the Boston Medical and Surgical Journal, Vol. CXLI, December 28, 1899, No. 26, page 645. THE FUNCTION OF THE SUPRARENAL GLANDS, AND THE CHEMICAL NATURE OF THEIR SO-CALLED AC- TIVE PRINCIPLE. JOHN J. ABEL, M.D., Professor of Pharmacology, Johns Hopkins University. 1. A Brief Summary of the Experimental Evidence in Proof of the Statement that the Suprarenal Capsules are Functional Organs. Although the suprarenal capsules have been known to medicine since 1563, when Eustachius first described them, their exact function is still shrouded in mystery. In 1855 Thomas Addison,^ with his careful scrutiny of clinical facts, his sound and dispassionate judgment, described the constitutional and local effects of disease of these organs and thereby added Morbus Addisonii to the list of maladies whose proximate causes are known. A j^ear later Brown- Sequard performed experiments proving that removal of these glands had a rapidly fatal result. Great progress in matters of detail has been made since the day of Addison and Brown-Sequard.^ Chemists, anatomists, histologists, physiologists, clini- cians, and pathologists have all made their contributions to our knowledge of the suprarenal capsules. How much light do the new facts throw on the function of the gland? Have we advanced so far beyond Addison and his contempo- raries that we can venture to assign to these structures their precise function and relative influence in the economy? The experiments of Brown-Sequard in 1856 led him to the conclusion that the removal of the glands from mammals is rapidly fatal, death resulting in about eleven hours, and that they are essential to life ; that the blood of such animals becomes poisonous to other animals of the same species, and that these organs have the office of removing toxic substances from the blood. Brown-Sequard 's conclusions were soon attacked,^ at first with apparent suc- cess. But in later decades, as a result of improvements in the technique of the 'Dr. Addison's Works, New Sydenham Society (1868), page 211. ^Compt. rend. Acad. d. So., Paris, 1856, pages 422 and 542; Arch. g^n. de m6d., Paris, 1856; Journ. de la physiol. de I'homme, Paris, 1858, Vol. I, page 360. * Philippeaux, Compt. rend. Acad. d. Sc, Paris, 1856; George Harley, British and For- eign Medico-Chirurgical Review, 1858, Vol. XXI, pages 204 and 498, and others as cited in the Goulstonian lectures on the Suprarenal Bodies by H. D. RoUeston, Brit. Med. Journal, March 23d to April 6, 1895, and in the comprehensive monograph of Hult- gren and Andersson. 138 JOHN J. ABEL. 139 operation and in consequence also of the recognized importance of tlie accessory- glands, his original assertion that these organs are n(>c(\swar3' to life has been verified. I need only name Tizzoni/ Langlois,^ liultgren and Andersson,' Boinet/ Gom-fein,^ Thiroloix," Strehl and Weiss/ Schafer,* Vincent," and Szymonowicz^" among the more recent in\'e,stigators who have established this point with certainty. Many species of animals have been used for double-sitled extirpations; among mammals, dogs, cats goats, rabbits, weasels, guinea-pigs, rats, mice, and porcupines; among amphibians, frogs; and among bony fishes, eels. With the exception of the eel, the suprarenal gland of these animals is a dou- ble organ, consisting, according to embryologists, of one organ, the so-called medulla, inclosed within another, the so-called cortex. As Guieysse" has put it, we are dealing with two tissues, which differ in their histological appearance, in their reaction toward chemical reagents, and in their mode of origin. Aichel,^^ one of the latest investigators in this field, holds the opposite view, that cortex and medulla have their origin in one and the same embryological tissue, and future investigators may possiblj- corroborate this conclusion. In detail the following points have been established : 1. If one gland only is removed the animal survives and no symptoms follow; the remaining organ, however, increases in size, showing a true physio- logical compensatory hypertrophy. In 1889 Stilling^^ performed a large number of experiments on rabbits, conclusively proving this point. This ten- dency to compensatory hypertrophy is so great that when in one instance both organs were removed except a small piece on the vena cava not larger than a pin's head, this was found after thirteen months to have grown to nearly the size of the original gland, being 11 mm. long, 7 mm. wide, and 5 mm. thick. StiUing's results have been repeatedly verified by more recent observers. Thus, Strehl and Weiss" in their recent paper state that they measured the intact capsule at the time of the removal of the other and found iZiegler's Beitruge Bd. VI (1889), page 1. ^'Compt. rend. Soc. de Biol., 1891, 1892, 1893, etc., Les capsules surr^nales, Paris 1897. 'Skand. Arch. f. Physiol., Vol. IX, page 73. st of the ileum. An abscess formed and broke into the rectum. She was ill for two years, but finally recovered. Five years ago she first noticed a small tumor the size of a hen's egg, low down on the left side of the abdomen. It grew slowly. Four years ago it had extended to the right side. At this time the mass moved on change of posi- tion. There was no pain from the growth, the only uncomfortable symptoms being extreme constipation, a slight irritability of the bladder, and a tendency to vomit when she stooped and the mass pressed on the stomach. Her gen- eral health has always been excellent with the exception of occasional attacks of distressing palpitation of the heart. Exam-tnation showed the abdomen filled with a hard, smooth, solid tumor partiallj^ divided into two lobes by a longitudinal sulcus, starting five inches above the pubes and extending to two inches above and one and one-half inches to the right of the umbilicus. Just above the pubes there was no break in the continuity of the growth. The left lobe was larger than the right and extended upward to the border of the ribs and laterally to within one and one- half inches of the left anterior superior spine of the ileum. The right lobe extended well into the flank, its upper border being one and one-half inches from the border of the ribs. There was dullness on percussion over the area outlined. The mass was only slightly movable and was non-sensitive. Vaginal examination revealed an atrophied cervix pushed to the left. Two small masses, one on either side of the cervix at the supravaginal junction, could be made out. Each was the size of an English walnut, hard, smooth, and non-sensitive. The sound passed two and one-half inches and the uterus seemed to be directly continuous with the tumor mass above, moving when it was moved. Diagnosis. — Uterine fibromyoma. Operation. — December 24, 1901. Median abdominal incision from one inch above the pubes to three inches above the umbilicus. Upon opening the peritoneum, a small amount of clear, yellowish ascitic fluid escaped. A whitish, shiny, fibrous tumor with two lobes presented, the smaller, somewhat softer and edematous, being to the right. Owing to its size, the tumor was delivered with some difficiilty, although there were no adhesions. The peritoneum ap- peared perfectly normal. Upon inspection, it was seen that the mass was a tumor of the left ovary, the tube, somewhat atrophied but of normal appear- ance, being separate from the tumor. The broad ligament formed the pedicle, which was thin and long enough to allow of the withdrawal of the tumor through the abdominal incision. The mass was entirely distinct from the uterus, the round Hgament having its proper relations to the latter. The right ovary was atrophic but otherwise normal, as was also the right tube. The appendix was normal and showed no signs of previous inflammation. The uterus was atrophic, and from its lateral and posterior surfaces were peeled three small 168 A CONSIDERATION OF OVARIAN FIBROMATA. fibroids, each about the size of pullets' eggs. The patient's convalescence was complicated by an irregularity of the pulse, otherwise it was uneventful. Pathologic report by Dr. A. S. Warthin : / '-^x X Fig. 1. Fibroma of Left Ovary. (Cross-Section.) Gross Appearance. — The tumor is a nodular mass, on the whole rather flat- tened, 25 cm. in its greatest length by 20 cm. in its greatest breadth (Fig. 1). On the under surface near the anterior border is a hilum-like groove in which there is attached, bj' a broad pedicle, a second nodular mass 20 cm. long by 10 cm. broad. The tumor weighs seven pounds four ounces. The REUBEN PETERSON. 169 greatly atrophic tube and broad ligament are attached to the main mass on the anterior sm-face near the upper third, running diagonally outward and downward toward the middle third of the anterior surface of the growth. The tumor has rather a dense capsule in which there are numerous dilated vessels. The upper part of the smaller mass contains a cyst the size of a hen's egg, filled with clear serous fluid. Smaller cysts are scattered over the surface of the smaller growth. The consistency of the larger growth as felt through the capsule is very irregu- lar, in places hard and bone-like, in others softer and more elastic. Different Fig. 2. Edematous Nodule of Ovarian Fibroid. (Cross-Section.) parts of the tumor can be made to move on others as if the growth were formed of separate nodules. On section the growth presents the appearance of a hard desmoid fibroma, cutting with great difficulty and creaking under the knife. The surface for the greater part is white and tendon-like, showing irregularly disposed whorls of coarse fibrils interspersed with softer, pinlcer areas. Over the cut surface are a number of ectatic blood-vessels, which are larger and more numerous in and belowthe capsule. Here and there over the cut surface are small nodules of softer areas, more pink in color, probably growing centers. There are a few small areas of calcification. No remains of ovarian tissue were found. Smaller Mass. (Fig. 2). — This is much softer and more elastic than the larger mass. On section it shows mottled red and yellowish-white. The surface is made up of irregular whorls of coarse fibrils. The cut surface is very edema- 170 A CONSIDERATION OF OVARIAN FIBROMATA. tous, exuding rather thick yellow serum. Throughout the surface are a number of small degeneration cysts filled with similar fluid. The large cyst at the upper pole is a multilocular cyst consisting of a num- ber of smaller cavities filled with clear yellowish fluid. The majority of these cyst spaces have a distinct smooth cyst wall; others are apparently only de- generation cysts in the tumor tissue. The blood-vessels of the mass are every- where deeply congested. No ovarian tissue can be found. The small nodules from the uterus are hard fibromata, one being almost calcified. Pathologic Diagnosis. — The large mass is a fibroma durum; the small mass is a soft edematous fibroma with areas rich in cells, probably growing centers. Case II. — Mrs. H. S., widow, aged 61, was admitted to the medical side of the University of Michigan Hospital on account of dropsical condition of the abdomen of four years' standing, accompanied at times by abdominal pain and chills. Her trouble began with an extreme soreness across the ab- domen which she attributed to a fall received about this time. Eight months later she had neuralgic attacks of pain in the back with nausea and vomiting. One month later she was tapped and fifteen quarts of clear yellow fluid with- drawn. She has since been tapped sixty-four times and varying amounts of fluid removed, averaging about twelve quarts at each tapping. Examination showed, besides ascites, an irregular tumor of pelvic origin, for which she was referred to the gynecologic service, April 25, 1902. Her menstruation had ap- peared at the age of fourteen, and had always been regular and without pain. The menopause occurred twelve years ago. She had been married thirty-four years and had borne six children, the eldest thirty-one and the youngest nine- teen. She also had had two abortions. Examination showed the abdomen occupied by a hard, irregular mass ex- tending in the median line from the pubes to within one and one-half inches of the umbilicus. To the right it extended nearly to the pelvic wall, its upper border being two inches below the level of the anterior superior spine of the ileum. On the left, the upper border was three inches below the level of the anterior superior spine. The mass was made up of a central and two lateral masses. It was somewhat movable and could be pushed upwards in the ascitic fluid shown by percussion to be present. Biinanual examination was unsatis- factory, because of sensitiveness. Under anesthesia, just prior to the operation, the uterus was found to be forward and two and one-half inches deep. It seemed to be continuous with the larger mass of the tumor and moved when it was moved. Diagnosis. — Fibroid of ovary or uterus. Operation. — ilarch 14, 1902. Upon opening the peritoneum in the median Une, about 700 c.c. of thin, reddish, ascitic fluid escaped. The irregular tumor mass was seen to spring from the left side of the pelvis. It was adherent to the lower portion of the pelvic and bladder peritoneum by easily REUBEN PETERSON. 171 loosened adhesions. The pedicle was long and thin and coiuposed of the left broad ligament, the normal left tube lying along its lower surface. The peritoneum was normal in appearance. The appendix was normal in size and position. The right ovary was the seat of a papuliferous cystoma the size of a hen's egg, and together with the tube was rcmo^•('d. The liver and gall-bladder were palpated, but no gall-stones were disco^■ered. The patient's convales- cence from the operation '\\ as une^•entful, except for a sharji attack of gall-stone Fig. 3. Fibroma of Left Ovary. cohc and the passage of four stones. Bacteriologic examination of the ascitic fluid showed staphylococcus pyogenes am-eus in a much attenuated form. Pathological report of Dr. A. S. Warthin: Gross Appear ance.^The tumor is an irregular, nodular mass, to which the atrophic Fallopian tube and a portion of the broad ligament are attached to its upper pole (Fig. 3). The lower portion of the tumor is an irregular nodule about the size of a cocoanut, the upper portion consisting of irregular smaller nodules fused together. The greatest length is 13 cm., the greatest width 11 cm., and the thickness 10 cm. The growth has a distinct thin capsule, over which are a number of stringy adhesions. Beneath the capsule are numerous distended veins and small ecchymoses. The consistency is very firm and elastic. On cutting, it creaks under the knife. There are a few small grating points, and calcification is present. The cut surface shows a structure of bundles of coarse, dense, white fibers, the bands running in different directions 172 A CONSIDERATION OF OVARIAN FIBROMATA; (Fig. 4). The color of the cut surface for the greater part is white, hke a ten- don. It has scattered bands and patches of dark red and a number of light, yellowish areas, the latter being softened and edematous. The blood supply is very poor. Fig. 4. Fibroma of Left Ovary. Microscopic Examination. — Fibroma durum, with areas very rich in cells, probably growing centers, areas of hydropic degeneration, edema and simple necrosis. Few small points of calcification. It has been thought best not to complicate the subject by classifying sepa- rately fibromata and myomata of the ovary. From a pathologic standpoint there can now be no doubt of the existence of true myoma of the ovary, but, as pointed out by Doran (33) there is at present no way clinically of differenti- ating myomata from fibromata; hence our purposes will best be served by considering the two growths together. Frequency. — ^While a fibroma is one of the rarest forms of ovarian new REUBEN PETERSON: 173 growths, it is by no means such a rarity as was formerly supposed. A com- mittee appointed to examine a tumor of the right ovary exhibited by Nunn in 1857 before the London Pathological Society as a fibroma, reported against the probability of its being of this nature^, on the ground that they w(n-e un- acquainted with any ovarian fibrous growths which had reached anything Uke its dimensions. The work of Doran^, Briggs'", Coe™ and others has shown conclusively that the disease, while rare, is common enough to necessitate consideration from the standpoint of diagnosis. Briggs' experience is some- what remarkable in that he met with eight cases in his own practice, and my two cases of large oA-arian fibromata were operated upon within four months of each other. Recently there have been reported a number of cases of ovarian fibromata. Fairbain, early in 1902, showed before the London Obstetrical Society five specimens of fibroid tumors of the ovary, varying froni a small growth the size of a hen's egg to a large tumor of over four pounds in weight. In the discussion eh cited by the reading of a portion of the present paper before the American Gjmecological Society, it was brought out that many of the mem- bers had met with timiors of this description in their practice, but had never reported the cases. It is fair to conclude that the large proportion of these growths were fibromatous, as their subsequent histories showed no return of the disease. Age. — It is interesting to consider the age of the patients with fibromata of the ovary. As we shall point out a little later, it is extremely rare to be able to determine definitely the beginning of fibromatous ovarian changes. The patients come under observation usually after the growth has reached a con- siderable size or has given rise to definite symptoms. Hence, age in relation to these growths must be computed from the time of the operation or the ap- pearance of symptoms which have comppUed them to seek the services of a physician. Dartigues*', who has wTitten the best monograph on solid tumors of the ovary, including fibromata, finds that in twenty cases of this disease coming under his observation, or which he has collected from literature, there were six cases during each decade, from twenty to thirty, thirty to forty and forty to fifty, while only two patients were over fifty. This, in a way, is con- firmatory of the prevalent opinion that ovarian fibromata occur as a rule in much younger women than do uterine fibromyomata. The latter are found most frequently in women from forty to fifty. Schroeder'" in 798 cases found fifty-one per cent occurring during this decade, as against 28.6 per cent during the decade from thirty to forty. My own statistics show that the age was mentioned in eighty-three out of the eighty-four cases collected. The following table shows the number of cases according to decades and the percentages : 174 A CONSIDERATION OF OVARIAN FIBROMATA. Years. Number of Cases. Percentage. 10-20 7 8.43 20-30 20 24.09 30^0 16 19.27 40-50 23 27.71 50-60 11 13.25 60-70 5 6.02 70-80 1 ■■ 1-20 Total number of cases, 83. While the largest number of cases, 27.71 per cent, were found occurring between forty and fifty, the percentage is only about half as large as Schroeder's for uterine fibromyomata during the same decade (fifty-one per cent). Between the years twenty and thirty we find twenty-four per cent of fibromata of the ovary against seven per cent of uterine fibromyomata during the same decade. Below the age of twenty, the difference is still more strik- ing, for we find eight per cent of fibromata of the ovary as compared with less than one per cent in the case of uterine fibro-myomata. Between fifty and sixty the percentages are about the same in both kinds of fibromata (thirteen and eleven per cent), but between sixty and seventy the percentage for ovarian fibromata is six, while that for fibro-myomata of the uterus is only one per cent. The youngest patient in my list was eighteen, the oldest seventy-seven. These figures show beyond a reasonable doubt : 1. That fibromata of the ovary occur earlier and in relatively greater numbers during early sexual life than do fibrous growths of the uterus. 2. That, while relatively the largest number are found during or just after the menopause, a considerable percentage of cases occur proportionately later in life than is the case with the uterine fibromyomata. Social State and Fecundity. — The social condition was mentioned in sixty-six out of the eighty-four cases. Of these, forty-four were married or widows, and twenty-two were single, making the disease twice as common in married as in single women. No mention is made of children in thirty-nine of the eighty- four cases. Of the forty-five remaining cases, thirty-three, or 73 J per cent, were mentioned as having had children. In five, however, the number was not stated definitely. Twelve, or 26| per cent, were sterile. Twenty-eight women had in aU 111 children, or an average of nearly four to each person. Both ovaries were the seat of the fibromatous changes in six out of the eighty-four cases. In Picqu^'s and Amann's cases (62) and (2) no mention is made of children. Kleinwachter (52) performed Cesarian section without removal of the fibromatous ovarian growth which was the cause of the dystocia. Briggs' patient (12) had five children, the youngest being sixteen. Terrier's patient (71) had had one child years before, and Wilks' patient (82) is mentioned as having had several children. REUBEN PETERSON. 175 The results of my researches do not bear out Dartigues"' statement that pregnancy prior to the development of the tumor is rather a rare occurrence. On the contrary, an average of four children to each of the twenty-eight patients is a good showing even for normal women. Menstruation. — Doran's^' carefully studied histories of eleven ovarian fibromata led him to the conclusion that the catamenia are not appreciably influenced by the presence of an ovarian fibroma. Bourgouin (85) is of the same opinion. Dartigues (87) , however, holds that while metrorrhagia is to be looked for as a result of ovarian fibromata, menorrhagia is a rare symptom. My own researches show that in forty-nine out of the eighty-four cases some mention is made of the menstrual epoch. Regidarity. — In twentj--three cases it is stated that the periods were regular. To these may be added six cases where the menstruation is referred to as being normal or not affected by the growth, maldng a total of twenty-nine cases, or forty-nine per cent, where the regularity was not affected. Three patients are recorded simply as being irregular, two cases are mentioned as having irregular periods before the advent of the growth, while in only two cases is it noted that the tumors gave rise to irregularity of the periods. Pain. — Six patients, or twelve per cent, are mentioned as having painful periods, a proportion no larger than one would expect in forty-nine women applying for treatment, and in no cases can it be attributed directly to the presence of the ovarian growth. Menstrual Flow. — A priori one would expect to find quiet a number of in- stances where the menstrual flow was increased by the ovarian growth, as this is true in ovarian cystomata in a certain proportion of cases. But my statistics do not show this to be a fact. In eleven out of the eighty-four cases it is stated that there was an increase in the menstrual flow, either in the form of menorrhagia or metrorrhagia. In cases (28), (47), and (32) the menstrual discharge is spoken of as being merely somewhat freer than normal. Menorrhagia is noted in cases (2), (25), and (66). Amann (2) mentions the menorrhagia as being coincident with the appearance of bilateral ovarian fibromata the size of the fist. The inference is that the increased flow has been brought about by the growths, since the uterus is mentioned as being small and unchanged. In Roberts' case (66), also, the uterus was found normal, the menorrhagia having existed for three years. The duration of the growth is not stated. The tumor was discovered in Cullingworth's case (25) while seeking an explanation for the menorrhagia. Metrorrhagia is noted in cases (23), (62), and (74). In two of these, Cul- lingworth's (23) and Picque's (62), both ovaries were fibromatous. The uterus in each of the three cases was normal and could not be held accountable for the increased flow. The third case, Villar's (74), is significant in that the dura- tion of the disease is stated as being ten years, with severe metrorrhagia during 176 A CONSIDERATION OF OVARIAN FIBROMATA. the last two years. The flow ceased after the removal of the growth, although the other ovary was not disturbed. From a consideration of these facts, I think it may be concluded that : 1. Neither menorrhagia nor metrorrhagia is a very constant accompani- ment of fibroma of the ovary. 2. It is more apt to occur where both ovaries are affected by the disease. Decreased Menstruation and Amenorrhea. — In four cases the statement is made that the menstrual flow was decreased, and in nine cases there was amenorrhea extending over various periods of time. In Carsten's case (18) the patient was pregnant, and in cases (11) and (78) simply the last period prior to the operation had been skipped. After the closest scrutiny of the cases, I am of the opinion that the facts do not warrant the assumption that an ova- rian fibroma causes a cessation of the menstrual function. Menopause. — In fifteen cases out of fifty-nine, or in twenty-five per cent, it is stated that the patients with the ovarian fibromata had passed the meno- pause. The latter occurred in two patients between thirty-five and forty; in five patients between forty and forty-five ; in five patients between forty-five and fifty, and in three patients between fifty and fifty-five. In four of these cases the time of origin of the ovarian growth cannot be estimated. In one it was three years before, and in the remainder at varying periods, after the menopause. It may be concluded that in patients with ovarian fibromata the time at which the menopause appears is somewhat later than under normal conditions. Duration and Rate of Growth. — It is extremely diffic\ilt to gather from my tabulated cases definite information regarding the duration and rate of growth of the ovarian fibroma. This is not to be wondered at when it is con- sidered that such tumors may escape the observation of the patient, from failure to produce marked symptoms, until after they have reached a consider- able size or given rise to enough ascitic fluid to cause an enlargement of the abdomen. I find that in thirteen of the eighty-four cases no reference was made to the probable age of the growth. In fourteen of the remaining seventy-one cases the statements regarding this point are so vague and unsatisfactory as to prove valueless for onr purposes. In fifty-seven cases it is either stated definitely that the duration of the growth was so many months or years, or the size of the tumor is given at the time of its discovery, or, finally, the beginning of the abdominal enlargement is recorded. We see from a consideration of the cases in the first category that the growth may exist for many years. Fleishmann (38) reports the case of a pa- tient who, during an attack of peritonitis, was seen to have a large abdominal growth. A twenty-pound ovarian fibroid was removed eleven years later. REUBEN PETERSON. 177 In Wilks' case (81) labor was retarded by a growth which was known to have existed at least ten years before. When shown, the tumor weighed three pounds. Piering (63) in his case is able to state definitely that an ovarian fibroma, the size of the two fists, which he removed from the right side did not exist four years before. Turning now to those cases where the size of the growth is described definitely at diiTerent periods, we find ourselves in a better position to estimate the rate of growtlr. Zumbusch (83) records a case where, at the birth of the last child, a hard growth the size of a walnut was discovered in the right lower abdomen. Wlaen it was removed four years later it measured 15 by 20 by 7 cm. Heinricus (48) reports a case where increase in the gro\vth in fifteen months was from the size of a hen's egg to that of a child's head. Cases (10) and (13), reported by Briggs, are interesting when considered together. The tumors when first discovered were both estimated to be the size of hens' eggs. One was removed five months and the other eighteen months later, yet the size of the tumors was about equal (7^x5-|^ inches respectively). Considering now the cases where the duration of the growth is estimated by the enlargement of the abdomen, we see at once that such estimates are of httle value in determining the duration of the growth, for we cannot say how long the tumor existed before producing ascites. By inference, we judge that the growth may exist for some 3^ears before it gives rise to this symptom, since in three cases (28), (62), (75), where the abdominal enlargement dates back two months only, the size of the tumors found at operation was six pounds, size of the fist and eight and one-half pounds respectively. In my first case the tumor had grown in five years from the size of an egg to a mass weighing seven pounds and this without producing symptoms which compelled the patient to consult a physician. We may, therefore, conclude that : 1. The growth of an ovarian fibroma may extend over many years before the tumor reaches large proportions or gives rise to marked symptoms. 2. On the other hand, we find a smaller number of cases where the growth is evidently more rapid, reaching considerable proportions and giving rise to noticeable symptoms in a comparatively few months. Pain. — It is the general opinion that ovarian fibromata give rise to little or no pain. While this may be relatively true as compared with other pelvic growths, I find it existing in a greater or less degree in thirty-seven of the entire eighty-four cases, or forty-four per cent. The symptom is mentioned in fifty-one cases. In fourteen of these, or twenty-seven per cent, it is distinctly stated that no pain existed. An interesting fact in connection with these fourteen cases is freedom from adhesions in all except two (65) and (71). These were both large tumors, one weighing three and the other seven and one- half pounds, and each had omental and intestinal adhesions. In the thirty- 178 A CONSIDERATION OF OVARIAN FIBROMATA. seven cases with pain, adhesions existed in ten, or twenty-seven per cent. Pain seems to be an accompaniment of the moderate and not the largest sized tumors. Dysuria was mentioned in eight out of the fifty-one cases, or fifteen per cent. It did not seem to be associated with the large tumors, and in three out of the eight cases adhesions were recorded. From these facts we may con- clude that : 1. Pain is present to a greater or less degree in nearly one-half the cases of ovarian fibromata. 2. Freedom from pain means, as a rule, the absence of adhesions of the growth to the surrounding organs. 3. Pain is more frequently met with in tumors of moderate size. 4. Dysuria is a symptom in about 15.6 per cent of the cases. Mobility. — Mention is made of mobility in forty-two, or one-half, the cases. Seven of these were recorded as being immovable and thirty-five as having some degree of mobility. Of these thirty-five, twelve were freely movable, eleven movable, seven limited in motion, four movable laterally, and one movable to the left and upward. The greatest interest centers in the relationship existing between the mobility of the ovarian fibroma and the presence or absence of adhesions to surrounding structures. The following table is of interest in this connection : Freely Motion Immova- Tumor. Movable. Limited. ble. Adhesions 5 8 4 No adhesions 9 5 1 No mention of adhesions 6 2 2 20 15 7 From this it is to be seen plainly that the mobility of the growth depends largely, though not entirely, upon its freedom from adhesions. In two of the freely movable tumors the adhesions were found to be recent and non-resist- ing. In a number of cases the size of the tumor and not the adhesions evi- dently explained the hmited motion (61). I could find no reason for thinking that the presence of an ovarian fibroma itself gave rise to adhesions. The origin was evidently infection from some neighboring viscus. In cases (36), (38), and (53) there was a history of a previous localized peritonitis. Cases (22) and (60) had been tapped for ascites, and in Kleinwachter's case (52) the tumor was wedged in the pelvis by the pregnant uterus, necessitating a Cesa- rean section. Uterus. — Inasmuch as the disease is outside of and entirely distinct from the uterus, we would not expect to find the latter enlarged in cases of ovarian fibromata. Some mention of the uterus was made in fifty-two cases. Its depth is recorded in sixteen cases and shows an increase in size above three REUBEN PETERSON. 179 inches in only three cases. Doran's case (34), a woman of forty-nino with a tumor weighing four pounds seven ounces, showeii a uterus measuring three and one-half inches in depth at the first examination. At tlie second exami- nation, a year afterward, it showed the hirgc-st depth of any in the series, four inches. Obviously, the position of the uterus is liable to be affected by the ovarian gi'O'wth. In the eighty-four cases position was referred to in nine out of the fifty- two cases where the uterus was mentioned, leading to the inference that there was nothing strikingly abnormal in this regard in the remaining forty-three cases. In five cases the uterus was in various stages of retrodisplacement. In only four cases was the uterus mentioned as being the seat of fibroid changes. From the standpoint of diagnosis of ovarian fibromata, it is of interest to study the relationship between the uterus and the growth. Experience in the two cases reported has shown me that it is by no means an easy mat- ter to decide definitely that the tumor mass, especially when of large size, is unconnected with the uterus. Just as is the case with ovarian cysts, an ova- rian fibroma as it increases in size becomes a central and not a lateral growth. If to this be added a uterus small in size and difficult to outline, it is no easy task to differentiate between a uterine and ovarian fibroid. The larger the tumor and the more it compresses the uterus, the more easily will each be af- fected by the movements of the other. In forty-two, or exactly one-half of the cases, some mention is made of the relationship of the uterus to the tumor, yet in only sixteen, and these, in the majority of instances, in the smaller growths only, did the examinations prior to the operation permit of the defi- nite statement that there existed no connection between the uterus and the tumor mass. In five cases the ovarian growth was thought to be connected with the uterus, and in quite a large number of cases, while no definite state- ments were made as to the bimanual findings, the tumor was considered uterine until its true relationship was revealed on the operating table. In only six cases was it stated that the growth could be moved independently of the uterus; in four cases movements in one resulted in motion in the other. In eight cases it was recorded that the tumor lay belfind, and in seven cases in front of, the uterus. Conclusions. — 1. Ovarian fibromata do not produce any changes in the size of the uterus. 2. The position of the uterus is but little affected by the ovarian growths but retrodisplacements may be produced by large, centrally situated tumors in a small proportion of cases. 3. While, even in the case of large tumors, the uterus will be found un- connected with the growth, it is not always an easy matter to determine this prior to the operation. In the absence of ascites, the smaller and more mova- ble the tumor, the easier will be the diagnosis. 180 A CONSIDERATION OF OVARIAN FIBROMATA. Ascites. — For the purpose of diagnosis and for comparison, I have made note of all the observations on ascites before and also after operation, if such was performed. Of course, even when a fairly accurate description of the case has been given, sources of error may creep in from the reporter's failure to mention that ascitic fluid was found at the time of operation. But the figiu-es as given below present a fairly accurate estimate of the frequency with which ascites is found in cases of ovarian fibromata. Frequency. — A study of a series of cases of ovarian fibromata shows be- yond a doubt that ascites is a quite frequent accompaniment of these growths. Deligny'^ in his thesis on ascites in connection with ovarian fibroids contends that ascites is less common than has been supposed. Picqu6 is quoted by Dartigues'^ and Teste'^ as stating that ascites is usually absent in fibroids of the ovary but present in mahgnant ovarian tumors. Doran^', however, found a peritoneal effusion in five of his eleven cases, and suspected that it existed in others in small quantities but was taken up unobserved by the sponges at the time of the operation. Ascites was a complication either before or after the operation in thirty-four of the eighty-four cases, or in forty per cent. In twenty-six cases, or thirty- one per cent, ascites is mentioned as being present before the operation. In nine cases it is stated definitely that no ascites existed before the opera- tion and that in seven cases none was found at the time of the operation. In twenty-four cases it was stated that a peritoneal effusion was found at the time of operation or post mortem. Amount of the Ascitic Fluid. — In eight cases the ascitic accumulation reached such large proportions as to necessitate its removal by tapping. In three cases there was but one tapping. One case was tapped five times and two cases (42) and (60) were aspirated many times, the first every two weeks for a year, the other sixty-five times. The last case is my own (Case 2) and is interesting as showing how much less harm results from tapping to-day, under aseptic precautions, than in the pre-antiseptic days. I found no evidences of infection of the general peritoneum and only a few yielding adhesions in the pelvis. The size of the tumors in the tapped cases varied from a half to six pounds, showing that the largest tumors did not produce the greatest amount of ascitic fluid. Cause of the Ascites. — As pointed out by Doran^^, we are in no better posi- tion at present to speak authoritatively as to the etiology of the ascitic accu- mulation than we were in 1886, when Ohlshausen'^ stated that the cause of the ascites was obscure. It is not my purpose to enter into any lengthy discussion of the different theories which have been advanced in order to explain the asso- ciation of ascites and ovarian fibromata. My collected cases show that while the largest amount of ascitic accumulation is not present with the largest growths, on the other hand ascites is not a very frequent compHcation of the REUBEN PETERSON: 181 smaller tumors, since fifteen of the twenty-four cases with ascites, where the weight of the tumor is stated, averaged over two pounds. A tumor of this size, while it may be movable is not so freely movable as to give rise to peri- toneal bruising, resulting in an effusion from the membrane. If this excessive mobility were a cause of the effusion, we would expect it to be an accompani- ment of small cystic and dermoid ovarian growths, which is not the case. Furthermore, if the ascites were due to an irritation of the parietal peritoneum, it would show evidences of the condition at the operation, and such does not seem to be the case. Uncomplicated uterine fibromata are rarely accom- panied by a peritoneal effusion. This holds true even when the fibroid is of the long pedicled variety and capable of free movement. It would seem, therefore, as if the explanation of the ascites on mechanical grounds alone must be abandoned. In only four of the thirty-four ascitic cases was mention made of any peritoneal changes. In Culhngworth's case (23), where the patient died without operation, from pleuritic as well as peritoneal effusion, the peri- toneimi was described as thickened and opaque. Case 64 had a peritoneum which was injected, while in case 65 it is described as red and vascular. The fourth case (42) was operated on in 1878 and the peritoneum was found hardened and thickened by repeated tappings. In the remainder of the ascitic cases, the appearance of the peritoneum is either not mentioned at all or described as normal. Character of Fluid. — The color and consistency of the ascitic fluid is men- tioned in only eight of the eighty-four cases, from which it may be inferred that there was nothing unusual in its appearance in the remaining cases. In three cases it is described as clear, twice as pale red, and once each as yellow, red, and yellow stained with bile. Conclusions. — 1. Ascites is present in about forty per cent of cases of ova- rian fibromata. 2. The peritoneal effusion may be large in amount and give rise to dis- tressing symptoms. 3. The largest tumors do not give rise to the greatest amount of fluid. 4. The cause of the ascites with ovarian fibromata is obscure. 5. The peritoneum does not show evidences of inflammation in the ascitic cases. 6. The ascitic fluid, as a rule, is clear and without much color. Operation. — Seventy-six of the eighty-four cases were operated upon; in three there were post mortems without operation and in five no mention was made of either operation or post mortem. There were seventj^-four ovariotomies with sixty-two recoveries and four deaths, a mortality of six per cent. There was also a Cesarean section (52) without the removal of the tumor and a hysterectomy (51) in addition to the ovariotomy, both cases end- ing fatally. The operations extended over many years, which accounts in a 182 A CONSIDERATION OF OVARIAN FIBROMATA. great measure for the high mortality. The operation as performed to-day should be almost free from fatality, except in neglected cases. Adhesions. — It has commonly been supposed that ovarian fibromata are but rarely adherent to the surrounding structui-es. This view is not supported by my statistics, since- there were adhesions in thirty out of the eighty-four cases, or thirty-six per cent. Dartigues*' found no intestinal adhesions in his twenty collected cases, while in my eighty-four cases the intestines were adherent to the tumor in nine instances. The omentum was most frequently adherent, twelve such cases being recorded. In four cases the tumor was adherent to the pelvic peritoneum; twice the cecum was involved, and the colon and bladder once each. In two cases the appendix was adherent to the growth, and the uterus and sigmoid involved once each. The close relation- ship existing between pain and the adhesions has already been referred to. Pedicle. — It is almost impossible to obtain from my list of cases statistics regarding the formation of the pedicle, as the statements in this regard have been exceedingly meager or indefinite. However, it may be stated that in a majority of cases the pedicle of an ovarian fibroma is made up of the broad ligament and of the Fallopian tube. The latter may be long or atrophied, according to the age of the patient. As a rule, the pedicle is of good length and is thin rather than thick. Twisted pedicle has been noted nine times out of the eighty-four cases, or approximately ten per cent. In three cases the pedicle is simply spoken of as "twisted." Once it was half a turn to the right, once it was twisted 180°, once it was twisted three times, once three full turns, and once five times. The remaining case is spoken of as "large and twisted. " The twisted pedicle was comphcated twice by peritonitis; in three cases adhesions were present, and in two cases extravasation of blood into the tumor occurred. The weight of the tumor is noted in six cases, two, three, four and one-half, and seven pounds respectively. In two cases it was as large as a child's and an adult's head respectively. Ovary Involved.~As regards the seat of ovarian fibromata, we find that the reporters have failed to record this fact in eighteen cases. Of the remaining sixty-six cases the left ovary was involved thirty-four times, the right twenty- seven, both right and left six times. , Opposite Ovary.—The condition/of the opposite ovary was not recorded in thirty-six cases; eleven of the regaining thirty-eight cases showed the oppo- site ovary to be healthy; in fourteen cases it was recorded as normal, in six it was atrophied, in three it was found enlarged, and in four it showed cystic disease enough to warrant its removal. Gross Appearances.— A more or less accurate description is given of the growth in sixty-two of the eighty-four cases. In thirteen it is described as irregular, rounded in seven, nodular in ten, and as having a regular outhne REUBEN PETERSON. 18:1 in three. It is described as lobular in seven and kidney-shaped in five. As regards color, it is described as white in ten, pinkish in lour; the other colors are yellowish-white, cream, a greenish-brown, and gray. The tumors were found to be smooth in fourteen cases. An edematous condition was noted in three, and in a number of cases it is obs(M'\'ed that the tissue creaked under the knife. Size. — The size was not mentioned in thirty-one cases. In some cases accurate measm-ements were made; in others the reporters in their descrip- tions likened the tumor to some familiar object. For instance, it was described as being the size of an adult's head in seven cases, of a child's head in three, the size of the fist in four, size of the two fists in two, size of a goose egg in two, ostrich egg, hen's egg, and orange in one each. As regards the length of the tumors we see much variation. In six cases they were seven inches long, in five cases six inches long, in three cases three and five inches, and one each sixteen, twenty-four, and thirty-three inches. The width varied from one to twenty-two inches. Weight. — No mention was made of the weight in thirty-eight cases; in the remaining forty-six cases the largest number, seven, weighed three pounds; next come four, six, and seven pounds, five cases each; in four cases the weight was two pounds. One, eight, twelve, and fifteen pounds were recorded in two cases each, while the remainder were one pound or under, with the exception of Fleishmann's case (38), which weighed twenty pounds. Calcification. — Calcification was noted in eleven cases, or thirteen per cent. Sometimes the entire tumor was affected so that a saw had to be used in its section; at other times, the denser parts were calcified. Irregular spicules were spoken of as located either on the inside or outside of the tumor. J. W. Wilhams'^ has gone into this degenerative change in ovarian fibromata very thoroughly and has shown it to be a calcification and not a bone formation, as formerly supposed. Cystic Formation. — A greater or less degree of cystic degeneration was noted in twenty-two of the eighty-four cases. As stated in the earher part of the paper, cases where the cysts were of such large size as to render the diagnosis between cystic degeneration and fibro-cystomata doubtful, were not included in the hst. The small cystic cavities were either on the outside or within the substance of the tumor. In some cases they were filled with either blood or pus. At times they were distended with clear fluid. Microscopic Examination. — As stated at the outset of the paper, no case was accepted which had not been subjected to microscopic examination, hence it becomes necessary to account for the eighty-four cases. The growths were described as fibromatous in sixty-three out of the eighty-four cases, or in seventy-five per cent. Forty-eight of these were fibromata, twelve pure fibromata, and fibroma with colloid infiltration, fibroma with mucoid degenera- 184 A CONSIDERATION OF OVARIAN FIBROMATA. tion, and fibroid with hyalin and myxomatous changes one each. Seventeen, or twenty per cent, were found to be fibro-myoma, two were called myo- fibromata, and there was one each accredited to the myomata and to the myo-fibromata with hyalin degeneration. BIBLIOGRAPHY. 1 to 84 refer to cases tabulated. 1. Addinsell, A. D. — A Solid Tumor of the Ovary Removed from a Woman Aged 36. Trans. Obst. Soc, London, 1900, Vol. XLII, page 139. 2. Amann, J. A., Jr. — Doppelseitiges Ovarialfibrom. Monatsch. f. Geburtsh. u. Gynakol, 1900, Bd. XII, S. 544. 3. Bagot. — Fibromyoma of the Ovary. Royal Academy of Medicine in Ireland, May 23, 1890. Abstract in N. Y. Med. Journ., 1890, Vol. LTI, page 248. 4. Baldwin, L. G. — Fibroid Tumor of the Ovary with Calcareous Degeneration. Brooklyn Med. Journ., 1896, Vol. X, page 254. 5. Bantock, G. G. — British Gynaecol. Journ., 1892-3, Vol. VIII, page 312. 6. Brothers. — Fibroma of Ovary. Amer. Journ. Obst., N. Y., 1900, Vol. XLI, page 194. 7. Borreman, C. H. — Fibromes de I'ovaire. Bull. Soc. belgique de Gynecol, et d'Obstetr., 1896-8, Vol. VIII, page 169. 8. Borreman, C. H. — Ibid. 9. Borreman, C. H. — Ibid. 10. Briggs, H. — Fibroma of the Ovary and Ovarian Ligament. Brit. Med. Journ., 1897, Vol. I, page 1083. 11. Briggs, H.— Ibid. 12. Briggs, H.— Ibid. 13. Briggs, H.— Ibid. 14. Briggs, H.— Ibid. 15. Briggs, H.— Ibid. 16. Briggs, H.— Ibid. 17. Briggs, H.— Ibid. 18. Carstens, J. H. — Removal of a Fibroma of the Right Ovary during Pregnancy. Trans. Amer. Ass. Obst. and Gynecol., 1889, Vol. II, page 151. 19. Carter, C. H.— Fibromyoma of Right Ovary Removed by Abdominal Section. Trans. Obst. Soc, London, 1887, Vol. XXIX, page 190. Ibid., 1896, Vol. XX'XVIII, page 204. 20. Carter, C. H.— Fibroid Tumor of the Right Ovary. Trans. Obst. Soc, London, 1882, Vol. XXIV, page 139. 21. Castaing, P.— Tumeur solide de I'ovaire. L'^cho medical Tolouse, 1900, Vol. XIV, page 16. 22. Coe, H. C. — Fibroma of the Ovary Complicated with Ascites: Laparotomy, Recovery, Amer. Journ. Obst., N. Y., 1890, Vol. XXIII, page 412. 23. CuUingworth, C. X.— Fibroma of Both Ovaries. Trans. Obst. Soc, London, 1879, Vol. XXI, pages 276, 314. 24. CuUingworth, C. N.— Fibroma of the Ovary. Trans. Obst. Soc, London, 1897, Vol. XXXIX, page 279. 25. CuUingworth, C. N.— Ibid, 1894, Vol. XXXVI, page 314. REUBEN PETERSON. 185 26. Dakin, W. R.— Ibid, 1896, Vol. XXXVIII, page 204. 27. Dannien, K. — CavernSsps Fibrom des linken Ovariums, Achsendrehung des Stiels. Blutung in den Tumor, Thrombose der Gefftsse, Nekrose, Probepunction, Verjauchung, Ovariotomie, Heilung. Arcli. f. klin. Chirurg., 1878, Vol. XXII, S. 973. 28. Delgrange. — Fibrome de I'Dvaire. Journale des sciences mgdicales de Lille, 1896. 29. Doran, A. — Cases of Fibroma of the Ovary and Ovarian Ligament Removed by operation, with a series of after histories of cases reported in "Transactions" since 1879. Trans. Obst. Soc, London, 1896, Vol. XXXVIII, page 204. 30. Doran, A.— Ibid. 31. Doran, A.— Ibid. 32. Doran, A.^Ibid. 33. Doran, A. — Ovarian Tumors, simulating Inflamed Ovaries, including a Case of Ovarian Myoma. Edinburgh Med. Journ., 1898, Vol. XLIV, page 454. 34. Doran, A. — Fibroma of the 0\'ary: Impaction; Ascites; Removal. Trans. Obst. Soc, London, 1897, Vol. XXXIX, page 37. 35. Doran, A. — Tumors of Ovary, Fallopian Tube and Broad Ligament. Trans. Obst. Soc., London, 1SS4, page 97. 36. Dubar. — De I'ascite dans les fibromes de I'ovaire. Deligny, These de Paris, 1898. 37. Feis. — Ein Fall von Fibromyoma ovarii. Centralbl. f. GynSk., Leipz., 1894, Bd. XAIII, S. 133. 38. Fleishmann. — Demonstration eines Ovarialfibroms. Centralbl. f. Gynakol., 1900, Bd. XXIV, S. 768. 39. Graefe, M.— Zwei Falle von Ovarialfibrome. Centralbl. f. Gynitk., 1895, Bd. XIX, S. 11. 40. Geissel. — Fibrom des linken Ovariums. Deutsche med. Wochenschr., 1877, S. 492. 41. Gibb. — Polycystic Ovarian Tumor of Right Side and Fibrous Tumor of Left Side Successfully Removed from the Same Patient at One Operation. Trans. Path. Soc, London, 1861, Vol. XII, page 154. 42. GoodeU, W. — Case of Ovariotomy with Fibroid Tumors of the Ovary. Amer. Journ. Obst., 1878, Vol. XI, page 152. 43. Goodell, W. — Fibroid Tumor of the Ovary. Amer. Journ. Obst., Supplement. 1882, Vol. XV, page 74. 44. GoodeU, W.— Ibid. 45. Hamilton, J. I. — Stormy Convalescence after Removal of Solid Tumor of Ovary. Peoria Medical Monthly, 1885, Vol. VI, page 1. 46. Haddfield-Joues, M. — Fibrosarcoma of the Right Ovary. Ti-ans. Obst. Soc, London, 1896, Vol. XXXVIII, page 204. 47. Hartmann, H. — Fibromyome de I'ovaire. Le Progres medical, 1884, Vol. XII, page 544. 48. Heinricus. — Fall von Fibroma ovarii. Finska LakaresSUsk, Handl., 1892, No. 1, S. 68. Abstract in Centralbl. f. Gynak., 1892, Bd. XVI, S. 459. 49. Hunter, J. B.— Fibroma of the Ovary. Amer. Journ. Obst., 1885, Vol. XVIII, page 1204. 50. Hofkmokl.— Fibroma Ovarii dextri. Wien. med. Presse, 1882, Bd. XXIII, S. 1416. 51. Jacobs. — Fibromatose g^nitale. Bulletin de la Society belgique de Gynecologie et d'Obstetrique, 1900, Vol. X, page 249. 186 A CONSIDERATION OF OVARIAN FIBROMATA. 52. Kleinwachter. — Kaiserschnitt bei normalem Becken, bedingt durch ein her- abgetretenes Fibroid des rechten Ovariums. Arch. f. Gynak., Berl., 1872, Bd. IV, S. 171. 53. Laidle}', L. H.— Fibroma of the Ovary. Amer. Journ. Obst., 1900, Vol. XLII, page 661. 54. Ledoux, M. — Un eas de fibrome pur de Tovaire. Journal de sciences m^dicales de LiUe, 1900, Vol. II, page 327. 55. Mann, M. — Removal of Solid Uterine and Ovarian Tumors by Laparotomy with Report of Nine Cases. Amer. Journ. Obst., 1887, Vol. XX, page 451. See also American System of Gynfecology, Vol. II, page 1039. 56. Montgomery, E. E. — Ovarian Fibroma. Amer. Journ. Obst., 1888, Vol. XXI, page 323. 57. Mor61y, P. — Fibrome de I'ovaire. Bulletin de la Soci^t^ anatomique de Paris, 1898, page 617. 58. Nunn. — Large Fibrous Tumor within the Abdomen suppdsed to be of the Right Ovary. Trans. Path. Soc, London, 1856-7, Vol. VIII, page 270. See also Ibid, Vol. IX, page 299. 59. Potel. — Voluminaux fibrome de I'ovaire. L'echo medical du Nord, 1898, Vol. II, page 306. 60. Peterson, R. — Present article. 61. Peterson, R.— Ibid. 62. Picqu^. — Bourgouin, Thdse de Paris, 1894. 63. Piering, O. — Fibrome des Ovariums. Prager med. Wochenschr., 1900, Bd. XXV, S. 501. 64. Purslow, C. E. — A Case of Solid Fibroid Tumor of the Ovary. Lancet, London, 1897, Vol. I, page 1398. 65. Qu6nu. — Lallemand, These de Paris, 1895-6. 66. Roberts, C. H. — A Case of Fibroma of the Ovary undergoing Calcareous Degenera- tion. Trans. Obst. Soc, London, 1897, Vol. XXXIX, page 8. 67. Rocher, L. — Fibrome de I'ovaire. Revue mensuel Gyn^cologie, Obstetrique et Pedriatriques, 1900, Vol. II, page 306. Also These de Bordeaux, 1900-1, No. 25. 68. Rueder, W. — Zwei Falle A-on Ovarialerkrankung. Cystosarcoma und Fibroma. Inaugural Dissertation, Wiirzburg, 1888. 69. Schachner, A. — Three Illustrati\e Cases of Abdominal Section. Amer. Journ. Obst., 1886, Vol. XXIX, page 1265. 70. Sims, H. M. — Specimens of Fibroid Tumor of the Ovary-Operation-Recovery. Amer. Journ. Obst., 1886, Vol. XIX, page 1265. 71. Terrier.— Lallemand, Thise de Paris, 1895-6. 72. Teufel, W. — Inaugural Dissertation, Miinchen, 1901. 73. Thomas, T. G. — Fibromyoma of the Ovaries. Amer. Journ. Obst., 1879, Vol. XII, page 350. 74. ViUar. — Contribution a I'fitude des fibromes de I'ovaire. Teste. Th^se de Bordeaux, 1900. 75. Warren, J. C— Solid Tumor of the Ovary. Boston Med. & Surg. Journ., 1893, Vol. CXXVIII, page 10. 76. Wells, Spencer.— Fibrous Tumor of the Ovary. Trans. Path. Soc, London, 1858-9, Vol. X, page 199. 77. Wiener, G. — Ein Fibromyom des Ovariums. Centralbl. f. Gynak., Leipz., 1900, S. 1111. 78. WiUiams, Sir J.— Fibrous Tumor of the Ovary. Trans. Obst. Soc, London, 1883, Vol. XXV, page 35; also, 1896, Vol. XXXVIII, page 204. REUBEN PETERSON. 187 79. Williams, Sir J.— Ibid. Trans. Obst. Soc, London, 1887, Vol. XXIX, pages 247, 513. 80. Willis, S. — Three Specimens of Fibrous Tumors of the Ovary. Tran.s. Path. Soc, London, 1857-8, Vol. IX, page 299. 81. Wilks, S.— Ibid. 82. Wilks, S.— Ibid. 83. Zumbusch, 0. — Fibromo Ovarii. Inaugural Dissertation, Bonn, 1896. 84. Medical Index, 18S4, Vol. V, page 288. 85. Bourgouin. — Des Tumeurs solides de I'ovaire k Evolution lente. ThSse de Paris, 1894. 86. Coe, H. C. — Fibromata and Cysto-fibromata of the Ovary.- Amer. Journ. Obst., Vol. XV, 1882, pages 561, 858. 87. Dartigues, L. — Des Tumeurs solides de I'ovaire. Rev. de gyn. et de chir. abd., 1899, A^ol. Ill, pages 601, 793, 1013. 88. DeHgny. — De I'ascite dans les fibromes de I'ovaire. Th^se de Lille, 1898. 89. Ohlshausen. — Krankheiten der Ovarien, 2d edit., page 418. 90. Schroeder. — Handbuch der Frauenkrankheiten, 1901, S. 307. 91. Test6, M. P. A. — Contribution a I'^tude des fibromes de I'ovaire. ThSse de Bordeaux, 1900. 92. Williams, J. W. — Calcified Tumors of the Ovary. Trans. Amer. Gynecol. Soc, 1893, Vol. XVIII, page 359. ON THE MOEPHOLOGY OF THE PYLORIC GLANDS OF VEETEBEATES. LYDIA M. DEWITT, M.D., B.S., Instructor in Histology, Department of Medicine and Surgery, University of Michigan. Until quite recently our knowledge of the morphology of microscopic ob- jects was gained entirely from the study of sections and of macerated and teased preparations. Neither of these methods could give a clear and accu- rate idea of the external form of the object studied, although we must admit that the older anatomists, in spite of the insufficiency of their methods, gained a wonderful insight into the structure of the body. The inaccuracy of these methods is shown especially by the fact that careful observers differed most radically in their ideas regarding essential facts of morphology. This is well illustrated in the literature of the so-called gastro-mucous glands of the pyloric portion of the stomach. Oppel, v. Ebner, Kuczynski, Schieflferdecker, Renaut, Watney, Ebstein, Bischoff, EUenberger, and many others regard them as tubular glands, in which the secreting portion consists of several tubules of qvdte uniform caliber opening into a common crypt or duct. These tubules may branch and be more or less convoluted, so that a section cuts them at various angles and suggests an alveolo-tubular, or even an alveolar, character. Sappey made a careful study of isolation preparations from the various species of vertebrates and has described and figured, especially for man and dog, very typical alveolar glands, in which the duct branches several times, the branches being surrounded at the sides and ends by numerous spherical or oval Vesicles, so numerous that they may overlap and entirely conceal the tubule. This concept of the alveolar character of the gland was also gained by the study of sections by Bruch, Frey, Cobelli, and others. That these glands are alveolo-tubular iii type is suggested by Glinsky and Maziarski, the latter basing his conception rather on the reported similarity of physiologic function of the pyloric and duodenal glands than upon careful study of the form of the pyloric glands themselves. These older methods of histologic investigation have, however, been ably supplemented by the wax-plate reconstruction methods with which the names of Born and His are inseparably connected, and which have already done much to settle disputed points in morphology, as well as to give clearer and more vivid pictures of structures which had been previously but little understood. The plate-modeling method, which has been used in my work, was first used by Born, in 1876, in the investigation of the nasal cavities and nasal ducts of 188 LYDIA M. DEWITT. 189 amphibia, and was described by hiin in detail in 1883. Although variously modified later by Strasser, Kastschenko, Keiljol, Alexander, Schaper, and by Born himself, yet the method is still, in all essential points, as it was first described. Since this method giA'es an exact, although enlarged, picture of the external form of the object studied, it seemed to be the method best adapted for the in- vestigation of the pyloric glands, regarding the morphology of which such vari- ous ^'iews have been expressed after investigation by the ordinary microscopi- cal methods. As used by me, the technic of the method was as follows : 1. The tissues were removed from the body as soon as possible after death. Small pieces of the mucous membrane were pinned out straight on paraffin plates to prevent the distortion of the glands which results from the contraction of the tissues during the fixation and dehydration. They were then fixed, usually in a solution of mercuric bichloride, dehydrated, and imbedded in paraffin. 2. Sections five m in thickness were then cut in series, fixed to slides, either by the water-albumin method or by the pure water method and stained in hematoxylin and eosin. 3. Bj-the aid of the camera lucida, outline drawings of the series of sections comprising a gland were made, each section being magnified four hundred times. Adjacent blood-vessels, glands, and connective tissue strands, as well as the duct of the gland itself, served as direction lines. 4. In making the wax plates I have used the apparatus recently devised by Dr. G. Carl Huber, which is represented in Fig. 1. It consists of a heavy Fig. 1. Wax-plate rolling apparatus. Steel plate, nineteen by twenty-four inches, to the sides of which are attached steel side pieces, which are moved up and down by micrometer screws. These are so adjusted that plates may be made varying in thickness from one to ten mm. A heavy steel roller fashioned on the model of an ordinary rolhng- pin completes the apparatus. The roller is heated in the water bath. The side plates were so adjusted that their upper surfaces were two mm. above the 190 MORPHOLOGY OF THE PYLORIC GLANDS OF VERTEBRATES. surface of the horizontal main plate, this being the thickness of the plates used in all my models. The melted wax was poured on the main plate, which had previously been lubricated with turpentine or a mixture of alcohol and glycerin. The wax was then rolled out with the hot roller until the entire plate was of the uniform thickness of 2 mm., removed from the apparatus and cooled. 5. The drawings were then transferred to the wax plates, cut out, piled in order and fused and finished in the ordinary way. I have used in this investigation the pyloric glands of man, dog, cat, rabbit, turtle, and frog. I have also traced the development of these glands in the cat and have reconstructed the Brunner's duodenal glands of man and cat in order to compare the type of these glands with that of the pyloric glands of the same animals. Man. — In the human stomach, the true pyloric glands occupy a relatively narrow zone, in which the glands are rather closely crowded, with but little interstitial tissue separating them. The result is, that it is often dif&cult to isolate a gland for the purpose of reconstruction. It has also been difficult to secure fresh, normal human tissue suitable for this purpose. For these reasons, I have been compelled to content myself with the reconstruction of human pyloric glands from a single individual. One gland, which seemed Figs. 2 and 3. Human pyloric gland, x 80; cr, crypt; mt, main tubule; st, secondary tubule; d. The crypts, however, of nu- merous glands often open together into a common large crypt or fold, giving a peculiarly complicated appearance to the surface of the mucous membrane. The number of crypts foimd in a sq. mm. of the surface of the muccnis membrane in this region averaged about 1,000. Each gland, however, represented only about 7.25 sq. mm. of secreting surface, so that, in spite of the great number of glands, each sq. mm. of the surface of the mucosa represented only about Figs. 15, 16, and 17. Models of pyloric glands of rabbit, x 80. 7,250 sq. mm. of secreting glandtdar surface. Figs. 15, 16, and 17 represent models of pyloric glands from the rabbit's stomach. The glands shown in Figs. 15 and 17 are simple, in that only one gland tubule opens into a simple crypt. The gland tubule is, however, somewhat tortuous and convoluted, which explains the fact that longitudinal sections of these glands frequently show transverse and oblique sections of the tubules. Fig. 16 represents a model of a gland in which three short, simple-, twisted tubules open into one crypt. Sappey, from his isolation preparations of the rabbit's stomach, also figured slender tubules, which sometimes forked at the extremity, but showed no marked branching, although several tubules often opened into a single crypt. Frog. — Partsch states that the pyloric glands of the frog represent the crypt and neck of the fundus glands, only the basal portion being lined by cells resembling mucous cells. Fig. 18 shows a model which seems rather to repre- sent three crypts or tubules opening together into a common crypt than a single gland. Only one of the tubules gives indication of branching in the form of a small bud at the right side of the middle tubiile. Tortoise (Emys Meleagaris). — Oppel states that in the tortoise the pyloric glands are very short, while the crypt presents almost no lumen. Two, four, or 198 MORPHOLOGY OF THE PYLORIC GLANDS OF VERTEBRATES. more cylindrical tubules may open into a single crypt. Fig. 19 represents the model of such a gland found at a short distance from the pyloric orifice of the stomach of the tortoise. In this gland, four short tubules open into the crypt, one of the tubules forking at the extremity, while the others are un- branched. Nearer the pyloric orifice, however, the crypts become longer and Fig. 18. Model of pyloric gland of frog. Figs. 19 and 20. Models of pyloric glands of tortoise, x 160. more numerous, but the glandular tubules are still simple and unbranched and of nearly uniform caliber, as shown in Fig. 20. Maziarski has recently made a classification of glands based entirely on the morphology of the secreting portion of the glands as shown by models reconstructed by him by the Born plate method. He states that the form of the secreting portion of a gland is closely related to the nature of the secretion, and therefore forms the only rational basis for a scientific classification. Ac- cording to his definition, a tubular gland is one in which the secreting portion has the form of a tubule, not necessarily of uniform caliber and often present- ing terminal enlargements, but in which there are no true cul-de-sacs connected to the tubule by a narrower neck. According to this definition, it seems best to classify the pyloric glands of the various vertebrates as tubular, although we must confess that in the dog some structures are found which might easily be interpreted as alveoli. One of the most interesting points in the investigation of the pyloric glands, and the one which has most attracted the attention of investigators, is the relation of these glands to the Brunner's glands of the duodenum. Oppel states that they are identical, at least in the dog, cat, and man, the only dis- tinction being that Brunner's glands project into the submucosa. This view is shared by Kuczynski, Schlemmer, Schwalbe, Watney, Schiefferdecker, Renaut, and many others, who regard both glands as branched tubular glands. LYDIA M. DE WITT. 199 Kuczynski states that in the dog, cat, and horse he saw only branched tubules of nearly uniform caliber, but that in the dog occasional enlargements oc- curred in the course of the tubules, but no true terminal vesicles such as char- acterize alveolar glands. Schwalbe, however, finds terminal enlargements in the duodenal glands of the dog, which he regard.s as alveoli, and therefore characterizes these glands as alveolo-tubular. V. Ebner, on the basis of similarity of function and histologic structure, finds a close relationship between the pyloric and Brunner's glands. Schiefferdecker says that in man, dog, and cat the pyloric glands are identical with those of Brunner. Both are tubular, in man the tubules being of quite uniform caliber with small lumen lined by clear cells resembling mucous cells. In the dog and cat, however, the tubules enlarge club-like at the extremity and the cells lining the tubules are different. In the rabbit the glands are essentially different. Ellenberger regards the Brunner's glands as alveolo-tubular, while the pyloric glands are usually tubular, although in some animals alveolus-like enlargements are found which make them resemble the alveolo-tubular glands of the duodenum. Castellant states that, while there is a marked similarity between the pyloric and duode- nal glands in the rabbit, dog, and man, they are essentially different in the rat. This absolute difference in the rat and slight differences in the dog led CasteUant to the conclusion that the secretion of the two glands is not identical, as is claimed by most physiologists, and that the glands of Brunner secrete a special digestive fluid. Heidenhain describes the glands of Brunner as consist- ing of branched, convoluted tubules, often much twisted and bent, with lateral processes and terminal cul-de-sacs. As these have usually a greater diameter than the tube, the gland is of the acinous type, but the duct is lined by the same epitheUum as the alveoli. Maziarski has given us our first positive knowledge of the morphology of the glands of Brunner, since he has reconstructed a portion of one of these glands from the human duodenum. He states that the gland spaces have the form of tubules winding in different directions. These tubules present on their walls and at their extremities numerous vesicular processes of spherical or oval form and the glands may therefore be regarded as alveolo-tubular. In investigating this question, I have reconstructed two small lobules from the duodenal glands of the cat and a portion of one of the lobules of a human Brunner's gland. Fig. 21 represents a model of a portion of one of the human Brunner's glands, in which two main tubules are seen which empty into a com- mon duct nearer the surface of the mucous membrane. These tubules pass down through the lobule, winding and twisting and changing markedly in call- ' ber and giving off numerous lateral processes, some of which are short, con- voluted tubules ending in an alveolar enlargement, while most of them are spherical or oval alveoli, having no tubular character. The main tubules, as well as their branches, end in very marked and characteristic enlargements, while 200 MORPHOLOGY OF THE PYLORIC GLANDS OF VERTEBRATES. in their course they present very typical vesicular dilatations connected by narrower portions of the tubule. Several shorter tubules, which branched from the main tubules above the level of the figure and served, with their alveolar lateral and terminal processes, to fiU out the upper portion of the gland and form the broad base of a cone-Shaped lobule, were removed in order to show more clearly the relation of the parts. This gland was very near the junction of pylorus and duodenum, was situated entirely in the mucosa, and seemed far simpler than those found at a greater distance from the pylorus and extending into the submucosa. The general type, however, was the same, as may be seen by comparing Fig. 21 with Plate XL, XII., Fig. 7, of Maziarski's work, which represents the model of a Brunner's gland from the submucosa of the human duodenum. Figs. 22 and 23 reproduce models of small lobules of the duodenal glands of the cat and show a dis- tinctly tubulo-alveolar character. In Fig. 22 espe- cially, the alveolar processes are so numerous as almost entirely to conceal the tubules and give the gland an alveolar character. If we compare Fig. 21 with Figs. 2 and 3 of the human pyloric glands, and Figs. 22 and 23 with Figs. 10 and 11 of the pyloric glands of the cat, it seems reasonable to conclude that, morpholog- ically at least, the pyloric and duodenal glands are not identical in man and cat, but present an essentially different type. Embryologically also, they appear to be essentially different, since in the new-born kitten, in which the pyloric glandi were represented by mere short. Fig. 21. Model of human Brunner's gland, x 80; mt, main tubule; bd, alveoli ; st, secondary- tubule ; ct, coiled tubule. Fig. 22. Model of lobule of Brunner's duodenal gland of cat, X 80; mt, main tubules ; bd, alveoli. Fig. 23.— Model of small lobule of cat's duodenal gland. Lettering as in Fig. 22, X 80. Figs. 24 and 25. Model of duodenal gland of new-born kitten, x 80. LYDIA M. DE WIT'L'. 201 crypts with forking buds at the extremity in phiei' of the future tubules, the Brunner's giands had already assumed a. distinctly ah'eolo-tubular type. This is seen in Figs. 24 and 25, whieh represent two views of a Brunner's gland of a new-born kitten, showing a duct di\-iding into two main tubules, each of which is provided with se^-eral lateral processes of spherical or oval or irregular form; the main tubules, as well as the branches, end in cul-de-sacs of much greater cahber than the main tubules. As to the identity of the cells lining these two kinds of glands, I do not desire to make any positi^-e assertion, since my investigation has not included the study of these cells under special stains intended to bring out their finer structure. It seemed to me, however, that in many of the animals studied, the nuclei of the cells were more spherical and the protoplasm of the cells more deeply stained with eosin in the pyloric glands than in the duodenal glands. No attempt has been made by me to investigate the physiologic function of either of these glands, but if, as jMaziarski asserts, there is a definite relation between the physiologic function and the form of the secreting portion of the gland, it seems probable that the secretion also is not identical, and the ques- tion is one which may prove worthy of further study. While it is evident that the pyloric glands, which are branched tubular glands in the animals studied, are not identical, and therefore not continuous with the Brunner's glands of the duodenum, yet small alveolo-tubular glands of the type of the Brunner's glands are occasionally seen, which have not crowded through the muscularis mucosae into the submucosa, but lie entirely in the mucosa. One of these glands is seen in Fig. 21. These are usually found in the upper portion of the duodenum, but may occasionally pass beyond its boundaries into the stomach, and crowding aside the true pyloric glands, occupy a narrow zone within the stomach near the pyloric orifice. This transition was noted also by Renaut, Castellant, and by Maziarski, who describes Brunner's glands of the mucosa as histologically identical with those of the submucosa. The occasional presence of these small alveolo- tubxilar ^ands within the pyloric orifice may explain a part of the confusion in regard to the morphology of the pyloric glands. From the data obtained by the reconstruction of these glands the following conclusions may be drawn : 1. The pyloric glands in the various vertebrates studied are essentially tubular in type, consisting of a crypt, which divides into two or more main tubules, which may divide still further into secondary and even tertiary tubules, and may become tortuous and convoluted near the lower extremity. In the dog, these tubules generally end in a marked and very characteristic dilata- tion, while shorter processes occur which might easily be interpreted as alveoli, but which seem to me to represent short, imperfectly developed tubules. 2. The pyloric glands of the dog are more complex than those of the other 202 MORPHOLOGY OF THE PYLORIC GLANDS OF VERTEBRATES. vertebrates whose glands have been reconstructed and represent a larger secreting glandiolar surface, but the number of glands in a given area of the human pylorus is greater, so that a square millimeter of the surface of the mucous membrane of the human pylorus represents a greater secreting surface than that of any of the other vertebrates studied. 3. The crypt of the pyloric glands attains its full development early, while the gland tubules are still comparatively simple. The gland tubules develop from buds growing out from the base of the crypt. The development of tubules seems to continue during the Hfe of the gland, new tubules budding out either from the base of the crypt or from the old tubules. These at first resemble alveoli, but soon increase in length sufficiently to assume a tubular character. 4. The duodenal glands of Brunner are distinctly alveolo-tubular in type, whether they are found in the mucosa or have crowded through into the sub- mucosa; they develop earlier than the pyloric glands, and present the alveolo- tubular character early in their development, even before the tubules of the pyloric glands have begun to form. The Brunner's glands are, therefore, es- sentially different from the pyloric glands morphologically, embryologically, and in all probability, also physiologically. In closing, I desire to express my sincere gratitude to Professor Huber, who suggested the subject of this investigation and has manifested a constant interest in its progress; I wish, also, to thank Dr. Katherine Rayl, whose draw- ings have so faithfully and accurately reproduced my models. REFERENCES TO LITERATURE. BL^choff. — Ueber den Bau der Magenschleimhaut. Miiller's Arohiv, Berlin, 1838. Bizzozero. — Ueber die Schlauchformigen Driisen des Magendarmkanals. Arch. f. mik. Anat., Bd. XLII, 1893. Bizzozero and Vassale. — Arch. f. path. Anat., CX, 1887. Cade. — Modifications de la muqueuse gastrique au voisinage du nouveau pylore, dans la gastro-ent^ro-anastomose exp^rimentale. Comptes rendus de la Soc. de Biol., Paris, T. LII, 1900. Castellant. — Quelques recherches sur les glandes de Brunner. Th&se pour le doc- torate en m^decine. Lille, 1898. Cobelli. — Le Ghiandole acinose della parte pilorica dello stomaco. Sitzungsber. d. k. Akad. d. Wissensoh. zu Wien., L, I, 1864. Ebstein. — Beitrftge zur Lehre vom Bau und den physiol. Functionen der sog. Magen- schleimdriisen. Arch. f. mik. Anat., Bd. VI, 1870. EUenberger. — A^ergleichende Histologic der Haussaugetiere, 1887. Frey. — Handbuch der Histologie und Histochemie des Menschen, 1876. Glinsky. — Zur Kenntniss des Baues der Magenschleimhaut der Wirbeltiere. Cen- tralbl. f. med. Wissensch., 1883. Heidenhain. — Hermann's Physiologic. Also, Arch. f. mik. Anat., Bd. VIII, 1872. Kuczynski. — Beitrag zur Histologie der Brunnerschen Driisen. Internat. Monatsschr. f. Anat. u. Physiol., Bd. VII, 1890. LYDIA M. DE WITT. 203 Maziarski. — Ueber den Bau und die Einteilung der Drusen. Anat. Hefte, Bd. XVIII, Heft 1, 1901. Oppel. — Vergleichende mikroskopische Anatomie dei- Wii-beltiei-e, Bd. I and II, 1876. Partsch. — BeitrSge zur Kenntniss des Voi'derdarms einiger Amphibien u. Roptilien. Arch. f. mik. Anat., Bd. XIV, 1877. Renaut. — Gaz. m6d. de Paris, 1879. Salvioli. — Quelques observations sur le mode de formation et d'acoroissement des glandes de I'estomae. Internat. Monatsschr. f. Anat. u. Physiol., Bd. VII, 1890. Sappey. — Traits d'anatomie descriptive, 1889. Also, Traits d'anatomie g(5ni5rale, 1894. Schiefferdecker. — BeitrSge zur Kenntniss der Drusen des Magens und des Duode- nums. Nachricht. d. Gottinger Gesellsch. d. Wissensch., No. 7. Schlemmer. — Sitzungsber, d. k. Akad. d. Wissensch. zu Wien., Bd. LX, 1869. Sehwalbe. — BeitrSge zur Kenntniss der Driisen in den Darmwandungen. Arch. f. mik. Anat., Bd. VIII, 1872. y. Ebner. — KoUiker's Handbuch der Gewebelehre des Menschen., 1899. Watney. — ilinute Anatomy of the Alimentary Canal. Phil. Tr. of Royal Society of London, Vol. CLXVI, 1877. THE MENTAL STATE OF ANAEOHISTS AND OF OTHEES WHO KILL OE ATTEMPT THE LIFE OF EULEES OE PUBLIC PEESONAGES. RICHARD DEWEY, A.M., M.D., Wauwatosa, Wisconsin. It is to be said at the outset that in inquiring into the mental state of the persons who commit or attempt public or political homicide, a very great vari- ety of mental "sorts and conditions" is met with. The question whether there is mental disease or unsoundness is in some cases far from being a simple one for two reasons; First, human judgm'ent is fallible and often placed at a great disadvantage in reading the hidden work- ings of the mind; second, the criteria of insanity are not established as yet upon an accepted physical basis which is capable of objective demonstration. The same phenomena may be diversely interpreted by persons of equal expertness. Let us consider the question whether the mental state of the anarchist is in itself an evidence of unsound mind. This may be considered in connection with a large body of strange and bizarre opinions found in numerous groups and bodies of men at the present day and in all the past history of human thought. Delusions and illogical and unsound opinions are found exercising their sway over the human mind in all ages and among all peoples. It is scarcely necessary to mention the various " crazes " which have swept their way through communities and nations in every age. Religious, political, social, and medical fallacies are always arising, ragingfor a time and then giving way to the truth which is mighty and in the end must pre- vail. In earlier times the Crusades, delusions of witchcraft, of " the elixir of life," of the "philosopher's stone, " of mesmerism, showed that contagion and propa- gation of diseases and disorders rule in the realm of ideas as in the physical world. In our own time the same thing has been shown by various aberrations seizing upon both the best and worst in the community — aberrations alike only in being more or less widely contrary to logic and reason. Adventism (Millerites), spiritism, and mediumistic performances in the domain of reli- gious mysticism, the "silver craze" in that of finance, Eddyismand Dowieism in the religio-medical world, illustrate my meaning, as do innumerable other fallacious systems of half truth or whole falsehood — like osteopathy, theoso- phy, hypnotism, and a legion of other fads and fancies, each, which it is evi- dent, makes a powerful appeal to large bodies of men, most of whom are sincere in their beliefs and normal in their practical relations, but who are misled by 204 RICHARD DEWEY. 205 credulity in one direction, usually in a field whc^re theories, capable neither of proof nor disproof, lead the judgment captive. In this fantastic region the question of Shakespeare's song, " Tell me, where is fancy bred Or in the heart, or in (he head? " must be resolved in favor of the heart, or the realm where emotion obscures intellect ; for logical thought can have nothing in common with these strange freaks of human judgment, among which must be classed the propaganda of Nihilism, " dynamitism, " and the more violent phases of socialism and com- munism. Bj' anarchists, I mean disciples of the doctrine that the social and gov- ernmental order of to-day is to be overturned and destroyed by carrying out a program of universal destruction, especially of rulers and the heads of gov- ernment, and it is only with reference to the outcome of anarchy in crimes of personal violence that I am here concerned. For my purpose the postulates of anarchy need not be more fully described. The ideas at bottom of this doctrine are by common consent opposed to reason as well as all justice and law. The essence of anarchy is unreason, and its ideas may be described as delusions. They differ, however, from insane delusions in not being necessarily the outgrowth of a diseased brain. A brain practically normal is capable of taking on the delusions of anarchy as it is of embracing those of ■n-itchcraft, hoodooism, mesmerism, Dowieism, or theoso- phy. Delusion is not an evidence of insanity when growing out of ignorance or evil passion; something more is needed to establish insanity. It is needful that an opinion in order to constitute an insane delusion shall originate in a diseased condition of the brain, and that the same brain in a state of health would not entertain such an idea. Many delusions are met with that are not insane, e. g., those pertaining to spirit communications, the notion that the earth is flat, or that the sun revolves around our world. Many persons may normally entertain these ideas, but if an educated and intelligent man believes them it is with him an insane delusion and shows his brain is working abnormally. So there are numerous persons who may normally entertain the ideas of anarchy and in whom these ideas are no indication of brain disease. The same may be said of many fanatical and mystical ideas. The mental state of the anarchist seems at first sight so unusual and un- natural that a tendency is shown in many minds to regard anarchical ideas as insane; and, on the other hand, the hatred and detestation of anarchy are such as often to prevent the recognition of mental disease when associated with anarchy, as is sometimes the case, in persons who commit or attempt political assassination. There is certainly no necessary connection between anarchy and insanity. 206 THE MENTAL STATE OF ANARCHISTS. and the anarchist, pure and simple, should be held legally responsible for his acts. There are cases, however, frequently engaging public attention in which persons mentally unsound and irresponsible espouse the cause of anarchy and commit deeds of violence ; such persons are, indeed, especially prone to extreme thought and action of every kind, by reason of a defective brain. The true view will recognize that there are persons primarily insane who, because they are insane, adopt the views of anarchy; in other cases, persons who are primarily anarchists may subsequently and secondarily become insane, but the simple holding of the opinions of anarchy cannot in itself be held to indicate an unsound mind or irresponsibility for acts committed. To take anarchical views or acts as a proof of insanity would be to assume that all fanatical and irrational opinions as those of Shakers, spiritists, Dowieites, etc., prove their advocates to be insane. Indeed, the only wayto arriveat the truth in any case is to study that case upon its merits; to understand the motives of the individual, and to learn his previous history physically and psychologically, as well as the present state. It is better to try to understand the degree of moral responsibility of each individual in the light of all obtainable facts than to inquire, " Is he, or is he not insane?" since insanity is at best an indefinite term. What constitutes insanity in a legal sense has never been determined, nor has any agreement of legal authorities been reached, their decisions being found constantly to conflict. Even the complete medical definition of insanity must await the elucidation of a physical basis upon which it rests. Furthermore, an added element of confusion has been introduced by the doctrines of degeneracy and decadence, about which we hear so much and really know so little, for our present knowledge admits of no safe deductions. We see the same physical stigmata in one case associated with soundness and brilliancy of mind, and in another with defect and marked abnormality, with- out being able to say what is the determining factor, although we may cheer- fully grant that the stigmata are more commonly associated with mental abnormality. My contention is, that in each case of crime of the kind under considera- tion we must determine what circumstances there are, if any, modifying or limiting moral responsibility, and further, that in the cases of a large ma- jority of so-called anarchists, it will be found that the accused is not to be regarded as irresponsible for his act in the eye of the law. A recent brilliant study of Spitzka,^ to which I am greatly indebted, gives 261 cases of pubhc assassination and assault in which, after ehminating every doubtful case, he finds eighty per cent of the assailants were sane. In order to contribute to greater clearness of view, I will now cite some ' Regentioides Not Abnormal as a Class. A Protest Against the Chimera ot "Degener- acy. " E. A. Spitzka. Phila. Med. Journ., February, 1902. RICHARD DEWEY. 207 instances illustrative of the mental state existing in a variety of cases of homi- cide or assault upon rulers or public personages. These will exhibit in some cases insanity as the sole cause of the act in question; in others an impaired or modified responsibility from mental defect is evident, and in others no evi- dence of mental defect was present, unless all criminals are to be considered irresponsible, or all persons having peculiar or untenable ideas are to be regarded as insane. In the year 1800 James Hadfield shot at King George III. in Drury Lane Theater. His attempt was without effect, but he was tried upon the charge of treason before Lord Cliief Justice Kenyon and the associate justices, and some of the greatest lawyers in England took part in the case. I will quote from the speech of Hon. Thomas Erskine, who defended him, the essential point of defense being that the prisoner's act was the direct result of an insane delusion. Lord Erskine stated, the prisoner "imagined that he had constant intercourse with the .-Umighty; that this world was coming to an end, and that, like our Blessed Saviour, he was to sacrifice himself for its salvation. * * * He went to the theater to perform, as he imagined, that blessed sacrifice, and because he would not be guilty of suicide, though called upon by the imperious voice of Heaven, he wished by the appearance of crime his life might be taken away from him by others. " In other words, Hadfield's power of reasoning and of self-control was im- paired by his delusions to such an extent that he could not justly be held responsible, and he was acquitted. It will be noticed that he was taking an indirect method of committing suicide, and suicide has many times been the outcome and intention of the regicidal acts, both in cases of sane and insane regicides, for suicide, although popularly supposed to be an insane act, is by no means to be so regarded in all cases. A case in point is the curious one of the Itahan anarchist Sperandio, who, like Bresci, the assassin of King Hum- bert, was a member of the Paterson, New Jersey, group and had been de- tailed for the very act which Bresci afterward committed, but, being "too lazy to cross the Atlantic," and knowing that either a stiletto between the ribs here or a dungeon for life on the other side was in store for him, killed a royalist fellow countryman and then ended his own life.^ Another case of assassination of a public official which was the direct outcome of an insane delusion, was that of McNaghton who shot and killed Mr. Edward Drummond in January, 1843, under the impression that Mr. Drummond was Sir Robert Peel. Mr. Drummond was in reality private secretary to Sir Robert Peel, and McNaghton, watching Sir Robert's house and seeing Mr. Drummond come out, followed and shot him on the street. It was shown in the trial that a certain man had,.in the November previous, pointed 1 The Mental State of Czolgosz, and of Assassins Generally. The Medical Critic, Novem- ber, 1901. E. A. Spitzka, N. Y. ' 208 THE MENTAL STATE OF ANARCHISTS. out the house of Sir Robert Peel to the prisoner, who had then said with an oath, "Damn him, or sink him!" and malice and premeditation of the act were shown; but it was also conclusively shown that the act was the direct outgrowth of insane delusions, hence that the prisoner was irresponsible. McNaghton imagined he was constantly followed by spies day and night. He was a victim of what are known as " delusions of persecution, " and had applied to magistrates to have the imaginary persecution stopped. He had traveled from Scotland to England, and from England to France, in order to escape these imaginary persecutions, but they followed him everywhere he went. He thought he saw men shaking fists and sticks in his face, and one brandished a handful of straw at him, signifjdng, as he thought, he should be given a "straw bed in a dungeon." The jury in this case acquitted the prisoner on the ground of insanity. A class of attempts upon public men may be here cited which do not come under the head of regicides or "magnicides, "these latter being persons who act impersonally, as they' believe, for some great cause, or to carry out some mission. The cases I refer to are those of "pseudo regicides" whose motive is some personal grievance, and who may be either sane or insane. Such a case was that of Jacob, an inventor of apparently unbalanced mind, who dis- charged a bullet into the air in front of President Carnot's carriage as it en- tered Longchamps during a military review in July, 1890. His motive was to attract attention to his inventions.' Such cases might be multiplied; some of them present no mental abnormality but malice or egotism. The case of Guiteau has been so much discussed it need not be here rehearsed, but I refer to it to sa}^ it is assimilated to the above cases in the fact that a per- sonal grievance and motive of revenge entered largely into his murder of President Garfield. That Guiteau was insane, and his responsibiUty decidedly curtailed by a diseased and defective brain, cannot be disputed ; he was not, however, to any extent infected with the views of anarchy. The above cited cases offer a strong contrast to those now to be presented of assassinations in which the motive was anarchy pure and simple. First, the assassination of President Carnot of the French Republic, at Lyons in June, 1894, by Caserio, an Italian anarchist.^ This act, Hke that of Czolgosz, was conceived and carried out by Caserio alone without the aid of accomplices, so far as known (though inspired by anarchistic lectures and writings, particu- larly those of female anarchists, who have often been shown to constitute a powerful influence in these fanatical minds). Caserio had premeditated the ' Cited by Regis in a paper entitled, " The Regicides. " Journ. of Ment. Pathol, Vol. I, No. 3, page 135. ^On a superficial glance one would say the Latin race had produced a large majority of the regicides. I ascertained the birthplaces of thirty-two, of whom twelve were natives of France, seven of Italy. RICHARD DEWEY. 209 deed for months, and learning that President Carnot was to visit Lyons, he planned to gain access to him there. He exeeuted his act with boldness and energy. He struck his A-ictim through the heart, in the public street in the midst of a procession and illumination, with a sharp poignard sixteen centi- meters long, driving it to the hilt. There was a remarkable similarity in the acts of Caserio and Ravaillac, who likewise stabbed Henry IV. of France in a carriage during a procession, and also of Bresci, who shot King Humbert at Monza. Caserio at the moment of his crime uttered a cry, "Vive la Revolution! Vive I'anarchie!" exulting in his act, and drawing attention to it with pride; and later, when brought to the guillotine, he shouted again, "Vive I'anarchie!" He was carefully examined for the government by Lacassagne and declared sane. Regis, an able French writer on insanitj-, sought to controvert this view that Caserio was sane. Noth- ing, however, it seems to me, was brought out by him which would prove insanity unless opinions and acts of anarchy in themselves form such a proof. Not unlike Caserio, Bresci, who shot and killed King Humbert of Italy in July, 1900, at Monza, drew his inspiration solely from anarchy, having been allotted to the act bj- the Paterson, New Jersey, group of anarchical conspira- tors. Bresci went to Italy, sought his opportunity, like Caserio, like Czolgosz, and so many others, in a public celebration, and fatally shot the king in his carriage while a public procession was in progress. No claim was advanced that his mental condition was anjrthing but sane and normal, and the motive of his act was the acknowledged one of anarchy — destruction of all government and all authority. _ It is true he afterward committed suicide in prison, but if this act grew out of insanity, there is no proof that the insanity existed previous to his act, but rather the contrary, while the development of insanity under such circumstances is not surprising. The act of Czolgosz is too recent, too familiar, and too deeply impressed upon all our hearts to require review in detail. I only wish to emphasize certain characteristics it has in common with the other purely anarchical attempts. It was carefully premeditated and planned with satanic ingenuity. Czolgosz has himself related how he went to Buffalo for this sole purpose; how he followed the president's movements for two days in a feverish state of excitement, meditating upon the, to him, inspiring eloquence of Emma Gold- man, his only fear being that fate might after all deprive him of the joy and glory of ending the life of a man honored and beloved by the people as few have been in any age or country. The diabolical "ruse" of enveloping his right hand, which grasped the pistol, with a handkerchief as if the hand were dis abled, and of giving the left to shake, thus leaving his victim more defenseless and giving himself a fatal advantage — all this is but too well remembered. His avowal all the way through that he was an anarchist, that he had no confederate, as if he feared part of the glory might be given others, his exhibition of cool- 210 THE MENTAL STATE OF ANARCHISTS. ness even in the chair of electrocution, the lack of any evidenceof insane delu- sions or mental unsoundness in his conversation or conduct, and the findings of the post-mortem examination of body and brain — all constitute a case in which there is nothing that suggests diseased mind, unless all abhorrent acts and all fantastic beliefs are proof of unsound mind. These cases of Caserio, of Bresci, of Czolgosz, are all typical of anarchy. They furnish unmistakable evidence of a murder conceived and committed with malice aforethought, by men capable of unusual self-control, free from passion or excitement, free from the symptoms of insanity. It was murder in the first degree; these men presented no evidence, mental or physical, of insanity. Caserio and Czolgosz were carefully examined by competent men, and Bresci 's career and all we know of him raises no presumption of insanity, but the reverse. Finally, their acts were avowedly prompted by the doctrines of anarchy. These doctrines are, indeed, delusional, but not insanely so, i. e., irrational ideas growing out of disease or defect in the brain. Again I em- phasize the difference between insane delusions and ordinary or sane delu- sions. The most ridic\ilous ideas may be perfectly sane, like the delusions of perpetual motion, or the sun revolving around the earth. Reverend Jasper of Richmond, who claimed "The sun do move" around the earth, and wrote a book to prove it, was not insane, he was simply ignorant. His brain was not diseased nor defective, considering his education and his opportunities. The ideas of the anarchists are delusions, but these misguided men are strictly accountable for their beliefs and the acts that grow out of them, unless there is something else to indicate insanity. It is true, many of them are not sincere, and many weak-brained and irresponsible dupes and tools who cannot be held responsible for their acts are dominated by the more masterful spirits; like Sipido, the boy assailant of King Edward VII., then Prince of Wales. The genus known as " cranks " are peculiarly prone to espouse these dangerous ideas; but before any anarchist can be deemed insane, it must be shown by other evidence than belief in anarchy. If such a person has inborn and heredi- tary weakness, or if degeneracy of brain or disease of brain is shown by other unmistakable symptoms, a case of insanity may be made out. One which I will now cite answers this description and differs widely from those of the anarchical murderers. It is a case which shows inborn defect of brain and dis- ease of reasoning power so great as to impair responsibility in the eye of ab- stract justice, although the fallible administration of courts and juries, as we so often find them, caused the madman to pay the penalty of his mad act with his life. I refer to the case of Prendergast, the slayer of the first Carter H. Harrison, mayor of Chicago in 1893. I will recall briefly the outlines of the case, which were as follows : Patrick Eugene Prendergast on October 28, 1893, shot and killed Mayor Harrison under the following circumstances: He was a newspaper carrier. RICHARD DEWEY. 211 twenty-five years old, born in Ireland. His father died of consumption while he was a boy ; a brother of his f atlier died insane in Ir(>land ; a brother of Pren- dergast died with consumption ; Prendergast himself was threatened with con- sumption during adolescence. At that time he became peculiar and fond of solitude and never was like other boys, or inclined to associate with them, being of a quarrelsome, morose nature. He was dull and backward in the parochial school, and yet very conceited, fancying, especially, tliat he was wise in matters of law and religion, and he was very fond of argument on these subjects, so much so that he wearied people with his talk. He espoused en- thusiastically the "single tax" ideas. His definition of a good Catholic was that "he is merciful, and believes in the 'single tax. ' " When about twenty he wrote out a form of prayer which he said had been miraculously given him, and sent it to a priest recommending its use as capable of producing miracu- lous resiilts. He ^^Tote letters to the Pope, to Cardinal Gibbons, and to the Archbishop giving advice about religious matters, and to the Secretary of the Treasury at Washington, and to other officials, advising about the nation's finances. After Carter Harrison was elected mayor in 1893, he told his mother that his influence had elected Harrison, and that he (Prendergast) was to be " corporation counsel " of the city. His only knowledge of law was gained from a book he had read, "Easy Lessons in Law. " He called upon Mayor Harrison, who possibly encouraged him to expect some office in a jolcing way; he called upon the incumbent of the office of corporation counsel and by him, also, was facetiously received, and introduced to subordinates as the "new boss. " All this seems to have been taken seriously by Prendergast. He had, as he thought, one especial object which he was to accomplish as corporation counsel — this was the elevation of the railroad tracks. The grade crossings of the city of Chicago were at that time the place of death daily throughout the year of an average of at least one person, and Prendergast believed himself divinely ap- pointed to correct this monstrous slaughter of human beings, concerning which _ the denunciations of the press, the indignation meetings of the people on the one hand, and the supineness and venality of the city government on the other, were sufficient to engender fanaticism in a weak brain. Add to this the office- holder's imagination that he and the office are created for each other, and finally the bitterness of an office-holder's "hope deferred," and you have the ground on which Prendergast 's diseased mind developed the homicidal pro- pensity. While the world looked upon him as an insignificant and even pusil- lanimous creature, he regarded himself as having a divinely conferred mission to elevate the railroad tracks, and the mayor as one who had broken faith with an important political supporter and an agent of the Almighty, and was thus an obstacle to the execution of the divine will which must be removed out of the way. He called at the mayor's residence on the evening of October 28th, carrying with him a loaded revolver. The mayor was resting, having been busy 212 THE MENTAL STATE OF ANARCHISTS. at the World's Fair all day, and he was told to return in half an hour, which he did, and was shown into Mayor Harrison's presence. He had been with him but a few moments when several shots were heard. Three of the bullets fired took effect, and the mayor fell and almost instantly expired. Prendergast took himself at once to the police station and gave himself up, announcing his crime. The next morning he asked for the newspapers and when given a supple- ment called for the front sheet which he knew would contain an accoui^t of his act. The prisoner when examined with reference to his mental state, exhibited an intensely self-satisfied and egotistical character. He said, " I have firmly believed there was something wrong with the world ever since I was fifteen. " He spoke of having helped men and causes by his prayers; said he "lived to do good. " When asked the reason for taking the mayor's life, he replied," Christ's laws were the only laws to obey. " And when asked if Christ told him to kill Carter Harrison said, "You talk to me now as if you thought I wasn't sane." When asked if he felt it his mission to elevate the tracks, he answered, "If I say it was my mission I don't think that shows that I am mentally unbalanced. ' ' In this he showed, as he did all the way through, a scorn of the idea that he could be insane, such as is common with criminals both sane and insane, who are proud of their act, and this pride in their crimes seems to be characteristic, both of paranoiacs like Prendergast and of anarchists. He said the mayor was at peace and he had felt better since the deed was done, as it was "like going through hell " for some time before. During the trial he insisted on being spoken of as "Mister" Prendergast. He frequently interrupted the lawyers on both sides to raise points he thought important. In his speech at the time sentence was passed upon him, he said, " It was the most infamous thing in history to kill Christ on the cross, and it will be as infamous to hang me. " Prendergast was hanged on July 13, 1894, nearly a year after the assassi- nation. Although he was insane, he was like many of the insane, partially responsible, and as a choice of evils, his hanging was apparently preferred by the jury to sending to an ordinary asylum for insane, the only alternative, as the criminal asylum of Illinois only received convict insane. Prendergast 's cranium was malformed, and marked physical as well as mental "stigmata of degeneration" were present in him. I do not emphasize the former, as we do not yet know how much significance to attach to them. Contrasting his case with those of Caserio, of Bresci, of Czolgosz, the difference between a sane and insane public assassin is plainly seen. I will now briefly review the circumstances of a number of cases typifying the varieties of so-called regicidal acts. In 1819 Karl Sand killed the Austrian minister, Kotzebue. He had meditated the act two years, and wrote in his journal the preceding year: "Lord, let me strengthen myself in the idea I RICHARD DEWEY. 213 have conceived of delivering humanity through the holy sacrifice of Thy son. Will that I may become a Christ to Germany, and that like Jesus, I may be strong and patient in my pain. " Sand's case, lilce many in early and recent times, is one of fanaticism, sometimes springing from exaggerated and un- healthy religious or mystical brooding, sometimes from meditation on social or political wrongs, real or imaginary. Such meditation in an egotistical or depraved mind or an essentially ill-balanced mind, or in a person of impassioned or impulsive temperament, may lead to acts of revenge and violence; and it is, perhaps, just to say that a person of such characteristics should not bear the full measure of responsibility and of punishment that an average individual sustains, and yet such a person may be responsible for cultivating, instead of repressing and restraining, overweening self-indulgence and conceit till these become diseased propensities. Louvel in 1820 assassinated the Due de Berri, heir to the throne of France, believing the " Bourbons to be the destroyers of France. " Three accomplices were executed with him and one imprisoned. In 1836 Alibaud attempted the life of Louis Philippe. He died by the guillotine, shouting, " I die for liberty, for the extinction of the infamous monarchy ! " He was one of six who, between 1835 and 1846, attempted the same monarch's life. I have been able to find Uttle which bears on the mental state of these regicides, for to this feature but scant attention was given until veiy recent times. In 1855 two attempts upon the life of Napoleon III. were made by Pinare and Bellamere. In 1855 Orsini in a conspiracy with others, attempted to ex- plode a bomb under the carriage of Napoleon III. in the streets of Paris. He was executed with three others. In 1840 an attempt was made upon Queen Victoria of England by an epileptic boy named Oxford, who was unquestionably imbecile in mind. In 1861 Becker, a nationaUst, assaulted King WiUiam of Prussia from political fanaticism. In 1878 the same monarch, then Emperor WilHam I. of Germany, was attacked by Hodel, a socialist, and in the same year Nobiling shot at the emperor and then shot himself. Nobiling's father was a suicide, and he was himself probably insane. In 1866 Bismarck was at- tacked by BUnd, a communist, and in 1874 by KuUman, an ultramontane, who saw in Bismarck an enemy of God. From about 1860 onward, in increasing frequency , these attempts were made. Alexander II. of Russia was the object of attempts in 1866, in 1867, in 1879 (two), in 1880, and finally, in 1881, the fatal bomb destroyed both himself and the nihiUst who hurled it. Five supposed accomphces were hanged two days later. The Winter Palace explosion in 1880, and the dynamiting of a railroad train supposed to have the Czar on board in 1879, both cost a large number of innocent lives. Little is known of the mental state of the Russian nihihsts. Humbert I. of Italy was assaulted by Passanante in 1878, and fatally shot by Bresci in 1900 as above detailed. Passanante became insane ; whether so or not 214 THE MENTAL STATE OF ANARCHISTS. at the time of the assault is not known. He had two sisters who were insane. Bresci, as stated above, committed suicide in prison. Two attempts were made upon Franz Josef of Austria, one by Oberdank, who was executed, the other by Ragosa; the latter was a member of a band of Italian anarchists.- King Amadeus of Spain was attacked by anarchists in 1872, and Alfonso XII. in 1878 and 1879. The present King Edward VII. was assaulted when Prince of Wales in 1900 by a boy fanatic, Sipido, while traveling through Brussels. In 1898 the Empress of Austria was fiendishly assassinated by an Italian anarchist, Lucheni. This man was subjected to life imprisonment, as the death penalty is not enforced in Switzerland. Lucheni was an anarchist, pure and simple. He wrote from prison to the press denying that he was a " degenerate " d, la Lombroso. Some of the remarkable cases of an earlier time were ttose of Henry III. of France, killed by Clement in 1589; Henry IV., as already mentioned, by Ravaillac in 1610 ; Louis XV. of France was assaulted by Damiens. Ravaillac and Damiens were both religious enthusiasts and both had "visions" of an hallucinatory character. Damiens was subjected to the most terrible tortures that could be devised, and was considered by Michelet the most remarkable recorded instance of withstanding such an ordeal stoically. Time -and space forbid further additions to these gruesome lists. We see a great variety of motives, and mental condition showing through these facts, meager though they are — anarchy, religious and political fanaticism, malice ■ and revenge, insanity in protean forms, paranoia, delusional mania, imbecility, the epileptic state, border-land cases, in which a satisfactory opinion often is unattainable. The sad experience of our own country in the occurrence of these catas- trophes deserves a word. Three times in less than forty years our chief magis- trate has fallen victim to the assassin's bullet. This fact seems to show that popular government is no safer than imperial, royal, or despotic rule, but a little reflection puts the matter in a different light. The martyrdom of Lincoln may be eliminated as growing out of the unprecedented conditions of the Civil • War. Furthermore, anarchy, properly speaking, had nothing to do with Garfield's death. Finally, great allowance must be made for the conditions of practically unlimited access that have prevailed. It is safe to say that if our presidents had adopted the same safeguards that prevail with European rulers, none of these three deplorable fatalities would have occurred. It is, indeed, remarkable that we have enjoyed so great an immunity; for anarchy, in its bhnd fury, is as ready to destroy a republic as a despotism, and it is worthy of remark that the victims chosen by anarchy have been almost without exception the most liberal, enlightened, and well disposed rulers — Alexander II., Carnot, King Humbert, McKinley, all benevolent and earnestly con- cerned for the welfare of their people: while the distressful death of the RICHARD DEWEY. 215 Austrian Empress reveals a depth of insensate hatred and cruelty almost beyond conception. Anarchy has grown mainly from centuries of wrong and oppression in other ages and countries than ours. It has been, to some large extent, the merciless extortion of the ruling classes of the old world in bygone times and still in our own day, that has produced this blind rage which chooses its victims so blindly. It must be admitted, however, that our own country is not free from the criminal rich and the injustice and cruelty that madden and craze their victims. The task is ours, not only of repressing this unspeakable and abhorrent machinery of murder, but of restraining the soulless greed and sel- fishness which produce it. Those who corrupt legislation at its fountain head, debauch executive power in its highest function, and infect the courts of justice with the poison of bribery and perjury, are anarchists as truly, though perhaps unwittingly, as the desperate spirits who employ the poignard and the pistol. It is our duty to give even the devil his due, and we must acknowledge that evil and hateful as is the work of anarchy, some of its converts have shown a courage and de^-otion to a bad cause which rise toward sublimity. These men are readj' to die with their victims or exult in their achievements even upon the bad eminence of the scaffold, or the guillotine. It is doubtless true, too, that this implacable and terrifying power of anarchy has served to curb the other- wise unbridled lust of the despot and the extortioner. Even the mad act of Prendergast may have served to hasten the day of track elevation in Chicago ; and when at last the indomitable courage and energy and passionate resist- ance to -nTong are turned into the peaceful channels of agitation, discussion, and persuasion, when the lesson is learned that violence and murder have no place in the life of the nation, and when, furthermore, the extortion and injus- tice that create anarchy are curbed, we shall see the end of this scourge which so sorely afHicts the nations. In conclusion, I remark that there is no necessary relationship between anarchy and insanity. The mental state of the typical anarchist must be re- garded as normal, and he is responsible for his acts; he is simply criminal, not insane. At the same time, there are persons inherently and primarily insane who adopt the principles of anarchy and belong with the class of paranoiacs, so-called degenerates, or " cranks. " Their insanity, however, exists independ- ent of their anarchy, and this latter is never to be taken as evidence of insanity. Finally, there are cases also where the slaying of public men is the direct out- come of insane delusions, hence not punishable as an intentional crime. These, too, have nothing to do with anarchy. And, lastly, there are acts of assassination of men in public places which are simply the outgrowth of private personal hatred and revenge, and which are to be regarded as common murder. There are, also, pseudo-anarchists who attempt some act of public assault to attract attention to themselves or secure redress for some grievance, real or imaginary. THE CHANGES PEODUCED IN THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT BY SPLENECTOMY. ALDRED SCOTT "WAE.THIN, M.D., Ph.D., Junior Professor of Pathology, University of Michigan. In the preliminary report of a study of the histology and pathology of the human hemolymph glands (Journal of Medical Research, July, 1901), the hypothesis was advanced that these glands under certain conditions may com- pensate for the spleen. This theory was based partly upon the general simi- larity in structure of hemolymph nodes and spleen, their common function of hematolysis, and the occurrence of transition-forms resembling accessory spleens, but chiefly upon the autopsy findings in a case of splenic anemia in which there was throughout the mesenteric fat a new-formation of hemo- lymph nodes resembling splenic tissue. In other forms of anemia, hyperplastic hemolymph nodes showing greatly increased hemolysis were also found. The evidence afforded by all these points was so suggestive of the existence of such a compensatory process that investigation along experimental lines was at once undertaken, with a view to the definite solution of the problem. Though the resemblance of the hemolymph nodes to the spleen in both struc- ture and function had been noted by Vincent and Harrison, and later by Drummond, the present paper records the first experimental research regard- ing the nature and function of these organs. (See notes.) The anatomical changes following splenectomy have been studied by numerous observers, but the results of these investigations are greatly at variance. This is particularly the case with reference to the question of total or partial regeneration of the spleen, and the occurrence of changes in the lymph nodes and bone-marrow interpretable as compensatory in nature. As early as 1680, Zambeccari observed in the mesentery of a dog four months after splenectomy numerous newly-formed nodules, isolated or collected into groups, of a yellow color, resembling lymphatic nodes. According to Lussana, Gerlach and Eberhardt were said to have observed complete reproduction of the spleen in the frog after total extirpation. Beclard, Verga, and Legros experimented with rats and dogs, and denied the possi- bility of such a regeneration after splenectomy. Legros disputed further the possibility of any compensation on the part of the thyroid, omentum, or mesen- teric glands. In 1859 Philipeaux performed three total extirpations in white rats, and found, eighteen months after, a new-formation of splenic tissue at the site 216 ALDRED SCOTT WARTHIN. 217 of the original organ. In 1861 Peyrani, after a number of experiments, attacked Philipeaux's conclusions as erroneous. The latter repeated his investigations, using young rats and rabbits, and was unable to find any evidences of regenera- tion. Explaining his first results as tkie to the fact that portions of splenic tissue had been left attached to the splenic vessels, he repeated his experiments with young rabbits, removing all but a small portion of the spleen. On au- topsy complete regeneration of the organ was found, the new tissue presenting the appearance of normal splenic tissue. Peyrani, carrying out further investi- gations, again attacked Philipeaux's work and denied the possibility of any form of regeneration. In 1880 Tizzoni and Pileti observed in splenectomized dogs an increase in size of the retroperitoneal and thoracic lymph nodes, the enlarged glands being of a red color. They attributed these changes to a post-operative lymph- adenitis. In the same cases there was found a transformation of the fatty marrow of the long bones into red marrow. In two cases they found a new- formation of spleen-like nodules in the great omentum, in one old dog fifty- four days after splenectomy, and in one young one three and one-half months after. The newly-formed nodules showed all stages of development, and from the appearances presented, Tizzoni and Fileti believed the new splenic tissue to be formed directly from the adipose tissue by a process of absorption of the fat, conversion of the fat cells into reticulum, followed by a leucocyte infiltra- tion and an active proliferation of the endothelium in the neighborhood of the small arterioles, leading to the formation of Malpighian corpuscles. Around these the continued proliferation of the reticulum and endothelial cells pro- duced a pulp-like tissue, between the cells of which there was an extravasa- tion of red blood-cells. Around the whole a connective-tissue capsule was formed. In 1882 Tizzoni published his work upon accessory spleens and the new- formation of splenic tissue following pathological processes of the primitive organ. • In numerous cases of indurative splenitis, in the dog, he observed a new-formation of splenic nodules in the gastro-splenic ligament, but rarely in the great omentum. On the other hand, in cases of splenectomy he found numerous newly-formed lymphoid nodules throughout the subperitoneal fat, over the diaphragm, and in the pelvic, sterno-abdominal, and subcutaneous adipose tissue. In the splenitis cases the new tissue resembled splenic pulp, containing occasional groups of more closely massed leucocytes suggesting to a certain extent the Malpighian corpuscles. In the splenectomy cases the new tissue was for the most part provided with Malpighian corpuscles and a distinct capsule. In the new splenic tissue formed after splenectomy, he found nucleated red cells ; these were not present in the newly-formed lymphoid nodules in the cases of indurative splenitis. Winogradow, in the same year, reported observations of the anatomical 218 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. and hematological changes occurring in splenectomized dogs. In one animal killed one hundred and thirty-two days after total extirpation of the spleen, the cervical, axillary, inguinal, and mesenteric glands were enlarged, soft, moist, dark or bright red on section, the cortical portions particularly resem- bling spleen tissue. The color he found on microscopical examination to be due to great numbers of red blood-cells lying in the reticular spaces and filling up the lymph-sinuses between the follicles and lymphoid cords, and also partly filling the peripheral sinus. Evidences of hemolysis were present, finely granular brown pigment being found in the sinuses and in the swollen cells of the reticulum. Winogradow regarded the finding to be similar to the ap- pearances described by Tizzoni. In two animals killed, five hundred and' seventeen and seven hundred and sixty days, respectively, after splenectomy, the same changes were found in all of the lymph glands, with the difference that they were firmer, their surface more irregular, and color brownish-red. The number of red blood-cells in the sinuses was not so great, but the amount of pigment greater, the pigment being present not only in the sinuses, but also in the lymphoid tissue. The trabeculse showed further a thickening resulting from the formation of new connective tissue. The presence of blood-contain- ing sinuses was explained by Winogradow as due probably to a diapedesis into the lymph-sinuses and not caused by artificial hemorrhage. He regarded it as very probable that the diminution of red cells in splenectomized animals during the first year after the operation could be caused by the continued diapedesis and destruction of the red blood-cells in the lymph-sinuses. The gradual return of the blood count to the normal he considers good evidence that the processes of blood regeneration are not interfered with, but are prob- ably increased. Further, Griffini in this same year observed in dogs a partial regeneration of the spleen after the removal of a portion of the organ. Tizzoni's observations were opposed by Foa in 1883, who denied wholly any regeneration of the spleen in the dog after total splenectomy. He further affirmed that the newly-formed nodules described by Tizzoni. were present before the operation, the difference in color, etc., being due to hyperemia or infarction, or other pathological changes, and that no compensatory function could be ascribed to them. In the same year Zesas experimented with rabbits and found in these animals, one week after splenectomy, that the mesenteric glands were slightly enlarged, but of normal color and consistence. In animals four weeks after splenectomy the liver was enlarged and hyperemic, the mesenteric and bron- chial glands greatly swollen, on section dark red, hyperemic, and firmer in consistence. In other animals seventeen weeks after the operation, the liver was hypertrophic and hyperemic, the bronchial and mesenteric glands heavily pigmented and of hard consistency. The kidneys were also hyperemic and pigmented. Zesas cites further experimental work of Hegar and Simon upon ALDRED SCOTT WARTHIN. 219 cats; in three instances, enlargement and pigmentation of the mesenteric glands were found after splenectomy. In 1884 Tizzoni made a series of total splenectomies on the rabbit, but could confirm no new-formation of the spleen nor any evidences of compensation on the part of thyroid, thymus, and lymph nodes. In the same year Hosier found in a dog, ten months after splenectomy, numerous spleen-like nodules of the size of a pea to that of a bean, dark red in color, and on section closely resembling spleen-pulp. They were scattered throughout the great and lesser omentum. I\Iicroscopically the nodules presented appearances identical with those in the cases of Tizzoni and Winogradow. Hosier, however, regarded them not as newly-formed splenic tissue, but as neoplasms, "hemorrhagic telangiectatic lymphoma. " Hyperplasia of the lymphatic glands was not constantly present in his cases. ^losler concluded from his investigations, that following splenectomy there is a compensatory action on the part of both lymph glands and bone-marrow, the latter appearing to play an important role, in one case resembling that of leukemia. These changes are not constant ; in one dog, killed eleven months after splenectomy, there was not the slightest trace of any alteration having taken place in any of the remaining lymphoid structures. In 1886 Gibson found, in splenectomized dogs, enlarged mesenteric glands containing both nucleated and non-nucleated red blood-cells in their sinuses. Eternoid, in 1888, found in a young fox, one hundred and sixty-one days after splenectomy, a splenic nodule in the omentum, and newly-formed lymphoid nodules in the mesentery, the other lymph glands of the body being enlarged and of a bro\Amish color. In 1894 Vulpius observed no enlargement of the lymph glands in splenectomized animals dying soon after the operation, or killed at the end of five months. Laudenbach, in 1895, found that in splenec- tomized dogs hyperplasia of the lymph glands did not occur constantly, and that signs of increased blood-formation were present in the bone-marrow onl}'. Ceresole in the same year found in splenectomized rabbits no hyper- plasia of the lymph nodes and no new-formation of the marrow. In the reported cases of splenectomy in man, enlargement of the lymph glands has been noticed but a few times. Schultz, in 1856, observed enlarge- ment of the axiUary glands in a splenectomized woman twenty-two years of age. Of the one hundred and seventeen cases of splenectomy collected by Vulpius in 1894, enlargement of the lymph glands was noted only in three, though special attention had been paid in a number of cases to the possibility of such an occurrence. Ceci, in 1889, observed a temporary increase in the size of the tonsils and thyroid after splenectomy. Czerny, Kocher, Lennan- der, and Riegner saw temporary lymphatic enlargement after splenectomy, and in 1900, Bolton reported a case in which after splenectomy for traumatic rup- ture there was a general enlargement of all the lymphatic glands. On the 220 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. Other hand, Czerny, Billroth, Albert, Trendelenburg, and others have failed to observe any change in the lymph nodes, tonsils, or thyroid after successful splenectomy. No autopsy reports of such cases exist as yet. Of special interest is the case of Hodenpyl, reported in 1898, of the autopsy findings in congenital absence of the spleen. Enlargement of all the lymph nodes of the body, axillary, cervical, bronchial, mesenteric, and the solitary follicles of the intestine, was found with a new-formation of lymphoid tissue in the adrenals and throughout the connective tissue of the liver. The case reported by Albrecht in 1896 of numerous small "accessory" spleens scattered through- out the peritoneum, the primary spleen being of the size of a small walnut, though explained by this observer as due to some congenital disturbance of the spleen anlage, is open also to the interpretation of a compensatory pro- cess. A number of cases have been observed of the occurrence of very large accessory spleens, the enlargement being interpreted as compensatory in nature. From this brief review of the literature it will be seen that none of thess investigators was aware of the existence of modified lymph nodes constantly containing sinuses filled with blood, and to which the term "hemolymph gland" has been applied. It is, however, very evident that the descriptions given by Tizzoni, Winogradow, and Hosier of the new-formations found in the omentum and mesentery of splenectomized animals apply so closely to these glands that it seems reasonable to conclude that these structures were in reality hemolymph nodes. Hosier's description in particular applies to the hemolymph nodes, and his diagnosis of "hemorrhagic telangiectatic lymphoma " is of peculiar interest. The animal chosen for the investigations recorded in this paper was the sheep, for the reason that the hemolymph nodes in this animal are very numer- ous in the prevertebral fat and are easily found because of their size and dis- tinctive dark-red color. Their normal histology has been very well worked out by Vincent and Harrison, and Drummond. To the work of these investiga- tors my observations have added a number of new points of importance. The normal histology may be briefly condensed as follows : Occurrence. — The more careful the dissection, the greater the number of hemolymph nodes found to be present in the animal; a rough calculation based upon animals examined at the slaughter-house makes the average occur- rence between three and four hundred. They are most numerous in the prevertebral fat of the retroperitoneal, thoracic, and cervical regions, but are present also in the pelvic fat, along the iliac vessels, in the mesentery, omentum, anterior mediastinum, in the fat about the anus, and very rarely on the diaphragm, in the subpleural tissue, and in the axillary region. Gross Appearances. — ^Their average size is that of a yellow mustard-seed or a small pea, but all sizes are found from a pin-head up to a coffee-bean, and rarely as large as a cherry. The smallest ones are usually slightly flattened. ALDRED SCOTT WARTHIN. 221 the intermediate ones more round, while the larger ones are somewhat flattened. Numerous blood-vessels pass to the node, in some oases surrounding it with a plexus, in other cases but few vessels are found. Some have lymph- vessels, others apparently none. The larger nodes have a distinct hilum into which the vessels enter. The exact relations of the vessels and the mode of circulation have not as yet been made out. The surface of the node is usually smooth, but occasionally small elevations are present giving the node the appearance of a minute raspberry. The color is usually- dark red, almost black; the nodes containing more lymphoid tissue and smaller blood-sinuses are lighter, other nodes are of a bi:own or chocolate color. The consistence is soft and elastic, the delicate capsule is easily ruptured, the node then appearing as a small blood clot. On section the appearance of the nodes varies greatly, as the rela- tive amount of lymphoid tissue to the size of the blood-sinuses is hardly ever the same. Smears. — The smears made from the blood contained in the sinuses of the dark-red glands resemble those made from peripheral blood. The propor- tion of leucocji;es may or may not be greater. From the lighter-colored glands containing more lymphoid tissue, and from the chocolate-colored ones, it is difficult to obtain the blood unmixed with the tissue-cells. Surface smears contain a great variety of cell-forms, a detailed description of which must be omitted here. Phagocytes containing pigment or red cells are constantly present. No nucleated red cells were ever found. Microscopical. — On microscopical examination the majority of the dark- red nodes present a thin capsule of connective tissue and unstriped muscle broken at intervals by blood-vessels. Beneath the capsule there is a blood- sinus of varying size from which similar sinuses run through the node in all directions, separating the cords or islands of lymphoid tissue which may be relatively large or small. In some cases the node is almost entirely made up of blood-sinuses. These are traversed by a reticulum of varying amount. A division into cortical and medullary portions may be made in the majority of the nodes, but the latter is much less developed than in ordinary lymph nodes. In the cortical portion, collections of cells resembling lymph follicles are usually present. In some nodes these are not found, the node consisting of delicate trabecules of lymphoid tissue traversing large blood-sinuses and held together by delicate threads of reticulum. The reticulum is apparently lined by endothe- lial cells; in its meshes are always present large phagocytes containing blood- pigment and disintegrating red cells. In the brownish-colored nodes these phagocytes are so numerous as to fill up the sinuses almost completely; at the same time there is apparently a thickening of the trabeculse and reticulum and an increase of the lymphoid tissue, the nodes coming to resemble more closely the ordinary lymphatic glands. All possible stages of transition are found, the appearances suggesting a constant new formation of hemolymph 222 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. nodes and their development into ordinary lymphatic glands, with a possible cyclical function of hemolysis. No evidences of blood-formation have been found under normal conditions. The thickness of the capsule and trabeculae, the size of the blood-sinuses, the amount of lymphoid tissue, and the amount of reticulum traversing the sinuses are not constant factors, hence the greatly varied microscopical pictures presented by the organs. A very striking and constant appearance is the great number of connective-tissue mast-cells found in the reticulum and stroma. Mononuclear eosinophils, are also numerous. In the smaller nodes containing but little lymphoid tissue the majority of the reticular cells may exhibit the mast-cell granulation; in the nodes where active hemolysis is present their number is not so great. Experimental Observations. The animals operated upon were allowed to fast for several days., preceding the operation, as experience showed in the case of one animal not so treated that the distended stomach greatly increased the difficulty of the operation. In the fasting animals the operation was performed with ease and without impor- tant loss of blood. Chloroform was used as an anesthetic, the animals not tak- ing ether well. The abdomen was shaved before operation and asepsis carried out as far as possible. The animals were fastened to the operating table in the right-side position, this being found more favorable for the operation than the dorsal position. The incision was made in the left hypochondrium in the splenic region, about a hand-breadth below and parallel to the edge of the ribs. The operations were kindly performed for me by Drs. Peterson, Morley, Griffin, and McNamara, and I take this opportunity of acknowledging my indebtedness for the favor. All the sheep operated upon suffered severely from shock dur- ing the first twenty-four hours following the operation, fifty per cent of the cases dying from this cause within this time. The survivors reacted quickly and by the third day were apparently as well as before the operation. Because of the great mortality it was decided to make use of goats instead, the two operated upon for this series of experiments showing no symptoms of shock and recovering quickly. Inasmuch as the hemolymph nodes of the goat show practically the same distribution and structure as those of the sheep, it was decided to use the former animal in the remaining experiments. With the exception of local abdominal hernia developing at the seat of operation and the rapid development of a goitre in one of the goats, no bad effects were ob- served to result from the operation. All the animals kept over a month after the operation showed a great increase of fat. The animals were killed by giving an anesthetic (chloroform), the abdomen being opened as soon as the animal was under the anesthetic. An opportunity was thus afforded of examining the glands while the blood was still circulating. On cutting through the diaphragm the thoracic glands were also examined before the heart had ceased to beat. ALDRED SCOTT WARTHIN. 223 The dissection was made while the tissues -wcmv still warm. All of the hemo- lymph nodes and lymphatic glands discoverable wc-re removed and fixed at once, in either mercuric chloride or Flemming's solution; the glands from the different regions were kept separate. In addition, portions of the other organs were taken for microscopical examination, smears of the glands and bone-mar- row were made, and portions of the latter fixed for special examination. The material fixed in mercuric cliloride was imbedded in paraflin, that fixed in Flemming's in celloidin. Sections were made from each block; hematoxylin and eosin, kresylechtviolett, the triacid stain, the iron reaction, and borax carmine were the staining methods commonly usetl. ExPERUiENT I, Sheep I. — TAvo-year-old merino grade wether. Splen- ectomy. Autopsy, twenty-four hours after operation. Autopsy Notes. — Wound negati\'e. Peritoneum in neighborhood of incis- ion slightly injected, roughened and cloudy. No hemorrhage from splenic vessels. Passive congestion of all organs, particularly marked in the case of liver and kidneys. The hemolymph nodes are unusually large and prominent, all dark red in color. Lymphatic glands, particularly the mesenteric and retroperitoneal, are enlarged, hyperemic, and softened. The glands at the brim of the pelvis, along the common iliacs, are particularly enlarged and of soft consistency, but are of lighter color. A few of the retroperitoneal glands appear hemorrhagic. Microscopical. — The blood-sinuses in the hemolymph glands are greatly distended, the cords of lymphoid tissue being apparently smaller and more widely separated. Very few lymphoid follicles are found in them, but small groups of more closely-packed lymphoid cells having a lighter central portion suggest their location. There is an apparent loss of leucocytes, the lymphoid tissue appearing poorer in cells. The mast-cells are greatly diminished in num- ber; those present appear larger than normal, more round as if swollen, their granules staining both reddish and blue with kresyl-violet. In some cells there is no distinct granulation, the cell protoplasm staining diffusely reddish. There is no evidence of increased hemolysis. The lymphatic glands show extreme congestion, the veins of the glands being greatly dilated. Hemor- rhages are present in many, the extravasation as a rule taking place into the stroma of the lymphoid tissue and not into the lymph-sinuses. The lymph- sinuses and the lymph-vessels are also greatly distended. Blood-cells in in- creased amount are found in these. The central portion of the follicles is in the majority of cases lighter than normal. The mast-cells in the lymphatic glands show changes similar to those in the hemolymph nodes. Experiment II, Sheep II. — Three-year-old registered Shropshire ram. Splenectomy. Autopsy three days after operation. Autopsy Notes. — ^Wound negative. Slight injection of peritoneum near incision. The stomach adherent to abdominal wall in neighborhood of opera- 224 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. tion. Adhesions easily separated. The liver is moderately hyperemic. The hemolymph glands are very prominent, but otherwise appear unchanged. The lymphatic glands are enlarged, softer, pinker than normal, the medullary portion of many showing a brownish tint. Microscopical. — The sinuses are less distended with blood than in the case of Sheep I. The most striking change is the light appearance of the lymphoid follicles, due to a diminished number of cells and to a vacuolization of those present. The follicles are prominent because of their lighter staining and not because of the reverse, as is the normal case. They are smaller, there is an actual loss of cells, the central portion or even the entire follicle being formed of rather large branching cells with swollen vacuolated protoplasm. Some of the cells contain very large vacuoles. Either hydropic or fatty degeneration, or both, is present. This could not be definitely decided, as unfortunately the osmic acid in the Flemming's fixed specimens did not penetrate to the foUicles. The large size and irregular shape of the vacuoles in many cases suggest hy- dropic change or edema rather than fatty degeneration. The same change is present in the follicles of the lymphatic glands, but to a much less degree. The number of cells present in the follicles in all of the glands is evidently greatly diminished. In the blood-sinuses of many of the hemolymph nodes' there is an increased number of pigment-containing phagocytes and evidences of increased hemolysis. The lymph-sinuses in the medullary portion of many of the lymphatic glands likewise are filled with pigment-containing phagocytes. The pigment in both cases for the greater part gives the iron reaction. The phagocjrtes which are stained diffusely brown with the pigment give an intense iron reaction, while those containing pigment granules show variation in this respect, some of the pigment granules giving the iron reaction, others not. These changes are more marked in the retroperitoneal glands than elsewhere, the thoracic glands coming next in point of change. The changes in the mast- cells of the hemolymph nodes are very striking. In some of the glands these cells appear as large, round, swollen cells containing coarse granules staining violet or blue with kresyl-violet. In other cells the granules are smaller and less distinct, while in other cases no distinct granulation can be made out, the cell protoplasm staining diffusely reddish or violet. Vacuoles are present in many of the mast-cells. In one hemolymph node containing many cells showing mast-cell granulation, phagocytes having this granulation were seen, containing red cells in various stages of disintegration. There is also an increase of eosino- philes in many of the glands. In the anterior mediastinum several large, deeply-congested lymphatic glands, as in Sheep I, were found. No nucleated red cells found in the blood of any of the glands or organs. Experiment III, Sheep III. — Two-year-old merino ewe. Splenectomy. Autopsy five days after operation. Autopsy Notes. — ^Wound negative save for a small stitch abscess. The ALDRED SCOTT WARTHIN. 225 stomach adherent to abdominal wall and diaphragm in the neighborhood of the operation; adhesions easily separated. Stump of splenic vessels negative, no lymphoid tissue present in it. The hemolymph nodes not so prominent as in the normal animal; they are less red and more brown in color, but on the whole are increased in size. The lymphatic glands are enlarged, the cortex pale, the medullary portion pink or brown. Bone-marrow negative. Mia-oscopical. — The same changes are present as in Sheep II, but are much more marked. The light appearance and the diminution in size of the follicles of the hemolymph nodes due to vacuolization are very marked, less so in the follicles of the lymphatic glands. There is an increased number of the hemolymph nodes showing hemolysis, the number of phagocytes in the blood-sinuses is everywhere increased, and there is a well-marked leucocytosis in the blood sinuses and vessels. The lymph-sinuses of the lymphatic glands also contain a greater number of pigmented phagocjrtes; as in Case II, the pigment in part gi^■es an iron reaction, in part does not. The changes in the mast-ceUs are more marked than in Sheep II, and the number of eosinophiles greater. No evidences of blood regeneration found; no nucleated red blood- cells found in any of the glands. Experiment IV, Goat I. — Black female goat, two years old. Splen- ectomy. Autopsy two weeks after operation. (7\jiimal developed large, soft goitre during this period.) Autopsy Notes. — Wound negative. Stomach adherent to diaphragm and abdominal wall in neighborhood of incision. No evidence of general peri- tonitis. No lymphoid tissue found in stump of splenic vessels. The liver somewhat browner than normal, not hyperemic. The hemolymph nodes are much diminished in number, particularly in the thoracic region, where they are apparently replaced by ordinary lymphatic glands of brownish color. The larger hemoljonph nodes are paler than normal or are more heavily pig- mented. Only a few of the larger ones containing large blood-sinuses are found. On section, all the hemolymph nodes show an unusual amount of lymphoid tissue. Numerous minute hemolymph nodes are found in unusual locations; they are apparently increased in number in the mesentery, pelvic fat, and cervical region. All lymphatic glands (retroperitoneal, mesenteric, thoracic, mediastinal, cervical, axillary, pelvic, inguinal, etc.) are enlarged, softer than normal, many presenting small elevations on their surface. On section the glands show great hyperplasia of the cortex, the medullary portion being softer, more translucent, and of brown color. A number contain red streaks or spots corresponding to blood sinuses or vessels. In the thoracic preverte- bral fat where hemolymph nodes are usually numerous, few of the latter were found, and large elongated masses of lymphoid tissue were present which on section showed a broad white cortex, and a brown translucent medulla. The superior mesenteric glands and the glands at the brim of the pelvis pre- 226 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. sented the greatest degree of enlargement and pigmentation, the peripheral glands much less. The bone-marrow showed no gross changes. Microscopical— The hemolymph nodes present hyperplasia of lymphoid tissue, in some cases nearly obliterating the sinuses. The follicles are for the greater part enlarged and form centers of lymphoid proliferation, masses of closely-packed lymphocjd;es surrounding the follicle and extending into the blood-sinuses, in some cases giving rise to an appearance suggesting splenic pulp. Numerous mitotic figures are present, especially at the periphery of the follicles. The mast-cells are diminished in number; numerous swollen ones are seen. Eosinophiles, particularly mononuclear ones, are found in abundance throughout the lymphoid tissue about the follicles, less commonly in the latter. There is also a proliferation of the reticulum, thickening of the trabeculse, and an increase in the number of the large mononuclear cells. Pigment-containing phagocytes are increased in number, and there is a great increase in the number of leucocytes present in the blood-sinuses. The lymphatic glands present a great hyperplasia of lymphoid tissue, chiefly in the cortical portion, which is enlarged out of proportion to the medulla. The ele- vations seen on the surface of some of the glands are due to hyperplastic folli- cles pushing out the capsule. All of the follicles are enlarged and apparently increased in number. Few show central vacuolization. Numerous mitotic figures are present, particularly about the periphery of the follicles. Through- out the lymphoid tissue between the follicles there are great numbers of eosino- philic cells. Localized collections containing hundreds of these are found scattered throughout the glands. These collections of eosinophilic cells are most numerous in the nodes showing the greatest degree of hemolysis; they lie usually in the zone between the cortical and medullary portions. The lymph-sinuses of the medulla are distended and contain many pigment phago- cj^es. Transitional forms between hemolymph nodes and the lymphatic glands are numerous. In some of these the proliferation of the lymphoid tissue into the blood-sinuses produces an appearance closely resembling splenic pulp. The connective tissue of the liver and lungs contains an increased num- ber of wandering cells with localized lymphoid collections of small size. Numer- ous pigmented endothelial cells are found in the liver capillaries. There is a general leucocytosis. No nucleated red cells were found in any of the glands. Sections of thyroid showed dilated vessels and increased formation of colloid. Experiment V, Goat II.— Black male goat, three years old. Splen- ectomy. Autopsy one month after operation. Autopsy Notes.— Ammal very fat. Thyroid somewhat enlarged. Lapa- rotomy scar negative. Slight adhesion of stomach to abdominal wall. No evidences of general peritonitis. Liver somewhat browner than normal. No lymphoid tissue in stump of splenic vessels. Thymus greatly enlarged. The hemolymph nodes are apparently decreased in number and size in the ALDRED SCOTT WARTHIN. 227 regions where they arc usually found in greatest abundance. Small ones are found in unusual locations, as the axillary and inguinal refrions, and are very numerous in the mesentery, about the jiancreas, and in the stomach and liver Ugaments. The largest ones are found in the pelvic fat. The small ones are dark red, the larger ones lighter colored or brown. All of the lymphatic glands of the body are greatly enlarged, particularly those of the mesenteric, posterior thoracic, and iliac regions. All show on section a great hyperplasia of the cortex with pigmented or reddish medulla. Numerous small glands are found resembling splenic tissue in whole or in part. Bone-marrow negative. Microscopical. — The changes in the lymphoid structures are of the same nature as those in Goat I, but are much more extensive. The hemolymph nodes from the usual locations show hyperplasia of lymphoid tissue at the expense of the blood-sinuses. In many cases the latter are almost completely obliterated, or so small that the tissue comes to xesemble spleen pulp. The smaller nodes found throughout the mesentery and retroperitoneal fat resemble normal ones. The lymphatic glands present great hyperplasia of follicles and lymphoid tissue at the expense of the medulla. The glands resembling spleen tissue have the structure of lymph glands with medullary sinuses filled with blood; they are probably to be regarded as modified hemolymph nodes. In many of the glands red blood-cells are found everywhere throughout the lymphoid tissue. No nucleated ones observed. The mast-ceUs are greatly decreased in number. Eosinophiles, mitotic figures, and pigmented phago- cj'tes are numerous, as in Goat I. The connective tissue of the liver and lung presents no greater infiltration than in that case. The number of pigmented endothelial cells in the liver capillaries is slightly greater; the cells of the convoluted tubules of the kidney also contain a small amount of hemosiderin. The sections of the thymus show great hyperplasia of lymphoid tissue, with an apparent new formation of lymph follicles in the surrounding fat. In the adipose tissue of the retroperitoneal and pelvic regions, small lobules of fat- cells are found whose capillaries are greatly congested and dilated. In some of these there is a lymphoid infiltration along the capillaries (early stage of hemolymph node?). ExPERiMEXT VI, Sheep IV. — Two-year-old merino grade ewe. Splenec- tomy. Autopsy two months after operation. Autopsy Notes. — Animal very fat. Laparotomy scar negative. Hernia of abdominal wall at scar. Stomach adherent to peritoneum at point of opera- tion. Small purulent focus in wall near scar (stitch abscess). No evidences of general peritonitis. Thymus enlarged. Liver small and browner than normal. No hemolymph nodes found in thoracic or anterior mediastinal regions ; numer- ous small ones, apparently newly formed, scattered throughout the mesentery, subperitoneal fat, pelvic fat, around the rectum, and in the cervical, axillary, inguinal, and subcutaneous fat. With each one of the enlarged peripheral 228 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. lymph glands there is a small hemolymph node Ijdng in its hilum or close against its capsule. Those of the inguinal and axillary glands are as large as small peas. Along the vessels leading to one of the inguinal glands a number of hemolymph nodes are present. All of the lymphatic glands are greatly enlarged, particularly in the anterior and posterior mediastinal, mesenteric, and iliac regions. Along the brim of the pelvis and extending upward along the abdominal aorta there is a double row of greatly enlarged glands, several of the size of a walnut. The superior mesenteric glands form a firm, solid cord about fifteen centimeters long and two centimeters in diameter. Similar elongated cords of lymphoid tissue are found in the anterior and posterior mediastinal regions. Many of the enlarged lymphatic nodes present the raspberry-like surface. On section the enlarged lymph glands present a greatly hyperplastic cortex, and a smaller, deeply-pigmented medullary portion. Many of the nodes appear to possess no medullary portion, being composed entirely of white, soft, cortical substance. The raspberry-like elevations on the surface of some of the glands are appar- ently due to hyperplastic follicles pushing out the capsule. Transition-forms between hemolymph nodes and ordinary lymphatic glands resembling spleen tissue are common. In the subperitoneal fat numerous minute red points are present. Small reddened areas are seen throughout the fatty marrow. Microscopical. — The sections of the hemolymph nodes and lymphatic glands present the same lymphoid hyperplasia as in the last case, but to a more marked degree. The cortical portion is enlarged at the expense of the medulla. Many glands appear to be composed of cortical portion only, the follicles being greatly enlarged and apparently increased in number. In other glands the hyperplasia of lymphoid tissue is more uniform, all trace of the follicles being lost, and the meduUarjr portion showing an extensive formation of lymphoid cords between the sinuses. In other nodes the lymphoid hyperplasia is stiU more marked, the distinction between cortex and medulla being entirely lost, and no follicles present, the sections presenting a uniform lymphoid hyperplasia resembling lymphosarcoma. In some glands the hyperplastic follicles are ten to twenty times the average size. Proliferation of the reticulum and thickening of the trabeculse are also present in many glands. The decrease of mast-cells, the numerous mitotic figures, eosinophiles, etc., are present as in the last case. There is, undoubtedly, a new formation of hemolymph nodes out of adipose tissue. All stages of this development may be seen. The process begins with the angiectatic dilatation of the capillaries of a fat lobule, the fat cells of which become enlarged and lighter in color. At the same time the lobule becomes fairly well set off from the surrounding tissue by a thickening of its capsule. The next step is an infiltration of lymphocytes along the walls of the distended capillaries coincident with an absorption of some of the fat, the conversion of the fat-cells into reticular cells, and proliferation of the endothelium into the dilated capillaries, dividing them up into blood-sinuses. Continued lymphoid ALDRED SCOTT WAUTHIN. 229 formation, developnunit of sinuses, and absorption of fat lead to the fully- developed hemolymph nodes or ordinary lymphatic glands. If the blood-sinuses persist, the structure of a hemolymph node is presented; if the formation of the lymphoid tissue is so great as to reduce the sinuses to capillaries, the node as- sumes the structiu-e of a lymphatic gland. Transition-forms of all kinds are seen. In some cases the follicles are formed first, in others they are developed later in the lymphoid tissue. The evidences of greatly increased hemolysis in the hemolymph nodes and lymphatic glands are present. The dark color of the medullary portion of the latter is due to the great number of pigment- containing phagocjiies present. Part of the pigment gives the iron reaction, part does not. Eosinophilic cells are most numerous in the glands showing the greatest amount of pigment. The sections of liver and lung show localized formation of lymphoid nodides in the connective tissue. There is a great increase of pigmented endothelial cells and also in the amount of the pigment contained in them. Apparently all the pigment in the liver gives the iron reaction. The cells of the convoluted tubules also contain small amounts of hemosiderin. In the liver the pigment is for the greater part in the peri- pheral zone of the lobule as in pernicious anemia, but in many lobules it is found in the central portion. In the long bones there is a beginning formation of red marrow. No nucleated red cells were observed either in the general circulation or in the hemolymph nodes or lymphatic glands. A general leuco- cytosis is present. Experiment VII, Sheep V. — Two-year-old merino grade ewe. Splenec- tomy. Autopsy two months after operation. A week before the animal was kiUed, two doses of three grams each of toluylendiamin were injected into the large superficial veins of the hind legs. The animal gradually became very weak, on the third day could not stand. The red blood-cell count fell in one week from eleven millions to less than seven millions. The hind-quarters be- coming paralyzed, the animal was killed. Autopsy Notes. — ^Animal very fat. Skin, conjunctivae, and fat of a decided brown tinge. Laparotomy scar negative. No evidences of general perito- nitis. Necrosis of thigh muscles and venous thrombosis at points of injection. The hemolymph nodes show similar distribution as in Experiment VI, and a similar new-formation in same regions, — one hemolymph node with each periph- eral lymphatic gland. Numerous small hemolymph nodes are present behind the pancreas and along the splenic vessels. The color of all is much browner than in any of the other cases. The lymph glands show the same extensive hyperplasia, but on the whole are larger, softer, and much more heavily pig- mented. One large gland at the brim of the pelvis is surrounded by a blood clot. Other small hemorrhagic glands are found in the retroperitoneal fat. The liver is not enlarged — soft and very brown. The kidneys are browner than normal. 230 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. Microscopical. — The microscopical appearances are similar to those in the preceding experiment, except for a great increase in the pigment and pigmented phagocytes found in the lymph glands and hemolymph nodes, liver, and kidney. A number of hemorrhagic glands are found. There is a similar new-formation of hemolymph nodes in the adipose tissues. The number of eosinophils is greatest in the glands showing most extensive hemolysis. The destruction of the red cells seems to be restricted wholly to the sinuses of the hemolymph nodes and lymphatic glands, no phagocytes containing red cells being found in the liver or elsewhere. The bone-marrow is hyperemic with probable begin- ning lymphoid change. No normoblasts were observed in the general circula- tion or in the lymphoid structures. Experiment VIII, Sheep VI. — Three-year-old merino grade ewe. Splen- ectomy. Autopsy five months after operation. Animal was very lean when first brought to the laboratory, but became very fat after the operation. Autopsy Notes. — Animal very fat. Small stitch abscess in laparotomy scar. Stomach adherent to the abdominal wall in the neighborhood of the operation. No evidences of general peritonitis. The liver is small, brown, and hard. The kidneys are very brown. Fatty marrow reddened. The hemolymph nodes and the lymphatic glands show a greater degree of hyperplasia than in any of the preceding experiments. Glands as large as walnuts are found along the brim of the pelvis and the abdominal aorta. The superior mesenteric gland is greatly enlarged, and there are long masses of lymphoid tissue in both ante- rior and posterior mediastinum. The cortical portion of the enlarged glands is white, the medullary portion deep brown. Many of the largest glands have the raspberry-like surface. On section the cortical portion is greatly hyper- plastic, white, soft, and opaque; the medullary p'ortions, brown and more trans- lucent. The enlarged follicles in many instances push the capsule outward, forming the little elevations on the surface of the gland. In some glands no medullary portion is present, the cut surface presenting a homogeneous appear- ance. No hemolymph nodes are present in the thoracic region. There is a new-formation of these along the splenic vess'els, about the pancreas, in the mes- enteric, subperitoneal, and pelvic fat; and, in association with the peripheral lymph glands, one hemolymph node to each lymphatic gland. Microscopical. — There is extensive hyperplasia of the lymphoid structures, new-formation of hemolymph nodes in the adipose tissues, and increased pig- mentation. The lymphoid follicles are greatly increased in size and appar- ently in number. The glands showing a homogeneous surface present a uniform lymphoid hyperplasia with loss of structure, suggesting lymphoma or lymphosarcoma. The same transition-forms are found as in the preceding cases. The medullary portion of the majority of the glands shows thick lymphoid cords along the sinuses. The number of eosinophiles is very striking, the con- nective-tissue trabeculse of many of the glands containing them in great num- ALDRED SCOTT WARTHIN. 231 bers. The localized collections of large numlxM-s of oosiiiophilcs are also jires- ent as in the other cases. These apparently bear sonic relation to the degree of hemolysis. In many glands proliferation of the stroma and thickening of the trabeculse are verj' marked. Lymphoid nodules are present in the con- nective tissue of the liver and lungs. The endoth(>lial ci'Us of the liver capil- laries contain a much gi'eater amount of hemosiderin than in the prec^'ding experiments. Moderate hemosiderosis of the kidneys is present. The bone- marrow is hyperemic and there is probabh' a ne\\'-f(3rmation of red marrow. No nucleated red cells were found in the general circulation or in the lymphoid nodes. There is a geiieral increase of leucocytes. EXAXU NATION OF BlOOD. The results of the blood study of the eases may be condensed as follows : The normal count of the red cells in the sheep was on the average twelve million five hundred thousand, the a^•erage leucoc}'te count about seven thousand. In the goat the red cells averaged about sixteen million, the leucocytes eight thousand. After splenectomy there was always a diminution in the number of red cells out of proportion to the amount of blood lost at the operation. The average decrease was two to three million. By the fifth month the number had increased, but was still lower than normal. No nucleated red cells were ever seen, and the red cells presented no morphological changes. The leu- cocytes in the peripheral blood were increased immediately after the opera- tion, this increase during the first ten days being confined to the polymor- phonuclear forms. After the tenth day there was an increase in the number of lymphocytes, the proportion gradually returning to the normal. At the ninth day the lymphocytes formed fifty per cent of the leucocytes. During the first few days following the operation the eosinophiles either disappeared entirely or were reduced to very small numbers, but after this time they showed a gradual increase. The number in the peripheral blood, however, was entirely out of proportion to the great numbers of eosinophiles observed in the lymph nodes. One of the most striking changes in the blood was the appearance of numerous atypical and degenerating leucocjrtes. These were most numerous during the first two weeks after splenectomy, but never entirely disappeared. Summary. The changes following splenectomy, as shown by the above eight cases, may be briefly summarized as follows : First Week. — Immediately following the operation there is an intense con- gestion of all the lymphoid structures (possibly due to shock?), a temporary decrease in the leucocytes in the vessels and sinuses of the lymph nodes, and changes in the mast-cells. By the third and fifth days the congestion has greatly lessened, and the chief change is a vacuolization (hydropic or fatty degeneration) 232 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. of the lymph follicles of both hemolymph and ordinary lymphatic nodes, and a withdrawal of leucocytes from these into the blood sinuses and vessels where there is a moderate leucocytosis. At the same time there are evidences of in- creased hemolysis, as shown by an increased number of pigment-containing phagocytes, increased number of eosinophiles, and a beginning proliferation of the lymphoid tissue of the hemolymph nodes. End of Second Week. — Active proliferation of all lymphoid tissues had be- gun, as shown by the numerous mitoses present throughout hemolymph and lymphatic nodes. In the former the proliferation into the blood-sinuses gave them the appearance of being transformed into lymphatic nodes. Evidence of greatly increased hemolysis in hemolymph and lymphatic nodes was shown by large numbers of pigment-containing phagocytes; increase of eosinophiles and moderate leucocj^osis were present. One Month. — At the end of one month the proliferation of lymphoid tissue, hemolysis, eosinophilia, and transformation of hemolymph nodes into ordinary lymph glands were all more marked. In addition there was a new-formation of hemolymph nodes in adipose tissue. Two Months. — At the end of the second month the findings were : advanced hyperplasia of all lymphoid structures, change of hemolymph nodes into lymphatic glands, new-formation of hemolymph nodes in the adipose tissues, in- creased hemolysis in all lymphoid structures, and marked formation of eosino- philes. Beginning pigmentation of the liver and kidneys, leucocytosis, and beginning new-formation of lymphoid marrow were also present. Five Months. — At the end of the fifth month great hyperplasia and new- formation of lymph nodes, new-formation of hemolymph nodes in adipose tissue, marked hemolysis, increase of eosinophiles in the lymphoid tissues, pigmentation of the liver, and slight lymphoid change in the fatty marrow were the most important changes. The leucocytosis was less marked than at two months. It is evident from the above that splenectomy in the sheep is followed during the first five months by a compensatory hyperplasia of the pre-existing lymphatic tissues, transformation of hemolymph nodes into ordinary lymphatic glands, and a new-formation of hemolymph nodes in the adipose tissues. No evidence of regeneration or new-formation of splenic tissue was found. The hyperplastic hemolymph nodes and lymphatic glands differ essentially from the spleen in the arterial relations of the follicles. In other respects the structure of the hyperplastic hemolymph nodes closely resembles splenic tis- sue, both to the naked eye and microscopically. There can be no doubt that the new-formations observed in the dog after splenectomy, and regarded as " newly- formed splenic tissue" by Tizzoni, "hemorrhagic lymph glands" by Wino- gradow, and "hemorrhagic telangiectatic lymphoma" by Hosier, were either hyperplastic or newly-formed hemolymph nodes. Tizzoni's description of ALDRED SCOTT WARTHIN. 283 their formation in adipose tissue is confirmed in e\-ery tletail liy tlic findinj;- in the above cases. His conclusions alone must be rep;arded as erroneous. The establishment of the fact that hemolyniijh nodc-s may be formed di- rectly from lobules of adipose tissue is of great importance in the light thrown upon manj^ questions pertaining to the lymphoid tissues, particularly with regard to the relationship i^xisting between the hcnnolymph nodes and the lymphatic glands. The new-formation of these structures repeats their em- bryonal development, and the earliest stages of development of both forms may run parallel. The hemolymph node must be regarded as a more primitive embryonal type than the lymphatic gland. It may develop into the latter through a further proliferation of reticulum and lymphoid cells. The first stage in the development of a lymph node from adipose tissue is a dilatation to the small capillaries lying between the fat cells, and an infiltration of lymphocytes along their walls. If now the capillaries become angiectatic and converted into blood-sinuses through the formation of a reticulum arising from endothelial proliferation, the hemolymph node is formed; if, on the other hand, the capillaries remain small or undeveloped, and the lymphoid tissue is increased at their expense while the lymph-vessels become enlarged and changed into sinuses, the structure is that of a lymphatic gland. The essential difference between the two is the degree of development of either blood capillary or lymph-vessel into blood or lymph sinus. Intermediate stages of development produce the great variety of transition-forms found. The development of the hemolymph nodes in fat bears also a very striking resemblance to the develop- ment of red marrow out of fatty. In connection with the question of lymphoid conversion of fat, it should be noted here that the work of Bayer on the regen- eration of lymph glands in adipose tissue is also confirmed by the above results. As to the question of the formation of red blood-cells in the lymph glands after splenectomy as held by Gibson, Laudenbach, and others, no evidence of such a formation was found. In so far as the question of compensation for splenic function is concerned, the findings would indicate that hemolysis and leucocyte formation are the two functions which are taken up to an increased degree by the hemolymph nodes and lymph glands after splenectomy. That the splenic function is not perfectly compensated is shown by the disturbed equilibrium of the blood in the excess of hemolysis over blood-formation. This might be explained by the hypothesis that there is some hemolytic agent formed in the body which is normally taken care of by the spleen, or that the spleen in addition to a hemolytic function has an influence also in the new-formation of hemoglobin. Bottazzi's theory of the " hemocatatonistique " function of the spleen may be referred to in this connection. The beginning lymphoid change in the fatty bone-marrow in the second and fifth months after splenec- tomy is to be regarded as compensatory only for the increased destruc- 234 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. tion of red blood-cells, not for any abrogated splenic function of red cell formation. Further, the vacuolization of the lymph follicles in the first few days after splenectomy might be taken as evidence of an intoxication by some substance drawing the leucocjrtes from the follicles so rapidly that degeneration of the cells of the germinal area results. The large number of degenerating leucocytes found in the circulating blood at this time may be also taken as evidence of this. The great increase in the number of eosinophiles in the pigmented glands is also of interest as indicating a possible relationship between these cells and the destruction of hemoglobin. The pigmentation of the medullary portion of the lymphatic glands may be explained by the assumption that the pig- ment is carried from the neighboring hemolymph node to the lymphatic glands through the lymphatics. The new formation of a hemolymph node in the hilum or along the vessels of the peripheral glands, whose medullary portion showed pigmentation, might be taken in support of this view, — ^the hemolymph nodes acting as hemoljrtic organs passing the products of blood destruction on to the lymphatic glands for further elaboration. It is also possible that the pigmentation of the lymphatic glands is the result of a constant diapedesis of red cells into the lymph-vessels and their destruction in these glands. Conclusions. 1. After total splenectomy in the sheep there is no evidence of regeneration of the primitive spleen or of the new-formation of splenic tissue. 2. The structural changes following splenectomy are: hyperplasia of exist- ing lymphoid tissues, transformation of hemolymph nodes into ordinary lymphatic glands, and a new-formation of hemolymph nodes out of lobules of fat tissue, and a later proliferation of the red marrow. 3. There is no e^ddence of the formation of red blood-ceUs in the lymph nodes after splenectomy. 4. The function of hemolysis is taken up first by the hemolymph node, later by the ordinary lymphatic glands. 5. The hemolytic function of the hemolymph nodes and hyperplastic lymph glands exceeds that of the primitive spleen, causing an excessive destruction of red cells. The resulting anemia is later compensated for by an increased activity on the part of the bone-marrow. It would appear, therefore, that the removal of the spleen leads to an increased production or retention of some hemolytic agent usually disposed of by the spleen. The effect of this hemolytic agent is either to stimulate the phagocytes in the hemolymph nodes to in- creased activity, or to change the red cells so that they are more easily destroyed by these phagocytes. 6. The presence of great numbers of eosinophiles in the glands showing ALDRED t^OOTT WAKTHIN. 235 great destruction of red cells seems to point to some relationship betwe(>n these cells and hemolysis. 7. The appearances described by Tizzoni, Winogradow, Eternod, Griffini, and Hosier, as occurring after splenectomy are confirmcvl by this work, but given a different interpretation. Bayer's work \\\wn the regeneration of lymph glands is also confirmed. Jly thanks are due to Dr. C. S. Bond, of Richmond, Indiana, for the accompanying photographs. Note 1. — The abo\'e woi'k had been entirely planned and largely carried out when the preliminary note by ^lorandi and 8isto (Contribution k I'etude des glandes h^molymphatiques chez I'homme et chez quelques mammif feres, Archives ital. de Biologie, 1901) appeared. These observers studied the effects of splenectomy upon the hemolymph glands of dogs. They noted the in- creased hemolysis taking place in these structures, and also concluded that the splenic nodules of Tizzoni and others were hemolymph nodes. Note 2. — Since the publication of the above I have discovered the exist- ence of the paper by Swale Vincent (Proc. of the Phys. Society, December 9, 1899). This observer found in one dog examined eighteen months after splenectomy, an increase of hemolymphatic structures. In five other dogs examined after splenectomy no increase in either lymphatic or hemolym- phatic glands was found. The fact that Vincent obtained positive results in only one case out of six, while the results obtained by Morandi and Sisto and myself are so uniform, accords with the varying results obtained by the earlier investigators, with reference to the changes produced by splenectomy (See above). It is possible that the different results obtained may be explained by a varying power of lymphoid regeneration or hyperplasia in individual animals of the same species. BIBLIOGRAPHY. Albert. Cited by Riegner. Albreeht. Beitrag. z. path. Anat., 1896, Bd. XX, p. 513. Bayer. Zeitschr. f. Heilkunde, 1885, Bd. VI. Beclard. Cited by Tizzoni. Billroth. Cited by Riegner. Bolton. Annals of Surgery, 1900, Vol. XXXI, p. 734. Bottazzi. Arch. ital. de bioL, XXIV, p. 462. Ceci. Wiener Idin. Wochenschr., 1889, p. 424. Ceresole. Beitrftge z. path. Anat., 1895, Bd. XVII. Czemy. Cited by Riegner. Drummond. Jonr. of Anat. and Phys., 1900. Eternod. Rev. m^d. de la Suis.se rom., 1888. Foa. Centralbl. f. klin. Med., 1884, Arch. ital. de Biol., IV, p. 299. Gerlach and Eberhardt. Cited by Tizzoni. Gibson. Jour, of Anat. and Phys., 1896, Vol. XX. 236 THE HEMOLYMPH GLANDS OF THE SHEEP AND GOAT. Griffini. Arch. ital. de biol., 1883, Vol. III. Arch, per le Scienze med., Vol. VI. Mem. dei prof. Griffini et Tizzoni, 1883. Hodenpyl. N. Y. Med. Record, 1898, Vol. LIV. Kocher. Cited by Warbasse. Laudenbach. Centralbl. f. Phys., 1895, Bd. IX, p. 1-4; Arch. f. path. Anat., 1895. Legros. Cited by Tizzoni. Lennander. Cited by Warbasse. Morandi et Sisto. Arch. ital. de biol., 1901, Vol. XXXV. Mosler. Deutsch. med. Wochenschr., 1884, p. 337. Peyrani. Cited by Ceresole. Philipeaux. Cited by Ceresole. Riegner. Berl. klin. Wochenschr., 1893, p. 177. Schultz. Deutsche Klinik, 1856. Cited by Zesas. Tizzoni. Arch. ital. de biol., Vols. I, III, and IV. Atti della R. Accad. dei Lincei, Vol. XIII. Accad. reale dei Lincei, 1882. Arch, per le Scien. med.. Vol. VIII. Tizzoni et Fileti. Atti deUa R. Accad. dei Lincei, Vol. X. Trendelenburg. Cited by Riegner. Verga. Cited by Tizzoni. Vincent and Harrison. Jour, of Anat. and Phys., 1897. Vulpius. Beitrage z. klin. Chir., 1894, Bd. XL Warbasse. Annals of Surgery, 1894, Vol. XX. Winogradow. Centralbl. f. d. med. Wissensch., 1882, No. 50, p. 900. Zambeccari. Cited by Ceresole. Zesas. Arch. f. klin. Chir., 1883, Bd. XXVIII, p. 157. Plate I. Fiar". I. Fig. 3. Fig. 5 Fig. 4. Warthir Fig. 2. Fip-. 6. Fig- B. Plate Fig. 9 Piff. iO. Warthin. Fig. I ;g. !2. Description of Figures. Plate I. Fig. 1. — Normal Type of Hemolymph Gland of Sheep. Dark areas, blood-sinuses; light areas, lymphoid tissue. X 35. Fig. 2. — Normal Type of Hemolymph Node of Sheep. Dark areas, blood-sinuses; light areas, lymphoid tissue. X 35. Fig. 3. — Hemolymph Node from Sheep, Three Days After Splenectomy, Showing Dila- tion of the Sinuses and Vacuolization of the Germ Follicles. Light, areas, blood-sinuses; dark areas, lymphoid cells. X 125. Fig. 4. — ^Hyperplastic Hemoljnmph Node of Sheep After Splenectomy, Showing Dilation of Blood-Sinuses. X 125. Fig. 5. — Hemolymph Node of Sheep, Resembling Spleen-Pulp. After Splenectomy; X50. Fig. 6. — Hemolymph Node of Sheep After Splenectomy, Showing Obliteration of Sin- uses, Resembling Spleen-Pulp. Dark areas, blood-spaces; light areas, lymphoid cells. X50. Fig. 7. — Beginning of Formation of Hemolymph Node in Retroperitoneal Fat of Sheep, Two Months After Splenectomy; Small Lobule of Fat Tissue with Angiectatio Capil- laries. Dark areas, blood. X 125. THE AERATION OF MILK. CHARLES E. MARSHALL, Ph.D., Professor of Bacteriology, Michigan Agricultural College, Lansing, Mich. {From the Hygienic Laboratory of the University of Michigan.) Although practical aeration of milk may be traced back to an indefinite past, where it is simply mentioned and recognized, there has never been any plausible explanation or demonstration of the results claimed for it. More- over, its application has always been conspicuously limited until within very recent years, and so far as its value is concerned at the present time, nothing has been satisfactorily established regarding it, other than the belief that it eliminates animal odors and removes, to a certain extent, tainted conditions of milk resulting either from physiological metabolism due to ingestion of aromatic food substances or produced by the growth of saprophytic bacteria in milk after the milk has left the udder of the cow. Even these animal odors and taints have been more or less undefinable and immeasurable; only daily practices in dairy operations have indicated and established beyond a doubt that both of these are greatly lessened by proper aeration. Many other beliefs have been associated with or attributed to the process of aeration, such as the production of more cheese and better cheese, the production of more butter and better butter, the more rapid rising of the cream, the more rapid churning of the cream, the better keeping qualities of the milk and butter — all of these advantages have been assigned to aeration over the non-aeration of milk. I. Review of AiiRATiON of Milk. Studying from the standpoint of the dairyman, Willard,' Arnold,^ Lewis,^ Cooke,^ Wing,^ Plumb,^ Dean,^ and the Danes under Storch* have contributed to our knowledge of its appUcation as it is at present practiced, but all of their work fails to explain the process itself or to state what aeration is. They ' WiUard's Practical Dairy Husbandry, 1871, p. 183. ^ Loo. cit. ^ U. S. Agric. Report, 1874, p. 397. ' Ver. Exp. Stat. Report, 1892, p. 123. = CorneU Exp. Stat., Bull. 39. « Purdue Exp. Stat., Bull. 44 . ' College Reports (Guelph, Out.), 1898, 1899, 1900, 1901. *48''« Beretning fra den Kgl. Veterinaer-og Landbokijskoles Laboratorium for Landkonomiske Farsog. 237 238 THE AERATION OF MILK. do not throw light on why aeration may answer any of the above questions named. An attempt is made to demonstrate whether butter made from aerated millc is better than butter made from non-aerated milk, whether more cheese has been made from milk that has been aerated than from milk that has not been aerated. By these methods, perhaps, some valuable points have been settled, and still the future may change the character of this work wholly, since it is true they have not answered the question, "What is aeration?" or more explicitly, "Were they really working with aerated milk or non-aerated milk?" It seems pertinent at the start, that before .we can understand or explain aeration of milk it is necessary we understand what aeration is, and we know whether we are producing aerated milk or not. All the practical experiments executed have been with aeration and non-aeration as they are known to us from present methods alone. The difference between the two may not be so marked as suspected and for this reason fail in producing the most noticeable effects. The starting point, non-aeration, mvist be first established and from that our conclusions should, in a way, be drawn; that is, we must determine any change in the carbon dioxide and oxygen content, with a decrease in the former and an increase in the latter as a process whose final limitations are not known or influences realized. It is regretted that the value of our present aerating methods cannot be stated more exactly at this time, rather than depend upon the virtually nega- tive results obtained by the individuals named in their endeavor to secure some practical explanation through their practical experiments. We hope to make this our next task ; for the present we must content ourselves with the references to be gained from the evidence at hand. The aerating methods now in vogue in dairy operations may or may not yield results. It is not neces- sary to say whether these methods are effectual or not ; for our purpose their value cannot be determined until each method has been worked out with painstaking care. As soon as there has been established a non-aerated milk and a satisfactory aerated milk, then we may hope to reach some definite con- clusion in its application to the dairy. Before this is done, the one hope that can be held out is that some of the aerating methods now employed may be giving some advantages over the non-aerating methods besides those which are recognized in the removal of odors and taints in a slight degree. The work which follows has for its immediate aim the demonstration of what constitutes non-aeration and aeration and how it may influence the condition of the milk itself and the germ content of the milk. After getting into the work it has been found necessary to eliminate for the present several side-problems which bear directly upon aeration and confine ourselves to the main line planned at the start. It should be noted, therefore, that all the aspects are not to be satisfactorily considered in this article, but rather, that CHARLES E. MARSHALL. 239 a path has been hewn out of an unknown forest of doubt and ignorance which we may follow with profit in the future and which may eventually lead to the solution of the whole question. II. Revieav of the Gas-Content op Milk. Dr. Felix Hoppe/ in 1859, made a study of the gases of milk taken from a goat. In his work, the milk from which the gases were pumped was ex- posed to the air and his results show that his analyses amounted to the same as our analj'ses after milking. Stechnow^ adopted another method of securing the milk. Olive oil, the best obtainable, was used to coA-er the milk. He inserted the end of the teat into the oil, then milked, thus maintaining a layer of oil over the surface of the milk. This, doubtless, had an influence in restraining the effect of th^ oxygen in the air by keeping the air away from the surface of the milk and also reducing the oxygen content which the milk as it passed in a stream from the teat of the cow to the receptacle necessarilj^ acquired. An objection, how- ever, should be offered to the use of oil for the exclusion of oxygen, because it is a well-kno'vs'n fact that oil contains oxygen in the form of super-oxygenated compounds, and also probably as free oxygen. In securing the gases from the milk a mercury pump was used, as seems to have been the general practice with Stechnow and those who followed him, to which little criticism can be offered as viewed from the results obtained. Pfliiger,' believing that Stechnow was open to error in his method of secur- ing the milk from the udder by shielding it as he did from the air by means of oil, attempted to overcome it by another device for conducting the milk from the udder to the receptacle without exposure to the air at all. To do this he employed a piece of rubber, fitting over the teat and udder of the cow as a glove would fit upon the finger; and, by means of a narrow rubber tube and glass tube was able to lead the milk into a vessel of mercury and collect it over mercury. He first expelled the air from the tube by milking for a time, filUng the tube with milk, and then placing the free end of the tube under the mercury in the container. This course may be open to criticism in that there would be some leakage about the surface of the teat over which the rubber was fitted, but the analyses of Pfliiger indicate that he was not troubled very much by the possible exposure to the air from this source, and that the milk secured by this method was practically free, if not wholly free, from air. Further than this, his resiilts are as satisfactory as any of the results obtained when other methods are employed for securing the milk, and which, in our case, we believe, did not allow of any exposure to the air at all. This will be shown later. ' Arch. f. path. Anat. u. Physiol, u. f. klin. Medicin, 1859, Bd. XVII, p. 417. ^ Zeit. f. rat. Medicin von Henle u. Pfeuffer, Bd. X, p. 285, 1861. « Arch. f. d. ges. Physiol., Bd. II, p. 166-173, 1869. 240 THE AERATION OF MILK. The analyses of all these- men were made directly after milking, and did not extend over a time sufhcient to warrant fermentation. Pfliiger was of the opinion that he had secured all the gas possible within one hour's time and made no allowance for a slow generation of gas after the hour when subjected to high pressure. This is not in accord with our experi- ence, for after the first ebullition of gas which takes place early in the pump- ing process, there follows a slight development of gas until fermentation sets in, when there is a marked increase. III. Description of Apparatus. For our experiments we departed from the methods used by Hoppe, Stech- now, and Pfliiger, in that we endeavored to eliminate any possible chance of exposure of the milk to the air. Pfliiger's method, the best previously adopted, is open somewhat to criticism in that there is a possible chance of the milk coming in contact with the air or being more or less churned with traces of air. His analyses, however, indicate that errors of this kind scarcely made any difference; in fact, he seems to have obtained about the same amount of oxygen in his experiments as we obtained in ours. In order to secure the milk from the udder of the cow without any exposure to the air, we employed a milking- tube 6.6 cm. long, attached and fitted to a piece of pressure tubing 80 cm. long and 5 ram. in diameter. The walls of this tube were 4 mm. thick. Over the aperture of the milking-tube was placed a piece of thick rubber tubing of small caliber so that it just fitted the tube. This served as a slide-valve in closing and opening the apertures. The other end of the rubber tube was at- tached to a glass tube, 5 mm. bore, containing a three-way stop-cock, and this in turn was attached to the receptacle used as a container for the milk by means of a short piece of pressure tubing. This will now be described. The container for the milk is of the nature of a large bulb of heavy glass fixed securely to a standard of wood. It has three openings, all of which are closed by means of a glass stop-cock at will. One of these openings is at the bottom of the bulb, another directly above at the top, and the other a side arm on the right shoulder. The top arm consisted of a capillary tube containing a three-way stop-cock an inch above the bulb, and above this a cup and the male portion of a mercury overlapping glass connection. The bottom opening ■and right shoulder opening were tubes 5 mm. in diameter, fitted with two- way stop-cocks. When ready for use, this bulb, with a capacity of 562 c.c, was filled with dry mercury. The side arm connected with the tube extending to the milk-duct of the udder was filled with mercury as far as the three-way stop-cock; the remaining distance was filled completely with boihng water, driving out aU the air, and when driven out, all of the apertures of the milking- tube were closed by means of a shding rubber valve. When ready for drawing the milk, the tip of the milking-tube was carefully inserted into the milk- duct of the udder, and as the teat was pushed on, the rubber valve over the CHARLES E. MARSHALL. 241 apertures in the tube was pushed down, thus allowing the milking-tube to be inserted without the escape or the entrance of any air. When all was in readiness the three-way stop-cock was so turned as to allow the water to run out and the milk take its place; then the three-way stop-cock was again so turned that the milk would run in the direction of the container. The stop- cock of the side arm of the container was also turned on. This made, therefore, a direct connection from the udder to the container of the milk without the presence of any air whatever. To draw the milk the stop-cock of the lower arm of the container was turned so as to allow the mercury to escape, drawing the milk from the udder as normally as it was possible. With the running out of the mercury the milk entered and filled all parts of the container, so that when the mercury had all escaped and the lower stop-cock turned off, the receptacle was completely filled without even traces of air. As soon as the container was filled with milk direct from the udder it was taken to the laboratory and immediately connected, by means of the mercury valve of the top arm, to a modified Hempel pipette specially made for the purpose, in which all rubber connections were excluded. The bulb of the pipette, the nearest the milk container when connected, was so made as to be above the second bulb, next to the Geissler pump. From the very top part of the bulb next to the container extended upright 5 cm. in length, a capillary tube; this then turned to a horizontal position for 10 cm.; finally turned down- ward in a vertical direction and connected with the mercury valve of the con- tainer, previously described, by means of its female portion. In the middle of the horizontal portion of this capillary connecting tube was another three- way stop-cock which faciUtated the filling of the tube with mercury. There- fore the container could be directly connected with the pipette through the capillary tube by means of the mercury connection. Before beginning to pump off the gas, the upper bulb of the pipette and the capillary arm were filled completely with dry mercury by means of the three-way stop-cocks. This was done so effectually that no traces of air were present. A Geissler pump was then attached to the upper stem of the lower bulb of the pipette, the air was removed, thus creating considerable pressure upon the mercury in the pipette. The stop-cock in the upper arm of the container was then turned so that the gas from the milk could escape and be drawn over into the upper bulb of the pipette and thus collected there. When the operation of pumping was completed, the stop-cock in the upper arm of the container was again turned off. To this three-way stop-cock was attached a Hempel burette, filled with mercury, by means of pressure tubing also filled with mercury. (It is under- stood that all rubber tubing connections throughout the entire experiment were further safeguarded by wiring.) As soon as the pressure of the pump was released and reduced to normal conditions, all of the gas was drawn over through the capillary tube and three- way stop-cock into the burette, where it was held for analysis. In carrying 242 THE AERATION OF MILK. out the analyses of the gases we used Hempel's methods entirely, employing his pipette and White's' burette corrected for temperature and barometric pressure. The carbonic acid gas was absorbed by means of potassium hydrox- ide, the oxygen was absorbed by means of alkaline pyrogallate, and what re- mained was considered as residual gas. In the case of previous experiments, the residual gas has been called nitrogen, to me a very presumptuous proceeding, inasjnuch as there is apparently much concealed in the residual gas concerning which we know little. In the lower curves of the pipette a little mercury was placed to act as valves to prevent diffusion. The three-way stopcock of White's burette enabled us to fill the connecting tubes between the pipette and burette completely. By this means all errors were reduced to a minimum. The milk used in these experiments came from one cow, and usually from a single quarter. In none of these analyses did the milk come from more than a single quarter, since a quarter contained sufficient milk to fiU the container, and the possibility of air entering in changing from one quarter to another was thus avoided. It was found in several cases, too, in our work that should the milk become exhausted in a single quarter the tissues closed over the apertures of the milking-tube and prevented the further flow of the mercury; this further demonstrates that no air could possibly enter by means of the apertures in the milking-tube. The whole apparatus, including tubqs, container, connections, and pipette, was severely tested by several preliminary trials, from the starting of the milk from the udder through to the securing of the gas from the milk by pumping. Everything had been verified by these preliminary trials. (In all of our gas analyses the barometric pressure has been reduced to 760 mm. at 0° C, unless otherwise stated.) (A uniform pressure during pumping was maintained by a large galvanized iron tank filled with water in which the entire apparatus was placed.) IV. Analyses of the Gas-Content of Milk befobe Exposure to Air. The following six samples were run through and the gases analyzed with these results : Table I. Sample. I. II. III. IV. V. Dec. 9th Deo. 11th Dec. 13th Dec. 14th Dec. 16th 8:20 a.m. 8:05 a. m. 8:20 a. m. 8:20 a. m. 8:15 a. m. 8:35 a.m. 8:20 a. m. 8:35 a.m. 8:35 a.m. 8:30 a m. 8:55 a.m. 8:40 a.m. 8:55 a.m. 9:00 a. m. 8:45 a. m. 9:55 a.m. 9:40 a.m. 9:55 a.m. 10.00 a. m. 9:45 a m. 38° 38.5° 37.5° 37.5° 37.5° 688.63 695.265 682.937 686.93 694.937 6.16 4.4 4.218 5.342 5.592 4.96 3.44 3.447 4.3788 4.4825 80.58 78.35 81.73 81.97 80.16 .117 .135 .1134 .1143 .0894 2.88 3.09 2.69 2.14 1.6 1.02 .816 .658 .8489 1.02 16.54 18.56 18.27 15.89 18.24 VI. Date Began milking at Finished milking at Began pumping at Completed pumping at Temperature of water in tank . Barometric pressure in mm. . . . Total gas obtained in c.c Free carbon dioxide in c.c Per cent of carbon dioxide .... Oxygen in c.c Per cent of oxygen Residual gas in c.c Per cent of residual gas Dec. 17th 9:00 a m. 9:15 a.m. 9:40 a.m. 10:40 a.m. 37.5° 690.937 4.256 3.667 86.19 .09 2.12 .499 11.71 ' Jour, of the American Chemical Society, Vol. XXII, No. 6, 1900. CHARLES E. MARSHALL. 243 The above six analyses furnish an average which is lower than that of Pfliiger, but this does not appear to detract from the value of the work of either after considering the possible variation occurring in the gas-content of milk from the same quarter of the cow. The amount of carbon dioxide obtained ■in the above analyses is in percentages: 80.58, 78.35, 81.73, 81.97, 80.16, and 86.19. The average of these percentages is 81.49f per cent. Pfliiger obtained in his analyses 90.45 per cent and 87.16 per cent. The average of Pfliiger's analyses, therefore, would be 88.80 per cent. Considering the extremes of the six analyses, there exists a difference of 7.84 per cent. In Pfliiger's the difference is 3.28 per cent. Between the two sets of averages, Pfliiger's and ours, there is 7.30 per cent difference. Thus the possible variation in six analyses is greater than the difference be- tween the average of the two sets of analyses. The uniformity obtained in the six analyses also indicates that there can be no serious discrepancies in the analytical work, and it is difficult to say how much variation Pfliiger might have had had he made six analyses instead of two; furthermore, there is probably as much variation in the gas-content of milk as there is in other constituents. This fact was brought out plainly in some of the prelim- inary tests which are not cited here because those tests served only to pave the way for these six analyses and to develop those points necessary for recording the systematic work. Pfliiger, further, does not state the temper- ature at which the gas was pumped from the milk, nor does he enter spe- cifically into a discussion of the influences which might alter further the gas- content of milk. He obtained a larger amount of gas from milk than we obtained. This alone may answer why more carbon dioxide was secured by him. In two of our preUmJnary tests we first pumped off the gas, which came with ease, collected it in a burette, and then began pumping again, when we obtained a larger percentage of carbon dioxide than we did in the first pumping; however, we did not carry out this work repeatedly, therefore cannot state with any degree of assurance that this will always follow. It was not our purpose to ascertain the amount of gas that may be obtained from milk, but rather to secure a qualitative estimation of any changes that may occur in milk under aeration. To do this it became necessary to hold to the same time and temperature and pressure, as far as possible, while pumping the gas from the milk. After trying the temperature of the room in several tests, about 22° C, the conclusion was reached that the yield in amount of gas in milk was reduced. This amount was greater than we could account for by the contraction of the gas due to a change in temperature; neither did we desire to heat the milk lest some change might occur. That heat increases the gas liberated from a 244 THE AERATION OP MILK. definite amount of milk is proved by Thorner's' work, where large amounts of gas were obtained from comparatively small amounts of milk. Taking milk at 37^° C. or thereabouts, it is believed that the milk does not undergo any change and that the greatest amount of gas could be obtained without altering the milk. Furthermore, it was also found necessary, after the preliminary tests, to limit the time of pumping because at no time was the milk totally exhausted, although most of the gas passed off within fifteen minutes after pumping be- gan. The time limitation was fixed arbitrarily at one hour, believing that this limitation practically precluded the possibility of fermentation. A pressure of 756 mm. to 759 mm. was recorded upon the manometer of the Geissler pump, and from these have been deducted the errors in measure- ment resulting from vapor tension, yielding the results recorded. With three constants, which influence the amount of gas under our control, we were able to obtain fairly uniform results. There was, however, an error creeping in occasionally, due to the presence of obstructions in the upright capillary tube connecting with the container. In this respect the apparatus designed is not as perfect as it should be, for mercury would sometimes fail to be drawn over from the capillary tube extending from the container to the horizontal portion of the arm, and thus interfere with the suction power of the pump upon the milk. I have given in two or three instances the analyses of very smaU amounts of gas; the explanation for these small amounts will be found in such obstructions. Whenever they occurred they co\ild easily be detected while pumping. Pfliiger in his studies has made reference to the possible atmospheric source of some oxygen and nitrogen as found in milk and places himself against such a view. Both in his work and in our analyses this relationship is not established, therefore his views are upheld. The percentage of oxygen falls far below that of the air and even below the solubility of oxygen in water. The percentages of oxygen obtained from unexposed milk are: 2.88, 3.09, 2.69, 2.14, 1.6, 2.12. These figures represent mere traces of oxygen when viewed in the light of percentage of gas volume obtained instead of milk volume, which has not been employed in the estimation of any of our percentages. The following are the percentages of oxygen in its relation to total gas less carbon dioxide ; however, were these figures the solubilities of these two gases, oxygen and nitrogen only, as in air, the oxygen should exist in the ratio of 33.9: 66.1 to nitrogen, according to Bunsen: 10.29, 14.09, 14.70, 11.86,8.06, 15.26. There is only one possible way of explaining the presence of oxygen in the gas-content of milk. It is barely possible that some oxygen may be accounted ' Chemiker Zeitung, Bd. XVIII, 1894. CHARLES E. MARSHALL. 245 for by its creeping in tlirough its adheixMico to the mercury and walls of the container; however, when an attempt is made to secure it in pumping in the absence of milk, not an appreciable trace of gas can be obtained. This ought to demonstrate that the oxj'gen is inherent in the milk. It may be also well worth considering the fact that the cow used for this purpose received very little exercise; she was confined to her stall most of the time. There may be some relation existing between the carbon dioxide and the oxygen of milk due to the amount of exercise of the animal caused by a reduction in the metabolism of the body cells. I call attention to this as a possible answer to the variation in the gas constituents of the milk. After ascertaining the gas-content of milk before it has been exposed to the air, it became necessary, in order to fully understand the changes in the gas-content of milk, to next studj- the gases of milk immediately after milking. V. Analyses of the Gas-Content op Milk Directly after Milking. The milk was secured in the ordinary way by milking into an open recep- tacle, thus allowing the milk to become exposed to the air in an open stream from the milk-duct to the receptacle; further, there was the churning action taking place as the streams impinged against the' surface, which also exerts a marked influence in bringing the air in contact with the milk, because by this method much air is carried down into the milk. It has been found that it takes a few minutes after milking before the perceptible air bubbles rise to the surface. The surface of the milk in the receptacle also offered an oppor- tunity for a considerable interchange of gases between the milk and the air. Immediately after milking, the milk was taken to the laboratory and placed in the same container as in the previous analyses, and the gas was obtained in identically the same manner. Thorner' has made some analyses of the gas obtained from milk imme- diately after roilking, but he employed heat to drive off the gases. The amount of gas which he obtained was much larger than in our case and his results were usually higher in carbonic acid gas, possibly due to his methods. The high percentages of carbon dioxide indicate that heat may have facihtated the dissociation of carbon dioxide, whether in solution or loose combination, to that extent that a higher percentage of carbon dioxide and a lower percentage of oxygen were the results. It is well known that carbon dioxide is with difficulty liberated in the presence of certain salts, as NajGOj, NajHPO^, and others (Setchenow). At any rate, as our tables will show, Thorner's average of percentages differs materially from ours, although it falls considerably below that of milk which has not been exposed to the air. Our analyses are as follows : 'Chemiker Zeitung, Bd. XVIII, 1894. 246 THE AERATION OF MILK. Table II. Sample. Date Began milking at Completed milking at Began pxunping at Completed pumping at .... . Temperature when pumping Barometric pressure in mm . Total gas in c.c Carbon dioxide in c.c Per cent of carbon dioxide . . Oxygen in c.c Per cent of oxygen Residual ga.s in c.c Per cent of residual gas I. II. III. IV. V. VI. Nov. 29th 7:35 a.m. 7:45 a.m. 9:00 a.m. 10:15 a.m. 37.6° 688.6 6.62 3.82 59.31 .91 13.79 1.78 26.9 Nov. 30th 7:25 a.m. 7:35 a.m. 8:35 a.m. 9:35 a.m. 38° 690.63 6.84 3.78 55.33 .86 12.66 2.19 32.01 Dec. 2d 7:30 a.m. 7:40 a. m, 8:30 a.m. 9:30 a.m. 38° 690.63 5.96 3.47 58.27 .84 14.17 1.64 27.56 Deo. 4th 7:35 a.m. 7 :45 a. m. 9:00 a.m. 10:15 a.m. 38° 696.6 2.70 1.78 66.1 .32 11.89 .59 22.01 Dec. 6th 7:35 a.m. 7 :45 a. m. 8:40 a.m. 9:40 a.m. 37.8° 698.6 6.64 4.35 65.65 .727 10.96 1.55 23.39 Dec. 7th :45 a. m. 7:55 a.m. 8:50 a. m, 9:50 a.m. 37.5° 692.84 5.5 2.93 53.28 .85 15.57 1.71 31.15 VI. Analyses of Gas-Content of Milk Directly after Aerating over Glass. To understand what is effected by aeration, a study of the gas-content of milk was made after it had been subjected to a specially devised process of aeration. The common methods of aeration were not employed on account of our inability to cope with the necessary requirements of such experiments with the facilities at hand. It would be absolutely essential that in such methods, the milk used should be under our control from the time it leaves the udder, and that the aeration should also be under our direct supervision. Since such methods were ruled out from necessity, we resorted to means which would simulate the methods in vogue as closely as possible. A piece of glass six feet long by two inches wide was placed on an inclined board at an angle of about twenty degrees. Over this the milk was run, drop by drop, making a film approximately of the extent of surface as indicated in the table. The amount of milk aerated during the process is also given. Room temperature varying from 18° to 20° C. was employed during aeration and was found to be the only temperature which could be satisfactorily used unless a special plant for aerating was constructed. The analyses of gases obtained from milk aerated over glass were conducted the same as with milk after milking. ^^^^^ jjj_ Sample. II. III. IV. VI. Date. . Dec. 3d Completed milking at . . 7-38 a m 7:45 a.m. Completed aerating at 8:50 a.m. Began pumping at 9:25 a.m. Completed pumping at 10:25 a.m. Barometric pressure in mm 692.632 Temperature of water in tank .... 38° Total gas in c.c 6.094S Free carbon dioxide in c.c 2.381S Per cent of carbon dioxide 39.08 Free oxygen in c.c 1.3743 Per cent of oxygen 22.55 Residual gas in c.c 2.3385 38 37 850 Amount of milk aerated in c.c Extent of surface approximately. . li inches Dec. 5th 7:30 a.m. 7 :40 a. m. 7 :50 a. m. 8:45 a.m. 9:10 a.m. 10:10 a.m. 698.632 38° 4.2344 1.4418 34.05 .9908 23.4 1.8017 42.55 700 li inches Dec. 12th 20 a. m. 30 a. m. 40 a. m. 40 a. m. 00 a. m. 00 a. m. 688.632 38° 5 . 3384 2.0670 38.72 1.1130 20.85 2 . 1583 40.43 570 1^ inches Jan. 3d 8:00 a.m. 8:15 a. m. 9 :00 a. m. 10:00 a.m. 10:30 a.m. 11:30 a.m. 703.772 36° 4.1727 1.4287 34.24 .9300 22.29 1.8138 43.47 700 li inches Jan. 4th 7:50 a.m. 8:00 a.m. 8 :45 a. m. 9:45 a. m. 10:22 a.m. 11:12 a.m. 702.632 38° 4.8757 2.4198 49.63 .8664 17.77 1.58 32.60 565 1^- inches Jan. 6th 7:45 a.m. 8:00 a.m. 9:28 a.m. 10:28 a.m. 11:15 a.m. 12:15p.m. 688.632 38° 6.9394 3.3107 47.71 1.1561 16.66 2.4725 35.63 700 1^ inches CHARLES E. MARSHALL. 247 VII. Analyses of Gas-Coxtent of Milk after Aeration over Tin, Copper and through Glas.s Wool and Copper Sieves. It was thought advisable to test different materials in place of glass, inas- much as it is well established that some metals have an influence upon oxida- tion. The next two tables will indicate the value of aeration over tin or copper. The last table is the result of an attempt to break up the drops of milk by passing the drops through copper sie^-es and through glass wool three inches thick. The drops pass in all through six feet of air before reaching the recep- tacle, the materials mentioned above intervening. It was found very difficult to operate this combination of air sieves and glass wool satisfactorily so as to obtain uniform results. The trouble lay in the fact that sometimes the drops could be thoroughly shattered and at other times small streams would form. T.\ble IV. (Over Tin.) Sample. I. II. III. IV. V. Date Began milking at Completed miUcing at Began aerating at Completed aerating at Began pumping at Completed pumping at Barometric pressure in mm Temperature of water in tank . . . Total gas in c.c Free carbon dioxide in c.c Per cent of carbon dioxide Free oxygen in c.c Per cent of oxygen Residual gas in c.c Per cent of residual gas Amount of milk aerated in c.c . . . Extent of surface approximately Dee. 20th 7:40 a.m. 7:50 a.m. 7:55 a. m. 8:55 a.m. 9:30 a.m. 10:30 a.m. 695.265 38.5° 4.4822 1.8874 42.11 .8541 18.61 1.7606 39.28 1000 iinch Dec. 21st 7:45 a.m. 8:00 a.m. 8:05 a.m. 9:05 a.m. 9:30 a.m. 10:30 a.m. 696.632 37.8° 6.4969 1.8561 28.57 1.4716 22.65 3.1691 48.78 600 2^ inches Deo. 23d 7:45 a.m. 8:00 a.m. 8:05 a.m. 9:05 a.m. 9:25 a.m. 10:25 a.m. 678.632 37.9° 6.7922 1.5900 23.41 1.7489 25.75 3.4531 50.84 565 2i inches Dec. 28th 7:35 a.m. 7:48 a.m. 7 :55 a. m. 8:55 a.m. 9:18 a.m. 10:18 a.m. 688.672 38° 6.9385 2.4485 35.29 1.4057 20.26 3.0841 44.45 800 li- inches ,Ian. 7th 7:40 a.m. 7:50 a. m. 8:45 a. m. 9:45 a. m. 10:15 a.m. 11:15 a. m. 687 . 242 37.2° 6.6669 2.9854 44.78 1.0320 15.48 2.6494 39.74 560 1^ inches Table V. (Over Copper.) Sample. I. II. III. IV. V. Date Dec. 10th 7:30 a.m. 7:40 a.m. 7:45 a. m. 8:40 a.m. 9:40 a.m. 10:40 a.m. 690.632 38° .6349 .136 21.43 .205 32.34 .293 46.23 1000 2i inches Dec. 12th 7:35 a.m. 7:45 a.m. 7 :48 a. m. 9:00 a.m. 9:30 a.m. 10:30 a.m. 696.937 37.5° 5.0379 2.638 52.25 .725 14.41 1.679 33.34 1000 li" inches Deo. 19th 7:50 a.m. 8:00 a.m. 8:10 a.m. 9:10 a.m. 9:35 a.m. 10:35 a.m. 696.937 37.6° 3.4637 1.516 43.79 .588 16.99 1.358 39.22 500 ^inch Deo. 24th 7:30 a.m. 7:45 a.m. 7:55 a.m. 8:55 a. m. 9:25 a.m. 10:25 a.m. 686.632 38° 4.4236 1.837 41.54 .589 13.33 1.99 45.13 500 1^ inches Dec. 26th Began milking at Completed milking at 7:35 a.m. 7:50 a. m. 8:00 a.m. 9:00 a.m. 9 :30 a. m. 10:30 a. m 690.632 Temperature of water in tank 37.8° 3.4411 Free carbon dioxide in c.c. . . 1.8110 52.63 .3169 9.21 Residual gas in c.c .... : 1.3131 38.16 1000 Extent of surface approximately li inches 248 THE AERATION OF MILK. Table VI. (Through Glass Wool.) Sample. Date Began milking at Completed milking at Began aerating at Completed aerating at Began pumping at Completed pumping at Barometric pressure in mm Temperature of water in tank . . . Total gas inc.c Free carbon dioxide in c.c Per cent of carbon dioxide Free oxygen in c.c Per cent of oxygen Residual gas in c.c Per cent of residual gas Amount of milk aerated inc.c... I. Nov. 11th 7:40 a.m. 7:50 a.m. 8:30 a.m. 9:30 a.m. 9:45 a. m. 10:45 a.m. 722 . 3932 22.8° 4.0405 .5385 13.33 1.0771 26.66 2.4247 60.01 560 II. Nov. 25th 7:35 a. m 7:45 a. m, 8:00 a. m, 9;00a.m. 9:25 a. m, 10:25 a.m. 690.632 37.8° 2.5399 .5440 21.42 .7711 30.36 1.2247 48.22 600 III. Nov. 26th 7:25 a. m, 7:35 a. m, 7:45 a. m 8 :45 a. m, 9:30 a.m. 10:30 a.m. 697.265 38.5° 4.9704 1.S810 31.81 .9940 20.00 2,3952 48.19 750 IV. Dec. 27th 7:45 a.m. 7:55 a. m. 8:00 a.m. 9:00 a.m. 9:35 a.m. 10:35 a.m. 694.632 38° 5.4427 1.9952 36.66 .8844 16.25 2.6629 47.09 1050 The studies of aerated as compared with unexposed milk and milk directly after milking can profitably be presented at this stage. The percentages of carbon dioxide are : In unexposed milk, 80.58, 78.35, 81.75, 81.97, 80.16, 86.19; in milk after milking, 59.31, 55.33, 58.27, 66.1, 65.65, 53.28; in milk after aeration : over glass, 39.08, 34.05, 38.72, 34.24, 49.63, 47.61; over tin, 42.11, 28.57, 23.41, 35.29, 44.78; over copper, 21.43, 52.25, 43.79, 41.54, 52.63; through glass wool and copper sieves, 13.33, 21.42, 31.81, 36.66. The averages of these are : in unexposed milk 81.496 per cent. in milk after milking 59 . 636 " in milk after aeration : over glass 40.572 percent. over tin 35.832 " over copper 42.328 " through glass wool and copper sieves 25 . 805 " In the case of oxygen the percentages are : in an unexposed milk, 2.88, 3.09, 2.69, 2.14, 1.6, 2.12; CHARLES E. MARSHALL. 249 in milk directly after milking, 13.79, 12.66, 14.17, 11.89, 10.96, 15.57; in milk after aeration : over glass, 22.55, 23.4, 20.85, 22.29, 17.77, 16.66; over tin, 18.61, 22.65, 25.75, 20.26, 15.48; over copper, 32.34, 14.41, 16.99, 13.33, 9.21; through glass wool and copper sieves, 13.33, 21.42, 31.81, 36.66. The averages of these are : in unexposed milk 2.42 per cent. in milk directly after milking 13 . 176 " in milk after aeration : over glass 20.586 overtin 20.55 over copper 17 . 256 " through glass wool and copper sieves 25.805 " The residual gas has the following percentages: in unexposed milk, 16.34, 18.56, 18.27, 15.89, 18.24, 11.71; in milk after milking, 26.9, 32.01, 27.56, 22.01, 23.39, 31.15; in milk after aeration over glass, 38.37, 42.55, 40.43, 43.47, 32.60, 35.63; over tin, 39.28, 48.78, 50.84, 44.45, 39.74; over copper, 46.23, 33.34, 39.22, 45.13, 39.16; through glass wool and copper sieves, 60.01, 48.22, 48.19, 47.09. The averages of these are in unexposed milk 16.535 per cent. in milk directly after milking 27 . 17 " in milk after aeration over glass 38.842 " overtin 44.618 " over copper 40 . 416 " through glass wool and copper sieves 50.8775 " 250 THE AERATION OF MILK. The relation of oxygen to residual gas is best shown by the determination of percentages of oxygen to the residual gas plus oxygen. This is done to bring out any atmospheric relationship. Percentages in case of unexposed milk, 10.29, 14.09, 14.70, 11.86, 8.06, and 15.26; milk after milking, 33.83, 28.19, 33.88, 35.16, and 31.96; milk after aeration over glass 37.01, 35.44, 34.00, 33.89, 35.45, 31.86; over tin, 32.14, 31.70, 33.61, 31.30, 28.05; over copper, 46.16, 30.15, 30.21, 22.83, 19.44; through glass wool and copper sieves, 30.75, 38.60, 29.32, 25.65. The average percentages are in unexposed milk 12 . 376 in milk directly after milking 32 . 53 in milk after aeration over glass 33 . 94 over tin 31.16 over copper 29 . 76 through glass wool and copper sieves 31 . 13 We gather from these studies that after the milk leaves the udder of the cow there is a diminution in the amount of carbon dioxide and an increase in oxygen to a certain low percentage, whi ch is dependent upon the thorough- ness with which the milk is brought in contact with the air. From the ex- treme results it is evident there has been a falling off in carbon dioxide of 73.86 per cent and an increase of 35.06 per cent of oxygen, thus indicating that there is no equi-volumetric interchange. Even the average results show no intimate relationship between the oxygen and carbon dioxide. This may, however, be due to the difference in the solubilities of the gases, carbon dioxide being dissolved volume for volume, while oxygen enters only comparatively meagerly into solution, about four per cent by volume. It would be possible by the direct determination of carbon dioxide in milk before the beginning of fermentation to say to what extent the milk has been aerated. Further studies of the gases of milk and their interchange will reveal other aspects of the subject. CHARLES E. MARSHALL. 251 VIII. No APPRECI.A.BLE INTERCHANGE OP G.VSES BETWEEN STERILIZED MiLK AND THE Air Confined over It. In six experiments an attempt was made to ascertain whether or not there was a decrease in the amount of oxygen in the air confined over sterilized milk and a development of carbon dioxide. Liter flasks were taken and 500 c.c. of fresh milk was added and sterilized on three consecutive days for one hour each day. They were then sealed with sterilized rubber stoppers through the single perforations of which passed two-way stop-cock tubes, by means of which the air-content could be drawn at will. Besides the rubber stopper seal, a mercury seal was made about the rubber stopper by wiring a rubber tube about the neck of the flask and allowing it to reach above the rubber stopper, where it was cut off. This formed a cup-shaped envelope for the stopper and could be filled with mercury to cover the rubber stopper com- pletely. Through the glass tube the air-content of the flask could be removed easily at wiU and analyzed without the possibility of extraneous air entering. Six analyses of air from different flasks made within a period of six weeks showed no appreciable difference from that of the extraneous air used as check. IX. Action of Specific Bacteria upon the Interchange of Gases between Milk and Air.^ It becomes necessary, after a review of the gas-content of milk when un- exposed to the air, when exposed to the air by milking, and by the process of aeration, to consider what influences fermentation may have. Of course, it is well understood that micro-organisms are capable of pro- ducing different gases, yet we have felt it obligatory to make some specific studies in this connection. Consequently several micro-organisms, isolated from milk and known by their laboratory numbers, were used to inoculate sterilized milk, and the flasks sealed as described under V. It will at once be recognized that no attempt has been made to determine any other gases which may be present than carbon dioxide and oxygen because with these alone are we mostly concerned. Hydrogen and marsh gas were determined in several instances and the results obtained were included with the total residual gas. In most analyses, it is worthy of emphasis, there have been noticeable decreases in the amount of oxygen and increases in carbon dioxide. This, however, is not universally true, because in No. 121 there has been no produc- tion of carbon dioxide and the oxygen has been but little reduced. In No. 126 there are no traces of carbon dioxide and still the oxygen has been prac- tically consumed. In short, there is a wide diversity of action; each micro- organism seems to be governed by its own peculiar functions. As a rule, ' See description of bacteria on page 257. 252 THE AERATION OF MILK. o O o g E 2; o O tc K tc < 2 >— I O O P o S3 lO CO CO o CO IX O 1?^ o IM CO CD CO c 00 oj e U5 O O CO Oi 00 IC CO r- ■^ O c^ CM CO (N (N o CO ^ 1-1 OS bD -e c CS] (N OS 6 3 C !zi <; IT « ^ 00 o6 CO o "* Ol (N CD lO O (N N o ir t^ CO C^ CD M -e o CS) P" M (N tr- d 3 O O iz; < ie u: O CM IT OS CO a C£ o o lO •<* CO OS o e^ O ixi o o If Tl^ d 3 o 2 C^ Tti lO Z << M iC lO ca O M CO 00 CSI c « CS i> CO to a> o o o d -5 c o cq 1-H >o ^ OS DC CSI CSI oc •* OS o a- ■* CI <— ' c IT IT CO OS ^ CSl I'. d o D- p^ 00 '3 o o 1 a £ 1 CI 3 'c 1 IS a i 1 o a 1 •s 1 a o 1 u O o 6 a 3 g > c 'c bO 1 1 o i la 3 £ "o c c ~p ■4^ 4^ i=! 1:^ c «- d =• a 3 c S 3 3 , 3 s. 3 e. 03 3 a 3 OJ 1 "5 i' g-i i-E a-s O n c e < < caiise the "spin" of the ball was nearly spent, or this latter force being weakened, and the ball striking obliquely, this combina- tion of influences determined the lateral passage of the projectile. I do not contend that ricochet shot wounds never occur, but that the ricochet theory need not be invoked when demonstrable physical laws are fully adequate to explain the conditions, and when the ball has been deflected by striking objects during its flight, it usually shows by the form of the en- trance wound that it has entered more or less sideways, and that this may, therefore, be due to ricochet. Again, its altered form, no bone having been encountered dm-ing its passage through the soft tissues, may prove that its lodgment was due to having had its flight artificially retarded. Under such circumstances the wound of entrance would probably show some pecuharity suggestive of a glancing ball. Even after a ball has been influenced by ricochet, further deflection of its course after contact with the tissues is, in most instances, due merely to the influence of the forces already described as operative upon a missile whose onward flight and rapidity of revolution have been materially lessened by the distance it has traversed in space. Finally, while the ricochet theory does account for the irregular entrance of some missiles, and because their "energy" is thus lessened by contact with extraneous objects, this theory may account for their lodgment and deflec- tion; in the majority of instances this is not the correct explanation, and at the ranges where a large number of missiles strike, many cases of lodgment .and deflection are to be expected, and accounted for on demonstrable scien- tific principles. THE TENDON ACTION AND LEVEEAGE OE TWO-JOINT MUSCLES OE THE HIND LEG OE THE EEOG, WITH SPECIAL REFERENCE TO THE SPRING MOVEMENT.' WARREN PLIMPTON LOMBARD, A.B., M.D., Professor of Physiology, University of Michigan. If a man had invented an apparatus for propelling a vehicle on the Hnes of the leg of a frog, I fear that he would be greatly criticised. It is not, at first sight, a very promising contrivance. The skeleton consists of fragile bones, loosely connected at their extremities by ligaments — a chain of levers, moved by muscles which' are elastic structures, capable of shortening and pulling on the bones to which they are attached, but when at rest, extensible, flexible, and flaccid. Not only does this apparatus differ in respect to stability from most of those devised by man, but it has a remarkable arrangement, viz., the ' presence of many muscles which pass over two joints. Most of these two- joint muscles flex one joint by one end and extend an adjacent joint by the other end. For example, the gastrocnemius, the most powerful extensor of the lower leg, and one which evidently plays a very important part in the leap, not only extends the ankle but flexes the knee. Excite its nerve so that it contracts alone — it is seen to produce both of these effects at the same time. Since all the joints of the leg must be violently extended when the frog leaps, how can it fail to be a decided disadvantage for the strongest muscles of the leg, those which manifestly act as extensors to propel the body, to be attached so as to act at the same time as flexors? A. Fick' points out many advantages which come from the presence of these muscles : — the lessening of the total mass of muscle substance which is required for movements of the different segments of the limb ; the consequent economy in the expenditure of energy; the fact, shown by Borelli,^ that the movement of one of the joints crossed may increase the tension of a muscle and so help it to work on the other joint; the fact that a contracted two-joint muscle may act merely as an elastic band, and transmit part of the energy developed by a one-joint muscle which passes over one of the joints, to the other joint, for example, when a man rises from the squatting position, a one-joint extensor of the hip, as the giutseus, may act through the rectus to ' Hermann's Handb. d. Physiol., 1879, Bd. I, II, S. 284; also Fick, Untersuch. u. Muskelarbeit, S. 39, Basel, 1867. ^ De Motu animalium, first edition, 1680. 280 WARREN PLIMPTON LOMBARD. 281 extend the knee.^ Other examples are the action of the gastrocnemius at the end of a step, to continue the plantar flexion of the foot, and at the same time to flex the knee; the closing of the claws of a bird when it squats on its perch or raises the foot ; extension of the claws of the lion when the fore leg is thrust out. In the case of the frog, all the largest and strongest muscles of the hind leg are two-joint muscles; no one, therefore, can work on any problem involving the co-ordina- tion of the action of the muscles of the leg without considering the method of action of these mus- cles. In an article on the " Centers and Paths of Transmis- sion in the Spinal Cor'd of the Frog, " Gad^ refers to the passive action of the two-joint muscles of the leg of the frog, and shows how it may play an important part in locomotion. While working under Professor Ludwig in the Physio- logical Laboratory in Leipzig, in 1883 and 1884, on the spread of reflex impulses in the spinal cord of the frog, I found, like Gad and his students, that it was necessary to consider the effect of the passive action of two-joint muscles upon the movements of the leg. In publishing the results of these experiments^ I called attention to the important part 'Eugene Pick; Arch. f. Anat. u. Physiol., Anat. Abthl., 1879, S. 201. ^ Verlil. d. physik.-nied. Ges. in Wiirzburg, 1884, N. P., Bd. 18, S. 129. ^ Die raumliche und zeitliche Auf- einanderfolge refleotorisch contrahirter Muskln; Arch. f. Anat. u. Physiol.. 1885, S. -108. Fig. 1 . — Diagrammatic re-presentation of the tendon action of the two-joint muscles of the hind leg of the frog. When the leg is extended, active contraction of a one-joint flexor muscle of the hip (a), and the passive tendon action of the two-joint muscles on the median side of the thigh (b) and on the lateral side of the lower leg (d) will cause complete flexion of the hip and knee, and partial flexion of the ankle. When the leg is flexed, active con- traction of a one-joint extensor muscle of the hip, (f ) and the passive action of the two-joint muscles on the lateral side of the thigh (c) and the median side of the lower leg (e) will cause the hip, knee, and ankle to be extended. 282 THE HIND LEG OF THE FROG. which these muscles must play in the movement of the leg, illustrating their action by the diagram shown in Fig. 1. Since that time, on several occasions, I have made a study of the anatom- ical relations of these muscles to the parts to be moved, and the effect of their peculiar method of attachment upon their functional activity. In the autumn of 1886, at the suggestion of Professor H. P. Bowditch, a series of experiments were made in order to ascertain, a propos of the Ritter-Rollet' phenomenon, whether the irritability of the muscles which flex and those which extend the leg is the same. These experiments failed to give a satisfactory answer to the question. The fact stood out prominently that the largest and strongest of the muscles of the hind leg of the frog are two-joint muscles, and produce opposite types of action at their two extremities, flexing the joint at one end, and extending the joint at the other; they are, in short, both flexors and extensors. It seemed hardly possible that, when functionally active by locomotion, they should produce these two antagonistic types of movement at the same time, but it did not seem safe to class them either as flexors or extensors until their normal action had been investigated. In a letter to Dr. Bowditch in November, 1886, in which the double action of two-joint muscles was described and their method of action was speculated upon, I ended by sajdng : " I have reached a point where I have shown that it is possible for an extensor of the knee to flex the knee, and it is, perhaps, as well that I should stop my theorizing. The mechanics of the muscles of the leg of the frog is not the less a tempting subject for research and speculation because it offers an excellent opportunity for reasoning in a circle. " The following spring a careful study of the anatomy of the muscles and joints of the hind leg of the bull -frog was made by me at Johns Hopkins University. The results obtained added greatly to my admiration and respect for these wonderful mechanisms, but I was only led back to the apparent paradox, that a two-joint muscle, because of the tendon action of a muscle on the opposite side of the leg, may flex a joint of which it is an extensor or extend a joint of which it is a flexor. Which of these it would do by a given movement would seem to depend on the relative leverage of the two extremities of the muscle by the position of the joints at the time. During the past year the meth- od of action of two-joint muscles of the hind leg of the frog has been again the subject of serious study, and especial attention has been given to the leverage of these muscles by different positions of the joints which they cross. This work is still in progress, but enough data have been gathered to justify the view expressed above with reference to the action of the two-joint muscles, and to remove one of the chief obstacles to an understanding of the action of these muscles by the leap. ^Sitzungsbr. d. k. Akad. d. Wissensch., Wien, 1876, Bd. VI, Abth. 3. WARREN PLIMPTON LOMBARD. 283 The Tendon Action of T\\'o-Joint Muscles. In the passage of the leg from the flexed to the extended position, as Hor ing' has shown, important rotation and adduction movements occur. The con- sideration of these will be deferred till another time, and in this paper attention will be paid oiil}^ to the ex- tension of the leg by the two-joint muscles. The terms flexion and extension are used diflerenth- by differ- ent authors, and so that there shall be no misunder- standing, I will state that in this paper the move- ments of the leg of the frog will be assumed to occur in a plane approximately par- allel to that of the earth, and that the term flexion mil be employed for movements in this plane of a type to cause the two bones entering into a joint to approach each other, and extension, for movements of a type to cause the bones to separate. In the case of the knee, the term over-flexion will be used to describe the position taken by the lower leg when it is carried by the flexing act past the thigh. The power of the spring of the frog is largely due to the method of attachment of the two-joint muscles and their resulting tendon action. Because of the peculiar relation of the two-joint muscles to the joints which they cross, any force which extends or flexes either hip, knee, or ankle, tends to cause like movement of all the rest. Fig. 2, A, shows the effect of the tendon action of » Arch. f. d. Ges. PhysioL, Bd. LXVIII, p. 9. 23 yff^Sy^s. Fig. 2. — A, the effect of passive flexion of the foot to flex the femur; B, the effect of passive exten- sion of the femur to extend the foot. 284 THE HIND LEG OF THE FROG. the gastrocaemius to flex the femur as the foot is flexed; and B, its effect to extend the foot as the femur is extended. Gad' observed, that when the frog is suspended vertically, passive flexion of the thigh on the belly will cause complete flexion on the knee, and draw the lower leg up to the thigh, but that the flexion of the knee will not cause complete flexion of the ankle, the foot being raised only so as to form something less than a right angle with the lower leg (see Fig. 2, A). Hering^ states that the foot can be flexed by the passive action of the two- joint muscles, in the case of a pithed frog, only about ninety degrees, and that the active contraction of flexor muscles is essential to the completion of the act. He emphasizes the fact, however, that under normal conditions the muscles are in reflex tonus, and that through this their tension is considerably increased. That one of the one-joint flexors of the ankle must act, is shown by the fact that if the peroneal nerve be cut, it is only occasionally that the foot flies into the sitting position by recovery from the spring. The passive flexion of the foot he finds to be largely the result of the tension brought on the tibialis anticus longus where it draws across the extensor side of the knee. If the tendon of this muscle is cut, the passive flexion of the foot is much less. Gaupp,' when describing the movements of the knee, refers to the passive action of the tibialis anticus longus and peroneus to extend the knee when the ankle is extended. Extension of the hip causes the triceps to be pulled on, and the knee to be extended ; extension of the ankle causes the tibialis anticus longus to be pulled on, and the knee to be extended; extension of the knee causes the semi- membranosus, gracilis magnus, and semitendinosus to be pulled on, and the hip to be extended; extension of the knee also causes the gastrocnemius to be pulled on, and the ankle to be extended. In certain positions of the joints, the ilio-fibularis and peroneus may also take part in producing these effects. In a similar manner, flexion effects are transmitted from joint to joint. This tendon action of the two-joint muscles can be readily seen in Fig. 1, and is well illustrated by a model, such as is roughly represented in Fig. 3. Since the extension of the hip will extend the knee, and extension of the knee the ankle, a one-joint extensor muscle of the hip can cause extension of the knee and ankle; since extension of the ankle will extend the knee, and extension of the knee will extend the hip, a one-joint muscle of the ankle can extend the knee and the hip; finally, a one-joint extensor muscle of the knee can extend both hip and ankle. The same is true of the one-joint flexors ' Verb. d. physik.-med. Ges. in Wiirzburg; N. F., Bd. XVIII, 1884, p. 171. ' Arch. f. d. Ges. Physiol., Bd. LXVIII, S. 11. 'Ecker's and Wiedersheim'& Anat. des Frosches, neu bearbeitet von Dr. Ernst Gaupp, 1896, p. 90. WARREN PLIMPTON LOMBARD. 285 of the hip and anlde; there are none of the knc-c. The one-joint flexor muscles of both anlde and hip can flex the ankle, knee, and hip. These effects are all the result of the passi^•e, tendon-like action of the two-joint muscles, and the amount that distant joints are actc-d upon dejiends on the resting length of these muscles. ■ T.AL. Fig. 3. — Model showing tendon action of two-joint muscles. II, one-joint flexor of hip, e. g., iliacus intemus; A M, one-joint extensor of hip, e. g., adductor magnus; T, two-joint flexor of hip and extensor of knee, e. g., triceps; G M, two-joint extensor of hip and flexor of knee, e. g., gracilis magnus"; T A L, two-joint extensor of knee and flexor of ankle, e. g., tibialis anticus longus; G, two-joint flexor of knee and extensor of ankle, e. g., gastrocnemius. If the one-joint iliacus internus is pulled on and it flexes the hip, the two-joint gracilis magnus flexes the knee and the tibialis anticus longus the ankle. If the one-joint adductor magnus is pulled on and it extends the hip, the two-joint triceps extends the Iinee and the gastrocnemius the ankle. 286 THE HIND LEG OF THE FROG. Not only the tendon action of muscles but the effects of inertia favor the simultaneous production of flexion and extension in adjacent joints. Otto Fischer^ found that muscles can act on a joint over which they do not pass. He writes : " It appears in general, that a one-joint muscle calls forth, as a rule, in an adjacent joint, a rotation the opposite to that in the joint which lies between its points of insertion. " Thus, a one-joint flexor of the elbow, which causes the lower arm to move forward, will act on the shoulder joint and cause the upper arm to move backward. In a like manner a one-joint flexor of the ankle of a frog, which causes the foot of a frog to move in the lateral direction, will cause the lower leg to move in the median direction and produce flexion of the knee. A flexor movement of the knee, which carries the lower leg in the median direction, will carry the thigh in the lateral direction, i. e., cause flexion of the hip. Similarly a one- joint extensor of the ankle will extend the knee and a one- joint extensor of the knee will extend the ankle. Fischer^ also found that a one-joint flexor or extensor of the shoulder of a man could produce an opposite rotation in the elbow. In a like manner a one-joint extensor of the hip of a frog would cause extension of the knee, and a one-joint flexor of the hip, flexion of the knee. Nor is this transfer of action from joint to joint confined to the one-joint muscles. H. E. Hering' ascertained by experiments on frogs in which all the muscles of the thigh and lower leg, except the gastrocnemius or tibialis anticus were cut away, if the gastrocnemius, which extends the foot on the lower leg and flexes the lower leg on the thigh, was excited by electricity, it at the same time flexed the thigh on the pelvis, and the tibialis anticus longus, which flexes the foot on the lower leg and extends the lower leg on the thigh, if excited, also extended the thigh on the pelvis. " Several-joint muscles work also on joints over which they do not draw, and in such a way that they call out an opposite rotation in a proximal joint. "^ I have repeated this experiment with success and, as Fischer has shown, it can be demonstrated on a model; he writes: "By the several-joint muscles, pulUng across the elbow and shoulder joints, the relations lie not so simply. For these, also, indeed, it has proved, that as the contraction of the muscles follows the period of rest, with an arbitrary tension for every special point of departure of the arm, there belongs a special relation of the degrees of rota- tion in the shovdder and elbow joints. " A two-joint muscle of the leg of the frog, therefore, like a one-joint muscle, by causing the extension of a joint, may produce extension in neighboring 1 Abhandl. d. k. Sachs. Ges. d. Wiss., math.-phys. Classe, 1895, Bd. XXII, No. 2, S. 193. 2 Abhandl. d. k. Sachs. Ges. d. Wiss., math.-phys., CI., 1897, Bd. XXIII, No. 6, S. 556. 3 Arch. f. d. Ges. Physiol., 1897, Bd. LXV, S. 629. ' Ibid., S. 630. WARREN PLIMPTON LOMBARD. 2S7 joints. These effects of inertia are in harmony with tlie effect produced by the tendon action of the muscles of the leg; they co-operate to cause all the joints to extend simultaneously by the leap. // the joints are bound together by the tendon action of two-joint muscles, how can independent movement of the separate joints be made? All the two-joint muscles of the hind leg of the frog are attached under very slight tension, varjang, of course, with the amount of reflex tonus present, and are all exten- sible. TMien they are not contracting, they readily yield to any stretching force sufficiently to permit of wide excursions of the joints. The one-joint flexors and extensors may, therefore, produce independent movements of any of the joints, if the corresponding movements of the other joints are prevented by the action of other muscles. There are one-joint flexors and one-joint extensors for the anlde and for the hip, and there is a one-joint extensor for the knee and another muscle, which, although having some action on the ankle, is principally a one-joint extensor of the knee. There are, strange to say, no one-joint flexors of the knee, and if the joint has to be flexed independently, it must be flexed by the two-joint muscles. If a two-joint muscle contracts alone and unopposed, it produces (the sartorius is an exception) two opposite effects by its two extremities : it extends one joint bj- one end, and flexes the other joint by the other end. If the action of one end is prevented by some opposing force, as by the contraction of antagonistic one-joint muscles, the principal effect of the contraction is observed at the other end. Thus, the two-joint flexors of the knee which lie on the median and under surface of the thigh, and which tend as well to extend the hip, if the extension of the hip is opposed by one-joint flexors of the hip, will act chiefly as flexors of the knee. The same is true, to give another example, of the gastrocnemius which extends the ankle, as well as flexes the knee; if the extension of the ankle is opposed by the one-joint flexors of the ankle, its con- traction will be manifested principally in a flexion of the knee. What is the effect of the simultaneous contraction of all of the two-joint muscles of the legf Independent movements of separate joints can only occur when the two- joint muscles, as a whole, are at rest, because most of these muscles are more powerful than the one-joint muscles, and when they contract they become tense-like cords, and as has been described, compel like movements to occur in all the joints. Most of these muscles tend to flex one joint and to extend the next. Nor does a two-joint muscle merely extend the joint of which it is an extensor, it makes use of the- tendon action of the other two-joint muscles to extend the other joints. The triceps, by extending the knee, pulls on the gastrocnemius and extends the ankle; the gastrocnemius, by extending the ankle, pulls on the tibialis anticus longus and extends the knee ; the tibialis anticus longus, by extending the knee, pulls on the semimembranosus, gracilis 288 THE HIND LEG OF THE FROG. magnus, and semitendinosus and extends the hip; and finally, the semi- membranosus, gracilis magnus, and semitendinosus, by extending the hip, pull on the triceps and extend the knee. Hering,^ when studying the action of the muscles by which the leg is drawn to the sitting position, was compelled to ask himself, how it is that a muscle, like the tibialis anticus longus, which extends the knee as well as flexes the foot, can be used with economy at a time when all the joints of the leg must be flexed? I, long ago, found myself confronted by like questions when studying the action of the muscles of the frog by the spring movement. How can the tibialis anticus longus, which flexes the ankle as well as extends the knee, be of use in the leap, when the ankle, hke the knee, must be vigorously extended? It is hard to believe that the muscle cannot be economically used, either when the leg, as a whole, is to be flexed or to be extended. One is the less ready to accept such a conclusion because it would apply equally well to the rest of the two joint muscles of the leg, which make up the bulk of the muscle substance. Nevertheless, the question has to be answered: Why, by the leap, when all the joints must be vigorously extended, does not the flexion action of these muscles interfere with the extension of the leg? This question leads to a para- dox. A two-joint muscle may act as an extensor of the joint of which it is a flexor. The following conditions are essential for the production of this effect with completeness : a. The muscle in question. A, must have a better leverage at the end by which it extends than at the end by which it flexes. b. There must be on the opposite side of the leg a two-joint muscle, B, which flexes the joint which A extends, and extends the joint which A flexes. c. The extensor leverage and strength of A must be sufficient to enable it to make use of the tendon action of B. d. B must be sufficiently contracted to act as a tendon. For example, the gastrocnemius, if all the other muscles of the leg be cut away, will, if it be excited, extend the ankle and flex the knee. Now, if a thread be fastened to the two ends of the tibialis anticus longus, on the oppo- site side of the leg, and be made sufficiently tense to produce the tendon action of this muscle, excitation of the gastrocnemius results in the extension of the knee, as well as of the ankle. This can be readily shown by a model in which a spring is used to represent the contracting gastrocnemius, and a string, the tendon action of the tibialis anticus longus. I of Fig. 4 shows the foot flexed and femur extended, and II the extension of the foot and flexion of the femur which occur when the parts are released, and the spring, which represents the contracting gastroc- ' Arch. f. d. Ges. Physiol., Bd. LXVIII, S. 13. WARREN PLIMPTON LOMBARD. 289 nemius, is allowed to act. Ill shows the foot flexed and femur flexed, and IV the extension of both foot and femur, which occurs when the parts are released, and the spring, by extending the foot, pulls on the string, which represents the tense tibiahs anticus longus, and through it extends the knc^e. I I m w Fig. 4. — a represents the foot; 6, the cms; c, the femur; d, the gastrocnemius, and e, the tibialis anticus longus. I, foot flexed and femur extended; II, extension of foot and flexion of femur caused by gastrocnemius; III, foot flexed and femur flexed; IV, exten- sion of femur, as well as foot, caused by strain brought by contracting gastrocnemius, through tibialis anticus longus on extensor side of knee. The gastrocnemius has a better leverage on the tarsus than on the femur, and hence its contraction will act more strongly to extend the ankle than to flex the knee. Moreover, the leverage of the gastrocnemius on the tarsus is better than that of the tibialis anticus longus, hence, when the gastrocnemius contracts, it extends the ankle and exerts a vigorous pull on the tibialis anti- cus longus; this pull is transmitted through the muscle across the extensor side 290 THE HIND LEG OF THE FROG. of the knee to the femur, and the knee is extended because the strength of the pull transmitted from the distal end of the gastrocnemius to the extensor side of the knee is greater than that applied by the muscle to the flexor side of the knee. The effect is still further increased by the fact that, when the leg is flexed, the extensor leverage of the tibialis anticus longus on the femur is stronger than the flexor leverage of the gastrocnemius. Inasmuch as the tibialis anticus longus is extensible, it would be necessary that it should be contracting if its full tendon action is to be obtained. The contraction of the muscle would not oppose the extension effects produced by the gastrocnemiixs, but on the contrary, favor them. If all the muscles are cut away from the leg excepting the tibialis anticus longus, and that be ex- cited, it will extend the knee and flex the ankle. If a string be substituted for the gastrocnemius, and this be made tense enough to produce its tendon action, excitation of the tibialis anticus longus will produce extension of the ankle, as well as of the knee. These two apparently antagonistic muscles, therefore, work in the same sense, each helping the other to produce extension of the ankle and knee. We can sum up the process by sajdng : The energy of the two-joint muscles of the lower leg of the frog is transmitted in a circle in the direction of the greatest leverage. The same principle would be effective in the production of extension by two-joint muscles of the thigh. As has been said, extension of the knee acts through the tendon action of the two-joint semimembranosus, gracilis magnus, and semitendinosus on the median and under surface of the thigh, to cause extension of the hip. So the gastrocnemius and tibialis anticus longus, muscles of the lower leg, can extend the hip. Moreover, since extension of the hip causes through the triceps extension of the knee, the energy of contraction of the gastrocnemius, for example, would be transmitted all the way around, through ankle, knee, hip, and knee to the muscle where it was developed. The same is true of any of these muscles; the energy is imparted to the bones through the end having the better leverage, and is transmitted thence through all the joints back to the point of origin. The energy of the two joint muscles of the hind leg of the frog is transmitted as by an endless chain , having the form of a figure eight with the crossing at the knee, and the effect progresses in the direction of the better leverage. H. E. Hering,^ in his study of the voluntary movements of the hand and fore-arm, observed that by flexion or extension of the fingers there is a move- ment of the middle hand in the opposite direction, caused by the simultaneous contraction of the hand extensors or flexors, as the case may be. Hering^ gives the rule, that, " In general, one can say that a muscle which draws freely ' Zeitschr. f. Heilkunde, 1895, XVI, S. 129. ' Arch. f. d. Ges. Physiol., 1897, Bd. LXV, S. 631. WARREN PLIMPTON LOMBARD. 291 across two joints will ahvays move the bone lying between these two joints, in the same sense as that on which it is inserted and on which it acts. " As Duchehene pointed out in the case of the hand, so with the foot, to get the strongest flexion or extension, the simultaneous movement of the lower leg in the same direction must be prevented, or better yet, the lower leg must be caused to make a movement of the opposite direction by the simultaneous action of the muscles. Hering calls these muscles pseudo-antagonists. "While the antagonistic muscles are inserted on the same bones, which they are able to move in an opposite direction, tlie pseudo-antagonists are inserted on two different bones. The pseudo-antagonists are antagonists in so far as they tend to move the bone in the opposite direction; but while the one muscle, namely, that inserted on this bone, rotates the same directly, the other rotates the same bone indirectly, in that, pulhng across the bone it moves it only by means of a neighboring bone to which it is inserted. Thus, while the flexors and extensors of the hand are antagonists, the flexors of the hand and exten- sors of the fingers are pseudo-antagonists. " Hering does not state in so many words in what light we shall regard two-joint muscles, like the gastroc- nemius and the tibialis anticus longus, which lie on opposite sides of a bone, and which, when acting alone, move each of the two bones adjacent to this, in opposite directions; but to judge from the sentence last quoted, they are to be- considered antagonist. I have shown, however, that these muscles, if contract- ing together, would help each other to move the bones to which they are attached, in the same direction, because they are capable each of employing the tendon action of the other. Hering' writes : " The view prevails that antagonists are simultaneously innervated when a movement results in the direction of one of the two antagonists. This, however, cannot be proved." "Antago- nists are synergetic only by the fixation of a bone; pseudo-antagonists are, also, synergetic by the movement of a bone. " It seems to me that in the light of my observations Hering's statement will have to be modified. I admit that, as yet, there is no direct proof that the tibialis anticus longus, for example, contracts by the spring movement synchronously with the gastrocnemius, but the value of a co-operation on the part of these muscles is so evident, that the probabilities are decidedly in favor of the supposition. There is no more reason for doubting the action of the tibialis anticus longus by the spring, than that the triceps, which appears to be the chief extensor of the knee and is also an opponent of the gastrocnemius, fails to contract by the spring. Apparently muscles which may be regarded as antagonistic, because of the opposite effects which they produce when acting under certain conditions, may act together; and further, that when acting together, they may so alter the effects which either would have produced when acting alone, that their ordinary antagonistic action may be temporarily 'Loc. cit., S. 632. 292 THE HIND LEG OF THE FROG. converted into one which is mutually helpful. In short, muscles Uke the tibiahs anticus longus and gastrocnemius, which, on account of the opposite effects which they ordinarily produce, are to be regarded as antagonists, may, under certain conditions, when acting together, become temporarily co-operators. The Le-\-erage op Two-Joint Muscles. It is evident that the effect of the contraction of a muscle depends largely on the leverage which it exerts on the bones to which it is attached. The leverage of each of the one-joint muscles of the leg changes with each new posi- tion of the joint which it controls, and the leverage of each of the ends of each of the two-Joint muscles changes with each position of each of the joints on which it acts. As this is the case, a knowledge of the leverage of each of the muscles of the leg, by every position possible to it, is indispensable to a com- plete understanding of the functional activity of these muscles. At present I am engaged in an attempt to determine the flexor and extensor leverage of each of the one-joint muscles of the hind leg of the frog at every ten degrees, at 10°, 20°, 30°, etc., as the joints are moved from extreme flexion to extreme tension. This work is very laborious, and is only partially completed, but the resiilts which have been arrived at are very satisfactory, and a preliminary set of curves have been obtained, showing the change in leverage of all the one-joint muscles of the ankle and knee, and of each end of each of the larger two-joint muscles of the hind leg. The method of measuring the leverage can be best described by taking a special case, e. g., the measurement of the leverage of the gastrocnemius on the ankle joint (See Fig. 5; Plate II). All the muscles going to the ankle, except the gastrocnemius, were cut away; the crus was cut in halves, and the lower half was fastened in a clamp, so that the bone and the foot were horizontal ; and the toes were cut off. The gastrocnemius itself was divided, and a strong, flexible silk thread was tied to the tendo Achillis. This thread was led over a vertical, easily moving pulley, and weighted with 50 grams. This weight was arbitrarily chosen to represent the contraction force of the muscle. The leverage was then measured by fastening one end of a thread to the end of the tarsus, 2 cm. from the ankle joint, and connecting the other end of the thread to a spring-balance devised to measure slight changes in tension and mounted on a movable iron standard. This balance was constructed by Mr. Miller, instrument-maker, of the University of Michigan. The construction of the spring-balance is shown in Fig. 6. It consists of a spiral watch-spring, a, fastened by its outer end to a pin driven into a brass plate, h, and by its central end to a steel axis which supports the hard rubber disk, c, which carries the scale. The steel axis has a fine point at each end, and the inner point is pivoted in the brass plate, and the outer point in a screw which passes WARREN PLIMPTON LOMBARD. 293 through a strip of brass, d, wUch is fastened to the top of the base plate, and extends forward a short distance and then downward in front of the hard rubber disk. The circumferenee of the dislc is grooved to receive a thread which is attached to the preparation. The strip d is prolonged as a pointer, and enables the number of grams, representing the leverage of the muscle, to be read off on the scale. This balance gives accurate readings to two- tenths of a gram. When the leverage was to be measured, the standard carrying the spring- balance was placed in such a position that the thread to the spring-balance B ^ Fig. 6. — Spring-balance which measures the leverage of the muscle. B, side view. A, front view; pulled at a right angle to the long axis of the tarsus, and then the standard carrying the balance was gradually moved away from the tarsus, until the strain exerted by the weight through the tendon on the tarsus was exactly balanced by the resistance of the spring. To facilitate the establishment of the angles to be measured and to insure that the thread leading off from the tarsus should always form a right angle with its long axis, pieces of vnre were bent to give each of the angles from 10° to 170°, and the wire at the end of one leg of each angle was bent to form a right angle. One of these wires was held directly above the preparation, and the position of the joint and of the thread, with respect to the balance, was adjusted to correspond. The amount of the tension produced by the muscle was read off on the scale of the spring- balance, this having been previously graduated to show, in grams, the tension put upon it. The strain exerted by the muscle was thus ascertained for every angle of ten degrees, 0°, 10°, 20°, 30°, etc., from complete flexion to complete extension, and the results plotted in the form of a curve, in which distances on the abscissa represented degrees of extension, and the distances on the ordinates represented in grams the strain exerted by the muscle at a point on the tarsus 294 THE HIND LEG OF THE FROG. 2 cm. from the joint. In measuring the leverage of the two-joint muscles acting on the hip, the femur was fastened horizontally and the pelvis sus- pended at an angle of about 30° to the earth. The amount of adduction and abduction of the femur has a marked effect on the leverage, a question which will be referred to at another time. The same method was followed in the measurement of all the muscles, in each case the pull of the muscle being assumed to be 50 grams. This assump- tion is, of course, incorrect, but makes it possible to compare the leverage of the different muscles, and the relative leverage of the two extremities of the two-joint muscles, for each of the positions possible to the joints. Errors are bound to creep into such measurements, and one ought to have enough measurements to be able to offer average curves. This failing, it would seem best to give for comparison only curves of like date. As a matter of fact, the time required for making the measurements is so long that it has not been possible to obtain a complete set from any one frog. Inasmuch as the curves which have been obtained from the muscles of different frogs have been found to resemble each other very closely, and as the number is not large {2 to 5 for each form of preparation), 1 have decided to present at this time only those curves which appear the most reliable and most characteristic. These curves (See Plates I and II) show the greatest differences. No two muscles give exactly the same curves, and the curves obtained from the two ends of the two-joint muscles differ greatly. Moreover, the curves show that we must be cautious of how we speak of the flexors and extensors of a limb until we have carefully studied the action of the muscles. Not only can most of the two-joint muscles produce opposite effects (flexion and extension) by their two extremities, the intensity of these effects changing with different positions of the joints, but they may produce opposite effects by the same extremity, at some position of the joint a reversal of the action of the muscle occurring, so that it may change from being a flexor to being an extensor. This reversal of action has been seen also in the case of one of the one-joint muscles. Some examples of the inconsistency of the action of muscles are the following : Ilio-fibularis ( flexes hip, 0°-70° extension, (flexor of knee) \ extends hip, 80°-170° extension. Semitendinosus ( flexes hip, 0°-40° extension. (flexor of knee) '( extends hip, 50°-170° extension. Glutteus magnus C extends hip, 0°-20° extension, (extends knee) ( flexes hip, 30°-170° extension. Peroneus ( extends knee, 30° over-flexion to 10° over flexion, (flexes ankle a short distance) . - flexes slightly knee, 0°-40° extension. ( extends knee, 50°-170° extension. Gracilis magnus ( flexes knee, 40° over-flexion to 110° extension. (extends hip) - extends knee, 120°-140° extension. ( flexes knee, 150°-170° extension. WARREN PLIMPTON L0M15ARD. 295 Semimembranosus C flexes knoo, 40° over-flrxiou to 90° extension. (extends hip) ^ extends knee, 100°-130° extension. ( flexes knee, 140°-170° extension. Tibialis posticus ^' extends anlde, 10°-60° extension, (one-joint muscle, lower leg) . . ^ no action, 70°-110° extension. ( flexes ankle, 120°-170° extension. The Leap. The only attempt to explain the g-reat force with which the body of the frog is propelled through the air when he leaps, that the writer has come across, is that of Gad.^ " If the leg is brought into the position ready for the spring, and the muscles named [the ilio-psoas (the iliacus internus), tibialis anticus (the longus is referred to) and peroneus] remain strongly contracted, the innerva- tion of all the other muscles of the leg can increase very markedly without its coming to a spring movement. If the extension of the foot is hindered more strongly (by the tibialis and peroneus) than the flexion of the knee, the gastrocnemius will act as knee flexor, as will also the two-joint muscles on the inner side of the thigh (semimembranosus, rectus internus, [the gracilis magnus,] semitendinosus, biceps [ilio-fibularis]), so long as their action on the hip joint is held in equilibrium by the ilio-psoas and triceps. As soon as the ilio-psoas relaxes, the muscles on the inner side of the thigh overweigh as hip extensors, wherefore they, in their knee-flexing action, will be overcome by the triceps the more easily, if their thus far ' Cumpan, ' the gastrocnemius, as a result of the simultaneous relaxation of the foot flexors, can transmit its principal effect from the knee to the ankle. The change from flexion to exten- sion is further thereby favored, that a part of the thigh muscles, as already stated (vastus externus and the adductors), from a given position of the thigh on, from hip flexors become hip extensors. " "The process by the ordinary locomotion of the frog consists, therefore, therein, that first, when the legs are extended, the position ready for the spring is produced through an out-weighing or separate action of the ilio-psoas and the foot flexors, that then the innervation of all the leg muscles rises uni- formly, until a condition of considerable tension is reached, and that, finally, the giving out of the tension which had developed in flexion, to produce the extension of the spring, is caused by the sudden relaxation of the ilio-psoas and the foot flexors. " There are several strong objections to be made to this theory. First, there is no proof that the one-joint flexors contract during the development of the early part of the spring movement; second, the muscles which are supposed to inhibit the extension of the leg are too feeble or have too slight a lev- ' After experiments in which Dr. O. Wegele, Hirsch, and Fuhr had taken part; " Einiges uber Centren und Leitungsbahnen im Riiokenmark des Frosohes, " — Verhandl. d. physik.-med. Ges. in Wiirzburg, 1884, N. F., Bd. XVIII, S. 173. 296 THE HIND LEG OF THE FROG. erage to do this. When the leg is in the sitting position, the two-joint tibiahs anticus longus and peroneus have a good extensor leverage on the knee, and the tibialis anticus longus has little flexor leverage, and the peroneus a shght extensor leverage on the ankle. Moreover, the gastrocnemius has no flexor leverage on the knee in the squatting position. Experiment shows that even when all the flexor muscles are acting, the extensors are capable of extending the leg, as is seen when all the nerves of the pelvic plexus are simultaneously excited. Third, even if, as is quite possible, the flexor muscles contract at the beginning of the spring movement, there is no proof that they suddenly relax and so release the tensed extensor muscles. Fourth, the theory apparently includes the idea that for a muscle to act well at one end, its action at the other end must be hindered in some way, and that, if the checking force be removed, the effect of the contraction will be suddenly transferred from the end formerly acting to the other. Jt is hard to see how this could take place. The tension developed in a muscle must be equally felt by both of the extremities of the muscle, and in the case of a two-joint muscle, when the conditions of leverage are suitable, movement must result at both ends, unless the parts to be moved offer more resistance at one end than the other. Relation of the Leverage of Two-Joint Muscles to the Action of these Muscles by the Leap. One would expect to find that the leverage of both the one-joint and the two-joint muscles would be more favorable for extension than flexion when the leg was in the squatting position. This is, however, not universally the case. The following table gives roughly in grams the leverage of the two-joint muscles when the leg is in the squatting position : Muscle. Hip. Knee. Ankle. Gastrocnemius Flexor — Extensor — 3 — 3 — 7 Flexor — 8 —10 —15 — 6 Extensor — 7 Flexor —3 —0 Flexor —0 Extensor — 2 " —3 " —0 Flexor —2 Gracilis magnus Spmitendinosus In the squatting position the gastrocnemius has a strong extension leverage on the ankle and no flexor leverage on the knee, so that the whole effect of the contraction would be expended in extending the ankle. The tibialis anticus has a comparatively small, but about equal leverage on the anlde and knee, a condition favorable to the use of the tendon action of this muscle by the gastrocnemius to extend the knee. The triceps closely resembles the gastroc- nemius; just as the latter has a strong extensor leverage on the ankle and no WARREN PLIMPTON LOMBARD. 297 flexion leverage on the knee, so the former has a strong' extension leverage on the knee and almost no flexor leverage on the hip. All this is favorable to extension. On the other hand, the strong group of muscles on the median and under side of tlie thigh, the semimembranosus, gracilis magnus, semitendinosus, and ilio-fibularis, has a stronger flexor lever- age on the knee than extensor leverage on the hip. This would seem an unfa- vorable condition to a sudden and powerful extension of the leg. It must be remembered, however, that if the knee were extended before or at the same time with the hip and anlde, the leg would be thrown out laterally instead of backward, and the effect of the stroke would be lost. To judge from the leverage curves, it would appear that the ankle begins to extend first, the hip next, and the Imee last. They also suggest that the resistance of the flex- ors of the knee at the beginning of the stroke has a further effect to put the extensors of the knee under tension, and to bring into full play the tendon action of these muscles, so that when they overcome the effect of the flexors of the knee, there would be a very sudden and powerful extension of the joint. The energy liberated by a contracting muscle is transmitted in both direc- tions throughout its length equally well. The fact that the leverage is better at one end of a two-joint muscle would not interfere with its action at the other end, and the fact that the gracilis magnus and semimembranosus and semitendinosus have, in the squatting position, a more powerful flexor lever- age on the knee than extensor leverage at the hip would not prevent these muscles from acting at the hip. In the squatting position the knee is over- flexed, and the strong flexor leverage of these muscles would tend to keep it flexed. The muscles are attached close to or in the knee joint and would practicaU}' act, as far as the hip is concerned, as one-joint extensors of the hip. The triceps does not have much flexor leverage on the hip in the squatting position, and would not greatly oppose the extensors of the hip. As the thigh was extended, the triceps would be put under more and more ten- sion, and by its tendon action would transmit the strain brought upon it to the extensor side of the knee. At the same time, the tibialis anticus longus through its tendon action would be transmitting to the extensor side of the knee the strain brought upon it by the extension of the ankle by the gas- trocnemius. The effect would be extension of the hip and ankle with ever- increasing tension on the extensor side of the knee. There can be little doubt that sooner or later the flexor action of the mus- cles on the median and under side of the thigh would be overcome by the rapidly increasing tension being brought to bear on the extensor side of the knee, by the contraction of the triceps, tibialis anticus longus, peroneus, and extensor brevis, and the strain transmitted by the tendon action of the tibia- lis anticus longus from the gastrocnemius, and by the tendon action of the triceps from those muscles themselves which lie on the under and median 298 THE HIND LEG OF THE FROG. side of the thigh. Of course, if this group of muscles on the inner and under side of the thigh continued, as the spring movement progressed, to exert a strong flexor effect on the Imee and only a feeble extensor effect on the hip, they would impede rather than aid the leap. This is not the case, however. The leverage of all these muscles changes as the leg is extended. In the case of the powerful semimembranosus and gracilis magnus muscles, the flexor effect on the knee rapidly lessens as it is extended, and at about 100° extension there is an actual change of the leverage on the knee from flexion to exten- sion. The leverage on the hip likewise undergoes a quick alteration ; the lever- age in favor of extension increases with great rapidity up to about 110° exten- sion, when it reaches its maximum, and is very powerful, the semimembranosus having an extensor leverage on the hip represented by 17 grams, the gracilis magnus 17 grams, the semitendinosus 15 grams. The ilio-flbularis reversing its leverage at about 70° extension, from flexion to extension, has at an increas- ing extension leverage to about 150°, where it is 9 grams. While the powerful group of muscles on the median surface of the thigh is gaining in extension power over the hip, the triceps group on the lateral surface is gaining in flexor leverage, but the gain is by no means as great, and probably acts not so much to impede the extension of the hip as to increase the tendon action of the muscle, and enable the muscles on the median side of the thigh, which extend the hip, to work through the triceps to help extend the knee. Concerning the one-joint muscles, it may be remarked that the extensor leverage of the tarsalis posticus and tibialis posticus, in the squatting position is good, but decreases as the ankle extends until 60°, after which, in the case of the former, it is small but constant until near the end of the extension movement, while in the case of the latter it is nul at 60°, and for some distance beyond this point. The flexor leverage of the tibialis anticus brevis and the tarsalis anticus on the ankle is weak when the ankle is flexed, but increases as it extends until 60°, when it is quite strong. From this point on, it remains quite constant, until near the end of the extension movement. The extensor leverage of the extensor brevis on the knee, though not very strong, and also undergoing fluctuations, exists in all positions of the leg. It has been said that the energy of the two-joint muscles of the hind leg of the frog is transmitted as by an endless chain having the form of a figure eight, with the crossing at the knee, and the effect progresses in the direction of the better leverage. The leverage curves would seem to show that the condition thus formulated does not exist by the spring until the muscles which extend the knee have been put under high tension and have overcome the resistance of the thigh WARREN PLIMPTON LOMBARD. 299 muscles which flex the knee. As the knco extends, the flexor leverage on the knee of the muscles on the median and under side of the thigh lessens, while at the same time, extension going on in the hip, the extensor leverage of these muscles on the hip increases. Wlren extension has progressed to a certain point, therefore, all of the two-joint muscles will have a good extensor leverage, and the direction of the transmission of the energy will be from gastrocnemius, on the median side of the leg, around the ankle to the tibialis anticus on the lateral side of the lower leg, from this around the knee to the group of muscles on the median side of the thigh, thence around the hip to the triceps group on the lateral side of the thigh, and so back around the knee to the gastrocnemius. Another point of interest revealed by the leverage curves is, that towards the end of the extension, the ends of the two-joint muscles which act as exten- sors begin to lose in extensor le^'erage, while those which act as flexors begin to gain in flexor leverage. Likewise, the energy of the one-joint flexors in- creases, and that of the one-joint extensors decreases. The extension effect of the semimembranosus, gracilis magnus, semi- tendinosus, and ilio-fibularis on the hip begins to lessen beyond 110° to 130° extension, while the flexor effect of the triceps is maintained. The extensor effect of the gastrocnemius on the anlde begins to lessen at 120°. On the other hand, the flexor effect of the tibialis anticus longus and the peroneus begins to increase at this point. A similar change is to be observed in the case of most of the one-joint muscles of the ankle. The extensor leverage of the tarsalis posticus lessens at 120°, and the leverage on the tibialis posticus, which is in favor of extension when the ankle is flexed, undergoes a reversal and becomes favorable to flex- ion at about 120°. The flexor leverage of the tibialis anticus brevis appears to be an exception, its flexor leverage lessening from 50° extension on. The extensor leverage of the triceps, tibialis anticus longus, peroneus, and extensor brevis on the knee lessens slightly towards the end of the exten- sion movement, while the leverage of the semimembranosus and the gracilis magnus on the knee undergoes, at 130° extension, a second reversal, these muscles again becoming flexors. The flexor leverage of the ilio-fibularis, which. has been lessening to 110°, here begins to increase; that of the semi- tendinosus, though less than when the knee is flexed, remains high; finally, the flexor leverage of the gastrocnemius increases as the extension of the leg progresses, being strong at the close. In short, towards the end of the extension movement there is a decrease of extensor leverage and an increase of flexor leverage. It is scarcely neces- sary to comment on the effect this alteration of leverage must have to protect the joints, and to favor a rapid recovery of the leg to the flexed position. 300 THE HIND LEG OF THE FROG. Summary. In this paper only the flexing and extending action of the two-joint muscles of the hind leg of the frog has been considered. Any force, whether acting upon the body from without or developed within the body by muscular contraction, which acts to extend or to flex either hip, knee, or ankle, because of the tendon-like action of the two-joint muscles, tends to cause a like movement of the two other joints. The one-joint exten- sors of the hip help in the spring movement to extend the knee and the ankle, and one-joint extensors of the knee and of the ankle also help to extend the whole leg. The resting length and the extensibility of the two-joint muscles is such that it is possible for quite extensive independent movements of any of the joints to occur when the other joints are fixed by the action of one-joint muscles. Nearly every one of the two-joint muscles, when acting alone, will flex one and extend the other of the joints which it crosses. If its action on one of these joints is prevented by any external force or by the contraction of one- joint muscles, the whole effect of its contraction will appear in a movement of the other joint. The fact that a two-joint muscle can make use of the tendon action of an- other two-joint muscle on the opposite side of the leg accounts for the para- dox that a two-joint muscle, when in a position to have a stronger extensor than flexor leverage, may extend a joint of which it is a flexor. Two-joint muscles can make use of the tendon action of other two-joint muscles to move joints over which they do not pass. When all the largest two-joint muscles of the hind leg of the frog are con- tracting simultaneously, the energy which they develop may be transmitted, as by an endless chain, having the form of a flgure eight with the crossing at the knee, and the effect progresses in the direction of the better leverage Thus, the total effect produced by a given two-joint muscle depends not only on what it individually might accomplish, but on the energy imparted to it from distant muscles, through the tendon action of two-joint niuscles. In the leap, the two-joint muscles, by their tension action, transfer the energy hberated by each of the muscles, which have a leverage in favor of extension, from the joint upon which it acts directly, to all the rest of the joints, thus enabling all the extending muscles to co-operate in the production of the extension of the whole leg. As the leg of the frog is gradually extended, the curve of leverage of the one- joint muscles, and of each end of each of the two-joint muscles, continually changes. The change may be so great that there is a reversal of the action of a muscle, its effect on a joint changing from flexion to extension and in some cases back again. Description of Plates I and II. Leverage of Muscles. — ^The ciirves given in these plates show the leverage of the one- joint muscles and of each end of each of the large two-joint muscles of the hind leg of the frog. The ordinates represent the leverage stated in grams, and the abscissa distances, the degrees of extension of the joint. The curve is drawn above the zero line when the movement is flexion and below this line when it is extension. Fig. 5. Method of Measuring Leverage. — -A diagram of the , position of the frog preparation, the spring-balance, the weight which represents the pull of the muscle, etc. o, spring balance; b, gastrocnemius preparation; c, clamp for bone; d, pulley; and e, weight. For construction of balance and method of experiment see pages 292 and 293. LOMBARD. I'l.ATE I. Leverage oii Hip. Leverage on Kiice. o' 20° 40° BO^ SO"" lOO'l 20' 40 160 . 180 20° o" 20' 40° 60" 80" loo" 120" 140'^ 160° 180° \r iiai WIG. V^'-'T' — .Qtiiinmes 1 1 1 1 1 1 1 - - .>' — hU-tenil /' ' \ J ! TrU'cpa n I>B ^» 1 _,._ Trice > 1 1 1 1 6— -8— r .'- — •— • —•- 1 I Kilend / -> ^•' / / ' i ■' ^ 1 — - A- 1 ■- - -- ; ; 1 1 • i^xirnd i ' 1 i . . 12 \ 1 ^ 4 ' r\ Gral ills^ Ia$ri lis 10 V ^\ Ora Bills Uagi US 1 J i\r 8 ^•. /n — ^ — :n / v 6 \, / :: : i\i ^ ! 4 s \ / 2 k / 1 ! \l 1 1 / I'-lex W( nj ft :x \ / / ^ 2 i--' V ./ 1 ^^1 \ ■ 1 Exl f.nd 10 • '- j ^ Sem: nicni i)ran »sns 8 A Sem nicm [)rau )SUB \ 6 \ \- / \ \ \ 4 '•v ^••« L / • \ : J 2 k / \ I'lex ^ ^ keiKi /V Flex \ / 1 • /^ • 2 > ■■ • •m ,-••■ ^, / /. 1 ^W — — \ h^-^ /\ 1 '. 1 1 1C ' \ 1 JIZctN F,x end —••^ ,■-. ^'^ K- -i»^ — — 1 \ V \ I .-•v '-•^■^ • • '■'1 — d— \ \ Sci litcn lino us Sci itcn lino u» 8- ■ V f' — W — \ / /— ' — _,..- -H > ,/ -0- Jica — Mp — N T — 1 A 2p ' / f7B.T ^^N F, hmd — 0- 1 \ i \ 2- 7 — N \ Tlin fihii ilTiX J \ Ili< fibii laris 4- / -V- \ f ~'\ t-.. 6- H • -. J / . '^•, ^., . '•■ ^•* 8- 1 -Lo- ^to 1__ LOMBARD. Platk 11. 2C o 0' 2C Leverage oii Knoe. ° 40° 60° 80° 100° 120' 140'- 160" 180° 20° 40° Leverage on Ankle 60° 80° 100° 120° 140° 160° 180° OVQ.) HIC5 ^- _.-' ^-t-N, Grammes itenk Ga^ tro( noiiil US \^'' B -4 6 •■" -.'' - ClUHt OCIH IllIlK / - ^'■ -- -fVc.r — - / — .•-' /! — •-•' -V «•— ' ^•*-i -." -.' " 1 Ti biall i Antlcns : jOng lis Tib alls lutlcus L u|!;u / fe end -.-i»«^ / 1 1 -'.-. ^- ^•' \. ,-.- ■».- / / / • — ^ /■ S_ -•v —%- / • Flex ■< k Scfeni Flex ^jtei Id irone US A Per )I1CU / /' \- N / ,.^ Flex / \ 'V,. / —2- ■' — , -.- • 1 ' Jlic( nd ^n " Ext nsor Brc( is 1 : T bialis An icus Brovls / .: y'^. . <" '~._ ~-^^. • / y -^ -^•—1 — -.-. / - • - 'N.,.^ ^•■* / / 1 j 1 1 1 J . / 1 hJ.Y ': Flex / -N 1 1 X Ta salis Aut fus c K ^ ^-^ .A s. 1 • \ ,0 "^nX '•% 1 \ ^ --a J k • / • ['*"!- f ---C / Flex - A y^y 1 i • ! \v ^^^y A Tib alia Posti L'lIS / r \^ ^ EtI find .^1^ ,•' Flex /i -- /-<•' / / V V • i ,-^t^ ,1^ Tar salis Post CU8 \ \ -s 1 u--' * J > ^.- — . ^•^ -._ 1^^ Fle.x •n 1 ' 1 1 WARREN PLIMPTON LOMBARD. 301 Although the tendon action of the two-joint muscles, together with the lev- erage of the two extremities of these muscles by different positions of the joints which they cross, go far toward explaining the mechanics of the leap, these are, of course, not the only factors of importance that enter into the production of this remarkable phenomenon. The rotation effect exerted by adjacent members of a hmb on each other, gravity and inertia, the relative strength and the shape of the different one-joint and two-joint muscles, their tension by different positions of tlie limb, their elasticity, the rapidity with which they contract and relax, and, above all, the order, strength, and duration of the nervous impulses reaching them from the spinal cord, must all be ascertained before we shall have a clear conception of the process. ON THE ACTION OF LITHIUM. CLARENCE A. GOOD, M.D., Superintendent of the Parker Memorial Hospital, Columbia, Missouri. {From the Pharmacological Laboratory of the University of Michigan.) Since the discovery of lithium by Arfvedson, in 1817, this substance has been found in small quantities in various minerals, in the sea water, and in the ash of plants, blood, milk, and the lung tissue of stone-cutters. Lithium belongs to the alkaline metals and has the extremely low atomic weight of seven. Bunsen and Kirchhoff' early studied the spectrum of lithium, show- ing that it colored the flame a purple red and showed a purple line near Frauen- hofer's lines C and D. They could demonstrate lithium carbonate in a dilu- tion of .000 000 000 9 g. per cc. Sir Alfred Garrod^ seems to have been the first to employ lithium to any extent in the treatment of disease, having employed it in the year 1860 for the cure of gout and rheumatism, although Lipowitz,^ in 1841, demonstrated the marked affinity of lithium salts for uric acid, and Dr. Alex. Ure,^ in 1843, wrote of experiments in which a urinary calculus, composed of alternate layers of uric acid and oxalate of lime, when placed in a dilute solution of hthium carbonate and maintained at blood heat for five hours, lost five grains in weight. Dr. Ure^ later advised the injection of solutions of lithium into the bladder for the cure of stone, but the first case in which the lithium solution was used as an injection proved the method to be unsuccessful, as the case came to operation and died. Dr. Garrod based his use of lithium salts in the treatment of gout and rheu- matism on the folloTOng experiments: A metacarpal bone, having the pha- langeal extremity completely infiltrated with gouty deposits, was placed in a dilute solution of lithixim carbonate and in two or three days no deposit could be seen, and the cartilage was apparently restored to its normal condition. Similar experiments made with dilute solutions of sodium and potassium carbonates showed no solution of the urates, and hence he concluded that by giving lithium carbonate one could dissolve the deposits in and around the joints and prevent their future occurrence. The fallacy of this treatment lies in the fact that lithium is a solvent for urates only when in concentrated solution, as is well shown by the experiments of Krumhoff," who placed 0.5 g. of pure uric acid in 200 cc. of distilled water, which was kept at a uniform tem- perature of 37.5 to 38.5 C. for six hours. He found, in several experiments, 302 CLARENCE A. GOOD. 303 that the average amount of uric acid dissolved by 200 cc. of distilled water was 0.0217 g. Similar experiments made with 200 cc. of distilled water, to which 0.5 g. of lithium chloride had been added, showed that the lithium solution would dissolve, on an average, only 0.0118 g. of uric acitl, as against 0.0217 g. dissolved by pure distilled water. That lithium salts are solvents for urates and uric acid only when in con- centrated solution is not generally understood, and the majority of those who now administer lithium salts do so with the idea that thej^ are uric acid sol- vents in the dilute solutions in which they exist in the tissues. In the literature one finds various statements as to the amount of urates and uric acid eliminated after the administration of lithium salts. Haig,' Oliver,^ and Levy^ found the uric acid diminished after giving lithium salts. Considering, however, that no constant diet was used and that the methods of estimation of the uric acid were faulty, these results cannot be considered conclusive. Dr. Garrod also claimed to have seen marked benefit derived from the external use of solutions of lithium salts, even to the disappearance of small tophi, but Hiifner,^" in some interesting experiments, showed that no lithium was absorbed through the skin. Hiifner had one of his pupils place both feet in a warmed one per cent solution of lithium carbonate for half an hour, and collected the urine following the experiment, but could not demonstrate lithium in it, even after the quantity passed in twenty-four hours had been evaporated to a few cubic centimeters. On the other hand, when lithium was added to the urine, the method permitted the detection of as little as nine-millionths of a milligram. Lithium salts are rapidly absorbed from the alimentary canal, as is well shown by experiments performed by Dr. Bence Jones," who gave an animal three grains of lithium chloride on an empty stomach and detected lithium in the aqueous humor of the eye and in the cartilage of the hip joint fifteen minutes afterwards. He records giving 7 gr. to a parturient woman eight hours before delivery and finding lithium in the umbilical cord. He also records giving twenty grains to a patient three and one-half hours before an operation for cataract and finding lithium in the aqueous humor at opera- tion. He was able in this case to detect lithium in the secretions up to the seventh day. That lithium salts begin to be ehminated quickly, though the eUmination extends over a prolonged period, I have shown by demonstrating hthium in the saliva within eight minutes, and in the urine and feces within ten minutes after the injection of 1 g. of lithium chloride subcutaneously in a cat. I have, also, been able to detect lithium in the urine twenty-three days after the injections were stopped. The effects of Hthium salts on animals were studied as early as 1868 by Rabuteau," who concluded that lithium salts were harmless, being less poison- 304 0]V THE ACTION OF LITHIUM. ous than potassium salts. He noted, however, that they caused vomiting and purging. In 1873, James Blake concluded from an experimental study of lithium that the amount necessary to kill rabbits was 1 g. per kilo of body weight, and Hesse," in 1875, found that lithium chloride injected into a vein caused diastolic arrest of the heart, decreased the excitability of the nerve centers, lowered the temperature, and often caused diuresis. Cash and Brunton," experimenting on frogs, found that lithium depressed the upper part of the spinal cord and paralyzed the motor nerves to a greater or less extent. Krumhoff, in 1884, made a careful experimental study of the effects of lithium salts on animals, and found that when injected into the blood they caused a fall of blood pressure and depressed the heart's action, and if the dose was large enough, stopped the heart in diastole, although the dose necessary to do this was larger than the dose of potassium salts necessary to produce the same effects. He also found that lithium salts caused vomiting, diarrhoea, and a fatal gastroenteritis, the lesions found at autopsy being reddening and swelling of the mucous membranes of the stomach and bowel and occasionally small hemorrhages into the heart muscle. Potassium salts caused no such effects. Binet,^' in 1892, found that in mammals lithium caused pronounced feeble- ness, with nausea, diarrhoea, dyspnoea, and death, preceded by convulsions. He ascribed death to a direct central paralysis of respiration, although the heart was markedly depressed and finally arrested in diastole. To carefully study the effects of lithium salts on animals, and to determine where it was excreted, I made thirty experiments on cats and dogs, adminis- tering the drug subcutaneously and by the mouth. The salt used in my experi- ments was the chloride, because of its easy solubility, and because the carbo- nate, which is usually used therapeutically, is in part, at least, changed to the chloride in the stomach. A pure salt was used and was examined quantita- tively for lithium, as well as tritrated for chlorides, so that the dosage might be accurate. The method of quantitative estimation which I used was as follows: The urine, saliva, or feces to be examined was evaporated to dryness and carbonized. The residue was then extracted with a large amount of water rendered acid by hydrochloric acid, filtered and the filtrate evaporated to about 50 cc. To this milk of lime was added and thoroughly rubbed up till alkaline in reaction. This was then filtered and thoroughly washed with water till the washings failed to show lithium by the spectroscope. To this solution ammonium car- bonate in solution was added as long as a precipitate formed, filtered, and the filter washed with water till the spectroscope failed to show any lithium in the washings. This filtrate was now acidulated with hydrochloric acid and evapo- rated to dryness in a platinum dish, and heated to drive off the ammonia. The residue was dissolved in water, acidulated with hydrochloric acid and CLARENCE A. GOOD. 305 evaporated to dryness. This residue consisted of potassium, sodium, and lithium chlorides, and as the chloride of lithium is soluble in alcohol, while the chlorides of sodium and potassium are nearly insoluble, the dry salt was extracted with equal parts of absolute alcohol and ether. Comparatively small amounts were used and the undissolved portion, after being washed with equal parts of alcohol and ether, was dissolved in water and examined by the spectroscope. If this showed the presence of lithium the process was repeated. In this way I was able to separate most of the sodium and potassium chlorides from the lithium chloride. The alcohol-ether solution was now evaporated to drjmess, dissolved in a little water, and a one per cent solution of sodium phosphate added, 1.1 cc. for each 0.1 g. of lithium expected. This was now evaporated to dryness, whereby the soluble lithium chloride was changed to the insoluble hthium phosphate. To this a small amount of water was added, heated sHghtly, and an equal amount of strong ammonia solution added, and let stand twelve hours in the cold. As the lithium phosphate is practically insoluble in the strong ammonia solution, while the phosphates of sodium and potassium are freely soluble, the undissolved portion was lithium phosphate. This was now filtered through a Gooch filter, washed with a solution of strong ammonia, and this first wash-water again run through for lithium. The Gooch filter was now placed in the oven, thoroughly dried, and the weight of lithium phosphate determined, which was calculated back to the chloride. By this method the lithium phosphate always shows a strong sodium flame, and the results are probably high, yet by becoming accustomed to the method fairly accurate results can be obtained. Several experiments will be quoted in detail and a detailed account given of the pathological changes found. Experiment I. — ^A healthy cat, weighing 2,150 g., was given 1 g. of lithium chloride hypodermically. Within ten minutes profuse salivation came on, and twenty minutes after the injection the cat vomited. During the next four hours she vomited several times and had four or five diarrhcEal stools. At this time she would not stir from her squatting position and seemed much stiffened in the hind parts, apparently being unable to raise the legs. She would frequently retch without being able to vomit. On the following morn- ing, or about thirty-six hours after the injection, the cat was found dead. At autopsy no pathological changes were found except a slight congestion of the entire gastro-intestinal tract more marked in the large bowel. A careful microscopical examination of the various organs showed nothing abnormal except a congestion of the gastro-intestinal mucosa. Particular attention was paid to the examination of the cord and brain. Thin shces were removed and rapidly hardened in absolute alcohol and embedded in paraffin. Some sections were stained with methylene blue and some with hsematoxyUn and eosin. No pathological changes were found. 306 ON THE ACTION OF LITHIUM. Experiment II. — March 21. A healthy cat, weighing 1,650 g., was given 1 g. lithium chloride hypodermically at 10:40 a.m. At 10:45, movement of bowels, hard formed; at 10:48 vomited. The vomitus consisted of mucus and undigested food. An examination of this by the spectroscope showed the presence of lithium. At 11:05, marked salivation; at 11:10 vomited; at 1:30, movement of bowels; at 2:30 the cat had passed two or three diarrhceal stools. She sat humped up in the cage and would not stir. By 5 :00 p.m. she had had two more movements of bowels and had vomited once. March 22d, 9 a.m., cat very weak. Unsteady on feet. Drinks, but will not eat. Several diar- rhceal stools during night. Has passed some urine, but not a great deal. Five P.M., cat in about same condition, although weaker. March 23. — Cat much weaker. Unable to walk vwthout tottering. Hind parts seem most affected. Will not eat nor drink. No stools, nor vomiting. About normal amount of urine. March 24. — Cat died during night. Autopsy Findings. — March 24th, 11 a.m. Weight, 1,405 g. All the organs were normal in appearance, except the stomach .and intestines. The mucous membrane of stomach was reddened, and in two places showed hemorrhagic spots, size of small pea. Mucous membrane of small bowel not much inflamed, but covered with a thick, tenacious, bile-stained mucus. The mucous mem- brane of large bowel distinctly reddened and thickened and contained two small ecchymoses. All the vomited matter, the stools, and the contents of the stomach and bowel were added together and from them 0.086 g. lithium chloride was obtained. A microscopical examination was made of pieces of the heart muscle, liver, kidneys, stomach and intestines. The heart, liver and kidneys showed no appreciable changes. Sections of the stomach and bowel showed a marked congestion. The mucous membrane was covered with a thick coating of mucus and in places contained small hemorrhages. Experiment III. — November 11, 1901. A healthy cat, weighing 3,700 g., was given, hypodermically, at 1:35 p.m., 2 g. lithium chloride well diluted. At 1 :40, watery stool. At 2 :20, washed out stomach. At 2 :00, passed urine. At 2:10, vomited. At 2:40, three watery stools. At 3:10, washed out stom- ach. At 4:00, washed out stomach. At 5:00, cat was killed with chloroform. Autopsy. — Stomach contained mucus and water which was added to vomi- tus and stomach washings. No marked changes in stomach walls. Bowels empty and showed no marked change. Liver, kidney, lungs, and heart appar- ently normal. The vomitus and stomach washings were kept separate from the bowel contents and stools and the lithium in each estimated. The former contained 0.125 g. hthium chloride and the latter 0.034 g. Experiment IV.— November 25, 1901. A healthy cat, weighing 1,320 g., was given hypodermically at 1 :20 p.m., 0.5 g. of lithium chloride well diluted. At 1:30, vomited several times; 1:40, profuse watery stool. Up to 4:00 had CLARENCE A. GOOD. ' 307 vomited several times and had several watery stddls. November 26, 1 p.m. Cat seems quite normal. Has not vomited nor had movement of bowels during night. Has passed considerable urine. When tak(>n out" of cage seems stiff, and perhaps somewhat unsteady on her feet. »She drinks, but will not eat. November 27th, 1 p.m. Cat more feeble than yesterday. When taken from cage she is so stiff and weak that she can hardly walk. Hind parts seem most affected. No vomiting. No stools. Will not eat nor drink. November 28th, 1 p.m. Cat same, but weaker. November 29th, 1 p.m. Cat extremely weak. Unable to walk. Seems to be unable to raise her hind parts. No stools nor vomiting. November 30th. Cat died during night. Aviopsy. — Cat weighed 1,080 g. The lungs, heart, liver, and kidneys were normal. Mucous membrane of stomach reddened and covered with a thick, bile-stained mucus. Two small ecchymoses into the mucous membrane, near the pylorus. Small bowel somewhat congested and covered with a thick tenacious mucus, but otherwise normal. Large bowel deeply congested. Sev- eral areas of hemorrhages into mucous membrane which is greatly thickened. The vomitus and stools were examined separately; from the former was ob- tained 0.021 g. and from the latter 0.026 g. lithium chloride. Microscopical examinations of the various organs showed no evident changes in the lungs, heart, liver, and kidneys, while the changes in the stomach and bowel cor- responded to those under Experiment No. II. Experiment A'^. — A healthy cat, weighing 1,600 g., was given hypoderm- ically daily for six days, 0.2 g. lithium chloride. After the first injection she vomited, and had two or three diarrhoeal stools. This diarrhoea became profuse on the third day and the animal was very_ weak, being only able to walk with difficulty. The hind parts seemed stiff and awkward, and would frequently fall clumsily, first on one side and then on the other. This weakness increased until just before her death, when she was unable to raise her fore parts. Autopsy. — No lesions were found other than congestion of the large bowel with a few small hemorrhages into the mucosa. Sections of the brain and cord stained in methylene blue, and in haematoxylin and eosin failed to show any changes. Experiment VI. — A healthy cat, weighing 2,650 g. , was given 1 g. of lithium, hypodermically. This cat did not vomit, but was profusely salivated. The salivation was, undoubtedly, due to nausea, which was frequently observed in other experiments before vomiting began. All the saliva was collected in a clean dish and as it showed lithium by the spectroscope, the amount was estimated and 0.021 g. was obtained from it. The cat died of gastroenteritis in three days, and from the stools and stomach and bowel contents 0.110 g. of lithium chloride was obtained. These experiments have been quoted in detail because they show the 308 ON THE ACTION OF LITHIUM. essential symptoms and cause of death from lithium salts, in whatever form they be administered. It will be seen that soon after the administration of the lithium salt the, animal manifests nausea, vomiting, and diarrhoea, and dies sooner or later of emaciation and weakness caused by a gastroenteritis. The stiffness and inability to use the hind parts is also a result of the gastroenteritis, and is not due to any changes in the cord or brain; it is also seen in poisoning from the heavy metals, colchicum and other drugs that cause fatal gastro- enteritis. From these experiments it will be seen that lithium is partiahy excreted in the saliva, and into the stomach and bowel, though the quantity that can be reclaimed from these secretions is usually not great. To show the effects of the prolonged administration of small doses hypo- dermicaUy, the following two experiments may be given : Experiment VII. — December 3d. A healthy cat, weighing 3,170 g., received, hypodermically, each day for eighteen days, 0.0567 g. lithium chloride well diluted. She had diarrhoea on the second day, which was severe on the fourth day, when the stools were blood-streaked. The urine and stools were kept separate. The diarrhoea was severe and she occasionally vomited, and on the tenth day she had lost 500 g., or about one-sixth of her original weight. Her coat had lost its gloss and her health and strength were distinctly lessened. By the eighteenth day she was very weak, weighed 2,650 g., had a severe diarrhoea, and as it was quite evident that she would soon die, she was killed by chloroform, as I wished to examine the tissues in a perfectly fresh condition. Autopsy Findings. — ^The lungs, kidneys, liver, spleen and heart were appar- ently normal. The stomach did not show any marked changes. The upper part of the small intestine was quite normal, but the lower one-fifth, together with the large bowel, was deeply congested. The mucous membrane was thickened and showed numerous small and large hemorrhagic areas. There were two small superficial ulcers in the large bowel, which was more severely affected than the small bowel ; and I may add that this is uniformly the case when small doses are administered, whether given subcutaneously or by the mouth. The stools and urine for the first ten days were examined for lithium. During this time she had been given 0.567 g. lithium chloride. From the bowels 0.039 g. and from the urine 0.228 g. of lithium chloride was obtained. Experiment VIII. — In order to be able to collect more accurately the urine, a healthy female dog, weighing 8.2 kg., had the perineum cut so that she could be easily catheterized. After the wound had thoroughly healed she was given daily 0.062475 g. of lithium chloride hypodermically. She was not allowed to pass any urine, being catheterized frequently enough to prevent this. The urine and stools were kept separate. At the end of ten days her weight was 8.1 kg., she was lively, ate and drank well, and apparently had not suffered from the injections. From the ten days' urine 0.383 g. of lithium chlo- CLARENCE A. GOOD. 309 ride was obtained, and from the feces 0.074 g. of lithium chloride. She had received in this time 0.62475 g. lithivun chloride. The same dose was contiuiuHl eight days longer and at this time she began to show symptoms, such as loss of appetite and slight diarrhcea. On the eighteenth day she weighed 7.9 kg. Her coat was roughened, and she was undoubtedly in poorer condition than at the beginning of the experiment. The injections were then stopped till February 4th, when they were again taken up, increasing the dose to 0.2479 g. From the second day on she vomited occasionally, the vomitus usually consisting mostly of mucus. There was a slight diarrhoea. Her appetite was A'ery poor. These injections were continued for six days. At the end of this time her weight was 7.1 kg. From the six days' urine 0.589 g. and from the feces 0.014 g. Uthium chloride was obtained. During the six days she had been given in all 1.5 g. This was continued for five days, the injections being increased to 0.62475 g. daily. During this time she vomited several times a day and had a severe diarrhoea, the stools being often streaked with blood. She ate scarcely anything, and at the end of the experiment weighed 7 kg. She was very weak and exhausted, coat rough and lusterless. From the five days' urine 1.49 g., and from the feces 0.128 g. of lithium was obtained. During the five days she had received 3.1 g. The injections were now stopped but the vomiting and diarrhoea con- tinued unabated. The appetite was completely lost and she became weaker daily. On the twenty-fourth of February she was so weak she was hardly able to stand, and as it was evident she would soon die, she was killed by chloroform. At this time the weight was 6.8 kg. At the beginning of the experiment she had weighed 8.2 kg., showing thus a loss of 1.4 kg. Autopsy. — ^The heart showed a few small hemorrhages under the endocar- dium, otherwise normal. The walls of the stomach were greatly thickened and reddened, and contained several small hemorrhages. The mucous mem- brane of the whole bowel was thickened, though this change was most marked in the large bowel. The lower part of the large bowel contained a few small hemorrhages. Lungs, liver," and kidney showed no marked changes. The examination of the stomach and bowel showed congestion and thickening of the mucous membrane with a superficial layer of mucus, and one or two small superficial ulcers. One hundred grams of the muscle from the ham was burned and the ash dissolved in 30 cc. of distilled water. This gave a distinct lithium spectrum even when diluted to 90 cc. One hundred grams of normal dog's muscle, when burned and the ash dissolved in 30 cc, gave no lithium spectrum, even when concentrated to 5 cc. The ash from 100 g. of blood taken from this dog showed lithium, even when diluted to 130 cc. That from normal dog's blood gave only a faint spectrum when dissolved in 5 cc. water. It is evident from this that lithium 310 ON THE ACTION OF LITHIUM. is slowly excreted, and is stored up in the body. The hver did not show as much lithium as an equal weight of muscle. When hthium salts are given by the mouth they produce essentially the same symptoms and changes as when given hypodermically, though, of course, larger doses are required. Cats, receiving 0.5 g. to 2 g. by the stomach, often vomit two or three times, but otherwise are very little affected, though fre- quently some diarrhoea is noticed. On the contrary, when small doses are ad- ministered daily by the mouth, cats and dogs show first diarrhoea with some blood in stools, vomiting, loss of appetite and weight and finally die from gastroenteritis, as the following experiment well shows : Experiment IX. — A healthy cat, weighing 2,000 g., received daily for the first six days 0.062475 g. of lithium chloride by the stomach. No symptoms appeared at this time, though she had lost 100 g. in weight. The dose was now increased to 0.125 g. daily, which was continued for five days. During this time she developed a moderately severe diarrhoea, and at the end of the six days weighed 1,800 g. The dose was now increased to 0.188 g. daily, and was continued for fifteen days. During the most of this time she had a severe diarrhoea, the stools often containing blood, and during the latter part of the period she was vomiting constantly, and finally died of gastroenteritis and exhaustion, weighing 1,500 g. Autopsy. — There was found a marked inflammation of the whole gastro- intestinal tract. This was most severe in the large bowel, which contained many large and small ulcers and ecchymoses. There was also some enlarge- ment of the abdominal lymph glands, and two or three small abscesses in the liver, undoubtedly secondary infections from the bowel. In some experiments the animals, when given lithium in small doses, lost over one-third of their weight and flnally died of gastroenteritis. Sixty miUigrams of lithium chloride per kilogram daily always killed dogs and cats, sooner or later, from gastroenteritis; frequently, especially in dogs, this was much delayed by the animal's vomiting the lithium solution soon after taking it. The cause of this gastroenteritis is undoubtedly connected with the excretion of the metal through the bowel wall. This action on the bowel is not peculiar to lithium, but is caused by many other substances, among which may be mentioned mercury, arsenic, colchicum, emetine, and aloin. These substances are excreted by the bowel, and also by the kidney, and induce irrita- tion and inflammation in these organs. This may be because they collect in larger quantities in the excretory organs, or because they are here freed from some harmless combination in which they have circulated iii the tissues. At times the bowels are most affected, and at other times the kidneys suffer most. Lithium salts are slowly excreted by the kidneys and do not seem to cause any appreciable amount of renal irritation. On the contrary, the excretion through the bowel causes marked irritation and inflammation. CLARENCE A. (iOOD. 311 The amount of lithium salts obtained from the feces in my experiments was always greater in those cases accompanietl by marked vomiting and diar- rhcsa. There are very few cases mentioned in the literature of poisoning from the use of lithium, and it is not generally considered as inducing any delete- rious symptoms. Wood, in Iris Text-book of Therapeutics, says he has seen a 20-grain dose of lithium carbonate produce severe general prostration in a feeble adult female; and Hare, in his Practical Therapeutics, says: "It is worthj' of note that in some cases lithium will disorder the stomach and pro- duce vomiting. " In a note to the French edition of Garrod on gout, the statement is made that small doses are readily borne, but doses of 30 to 50 grains of the carbo- nate give rise after a few days to cardialgia and dyspepsia. Rabuteau" records similar results in his own person, and Althaus states that lithiated waters, if taken in large amounts, give rise to sickness and diarrhoea. Koli- pinski" reports two cases of marked tremor following the use of lithia tablets. This was also accompanied by marked general prostration and weakness. The condition disappeared in three or four days. Considering the marked effect of lithium salts on animals, it is surprising that symptoms on the part of the gastro-intestinal tract have not been noticed more frequently. It is not improbable that such symptoms have been noticed but have not been reported, or have been considered due to other causes. It might be stated here, however, that lithium is frequently given as the natural lithia waters, which, as shown by the in-\-estigations of Harrington,^^ Waller,^" and others, contain very small amounts of lithium, many containing none at all. Lithium salts are frequently credited with possessing some diuretic action. To determine whether this was correct I made several experiments on rabbits, as follows : Large rabbits were anEesthetized with urethane and paraldehyde, a canula was introduced into the jugular vein, and one into the bladder, so as to be able to inject the fluids into the blood and to collect the urine. After determining the normal secretion, warmed solutions of lithium chloride and sodium chloride were injected in succession into the vein and the urine collected and measured. In order to have the solutions contain an equal number of molecules, they were so prepared that their freezing points were the same. Solutions containing about 3 g. and 6. g. of lithium chloride in 100 cc. distilled water were used, and from 30 cc. to 50 cc. were injected at the rate of 15 cc. in five minutes. These experiments did not show that solutions of hthium chloride caused any greater diuresis than solutions of sodium chloride. The salt of lithium used in medicine is the carbonate. It is much less soluble than the chloride, but experiments on animals showed that it produced the same effects. The bromide of lithium has been used in the treatment of gout and epilepsy and other nervous diseases. Dr. Weir Mitchell,^" after 312 ON THE ACTION OF LITHIUM. using lithium bromide in several cases, concluded that it was as efficient as sodium or potassium bromide, and that its influence in insomnia was greater. Fontan^^ found lithium bromide to possess the same sedative action as potas- sium bromide, but he thought it less liable to cause untoward symptoms. Dr. Mitchell, however, saw skin rashes follow the use of lithium bromide. L6vy^^ found lithium bromide to possess marked sedative action and to be less depressing to the heart than potassium bromide. He could not find lithium bromide to be of any special value in gout. Since the atomic weight of lithium is so small, there is, of course, more bromide in Uthium bromide than in an equal weight of potassium bromide, and the action of the bromide ion completely overshadows the action of the lithium ion. Lithium bromide does not possess any advantages over the commonly used potassium bromide. Conclusions. Lithium is excreted in the saliva, into the stomach and bowel, and in the urine. The greater amount is excreted in the urine, though more appears in the stomach and bowel when nausea, vomiting, and diarrhoea have been pro- fuse. It can usually be demonstrated in the secretions within ten minutes after a hypodermic injection, though its excretion proceeds slowly, for I have found it twenty-three days after the injections were stopped. 2. Lithium salts given to animals hypodermically, or by the stomach, cause, sooner or later, fatal gastroenteritis. This gastroenteritis is probably connected with the excretion of the metal through the bowel wall. 3. Lithium salts do not possess any diuretic action that cannot be accounted for by their salt action. 4. Lithium carbonate given in 15-20 gr. doses and lithia tablets have been known to cause gastro-intestinal symptoms in man. 5. Dilute solutions of lithium salts are not solvents for uric acid or urates. BIBLIOGRAPHY. 1. Ann. d. Chem. u. Pharm., Bd. CXVIII, p. 353. 2. Gout and Rheumatic Gout by Garrod, 2d edition, pp. 419-428. 3. Ann. d. Chem. u. Pharm., Bd. XXXVIII, p. 348. 4. Pharmaceutical Journal, Aug., 1843. 5. Lancet, London, 11, 1860. 6. Wirkung des Lithium — Inaugural Dissertation. Gottingen, 1884. 7. Uric acid in the causation of disease, p. 53. 8. Lead poisoning, p. 107. 9. Gazette Medicale de Paris, Nov. 27, 1875. ,10. Ztschr. f. physiol. Chem., Bd. IV, p. 378-81. 11. " Lectures, " p. 16. Quoted in Phillip's Materia Medica, Vol. II, p. 228. CLARENCE A. GOOD. 313 12. fitudes exp6rimentals sur les efFets physiologiques des fluorures ot des compos(5s mStalliques en g6n6ral. Paris, Germer Bailliere, 1867. 13. Inaugural Dissertation, " Lithion. " GOttingen, 1876. 14. Phil. Trans. Roy. Soc, 1884, p. 222. 16. Revue Med. d. 1. Suisse Romande. T. VIII, 1892. 16. Quoted in Phillip's Materia Medica and Therapeutics, Vol. II, p. 228. 17. Maryland Med. Journ., Vol. XI, No. 1. 18. Boston Med. and Surg. Journ., Vol. CXXXV, p. 04-1, 1896. 19. Journ. Amer. Chem. Soc, Vol. XII, p. 214, 1890. 20. Amer. Journ. Med. Scienc, Vol. LX, p. 413. 21. Du Bromure de lithium, 4°, Nancy, 1875. 22. Essai sur I'action physiologique et th^rapeutique du bromure de lithium, 4°. Paris, 1874. ON THE SECEETION OF THE UEINE AND SALINE DIUEESIS. ARTHUR R. CUSHNY, A-M., M-D., Professor of Materia Medica and Therapeutics, University of Michigan. {From the Pharmacological Laboratory of the University of Michigan.) The blood-vessels carry to the kidney a dilute solution of salts and organic substances, while the ureter carries from it a concentrated solution of certain organic bodies and of inorganic salts. Unlike other glands, the kidney forms little or none of its secretion, its activity being evidenced only in the change in the character of the fluid, which it eliminates from that which it receives. It has accordingly been regarded as possessing a more simple function than such glands as those of the alimentary tract, and the hope has been entertained that its activity might prove to be explicable by the physical and chemical forces already known. This hope has hitherto proved delusive, but on each advance in knowledge of the physical forces, the question has to be reinvesti- gated. How far these explain the processes at work in the kidney, and how much must still be relegated to the unknown or vital activity? The recent progress in physical chemistrj^ has again brought the question into prominence, and a considerable number of laboratories have been occupied in the last few years in applying to this question more exact methods than were formerly at our disposal. The results have appeared in the papers of Starling,' Koranyi,^ Magnus, and Gottlieb,''' Sollmann,'' Spiro,'' Filehne,^ and his pupils, and Cushny.'' On approaching this subject, one is met by two fundamentally opposite views, each of which has numerous adherents — the theory of Bowman and Heidenhain and that suggested by Ludwig. The former holds that the glomerulus "secretes" the water and salts of the urine, the tubules, the urea and some other organic constituents. Ludwig originally held that the glomerulus is a filter which allows the passage of all the constituents of the blood, except the proteids, and that most of the water of this filtrate diffuses back into the blood through the walls of the tubules. He, therefore, required 1 Journ. of Phys., XXIV, p. 317. 2 Ztschr. f. klin. Med., XXXIII, p. 1; XXXIV, p. 1. 3 Arch. f. exp. Path. u. Pharm., XLIV, pp. 68, 396. XLV, pp. 210, 223, 248. 'Ibid., XL VI, p. 1. = Ibid., XL VI, p. 148. « Arch. f. d. ges. Physiol., XCI, p. 565. ' Joum. of Phys., XXVII, p. 429; XXVIII, p. 431. 314 ARTHUR R. CUSHNY. 315 nothing except the known physical forces — tlic pressure of the blood causing the filtration, and diffusion explaining the reabsorption in the tubules. Heiden- hain, on the other hand, neglecting the physical forces, regarded the whole secretion as a process inckieed by luiknown or " \-ital " forces. Neither of these explanations appears to be held in its entirety at present ; it is generally allowed that the physical forces have some influence on the secretion by those who favor Heidenlrain's position, while, on the other hand, Ludwig's followers have franldy conceded that diffusion is inadequate to explain the absorption in the tubules, and that some unknown force must induce a current from the lumen of the tubules towards the blood-vessels. The most acute differences between the two schools appear to be whether filtration is sufficient to explain the passage of fluid through the glomerulus, and whether the tubules are absorbing or secreting organs. In regard to the glomerular function then, there is to be determined whether the passage of the fluid is a mechanical process merely, and if the physical forces prove inadequate to explain the phenomena, a further problem remains — the disentanglement of that part which is physical from what maj^, for brevity, be termed vital. In the tubules the fimdamental question is still in dispute — ^whether the current passes towards the lumen from the blood-vessels or vice versa. When this is settled, the process has here again to be resolved into its physical and vital compo- nents. It is no longer sufficient to show that the renal secretion involves the activity of unknown forces, for this is generally admitted. But the question as to how far these forces are capable of being modified by physical condi- tions is hardly entered upon, and its solution will tend to delimit the territory occupied by the unknown activities, and to define the nature of these activities. One of the most promising ways of approaching this question is that offered by the saline diuretics, which not only cause diuresis, but also alter the physi- cal condition of the blood, and a large number of investigations have, therefore, been carried out of late years to determine whether this special form of diuresis is due to the change in the physical conditions, or to the increased activity of the unknown factors in secretion. My attention was drawn to the subject by observing that the chlorides of the urine were markedly increased in phlorid- zin glycosuria, as well as in various other forms of diuresis (Katsuyama). It seemed necessary to determine the permeability of the kidney by different salts, and I made a series of experiments to estimate not only the amount of urine in saline diuresis, but to determine whether two salts injected in equal quantity into the blood appeared in equivalent amounts in the urine. General Course of the Experiments. The secretion of the salts by the kidney has recently been investigated by Magnus and Gottheb, who compared the behavior of sodium chloride solu- tion injected intravenously in one series of animals with that of the sulphate 316 SECRETION OF THE URINE AND SALINE DIURESIS. in another series. In order to eliminate differences in the degree of hydrsemia induced, Magnus used solutions of equal osmotic pressure, his haemoglobin estimations showing that in this way approximately the same dilution of the blood was obtained. More recently Sollmann has confirmed and extended Magnus's observations by a similar method. The actual amounts of chloride and sulphate were not equal in these experiments, and in fact, approximated the proportion of three to two. Before I became acquainted with Magnus's results, I had adopted a method which eliminates the differen,ces in the degree of hydrsemia induced by the salts, and at the same time, permits of the injection of equal quantities — ^namely, the simultaneous injection of the two bodies under examination. And as the difference in the behavior of two such bodies as the chloride and sulphate of sodium^ are obviously due to the acid radicles, the solution injected was made up so as to contain an equal number of these. The standard was a normal solution of sodium chloride (CI. = 3.55 per cent), and the sodium sulphate solution was made up to contain an equal number of sulphate radicles; i. e., SOi = 9.6 per cent, or sodium sulphate 14.2 per cent. Similarly, a solution of sodium phosphate was formed, neutralized with phosphoric acid and diluted until it contained 9.5 per cent of PO4. The urea solution ordinarily used was six per cent. When equal quantities of two of these solutions were mixed, the resulting fluid contained an equal number of the bodies under consideration, the a,nions in the case of the salts, or urea. Of course, when such a mixture was injected into the blood, no such equality was maintained, but the excess of chloride and of sulphate in the body was at any rate equal and the results of this excess were the first objects of inves- tigation. The rabbits used in my experiments were anaesthetized with urethane (1.5 g. per kg.), or when the urea was to be estimated, with paraldehyde (1 g. per kg.). A Naunyn-Pfaff cannula was bound in the bladder and a ligature tied round the bladder between the orifices of the ureters and of the urethra. The animal was kept warm and the urine collected for two to three hours in order to ascertain the normal rate of secretion and where possible to obtain enough for chemical examination. When, as not infrequently happened, the secretion was insufficient for this purpose, some of the urine obtained from the bladder at the operation was added to it. After the normal secretion had been ascertained, a cannula was tied in the jugular vein and attached to a burette containing a warmed mixture of the two solutions under examination, which was allowed to flow in at the rate of 1-4 cc. per minute, according to the toxicity of the salts used. The first 2-4 cc. of urine collected after the infusion commenced was not used for analysis, as it contained some ' Most of my experiments were done with these two salts, as their behavior towards other organs suggested that the kidneys would react very differently towards them, and at the same time, the sulphate is only slightly poisonous. ARTHUR R. CUSHNY. 317 normal urine from the bladder and cannula. Thirty to fifty cc. of solution was injected, and the urine was collected every fiftoou minutes during the diuresis, and afterwards every half-hour or hour. The first experiments were performed vath chloride and sulphate of sodium, the mixture actually injected containing 1.775 gram of CI. and 4.S gram of SO^ in 100 cc. The chloride of the urine was deterniined by Volhard's method, the total sulphate by precipi- tation with barium chloride and weighing. Experiment I. — A rabbit of 1,500 g. weight was anajsthetized with 3 g. urethane. 12 :30 P.M.-2 :30, urine collected 4 cc. ; 2 :35-3 :00, infusion into the jugular vein of 50 cc. of the solution of chloride and sulphate of sodium. a •a 1 '3 2 (U i > So a <<-i o a a a !s a .s a c o;^ (U (U o^ .FH •S T3 -o a ■^s "rt "c3 "^ • S u . 'C ■eg XI ji •^'2 3 g p i 'B 6s 2 o ■2 -a OS as ^ m a CO m D. 12:30-2:30 ... i .0031 .078 .0012 .031 (Per 15 mm.) . 0.5 .0004 .078 .00015 .031 2:35-2:50 .... 3 ) 69 0.639 .258 .0073 .373 1.728 .349 .0036 .606 3:05 1' 32 0.2485 .1039 .0029 .3248 0.672 .352 .0037 1.1 3:20 11.5 .0214 .0006 .1864 .201 .0021 1.75 3:50. 5.5 4.0 3.5 r .0067 .0002 .071 .256 .0026 2.68 4:05 4:20 3 2.5 .338 .0035 4:35 2.94 4:50 2.5 Total 2:35-^:50 .... 5( ) 133.5 .8875 .3900 .011 2.4 1.495 .0165 In the above analysis, as in all others in this paper, the chloride and sul- phate are stated in terms of CI. and SO4 neglecting the base. In the columns marked '"chloride equivalents" and "sulphate equivalents," the total amount in the urine is divided by the atomic weight of chlorine and the molecular weight of SO4 in order to permit of a comparison of the number of chloride and sulphate radicles present in the urine. In the accompanying curves of excretion, the amounts are those of these two columns. This experiment was repeated a number of times with practically the same results, the only difference of importance being that the chloride did not generally disappear entirely from the urine as in Experiment I. It always fell to a mere fraction of that present in the earlier part of the experiment, however, and the percentage also diminished. The uniform result in this series was a marked diuresis, the fluid excreted much exceeding that injected; the diuresis began within a few minutes of the commencement of the infusion and diminished after its termination, but the secretion did not fall to the normal during the observation, i. e., within two and one-half hours after the injection began. The chloride of the urine rose 318 SECRETION OF THE URINE AND SALINE DIURESIS. with the injection not only in absolute amount, but also per cent, attained its maximum with the maximal diuresis, and then rapidly fell in absolute amount and per cent. In the course of two and one-half hours the percentage reached 1 l ij Y 7 ,jf 1 y \ \ \, ,j / v\ j / \ \ j / V s. il \ \ \, If w S, N ^ [_ V k^ s Ijl B i_ >^^ zc. .— . -.. ""~ ^■* 15 30 4-5 60 75 90 lOS 180 13S Fig. 1. — Curve of excretion in Experiment I. The heavy, unbrolten line represents the urine, the light line the sulphate equivalent ( =80/96). The broken line the chloride equivalents (=Cl/35.5). The duration of the injection is indicated by the heavy line along the base. that of the normal urine or below it and in some cases towards the end of the experiment no reaction was obtained on the addition of silver nitrate. The sulphate of the urine rose in absolute amount immediately after the injection began and reached its maximum about the same time as the chloride; but it fell more slowly and never disappeared from the urine within two to three hours of the injection. The percentage of sulphate rose during the injection and continued rising throughout the time of observation, the urine finally containing two to three per cent of sulphate. The whole of the chloride and sulphate injected was not excreted when the experiments were broken off. This is the more remarkable because there must have been, in Experiment I, an excess of chloride in the tissues at a time when there was no chloride in the urine. Magnus noted similariy, that after an infusion of dilute salt solution, there may be a large surplus of water and salt in the tissues, while the urine has returned to its normal quantity. ARTHUR R. CUSHNY. 319 The total weight of chloride excreted from the beginning of the injection to the end of the experiment was less than the total weight of sulphate, this agreeing with the results of Magnus. Even -when the quantities were reduced to equimolecular values, the sulphate radicles were considerably more numerous than the chlorides in experiments lasting two hours or more. In the earlier stages of the experiments, however (marked diuresis), the chloride ions exceeded the sulphate. Thus, in Experiment I, between 2 :35 and 3 :20, the equimole- cular ratio of the total chloride and sulphate excreted was 108:94. In the course of the next one and one-half hours it changed to 110:156. The great mass of the chloride thus appears early in the experiment, while the sulphate is more evenly distributed throughout the time of observation. Magnus's statement that the sulphate is more readily excreted by the kidney than the chloride (harnfahiger) thus seems to require some qualification, for in the stage of marked diuresis the chloride appears in larger quantities. An invaria- ble feature in my experiments was the parallelism of the water and chloride of the urine; they rose and reached their maximum together and then fell together, though, in some cases, the chloride disappeared while there was still a small secretion of fluid. The sulphate also ran parallel to the others in the first stage of marked diuresis, but then fell much more slowly than either, as is shown by the graphic representation in Fig. 1. A somewhat similar result was obtained in some experiments in which sodium sulphate was injected alone. Here, also, the chloride equivalent of the urine exceeded the sulphate during the stage of maximum diuresis, while it afterwards fell to zero as the diuresis passed off. The sulpha'te maxi- mum was attained at the same time as the maximum diuresis and maximum chloride, and the sulphate then fell in absolute amount though very slowly. The sulphate percentage rose as long as the experiment lasted, while the chloride per cent fell as the diuresis passed off, and soon no reaction was ob- tained with silver nitrate. Experiment II. — A rabbit of 2,100 g. weight ansesthetized with 3 g. urethane. Urine collected from 12:00-2:30 = 14 cc. (normal) ; 2 :35-2 :57, infusion of 30 cc. sodium sulphate solution (14.2 per cent). '6 6 G -6 6 CD s P n 3 2 = s _C3 o a a a m si s^ ■D > to 1 'a 3 5.S 3 c .s a; OJ 3 ■;1 o CO 05 r>- .— 1 1^ 00 *"* ^^ IN OJ IN o 13 T3 u >.S 03 lO CO 03 03 lO IN o s d aS.S o b o " c S o o 4-i « o« — 1 o c^ •^?i s ro CO CO m CM r^ t^ to 1— 1 oS- o lO TJH lO IN >o CO oa"^^ rH -1 --I a 1= • »o '>'t S lO cs C 5 (M 00 S^ 5. c'E O i-i o ^?i!3 ^. ^ q q ° ^ ^ q 1> -IT W) ^_^ >-H to »o CO ^ lO CO CO o IN (XI i-H CO Tt< m Ol C r^ CO eg CO o o: ^.I _; lO m ■ _ c o »o 03 O 05 01 IN O i: '^ .— 1 lO CO lO O CO CO p_ - o I-H T— 1 ^ nils CO on T— 1 r^ CO CO l> IN 1— 1 o T— 1 r- C5 o O o o P-Sll '~ H q q q q p q CO CO CO lO o ^ .= © a: CO t^ CO >o m 00 CO (N o I> o CO IN r^ Ol lO o-E o o (N r-i o o i> CLhS ■ .i^'i o 2^o I> o t^ lO j?^ s lO o 05 .-H (N lO (N CO O o: (N ^ ,_( CO I> 1< CO ^ CO P T— ( T— ! lO CO IN 1-T3 -*» -*1 CO T— 1 o ^1 i-H (N O o a> CO s ^^ P t- rl Tf( <-l (N a ™ 1 co T— I CO 1 •o ^ lO o o o o 1 CO 1-H -!t< o •71 1^ o 1^ ° CO CO -* ^ Tjl lO •a CO 322 SECRETION OF THE URINE AND SALINE DIURESIS. This experiment confirms the inferences drawn from a comparison of Experi- ments I and II, that the phosphate resembles the sulphate very closely in its excretion by the kidney. Rather more sulphate than phosphate occurs in the later stages. It is to be remarked that during the diuresis more chloride equivalents are excreted than sulphates or phosphates, although no chloride — r J V 1 //■ V > L / / \ / / ^ \ / y N \ * / / / V \ L ' / N^ \ / ^ I V ^ — u H H > — ■ -_ "^ - — — — -. IS 30 45 SO 75 90 lOS 120 135 ISO Fig. 2 — The curve of excretion of Experiment IV. The lines correspond to those of Fig. 1. The phosphate secretion runs so near that of the sulphate that it was impossible to reproduce it. was injected; in fact, at the height of diuresis (3.45), almost as many chloride ions appeared as both of the other constituents together. As the diuresis passed off, the chloride fell to a mere trace, while the sulphate and phosphate lessened more slowly in absolute amount and increased in percentage. A similar increase in the chloride of the urine without any injection of this salt was discussed in reference to Experiment II. In a number of experiments a normal solution of urea (six per cent) was mixed with an equal quantity of normal sodium chloride solution (5.85 per cent) . The urea of the urine was estimated by the Morner-Sjoqvist method. In Experiment V, the urea and chloride of the urine run parallel during the diuresis, attain their maximum with it, and commence their descent to- gether. The fall in the urea soon becomes slower than that of the chloride, however, and the former remains at a higher level. The percentage of chloride undergoes changes similar to those seen in Experiment I, except that it is unusually high for some time after the diuresis had practically disappeared, a point noted in several other experiments, in which urea was injected with chloride. The percentage of urea was high in the normal urine and fell ab- ARTHUR R. CUSHNY. 323 mptly during the diuresis to rise again gradually as the quantity of urine returned to the normal. The urea thus resembles the sulphate and phos- phate. Experiment V. — A rabbit of 1,300 g. ansesthetized with 2.0 g. of paraldehyde. 12:00 to 2 :30, 4 CO. of urine collected ;^ 2 :35 to 2 :57, infusion of 50 cc. of the urea and chloride mix- ture. a Li fi P. -a u 2 8 -a 3 > '3 . (a 'a .9 a a .9 3 a 6 '3 D .9 o '". d as £ P OJ S a P o 12:00-2:30 ... 4. .0011 .028 .108 .0018 2.7 (For 15 min) . . .4 .0001 .028 .0108 .00018 2.7 2:35-2:50 . . . . 384 30. .6834 .1416 004 .472 1.155 .1197 .002 .399 3:05 114 48. .2041 .2573 0072 .536 .345 .203 .0034 .423 3-20.. 14. 12.5 3.4 .0758 .0687 .0122 0021 0019 0003 .541 .55 .358 .0672 .1787 .0692 .0011 .003 .0012 .48 4:20 1.43 5:50.. 2.04 Totals 1 2:35-5:50 .. J 50 107.9 .8875 .5556 0155 1.5 .6378 .0107 ' Some urine found in the bladder was added to this for analysis. Jl / « /--I / I f 1 1 y / 1 « , =-cz:tc:;;;;:::ii:i::i i:p|:i::iziiii:iii: -f/ ^^v_ I|__ ^ >_ 15 30 45 60 75 90 105 120 135 150 Fig. 3. — Curve of excretion of Experiment V. The heavy line represents the fluid, the light the urea equivalents, and the broken, the chloride equivalents (the urea of the urine divided by 60, and the chloride by 35.5). 324 SECRETION OF THE URINE AND SALINE DIURESIS. The results of these experiments may be summarized as follows : The injection of the solutions used was always followed by marked diuresis, which commenced soon after the injection and gradually diminished after it ceased. The total chloride of the urine rose with the rise in the fluid, reached the maximum at the same time, and fell with the ebb of the diuresis, but often disappeared before the amount of urine reached the normal; the sulphates, phosphates, and urea increased with the diuresis and reached their maximum at the same time, but then fell more slowly than the fluid and chloride, so that their percentage in the urine rose continuously until the end of the obser- vation. The chloride and water ran an approximately parallel course through- out, while the other constituents examined diverged after the maximum was reached. When the solids of the urine were reduced to their molecular equiva- lents, it was found that the chloride was invariably present in larger propor- tion during the stage of marlied diuresis, but that as the fluid diminished the other constituent injected soon came to exceed the chloride. This was true in many cases even when no chloride was injected. The Composition of the Urine During the First Stage^ is Largely Determined by the Composition op the Plasma. The first question that arises in the attempt to explain these results is whether they necessitate recourse to a special secretory function. This may, of course, be introduced as an " explanation, " but the difficulty is met at once that this involves a change in the form of the activity of the secretory activity, for at first it prefers chlorides to sulphates, and later, extrudes the latter to the entire exclusion of the former. Moreover, it pursues the same capri- cious policy when sulphates alone are injected, and when it is, therefore, to the advantage of the body to retain the chlorides which are present in only the normal proportion. This view appears so improbable that it could only be accepted in default of any other explanation whatever. A second inter- pretation may be sought in the composition of the blood, namely, that the rela- tive proportions of the constitutents of the urine correspond with those existing in the blood. A number of estimates of the chlorides and sulphates of the blood were made,^ and it was found, as was to be anticipated a priori, that the chlorides in every case and throughout the experiment exceeded the sul- phates. In the urine, on the other hand, the chlorides exceed the sulphates only during the stage of marlced diuresis, while the latter predominate during 1 In the rest of this paper the " first stage " will be used to indicate the phase of maximum diuresis and maximum chloride excretion; the "second stage" will be used to indicate the phase during which the diuresis fell markedly, the chlorides lessened or disap- peared, while the sulphates, phosphates, etc., increased in percentage. No absolute line of demarcation could be drawn between the two phases. ' See Journ. of Phys., XXVII, p. 429. ARTHUR R. CUSHNY. 325 the later phases. If the composition of the blood determines that of the urine during the first stage, then some other factor must intervene during the later stages which is in abeyance during the diuresis, but tends to modify the influ- ences of the composition of the blood when the excessive flow is passing off. The view that the constitution of the blood determines that of the urine might be supported by some experiments in which some other constituent is injected into the vessels in large amount so as to exceeil the blood chloride. It is impossible to use the sulphates and phosphates for this purpose, as they are poisonous when injected in large enough quantities, but urea is compara- tively harmless. In several experiments in which urea was injected in such quan.tity as to exceed the chloride of the blood, the urine actually contained more urea than chloride in the stage of diuresis. Experiment VI. — A rabbit of 1,325 g. weight was anEesthetized with 2 g. paraldehyde. 2:30 to 3:15 p.m., the urine amounted to 6 cc, to which some urine, previously collected, was added for analysis. 3:15 to 3:40, infusion of 60 cc. of a mixture of equal parts of solution of urea (20 per cent) and sodium chloride (normal) . 1 6 c d ■6 ■s 'a 3 'U a CD 6 a 'E c a) »3 •o a 'a 3 . 0, a ti (D ~ 2 — O Si O OS Oft aS d o e _3 p Q a; 5.S 2:3(>-3;15 ... 6 .0036 .0001 .06 .032 .0005 .534 (Per 15 min.). . 2 .0012 .06 .011 .534 3:l&-3:30 ... 30 38 .5325 .1470 .0041 .387 3.0 .417 .0069 1.098 3-45. 30 114 22 15 8 oi .5325 .5670 .1015 .0495 .0128 .0074 .0160 .0029 .0014 .0004 .0002 .497 .4615 .33 .16 .135 3.0 1.38 .31 .294 .206 .085 .0230 .0051 .0049 .0034 .0014 1.21 4:00 1.41 4:30 1.96 5:00.. 2.57 5:30 1.54(?) Total. 1 3:15-5:30 .. J 60 202.5 1.065 .8852 .0250 6.0 2.692 .0447 Experiment VI differed from No. V in the chloride and urea not being injected in equimolecular solution but in the proportion of 3:10. If the rab- bit's blood be taken at 1/13 of the body weight, the whole blood of the animal in this experiment would contain 0.388 g. of chloride,' and probably the whole available fluid of the body would scarcely contain more than 1-2 g. at a liberal estimate. To this was added 6 g. of urea and about 1 g. of CI. The urea must obviously have exceeded the chloride in the blood and lymph. The only essential difference between the results of Experiments V and VI is, however, the excess of urea molecules over the chloride, which is present in the stage of diuresis as well as later. This experiment and another performed in precisely the same way are the only ones in which the chloride did not exceed the other constituents injected in the stage of diuresis. They are also the only ' Abderhalden. Zeitschr. f. physiol. Chemie, XXV, p. 65. 326 SECRETION OF THE URINE AND SALINE DIURESIS. ones in which another body was present in the blood in larger amounts than the chloride. The inference seems justifiable that the excess of chloride in the urine during the diuresis in other experiments is due not to any special affinity between it and the renal cells, but to its preponderance in the serum. 1 / \ / \ // \\ 1 / / \ // v\ \ , / V \ J \lj f \ \\ \ 1/ i \ / / r \ \ m H. •»* "«^ — . ;^ — 15 30 4-5 60 75 90 105 Fig. 4. — Curve, of excretion in Experiment VI. The interpretation is the same as in Fig. 3. Fig. 4 is reduced to one-quarter the height it would occupy on the same scale as Figs. 1-3. Doubtless, if sulphate could be injected in such quantity that it exceeded the chloride in the blood, there would prove to be more sulphate than chloride in the urine during the diuresis. The Second Factor in the Composition of the Urine is Differential Absorption in the Tubules. Accepting then the view, that during the diuresis the proportion of the salts in the urine is determined by their proportion in the blood, what is the factor which interferes in the later phases and induces a complete reversal of these proportions? After weighing a number of possible explanations, I found that the one which had least apparent objection was that during the second stage of ebbing diuresis, the water and chloride were largely absorbed in the tubules, and the sulphate, urea, and phosphate allowed to pass through unabsorbed. This would indicate that the epithelium of the tubules had greater avidity for water and chloride than for the other constituents, and this would only be in accord with what is known in regard to the behavior of the sul- phate and phosphate towards other living cells. Their slow absorption in the intestine has led to their use as saline cathartics and appears to be due to their ARTHUR R. CUSHNY. 327 failure to penetrate the epithelial cells. The red blood corpuscles similarly are permeated much more slowly by the sulphate than by the chloride of am- monium/ and gelatin discs take up less fluid from sulphate than from chloride solutions.^ Many colloids, such as the globulins, are thrown out of solution more readily by the sulphates than by the chlorides,^ and some crystalline substances resemble them in this/ On the other hand, certain membranes seem to absorb sulphate solutions as readily as chloride, as was shown for the the pleura by Leathes and Starling ,° and was rendered probable for the capil- lary walls by Magnus" and Sollniann. Urea is generally regarded as one of the most readily diffusing constituents of the body, and it seemed improbable that it would be rejected by these cells; but it has been shown by Overton that urea permeates vegetable cells' only very slowly, and that when tadpoles are put in a weak solution, it passes into the blood and tissues so slowly that twenty-four to thirty-six hours, or more, are required to establish equilibrium between the animal and its environment.' Somewhat stronger solutions actually withdraw water from the tadpole. So that the behavior of the epithelium of the renal canals towards urea is not without analogy in other living cells. In the ureter, bladder, and urethra, the mucous membrane must possess a similar lack of affinity for urea, otherwise much of it would be ab- sorbed from the urine. The absorption of chloride and water in the tubules in preference to sul- phates, phosphates, and urea would, therefore, not be devoid of analogy in the behavior of other animal cells. On the other hand, it would suffice to explain the phenomena observed. For it is readily apprehended that when a large amount of fluid is poured into the tubules from the glomerulus it must pass through the tubules rapidly and very little time is allowed for absorption. The fluid, therefore, issues from the tubules into the ureters but little changed in composition; that is, it resembles more or less closely the composition of the plasma. As the diuresis abates, however, the flow through the tubules be- comes slower, more time is allowed for absorption, and those substances for which the epithelium has an affinity — water and chlorides — are taken up, while the less permeating sulphates, phosphates, and urea are permitted to find their way into the ureters. 'Giyns. Arch. f. d. ges. Physiol., LXIII, p. 86; Hedin. Ihid., LXVIII, p. 229, and LXX, p. 525. 2 Hofmeister. Arch. f. exp. Path. u. Pharm., XXVIII, p. 210. ^ Lewith. Arch, f . exp. Path. u. Pharm., XXIV, p. 1 ; Hofmeister, Ihid., p. 247. ^ Pohl. Ztschr. f. phys. Chem., XIV, p. 151 ; Young. Journ. of Phys., XXI, p. XVI. * Ibid., XVIII, p. 106. « Arch. f. exp. Path. u. Pharm., XLIV, p. 396. ' Ztschr. f. physik. Chem., XXII, p. 189. ' Studien ueber die Narkose, p. 120, Jena, 1901. 328 SECRETION OF THE URINE AND SALINE DIURESIS. The Tubules of the Kidney Absorb Water and Chloride More Readily THAN Sulphates and Urea. The only method of measuring the absorption of the tubules that suggested itself to me was to retard the passage of the urine into the ureter and thus raise the pressure in the tubules and promote the absorption. The urine secreted under pressure ought then to show less of the water and chloride than that secreted under ordinary conditions, while the less readily absorbed substances would show less variation.^ The rabbits used in the following experiments were anaesthetized in the same way as in my earher experiments. A cannula was inserted into the jugu- lar vein and connected with a burette containing the warmed solution to be injected. The ureters were exposed by an incision in the abdomen and into one a long glass cannula was passed and tied as far up as possible, while the cannula of the other was of a T-shape. One horizontal arm of this cannula was tied into the ureter, the other was provided with a short India-rubber tube, which could be compressed by a finely graduated screw clamp; the per- pendicular arm was bent into a small manometer containing mercury. The urine from the ureter ran through the horizontal part of the T, and was re- ceived in a capsule placed below the orifice of the India-rubber tube.^ When the clamp was tightened, the urine escaped only under pressure, which was measured by the manometer. As a general rule, the clamp was closed at the beginning of an experiment until the manometer showed the desired pressure, and the solution was then infused into the vein while the pressure in the ureter was kept constant by tightening or loosening the clamp as required. In the beginning of an experiment the pressure rose and fell, obviously with the pas- sage of peristaltic waves along the ureter. This soon disappeared, however, as the ureter became dilated and the pressure afterwards showed very httle variation. Even at the end of an experiment, after the clamp had been removed, no evidence of the contraction of the ureters was offered, the urine escaping in a regular stream, while that from the other ureter showed the usual intermittent flow arising from the ureteral peristalsis. As a general rule, the clamp had to be loosened during the diuresis and tightened as it passed off, in order to maintain the pressure at a constant level. The solutions injected were composed of equal parts of equimolecular solutions of chloride and sulphate of sodium, or of the chloride and urea, and the perfusion was carried out in the same way as in the experiments already described. The pressure in the manometer was generally 20 mm. Hg., rarely ' The objections which may be raised to this mode of experimentation are discussed in the Journ. of Phys., XXVIII, p. 431. ' The cannula and ureter when the manometer indicated 30 mm. pressure were found to contain 0.3 to 0.4 cc. ARTHUR R. CUSHNY. 329 30 mm., and in some experiments 15 mm. In the first experiments, a mixture of equal parts of sodium chloritle solution 5.85 per cent, and of sodium sul- phate solution 14.2 per cent, was injected into the jugular vein. The chlorides of the urine were estimated by ^^olhard's method, the sulphates by precipita- tion with barium chloride and weighing. The quantities given in the tables are the absolute amounts of chloride (CI.) and sulphate (SO4) contained in the urine. Experiment VII. — October 21, 1901. A rabbit of 1,400 g. anjesthetizod with urethane. The manometer cannula was inserted in the right ureter. 3 :30 to 3 :S0, infusion of 30 cc. of solution in the jugular vein. The screw clamp was tightened at 3 :30 until the manometer gave a pressure of 30 mm. Hg., which was maintained throughout. Urine. Chloride. Sulphate. 3:37-3:47 Left kidney 24 0.0809 0.1080 Right kidney 8 0.0142 0.0667 Difference (absorbed) 16 0. 0667 . 0413 Assuming the glomemlar secretion to have been equal on the two sides, 66 per cent of the fluid, 82 per cent of the chloride, and 38 per cent of the sulphate was reabsorbed in the tubules of the right kidney. This indicates that more of the chloride permeated the epithelium than of the water, while less than half as much sulphate as chloride escaped in this way. Before the clamp was applied the two kidneys were secreting at an equal rate. ExPEHiMENT VIII. — October 7, 1901. A rabbit of 1,400 g. was anaesthetized with urethane. The manometer cannula was inserted in the left ureter. 3:10 to 3:40 p.m., 50 cc. of the solution was infused; 3:10 to 3:15, urine from left ureter, 1.54 g., from right, 1.35 g. At 3:15 the clamp was tightened until the manometer showed a pressure of 20 mm. of mercury. From 3 :15 to 3 :20 the mercury oscillated from 15 to 25 mm., but after- wards was maintained practically stationary at 20 mm. by very minute changes of the screw. Time. Urine. Chloride. Sulphate. 3:15-325 Right 10.4 0.0328 0.084 Left 7.0 ,0.0173 0,083 (?) Reabsorbed 3.4 (33%) . 0155 (47%) , 001 (1%) 325-3:30 Right 12.0 0.0456 0.086 Left 7.6 0.0256 0.066 Reabsorbed 4.4 (37%) 0.02 (44%) 0.02 (23%) 3:30-3:35 Right 12.2 0.0486 0.0842 Left 7.5 0.0260 0.0643 Reabsorbed 4.7(38%) 0.0226(46%) 0.0199(24%) 3:35-3:40 Right 9.8 0.0382 0.0755 Left 5.2 0.0168 0.0499 Reabsorbed 4.6(47%) 0.0214(56%) 0.0256(34%) 330 SECRETION OF THE URINE AND SALINE DIURESIS. Time. Urine. Chloride. Sulphate. 3:40-3:55 Right 8.9 0.0192 0.1317 Left 3.6 0.0047 0.0763 Reabsorbed 5.3(59%) 0.0145(75%) 0.0554(42%) 3:55-4:15 Right 6.8 0.0090 0.1129 Left 3.7 0.0007 0.0821 Reabsorbed 3.1(46%) 0.0083(92%) 0.0308(27%) 4:15-5:5 Right 9.0 0.0108 0.1683 Left 2.8 0.0 0.0532 (?) Reabsorbed 6.2(69%,) 0.0108(100%) 0.1151(68%) In these experiments the sulphate absorbed in the tubules forms a smaller proportion of that leaving the capsule than is true of the chloride and water, if, as seems probable, the same amount of glomerular secretion was formed on the two sides. The epithelium of the tubules takes up more chloride than water and least of all the sulphates. The course of the excretion by the kid- ney which is free from resistance, is similar to that described in the earlier part of this paper, the chloride ions exceeding the sulphate as long as the diure- sis is marked, that is till 3 :40 in Experiment VIII, but the sulphate ions after this point prepondering over the chloride. In the urine from the left kidney, on the other hand, the sulphate proves to be in excess throughout the experi- ment, which demonstrates sufficiently the correctness of the explanation al- ready suggested, namely, that the change in the relative proportion of sul- phate and chloride in the first series of experiments is due to the absorption being imperfect in the first stage of diuresis and more efficient in the second. In the next series of experiments chloride of sodium and urea were in- jected in equimolecular parts by mixing equal parts of a solution of sodium chloride 5.85 per cent, and one of urea six per cent. The urea of the urine was estimated by the Morner-Sjoqvist method. Experiment IX. — April 16, 1901 . A rabbit of 2,000 g. anaesthetized with paraldehyde. The manometer cannula was placed in the right ureter and the clamp was screwed tight until the pressure amounted to 20 mm. of mercury. 3 :20 to 3 :38, infusion of 50 co. of the mixture into the jugular vein. Time. Urine. Chloride. Urea. 320-3:30 Left ureter 8.4 0.0379 0.0517 Right ureter 5.0 0.0204 0.049 Reabsorbed 3.4 (40%,) . 0175 (46%o) . 0027 (5%) 3:30-3:40 Left ureter 22.8 0.1085 0.1121 Right ureter 12.3 0.0572 0.08 Reabsorbed 10 . 5 (46%) . 0513 (47%,) . 0321 (28%) ARTHUR R. CUSHNY. 331 Time. Urine. Chloride. Urra. 3:40-3:60 Left ureter 17.4 0.0821 0.0814 Right ureter 6.8 0.033 0.037 Reabsorbed 10.6 (61%) 0.0491 (60%) 0.0444 (54%) 3:50-4:00 Leftureter 8.2 0.0398 0.0535 Right uterer 2.3 0.0102 0.0310 Reabsorbed 5.9 (72%) . 0296 (74%,) . 0225 (42%) 4:00-4:30 Leftureter 14.4 0.0736 0.0847 Right ureter 3.6 0.0146 0.0443 Reabsorbed 10 . 8 (75%) . 059 (80%) . 0404 (48%o) At 4:30 the clamp was removed and a solution of lithium chloride was run into the vein. The urine secreted by the right kidney in fifteen minutes was 28 cc, that by the left 26.5 ec, so that no injury had been done by the manipulations dm-ing the experiment. This experiment was repeated a number of times with the same result. The deficit of urea on the side on which resistance was offered was always less than the deficit of fluid and chloride, so that urea resembles the sulphates in being less readily taken up by the tubules than chlorides and water. The chloride was absorbed in larger proportion than the water, though the differ- ence is less than in Experiments VII and VIII. The fact that the urea does not disappear so readily as the fluid and water may, of course, be interpreted as indicating that the urea is secreted by the tubules, but the logical sequence of this \iew would be that the sulphate is also secreted here, for the urea and sulphate of the urine run parallel courses. The excretion of the phosphates followed that of the sulphates so closely in my former experiments, that I have not considered it necessary to perform experiments in regard to their excretion under resistance. One other constit- uent of the urine that is obviously absorbed with difficulty is the pigment. When diuresis occiu's, the urine loses its coloration and becomes clear and limpid ; but when the flow is retarded by obstruction of the ureter, the color departs much less from the normal, so that if the pigment is excreted by the glomerulus, the tubules have difficulty in taking it up. In all of these experiments the percentage of chloride absorbed was greater than that of the water ; in other words, the fluid entering the epithelium con- tained a higher percentage of chloride than the original glomerular secretion, or the cells were permeated more readily by chloride than by the water in which it was dissolved. This would appear at first sight improbable, were not a similar phenomenon known in unorganized colloid bodies. But Hof- meister^ has shown that a wet gelatine plate placed in a solution of sodium chloride takes up more of the salt than of the water, exactly as the renal epithe- lium appears to do. In my experiments the amount of water absorbed was ' Arch. f. exp. Path. u. Pharm., XXVIII, p. 210. 332 SECRETION OF THE URINE AND SALINE DIURESIS. probably limited by the osmotic resistance offered by the non-permeating sulphate and urea, and three experiments were, therefore, performed to find whether more chloride than water would be absorbed, when sulphate or urea was not present in excess. For this purpose a solution of sodium chloride, 5.85 per cent, was infused into the vein. Experiment X. — January 30, 1902. Rabbit of 2,200 g. ansesthetized with paralde- hyde. Manometer cannula in the left ureter. Fifty-three co. of a 5.85 per cent solution of sodium chloride were injected into the jugular vein between 4:05 and 4:30 p.m. The clamp was screwed down until the resistance was 20 mm. mercury. Time. Urine. Chloride. 4:05^:15 Right ureter 17 0.1125 Left ureter 5.5 0.0451 Reabsorbed 11.5 (68%) . 0674 (60%) 4:15-425 Right ureter 16 0.1049 Left ureter 4.4 0.0358 Reabsorbed 11.6 (72%) 0.0691 (66%) 4 25'-4:35 Right ureter 5.4 0.0427 Left ureter 2.1 0.0169 Reabsorbed 3.3 (61%) .0258 (58%) 4:35-4:45 Right ureter 3.7 Left ureter 0.3 3.4(91%) 4:45 The clamp was removed from the cannula. 4:47^:57 Right ureter 3.4 0.0251 Left ureter 4.2 0.0298 4 :57-5 :17 Right ureter 3 . 0194 Left ureter 3 0.0206 In these experiments the chloride was reabsorbed in rather lower propor- tion than the water, indicating that a somewhat more dilute fluid returned to the blood than that in the tubules. The presence of sulphate and urea in large amounts thus appears to retard the absorption of water and not of chloride, while the cells of the tubules are ordinarily not more permeable by chloride than by fluid. The result of these experiments is to show that under circumstances in which absorption is admitted to occur by Heidenhain and his followers, the chlorides and water return to the blood much more readily than the sulphates, phosphates, urea, and pigment. And this differential absorption is all that is required to explain the course of the excretion of these bodies, when they can pass into the ureters without resistance. It, therefore, seems justifiable to assume that here also the phenomena are due to the reabsorption in differ- ' During this time there was partial asphyxia. ARTHUR R. CUSHNY. 333 ent degrees of the constituents of the glomeruhir fluid. It remained only to show that this reabsorption takes place in the renal tubules, and not in the ureter and renal pelvis. For this purpose a cannula was inserted in the ureter and filled with sodium chloride solution (0.8 per cent). It was then connected to a tube filled with the same solution and this was raised until the column of fluid pressing on the ureter and pelvis amounted to 1,200 mm. Beyond the first fall in the fluid, which indicated the distension of the ureter and jDelvis, practically no change in its height occurred during thirty minutes, indicating that no absorption took place. It has long been known that the tubules can- not be injected from the ureter; this experiment indicates, therefore, that the ureter and pelvis fail to absorb salt solution. I had hoped to paralyze the absorption in the tubules by allowing quinine nitrate to act on them, but a solution of this salt with methylene blue injected into the ureter, and held there for half an hour, failed to color the tubules, so that the attempt had to be abandoned. The relative proportion of the solids of the urine during saline diuresis may thus be explained by two factors, the relative proportion of these constitu- ents in the blood and the differential absorption in the tubules. These both act throughout the diuresis, but in the earlier phase the former predominates, in the later the latter. For if the composition of the blood alone determined that of the m-ine in the first phase the proportions of sulphate and chloride , , , , . , ., r .■ sulphate of blood per cent , , would be equal m each, or the fraction — —-: — : ^ — -. , , would = sulphate of urine per cent chloride of blood per cent -r, j. ^i • • ^ ^ r ^.i ^ j. r x- -n — n — 5 — -■ 7 • -Dut this is not correct, for the first traction is chloride of urine per cent always less than the second, because some of the chloride is reabsorbed in the tubules. That the composition of the blood exercises an influence on that of the urine in the second stage scarcely requires demonstration. Another point which is elucidated by this differential absorption of the salts in the tubules is the slight amount of diuresis which follows the injection of chlorides compared with that after an equivalent quantity of sulphate. This has been demonstrated by Magnus and more recently by Haake and Spiro.' It is a well known physical law that if a salt fails to permeate a diffusion mem- brane, it prevents the water in which it is dissolved from passing through the membrane. Accordingly, a sulphate solution placed in the intestine^ is not • Hofmeister's Beitrage zu chem. Phys. u. Path., II, p. 149. ^ Some months after I first pointed out the analogous behavior of the intestine and the renal tubules towards the chlorides and sulphates, and asserted that they were both ab- sorbing organs, I found an instance of the intestine actually partaking in the function of absorbing the fluid of the urine. Wiener (Hofmeister's Beitrage, II, p. 56) finds that when the cloaca in fowls is tied oil from the gut, enormous amounts of urine are passed, while normally no fluid escapes from the cloaca. He accounts for this by supposing that the fluid of the urine is absorbed in the gut in the fowl. 334 SECRETION OF THE URINE AND SALINE DIURESIS. absorbed, because the sulphate permeates the membrane with difficulty, while a chloride solution is absorbed readily because the chloride penetrates the cells. In the same way a sulphate solution reaching the renal tubules pre- vents the absorption of fluid by the epithelium, and more fluid reaches the ureter than is the case when the permeating chloride aloiie is present in the tubules; the sulphate, therefore, proves the better diuretic when it is injected into the blood-vessels. The presence of sulphate in a diffusing fluid affects only the water and not the chlorides present, and accordingly, I found that more chloride was absorbed than water in the experiments in which sulphate was injected, while the outflow from the ureter was retarded (Experiments VII and VIII). The difference in the absorption of sulphates, chlorides, etc., is obviously due to some difference in their physical properties,' for the same variation occurs in their absorption into gelatine discs (Hofmeister). The simplest view of the epithelial absorption in the kidney, and also in the intes- tine, demands only a blind force which causes a current from the lumen of the tubules towards the blood-vessels. This force acts equally upon all the constituents of the glomerular fluid, when they have once succeeded in pene- trating the cell; but some of them, notably the water and chlorides, permeate it much more readily than others, such as the sulphates and urea. No more complicated activity has been demonstrated as yet, although it may prove necessary to assume this in the future. The Glomerular Secretion. The predominating factor in the earlier phases of saline diuresis is the glomerular activity, while the absorption in the tubules is comparatively unim- portant in determining the relative proportions of the urinary constituents. The question of the character of the force involved in the glomerular secretion does not come within the compass of my experiments. On the other hand, it is important to know whether the process of secretion alters the character of the fluid on the two sides of the capsular membrane; that is, whether the cap- sule pennits the passage of the fluid of the plasma unchanged, or whether it differentiates in favor of some of the constituents. In other words, is the capsulat membrane equally permeable by the sulphates, chlorides, urea, etc., or does it strain off some of these while permitting the passage of others. This could be determined absolutely only by comparing the composition of ' Loewi (Arch, f . exp. Path, u Pharm. XLVIII, p. 410) points out that the absorption of the chloride varies in different animals and under different conditions; this he attributes to varying activity of the renal cells. This is quite in accord with my own experience, and I think, indicates that while the relative proportion of chloride and sulphate absorbed is mainly determined by their relative permeating qualities, the absolute amount of either salt absorbed varies with the activity of the renal cell, while this in turn may depend upon the physical condition of the blood. Loewi's article, unfortunately, reached me too late to permit of its adequate discussion in the text. ARTHUR R. CUSHNY. 335 the plasma and of the glomerular fluid, but the latter is unobtainable in a pure form, since it is altered in passing through the tubules by the absorption of some of its constituents. But, as has been stated, the urine, during diuresis, resembles the plasma more closely than at other times, as would be the case if the capsule transmitted the inorganic constituents in the same proportions as it receives them from the plasma. And the deviations in the composition of the urine from that of the plasma are such as would be inferred from the selective absorption in the tubules. Thus the percentage of sulphate is higher than in the blood because the sulphate is reabsorbed with difficulty. These experiments thus fail to indicate that the capsule differentiates between differ- ent salts ; that is, the membrane is passive in regard to the nature of the salts it transmits, although it may verj^ probably exercise force in transmitting them. This may not be universally true. I am at present engaged in a re- search as to the permeation of some other constituents of the plasma. The Increased Flow of Urine in Saline Diuresis is not Due to Stimu- lation OF the Secretory Cells. SaUne diuresis has generally been recognized as due to stimulation of the renal cells by the presence of water or salts in excess in the blood; but another view advocated by Starling,' on the basis of Ludwig's theory, attributes it to changes in the renal circulation, and thus regards the augmented activity of the kidney as largely the indirect result of changes induced by the salts in other parts of the body. Starling's experiments, in support of this view, consisted in keeping the volume of the kidney constant by bleeding from the carotid, when it tended to increase after the injection of dextrose; he found that as long as the blood-flow through the kidneys (measured by the oncometer) did not rise in amount, no diuresis resulted, and concluded that the diuresis ordinarily following dextrose injection is the result of hydrsemic plethora, along with some dilatation of the renal vessels through the action of the foreign substance in the blood. The objection may be made to this experiment that the hemor- rhage, besides keeping the blood volume constant, may interfere with the activity of the kidney in some other way, and I have accordingly attempted to test the theory in a different way. The volume of the kidney was meas- ured as in Starling's experiments by the oncometer, but instead of withdraw- ing blood to maintain the volume at a constant point, I compressed the renal artery by a screw-clamp made for the purpose. The other kidney continued to secrete normally, and the urine on the two sides could be contrasted, with certainty that it was secreted under the same conditions, except as to the quantity of blood flowing through the two kidneys. The left renal artery in the rabbit runs at a little distance from the vein for part of its course, and special ' Joum. of Physiol., XXIV, p. 317. 336 SECRETION OF THE URINE AND SALINE DIURESIS. precautions were taken to avoid any pressure on the vein by the clamp. It was found necessary to immerse the animal in a bath of warmed physiological salt solution (0.8 per cent) in order to avoid the effects of air and cold on the abdominal viscera. The ureter cannulas were long, thin, glass tubes and the free ends were brought above the surface of the solution and led into weighed glass capsules. The oncometer communicated with a horizontal tube and the changes in volume were read off on a scale placed behind this. Experiment XL — January 23, 1902. A rabbit of 950 g. weight was anEesthetized with paraldehyde 1.5 g. Laparotomy, cannulse in the ureters, oncometer on left kidney, and clamp round the left renal artery as described. The fluid extended to 52-55 mm. along the oncometer scale ; pulsations of the heart and respiratory movements were well marked. At 3 :40 an infusion of three per cent sodium chloride into the jugular vein was commenced and continued to the end of the experiment, about 2 cc. being run in per minute. The changes in the volume of the kidney and in the amount of urine are given in the following table. The urine is estimated in drops per minute, the drops from the two cannulas being practically equal in size. The drops were counted by two assistants, and the numbers given are generally the average of several counts. The arterial pulsations remained visible in the scale throughout the experiment. Volume of Time. kidney in mm. on sS.2 Scale. o c as o»3 3:30 52- 55 1/4 -Vs Vs Normal secretion. 3 •.40-3 :50 80-190 6 5 Infusion of salt solution began at 3 :40, and the vol- ume of the kidney at once begins to swell. 3:51-4:00 55 20 i The clamp on the left renal artery was tightened un- til the kidney resumed its normal volume. 4:02 200 20 17 Clamp loose at 4 :01 . 4:04 200 20 23 4:07 200 25 15 4:10 120 23 4 The clamp was partially closed at 4 :09. 4:14 120 20 2 Clamp in same position. 4:15 55 20 1/2 Clamp tightened until kidney of normal volume. 4:20 140 20 11 The clamp loosened, but not completely. In this series of experiments, the blood flowing through the two kidneys had, of course, the same constituents, and as long as the circulatory conditions remained the same on the two sides the urine was practically equal in amount, , the left kidney secreting rather less than the right in Experiment XL When, however, the changes in the circulation induced by the saline injection were excluded on the left side, the secretion fell to practically the same quantity as was passed before the injection. The constriction of the renal artery was sufficient to counterbalance the effects of the saline injection, although it did not prevent the passage of blood, for the kidney volume was the normal, while complete closure of the artery at the end of the experiment reduced it ARTHUR R. CITSHNY. 337 considerably. Moreover, the pulse was marked in the oncometer tube through- out the experiment, and a normal secretion persisted during constriction, lasting for over half an hour in some experiments. Again, in Experiment XI, there could be no question that blood was passing through the kidney in large amount when the clamp was only partially tightencil at 4:10 and 4:20, yet the secretion was markedly diminished. If the diuresis, after saline infusion, were due to stimulation of the renal cells by the abnormal constitution of the blood, it would probably be much less affected by such a slight diminution of the blood supply. On the other hand, these experiments confirm Starling's view fuUy, that the saline diuresis is due to the hydrseraic plethora, which may be accompamed by a direct dilator effect in the kidney itself. Anything which interferes with this excessive supply of blood to the kidney lessens the diuresis, and when the blood volume in the kidney is reduced to the normal, the secretion of urine also becomes practically normal. This conclusion is opposed to that reached by Magnus* after a series of experiments in which the effect of intravenous injections of strong salt solutions was contrasted with that of the plethora induced by blood transfusion. No diuresis followed in the latter, and Magnus infers that plethora alone is not sufficient to increase the activity of the kidneys, and that some other factor is involved in saline diuresis. But much of the plasma of the transfused blood escaped as lymph, so that although the increased volume of the circulating blood led to augmented tension in the arteries, veins, and capillaries, the volume was increased mainly by the mass of corpuscles, and the vessels contained less plasma in proportion to the cells than before the transfusion. When strong saline infusions were injected, on the other hand, and diuresis followed, the blood volume was in- creased only in plasma, this increase being due in part to the fluid injected, in part to the inflow of lymph from the tissues into the blood-vessels. The transfusion experiments thus resembled those with saline injection only in the increased bulk of the blood but differed entirely in the relation which the total volume of the blood bore to the fluid, which alone can pass through the renal cap- sule. In plethora from transfusion, the excessive fluid escapes as lymph, while in plethora from strong saline injection it escapes as urine. The explanation may be that in the latter case escape into the tissues is blocked by the counter- current of Ijrmph poured into the blood-vessels, owing to the high osmotic pressure of the solutions injected; the fluid, therefore, follows the only other way of escape, that by the glomerular capsule. In short, while Magnus' experiments indicate that all forms of plethora do not induce diuresis, one cannot safely deduce from them that the saline diuresis is independent of the hydrsemic plethora. In addition, the blood after transfusion was in a state of marked anhy- draemia, as was indicated by the very high percentage of haemoglobin (e.g., ' Arch. f. exp. Path. u. Pharm., XLV., p. 210. 338 SECRETION OF THE URINE AND SALINE DIURESIS. 19.78 per cent as compared with 15.38 before transfusion). It is to be an- ticipated that absorption was accelerated from all available supplies of fluid and the reabsorption in the tubules may have been sufficient to conceal any augmentation of the glomerular fluid. Magnus regards the saline diuresis as due to specific stimulation of the renal cells which is not induced by plethora, while I would suggest that the diuresis from saUne injection is largely physical in origin, and its absence after transfusion is due to the fact that in the latter case the fluid of the blood can escape into the lymph spaces, and in anhydrsemic plethora (such as in Magnus' experiments) there may be a further hindrance to diuresis through an augmented absorption of the glomerular filtrate in the tubules. According to Ludwig's view, the balance between the different constituents of the blood must be maintained for the most part by varying activity of the epithelium of the tubules. For example, when large amounts of fiuid are lost from the body by other channels, the activity of the renal epithelium must be increased, in all probability by the physical effects of the alterations in the blood and lymph surrounding them. An exact parallel is offered by the behavior of the intestinal epithelium when fluid is withdrawn from the body in large amount, for it then absorbs solutions of saline cathartics which would ordi- narily lead to the passage of fluid from the blood into the lumen of the bowel. Magnus' plethora appears in reality to induce the same effects as the with- drawal of fluid, for the haemoglobin percentage of 19.78 per cent, as compared with the 15.38 before the transfusion, indicates a change in the relative propor- tion of corpuscles and plasma, such as would follow the withdrawal of eleven per cent of the latter. I have performed several experiments in which serum was injected instead of blood. Experiment XII. Several rabbits had been fed on carrots for two weeks. Two of them were bled and the serum separated by the centrifuge. It was quite clear and con- tained 0.387 of chloride per cent. (Abderhalden' gives 0.3883 per cent.) A third rabbit of 965 g. weight was anaesthetized with urethane and a bladder cannula was tied in. From 4:15 to 4:26, 40 cc. of serum was infused into the jugular vein. Time. G. Urine. Time. G. Urine. 3:55-4:15 0.7 5:35-5:55- 13.87 4:15^:35 1.86 5:55-6:15 17.98 4:35-4:55 2.46 6:15-6:35 8.02 4:55-5:15 2.84 6:35-6:55 2.24 5:15-5:35 7.6 If these res\ilts be compared with those obtained by Thompson^ after the injection of dilute sodium chloride solution, a striking similarity will be recog- nized at once. The blood of this rabbit may be estimated at 75 cc, and its plasma at 47 cc. (Abderhalden). The plasma was practically doubled by 1 Ztschr. f. phys. Chem., XXV, p. 65. 2 Journ. of Physiol., XXV, p. 487. ARTHUR R. CUSHNY. 339 the infusion and this led to a marked diuresis, which, however, reached its maximum only during the second hour, exactly as in Thoniiwon's experiments. This late diuresis may be due to the injection first escaping as lymph from the blood-vessels. Experiment XIII. Three rabbits fed on dry hay for a week and then starved for twenty-four hours. Two of them were bled and the serum separated by the centrifuge. The third (1,320 g.) was anaesthetized with paraldehyde, and a bladder cannula was in- serted. Fifty CO. of serum was infused into the jugular vein between 3:45 and 4:05. Time. G. Urine. Time. G. Urine. 320-3:40 0.51 4:40 0,37 3:40-3:50 0.29 4:50 0.26 4:00 0.26 5:00 0.21 4:10 1.45 5:10 0.28 420 5.93 5:20 0.35 4:30 3.4 An infusion of strong sodium eUoride at this point caused profuse diuresis. In this experiment the diuresis was comparatively slight, perhaps because the fluid of the blood was reduced by the feeding with dry fodder during warm weather. The preliminary treatment is similar to that advocated by Magnus. Serum, therefore, seems to cause diuresis when injected intravenously, al- though no marked change in the constitution of the plasma could be induced by it, but only in the volume of the blood. It is impossible to state, however, that the serum injected was identical in composition with that of the animal experimented on, and comparatively slight deviations may be sufficient to induce diuresis, as Magnus has shown. It appeared impossible to obtain evidence of the diuretic effects of serum infusion to which this objection could not be made, and this method of attacking the problem was, therefore, aban- doned. The few experiments made, however, suggest that serum has diuretic properties similar to those of dilute salt solution, and in consequence support the view that the diuresis following the injection of saline diuretics is not due to stimulation of the secretory renal cells by the presence of abnormal con- stituents in the blood, but is the result of the increased volume of the blood. Summary and Conclusions. The results of these. experiments seem to supply answers to, at any rate, some of the questions which were put in the earlier pages of this paper. There can, I think, be no question thatthe epithelium of the tubules has the function of absorbing some constituents of the glomerular fluid, notably the water and the chlorides.' Whether in addition to this activity it excretes some of the constituents of the urine, as has been stated by Bowman, Heidenhain, Koranyi, and others, is unknown; but there is, so far as I can find, no evidence in sup- port of this view. ^ Brand (Pfluger's Arch. f. d. ges. Phys., XC, p. 491) has recently concluded that absorption of the bile normally occurs in the gall bladder. 340 SECRETION OF THE URINE AND SALINE DIURESIS. This absorption cannot be explained by the ordinary process of diffusion. Some unknown form of energy is in action here as in the intestine, and as in the case of the latter, causes a current passing from the lumen towards the blood-vessels. This force must be of considerable magnitude, for the con- centration of a dilute solution of sulphate, such as occurred in the blood and glomerular fluid, to the 2-3 per cent, met with in the urine necessitates the per- formance of a large amount of work in overcoming osmotic pressure, as Dreser pointed out in the first paper in which physico-chemical methods were applied to the problems of the kidney. The unknown form of energy involved in the absorption in the tubules is subject to very material modification by condi- tions which are found in action quite apart from living cells, and organized matter. Thus the chlorides are absorbed by the epithelium of the tubules more readily than the sulphates, and Hofmeister has shown the same to be true for gelatine discs. Moreover, the presence of sulphates in the tubules Umits the absorption of the fluid through the osmotic resistance of the solution opposing the cellular activity. The difference in the diuretic value of some salts is thus due to a physical factor. As regards the nature of the force involved in glomerular secretion, my results give Ho definite evidence, for they mS,y be explained equally well by the theory of physical filtration, or by the view that the capsular membrane is the seat of an unknown force which propels diffusible bodies and water entering it in the direction from the blood toward the lumen. My experi- ments can be interpreted without such a force being assumed of course, but offer no evidence against its existence. But whatever the nature of glomerular secretion, it is capable of marked modification through changes in the physical conditions, notably by changes in the volume of the blood and the pressure exerted by it; this has been shown by Starhng's experiments, and by the results recorded in the last division of this paper. The change in the physical conditions caused by the injections of the salts, etc., used in my experiments, is sufficient to explain the diuresis without having recourse to any specific action (stimulation) exercised by the salts or water on the renal cells. Addendum. ' The first part of this paper, that dealing with the general course of the excretion of chlorides, sulphates, etc., appeared in the Journal of Physiology, XXVII, p. 429, 1902. The results of the earlier experiments in which the ureter was partially obstructed were given before the American Association of Physiologists in 1901, and abstracted in the Amer. Journal of Physiology, VI, p. XVII (March, 1902). Two criticisms have since appeared, the one by Filehne,* who refuses to admit the absorption in the tubules on the evidence • Arch. f. d. ges. Physiol., XCI, p. 565. ARTHUR R. CUSHNY. 341 of my first paper, the other by SoUmaun/ who, accepting absorption on the evidence of my paper and abstract, cannot believe that it is the only factor in the concentration of the urine. The arguments used by Filehne and Soll- mann are very similar and may be shortly discussed. They both quote an experiment of Magnus, and Sollmann gives two others of his own which sub- stantiate it. In this experiment sodium sulphate was injected intravenously in a dog in considerable quantity, and the chloride of the urine fell to a very low percentage (0.067 to 0.073 per cent). The sodium sulphate of the urine was 2.159 to 3.124 per cent, while that of the blood appears to have been less than 0.272 per cent. My critics seem to consider that, according to my view, the chloride of the urine should have been very much higher, but this is a misconception. The glomerular filtrate would, in this case, contain 0.272 per cent of sulphate, and in order to bring this up to the concentration found in the urine, about seven-eighths of the water had to be absorbed in the tu- bules, and along with this water the chloride disappeared. The urine in this experiment, in fact, corresponds to that observed in the later stages of diuresis in my experiments I and III, and notably with that obtained from an obstructed ureter in Experiments VII and VIII. In these last two experi- ments, it was shown conclusively that chloride is absorbed in excess of water, and Magnus and Sollmann 's experiments thus offer no insurmountable argu- ment against the ^'iew that the difference in the relative proportions of the sulphates and chlorides of the urine from those prevailing in the blood is due to the absorption in the tubules. The other arguments cited by Filehne and Sollmann are that it has been shown by Dreser and by Ruschhaupt that copious drinking of water causes diuresis, in which the urine is very low in solids and in particular in chloride. Sollmann cites a similar instance, the fall in the chlorides of the urine in star- vation without any corresponding diminution in the amount of the urine. Both of these arguments would be fatal to my view, were it such as Filehne and Sollmann conceive it. But they have entirely misapprehended my mean- ing, or rather have interpreted my remarks concerning saline diuresis as an attempt to explain the whole renal function. One need scarcely go to such exceptional conditions as those they have cited to find the chloride of the urine differing from that of the blood in percentage, for large deviations occur in perfectly normal life, as may be shown by any clinical estimation. But I have never put forward any statement which can be interpreted as meaning that in diuresis or in any other condition the percentage of chloride in the urine is equal to that of the blood. In the profuse diuresis obtained in rabbits from intravenous injection, the percentage of chloride of the urine approximates that of the blood more closely than usual, as I have stated. But in my experi- ments with obstruction of the ureter, it was shown distinctly that the absorp- ' Amer. Joum. of Physiol., VIII, p. 155. 342 SECRETION OF THE URINE AND SALINE DIURESIS. tion of the chloride is independent of that of the water, that in fact the epithe- lium of the tubules differentiates not only between chlorides and sulphates, but also between chlorides and water. Curiously enough, SoUmann cites my abstract of an experiment in which this is clearly shown, but fails to appreciate its import. The occurrence of urine with a low percentage of chloride may, therefore, be due to a reabsorption of chloride without a corresponding absorption of water. I think it is very probable that the activity of the epithelium may vary with the condition of the blood, in fact this is almost necessarily true. But I have not yet obtained unequivocal evidence of this. SoUmann, finding these difficulties in explaining his observations by the reabsorption in the tubules, has recourse to another view — that the relative amount of the urinary constituents is, in part, due to what may be termed a discriminating excretion by the glomerular capsule, which in the sulphate experiments he cites would concentrate the sulphate and reject the chloride into the blood.^ In fact, he believes that part of the process which I have located in the tubules is carried out in the excretory cell. It is by no means impossible that the capsule has this function. Some results, I have recently obtained seem to indicate that it is not a simple filter. But it seems undesirable to introduce this complication unless it is founded on some posi- tive evidence. I claim that the selective absorption in the tubules has been established as a vera cavsa. The same cannot be stated as regards the dis- criminating power of the glomerular capsule; the only grounds on which this has been suggested is a supposed failure of the absorption in the tubules to explain observed facts, and this failure I cannot accept as proved. ' It may be remarked in passing that this would be contrary to all that is known' in regard to the behavior of sulphates and chlorides toward colloids; for wherever there is any difference in the penetration of chlorides and sulphates, the latter is always dis- criminated against. ON THE DIUEETIC ACTION OF ALCOHOL. CHARLES W. EDMUNDS, M.D., Assistant in Pharmacology, Department of ^[edicine and Surgery, University of Michigan. (From the Pharmacological Laboratory of the Unioersily of Michigan.) A widespread popular belief exists that alcohol possesses a specific action on the kidneys which induces a marked increase in the secretion of urine. The truth of this view has seldom been investigated carefully, the only work on the subject that I have been able to find being that of Simanowsky' and Mori.^ Simanowsky's experiments, carried out on himself and two others, were as to the effect of different grades of beer upon health, and during his experiments he studied the effects upon the amount of urine. His conclusions were that good beer is slightly diuretic unless, as in his own case, it causes such a sweat- ing that less urine was passed than would have followed an equal volume of water. Mori, in a large number of experiments carried out upon himself and others, investigated the diuretic action of beer, wine, and of alcohol. His results show that alcohol (four per cent) is a diuretic, causing a flow of urine about two and a half times as great as would be caused by an equal volume of water. Beer is slightly stronger, and wine still more powerful, producing more than fom- times the amount of urine that would be caused by an equal amount of water. The constituent of the alcoholic beverages causing the diuresis, he says, is the alcohol, although CO2 is also diuretic to a lesser degree. His results point to a distinct influence exerted by alcohol on the secretion of urine, and it appeared desirable to investigate how far this is due to a specific action of the drug and how far it is due to the physical changes introduced in the fluids of the body by its presence. Mori states that he believes the most important cause to be a direct action on the kidney epithelium, although he had not specially investigated this aspect of the question. On the other hand, common salt, sodium sulphate, and some other similar bodies appear to owe their diuretic action mainly to their physical properties (salt action), accord- ing to recent researches by Starling and Cushny.' The fact that alcohol caused more diuresis than water in Mori's experiments does not indicate a specific action exerted by the drug, as he seems to consider; this could have been established only by comparing his alcohol results with those obtained ' Arch. f. Hygiene, IV, p. 1, 1886. ^ Ibid., VIII, p. 354, 1887. ' For the references to the literature, see the preceding paper. 343 344 ON THE DIURETIC ACTION OF ALCOHOL. from equimolecular solutions of some substance resembling alcohol in its behavior in the body. If alcohol proved to cause a greater diuresis than such a substance, the fair inference would be that it possessed some specific action on the kidney. Another way of attacking the problem was to compare in animals the diuresis from alcohol with that from other bodies, and this method was adopted in the following experiments, rabbits being selected for the pur- pose. The animals had been kept in the laboratory for some weeks and had been fed on hay, beets, and grain, in order to get them as nearly as possible in uniform condition. Efforts were made in the beginning to feed them on hay and grain alone, but as many died, probably from lack of fluid, beets were added to their diet. The anaesthetic employed was paraldehyde, 1.5 to 2 cc. per kg. of weight being given. The drugs were injected into the jugular vein at the rate of 3 to 4 cc. per minute, and the urine was collected by a Pfaff- Naunyn cannula in the bladder. .The urine was weighed in all the experi- ments, as in some the quantity obtained was too small to measure. As will be seen in the table below, in the earlier experiments weighings were made every ten minutes, while in the others, at fifteen-minute intervals, but in either case the urine was collected and weighed for two periods before the injection was begun in order to get a normal rate of excretion. In the first two experiments two and one-half per cent alcohol was used, the total amount injected being 50 cc. (i. e., 1.25 cc. absolute alcohol). The fol- lowing table shows the rate of excretion : Experiment I. — Rabbit; weight, 700 g. Alcohol injected during interval between 10 and 20. Time . Urine . . 10 . .15 20 .12 30 .06 40 50 60 0. 70 80 .24 90 .22 100 .27 110 .32 120 .32 130 .32 140 .38 150 .30 Experiment II. 10 and 20- —Rabbit; weight, 770 g. Alcohol injected during interval between Time . . Urine . . 10 .26 20 .16 30 .06 40 .02 50 .18 60 .33 70 .22 80 .35 90 .50 100 .59 110 .60 120 .48 It will be noticed that there is a marked fall in the excretion during and after the injection, the fall being most noticeable in Experiment I. The curve of excretion is essentially the same in both, the maximum being reached in the latter half of the second hour following the injection. In both experiments, one and one-fourth hours after the injection began, the urine became blood tinged and continued so to the end. The animal, in Experiment II, was bled after the observations were over, and the serum which separated, was markedly tinged with pink, explaining the hemo- globinuria which had occurred. To avoid this laking of the blood and resulting hemoglobinuria, in the next series of experiments ten per cent, alcohol was used and only 15 cc. injected (1.5 cc. absolute alcohol). In this series the effect of alcohol was compared with that of a known diuretic, dextrose, recrystallized once from eighty per cent, ethyl alcohoL CHARLES W. EDMUNDS. 345 The two solutions were made equimolecular, alcohol as stated, ten ])pr cent, and the corresponding dextrose solution being 39.12 per cent. The following tables will show the results obtained from the seven experiments in which alcohol was injected first and then dextrose, and also in two experiments in which the dextrose was injected first: Experiment I. — Rabbit ; weight, 800 g. Alcohol injected during interval 20-30 ; dex- trose during interval 150-160. Time 10 20 30 40 50 60 70 80 90 100 Urine 31 .14 .24 ,51 .62 .79 1.03 1.21 1.14 1.09 Time 110 120 130 140 150 160 170 180 190 Urine 1.16 1.19 .98 1.04 1.14 7.82 6.93 5.92 5.57 Experiment II. — Rabbit; weight, 950 g. Alcohol injected in interval 20-30; dextrose, 160-170. Time... 10 20 30 40 50 60 70 80 90 100 -110 120 Urine... .05 .14 .04 .04 .21 .51 .51 .49 .47 .55 .54 .63 Time... 130 140 150 160 170 180 190 200 210 220 230 240 Urine.. .58 .59 .57 .59 1.18 11.40 7.13 4.66 3.41 2.33 1.85 1.54 Experiment III. — Rabbit ; weight, 1,300 g. Alcohol injected between 45-60 ; dextrose, 180-195. Time.... 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 Urine... .68 .53 .71 .21 .14 .44 2.35 2.86 2.78 2.97 3.15 2.31 11.78 14.52 7.57 Experiment IV. — Rabbit; weight, 1,710 g. Alcohol injected between 30-45; dex- trose, 165-180. Time 15 30 45 60 75 90 105 120 135 150 165 180 195 210 Urine 12 2.56 1.60 1.45 1.52 1..33 1.73 1.81 1.79 1.96 1.99 9.02 15.78 4.90 Experiment V. — Rabbit; weight, 1,450 g. Alcohol injected between 30-40; dex- trose, 150-160. Time 10 20 30 40 50 60 70 80 90 100 110 Urine 17 .21 .18 .14 .24 .07 .09 .06 .15 .14 .12 Time.. 120 130 140 150 160 170 180 190 200 210 Urine 07 .05 .04 .00 .16 .15 .31 .19 .12 .09 Experiment VI. — Rabbit; weight, 1,120 g. Alcohol injected between 30-45; dex- trose, 150-165. Time 15 30 45 60 75 90 105 120 135 150 165 Urine 50 .62 .55 .00 .21 .19 .14 .07 .10 .19 1.82 Animal died. Experiment VII. — -Rabbit; weight, 1,165 g. Alcohol injected between 45-60; dex- trose, 120-135. Time 15 30 45 60 75 90 105 120 135 150 165 Urine 1.95 2.25 1.21 .52 .00 .04 .00 .05 .62 .84 .70 Experiment VIII. — Rabbit; weight 940 g. Dextrose injected between 20-30; alcohol, 150-160. Time 10 20 30 40 50 60 70 80 90 100 110 Urine 76 .49 3.89 22.79 13.67 8.33 4.00 2.68 2.26 2.01 1.34 Time 120 130 140 150 160 170 180 190 200 210 Urine 99 1.26 1.03 .96 .51 .71 .18 .17 .11 .13 In Experiment IX the urine fell to zero after alcohol, and even dextrose caused no diuresis. It would seem as if we should take the first four experiments as being those in which alcohol exercises its typical action not only on account of the great 346 ON THE DIURETIC ACTION OF ALCOHOL. similarity of the alcohol curves, but also because in these the dextrose diuresis is typical. In three of these there is a slight diminution in the urinary secre- tion immediately after the alcoholic injection, followed by a gradual increase until the highest point of diuresis is reached in from an hour and forty minutes to two hours after the injection, when the amount excreted varied from four to six times the normal. The dextrose diuresis in these experiments, as is seen, was prompt and marked. In Experiments V, VI, and VII, there was an exaggeration of the fall which takes place after the alcoholic injection, culminating in complete anuria, which lasted for some time. After the injec- tion of dextrose there was a slight excretion of urine. In Experiment VIII, when dextrose was injected, there was a well-marked diuresis from it, but when alcohol was injected later, no diuresis occurred from it. In the first four experiments there exist not only a marked difference between the extent of diuresis caused by the two drugs, but also a long interval before the height of alcoholic diuresis is reached. This is also shown in the two experiments in which two and one-half per cent alcohol was used, in which it took place more than an hour and a half from the time of injection. The results are remarkably like those obtained by Thompson^ in his experiments with dilute sodiimi chloride solution, the maximum diuresis in these being reached in the second hour, as in my experiments with alcohol. Mori gets the same delayed diuresis in his experiments as shown in the following table giving the average time at which the changes took place : Diuresis. Appeared. Reached Maximum. Lasted. With 4 per cent alcohol . . " Beer " Wine . 65 min. 59 min. 90 min. 82-109 min. 80-127 min. 112-187 min. 204 min. 210 min. 216 min He also says water gave much the same qualitative resxilts. The action of alcohol on the kidney differs not only from dextrose, but also from NaCl in strong solution, which causes an immediate increase in urine, while even ten per cent alcohol has no such effect. The late diuresis from alcohol and also from dilute NaCl is probably due to the increased fluid of the blood, for a similar effect occurs from the injection of serum (Cushny). Alcohol then in these experiments, not only seems devoid of any specific action, but fails to induce the increased diuresis which follows the infusion of indifferent bodies from the physical changes induced in the blood. In order to get some Ught, if possible, on the subject, the blood changes were studied in several cases to see the extent of hydrsemia following each injection, as several investigators have noted hydrsemia when diuresis follows the injection of salts (Magnus and Thompson). In two of the experiments blood counts were made both before and after each injection. The results » Joum. of Physiol., Vol. XXV, p. 490. CHARLES W. EDMUNDS. 347 showed the blood count after alcohol injection was practically the same as before, while after dextrose injection the count fell over twenty-eight per cent, indicating an increase of twenty-eight per cent in the fluid of the blood. As some of the changes were so slight that they might have been within the limits of error admissible in the blood counts, in the next four experiments a small amount of blood, varying from 2 g. to 4 g., was collected from the carotid in a weighing bottle, weighed, dried at 84° C, weighed again and per cents of liquids and solids estimated. Lest exception should be taken to this method of estimation on the ground that the heat employed to dry the blood would drive off any alcohol that might have been present, but would leave any dex- trose, which being weighed would add to the amount of solid, the chlorides in Experiment IV were estimated as a control, but the results showed that this possible objection could be disregarded. The results obtained in the blood examinations are shown in the following table. The time at which the blood was drawn after the injections were made is given below the estimation. As in some cases the diuresis also plays an important part in the concentration of the blood, the approximate amount of urine secreted in the interval is also given. 1 er cent of solids 3efore alcohol injection. er cent of solids after alcohol injection. II. cent solids before Jctrose injection, ime " is length of tervalll to III. III. r cent solids after xtrose injection IV. •So .a" O N . S a > 1 Oi (^ ^^^•2 ^■S f^ in 16.52 15.85 20 min. .24 G. 14.72 105 min. 14.78 20 min. 24.3G. 15.42 7.57 Gin 15 min. IV 18.64 18.71 10 min. 1.05 18.98 120 min. 17.98- 10 min. 15. 17.98 + 12. vn 18.27 17.61 10 min. .34 17.27 60 min. 16.25 10 min. .65 16.26 1.05 IX 14.50 14.01 20 min. 2.10 14.70 45 min. 11.35 20 min. .00 10.90 .00 The greatest fall in the percentage of solids caused by alcohol was in Experi- ment III, .67 per cent, with only about .2 g. of urine excreted in the interval, while in the same experiment after dextrose, the solids remained constant, but there had been 24 g. of urine excreted in the corresponding time, and the withdrawal of this amount of fluid would be sufficient to remove the hydraemia. In Experiment IX there was practically no diuresis to interfere with the blood changes, and here alcohol caused a fall of .49 per cent and dextrose 3.35 per cent. The same thing is noticed in Experiments IV and VII, where the intervals were only ten minutes between the infusion and the bleeding. In Experiment IV a rise of .07 after alcohol, but a fall of one per cent after dex- trose, with the addition of 15 g. urine excreted in the interval after dextrose. 348 ON THE DIURETIC ACTION OF ALCOHOL. In Experiment VII a fall of .66 per cent after alcohol, but 1.02 after dextrose. The differences here noted in the behavior of the two substances in the body- may partially explain the difference in the diuretic action. As shown by the blood examinations, alcohol leaves the blood quickly, escaping into the lymph spaces, so that only a slight hydraemia occurs, while the dextrose solution remains in the blood or escapes by the kidney, causing diuresis. But even the comparatively sUght hydrsemia following the alcohol injections might be expected to cause some diuresis, and I have, therefore, attempted to discover what interferes with this. In order to do this, three experiments were per- formed in which the volume of the kidney was measured by the oncometer during the infusion of alcohol and dextrose. Starling, Thompson, and Magnus have shown that there is a certain rela- tion between diuresis and the volume of the kidney, which tends to increase during the increased activity of the organ, although this is not invariably the case. Experiment. — Rabbit, weight 1,110 g., 2 cc. paraldehyde. Solutions of alcohol and dextrose, same as heretofore injected into the jugular vein and the urine collected by a cannula in bladder. As the left kidney was to be used, the right ureter was tied off close to the bladder and cut on the kidney side of the ligature so that the urine from the oncometer kidney alone passed into the bladder and was collected. The oncometer was connected with a horizontal tube with a scale behind it and the volume read off this. A tube was inserted into the trachea and the animal then placed in a salt solution bath, kept during the experiment at 36° C. The results will be seen in the fol- lowing table : Time. 2:50 2:55 3:00 3:05 3:08 3:15 3 20 3 25 3:30 3:35 3:50 4:05 TMne .... 09 10 19 .22 .10 Vol. .. 59. 57.2 56.6 55.9 54.6 54. 40. 49. 49. 49. 46.4 47. Ale. Inj., 3:14-3:18J. Time .. 4:20 4:35 4:50 4:55 5:00 5:05 5:10 5:15 520 525 5:30 5:35 Urine. .13 .10 .10 90 2,87 1.20 Vol... 47. 47. 46.4 45.7 45.4 50.8 49.6 47.6 47.3 47.6 47. 47. Dext. Inj., 5:00-5:05. Observations were taken every five minutes on the volume of the kidney, but the figures are not given above where the readings were nearly uniform. The most interesting part of the above experiment was the effect on the volume during the time of injection. To show this better readings were taken every minute during injection, with the results given below. Volume changes during alcohol injection. Time 3:14 3:15 3:16 3:17 3:18 3:19 320 322 325 Vol 54. 54. 52.1 49.6 45.7 40.6 40. 45.7 49. Ale. Inj., 3:14-3:18i. Changes during dextrose injection. Time 5:00 5:01 5:02 5:03 5:04 5:05 5:08 5:10 Vol 45.4 49-.6 50.8 51.1 50.8 50.8 49.9 49.6 Dextrose Inj ., 5:00-5:05. CHARLES W. EDMUNDS. 349 As shown by the tables with the one-minute readinfjs, there was a rapid and marked decrease in the size of the kidney during the iilcohol injection, while under dextrose an equally rapid but less marked increase in volume occurred, this increase being caused as is well known by dilatation of the renal vessels resulting in more blood being carried to the organ. In the other two experi- ments with the oncometer, the same results were obtained, but as the scale on the oncometer tube was not long enough to record the changes in these cases, the results tabulated are not as accurate as in the experiment cited. However, the volume changes were even more marked. The diminution in volume during alcohol injection must be due to one of two factors, either a constriction of the renal vessels or lessened efficiency of the heart, due to the alcohol injected. It hardly seemed likely that the amount of alcohol used could have affected the heart to this degree, but to make certain it did not, two experiments were done in which the blood pressure in the carotid was recorded. It was found that during the alcohol injection the blood pressure was raised the equivalent of 4 mm. of mercury, showing clearly that the decrease in renal volume was not caused by weakening of the heart but by a narrowing of the kidney vessels. The decreased blood supply caused by this constriction during alcohol injection accounts for the fall in the amount of urine following the injection, which has been noted above. The last experiment carried out in this work was to compare the action of whiskey and sodium chloride solution with the alcohol and dextrose used in the earher experiments. The two new solutions weje made up so as to depress the freezing point to the same extent as the ten per cent alcohol solution formerly used. The same amounts were injected with practically the same results as seen in Experiments V and VII, the excretion falling to zero after the whiskey injection and remaining there until sodium chloride was introduced which caused a mild diuresis immediately. The blood changes were almost identical with those obtained under alcohol and dextrose. The results are given in the following table. Experiment. — Rabbit; weight, 1,180 g. Whiskey injected between 30-45; NaCl, 105- 120. Time 15 30 45 60 75 90 105 120 135 150 165 Urine 21 .11 .10 .42 .66 .05 The per cent of solids in the blood before the whiskey injection was 18.21 per cent; 10 mins. after injection, 18.31 per cent. Before NaCl injection, 17.62 per cent, and 10 mins. after, 15.68, and 20 mins. later, 16.31 per cent. It would appear from the blood examination that whiskey' acted in the same way as alcohol, escaping into the lymph spaces rapidly and causing no immediate diuresis, while sodium chloride, unable to escape so readily, induced diuresis. Alcohol would thus seem to differ in its diuretic action from salts in strong 350 ON THE DIURETIC ACTION OF ALCOHOL. solution in three ways: First, in not causing immediate diuresis; second, in causing little or no hydraemia ; third, in contracting the renal vessels, lessening the blood supply, and therefore the kidney volume. It differs also in its action from the weak saline solutions used by Thompson in the third respect given above, but resembles them in the first two, as the blood changes found by Thompson were small and the diuresis delayed. Alcohol thus differs from salts, urea, and dextrose in its action on the kidney, and my results fail to decide the question whether it has a specific action on the epithelium or not. The fact that its injection intravenously in strong solution is not followed by diuresis, resembling that induced by dex- trose, is sufficiently explained by its rapid disappearance into the tissues and by the constriction of the vessels, and this cannot be used as an argument for a spe- cific action. On the other hand, the late diuresis following its use intravenously, or by the mouth, resembles that following the injection of dilute salt solution or serum, and therefore supports the view that its action is physical. As stated earHer, Mori believed it to be a specific action on the renal epithelium, but he did not eliminate salt action by comparing the alcohol results with those of an equal volume of a solution equimolecular with the alcohol and beer used. ON NUTMEG POISONING. GEORGE B. WALLACE, M.D., Instructor in Pharmacology, University and Bellei'ue Medical School, New York. (From the Pharmacological Laboratory, University of Michigan.) The researches of Binz, Heffter, Lindemann and others on the various members of the group of volatile or ethereal oils, have shown that these sub- stances produce effects of considerable pharmacological interest. At the present time their use as medicines is not an important one, it is true, since they are applied chiefly for local effects and have no general or systemic action unless taken in dangerous amounts. The frequency with which poisoning has occurred after large doses of such oils as tansy, nutmeg, and pennyroyal, however, makes a study of their action of some practical importance. Among these poisonous oils, that of nutmeg, oleum myristicae, is not generally included, although the cases of poisoning recorded probably equal in number those of oils considered to be more dangerous. Cases of poisoning almost invariably arise from the use of the crude fruit, its general use in cookery lending an opportunity which is wanting in many other oils; and in a large number of the cases it has been used either as an abortifacient or emmenagogue, uses which the results show to be unwarranted. Toxicology. In the older literature on nutmeg poisoning, Lobelius' mentions a case in which a large woman, pregnant, took 10-12 nutmegs, these producing only delirium. Sohmid (quoted by Hughes^ cites the following : " A man, aged 36, took four nutmegs, which completely overwhelmed his nervous system. He remained in a state of ' coma vigil ' for three days, and emerged with his memory completely lost. A continued fever super- vened, with insomnia and palpitation, and at length paralysis of all the limbs. Reason and recollection did not return until after eight days. " PauUinus' states that a woman had bloody sweats after taking nutmegs, while CuUen* mentions a case in which the patient, after taking about two drams of powdered nutmeg, became drowsy, and then passed into a condition of complete stupor and insensibility, followed later by alternating delir- ium and stupor, with recovery in twenty-four hours. Another instance of poisoning is described by van Bosch. *^ After taking seven nuts, the symptoms were pain in the stomach, headache, dilated pupils, unsteady gait, and labored breathing. The tempera- ture was below normal, and the pulse quite weak. There was marked drowsiness for four days. Some other cases are mentioned by Gmelin*' as having been observed by Rumph, Thunberg, and Radloff. The more recent cases reported are as follows: Case I. — Reported by Barry.* Woman, aged 38, shortly after labor took one and one- half nutmegs as ' nutmeg tea, ' to relieve a spasm of the uterus. This was taken late in 351 352 ON NUTMEG POISONING. the afternoon. At 10 p.m. she began to get drowsy, and by 4 a.m. was in a profound stupor. At 10 A.M. the narcotic effects began to pass off and by 4 p.m. the patient had practically recovered. The symptoms and treatment, Barry states, were the same as in opium poisoning. Case 11. — Reported by Matthews." Girl, aged 9, ate one-half small nutmeg during the afternoon; she became sick during the night and fell into a stupor from which she could be aroused with difficulty ; did not recognize her parents or surroundings, but could be aroused sufficiently to answer questions; she complained of dryness of the thi-oat and the pupils were somewhat dilated. The treatment consisted in giying strong coffee, and castor oil. The stupor was much less next morning, but the girl was very sleepy, and could not see. The treatment was continued and complete recovery occurred. During the whole case the pulse and respiration were normal. Case III. Reported by Gaulke.' Man ate one nutmeg, cut up into small pieces, for pain in the stomach. After an hour he became delirious, opening his mouth and making biting movements at the persons about him. When seen by Gaulke some hours later the pulse and respiration were normal and there was no pain present. The biting movements continued, with short periods of quiescence. For treatment, morphine was given, and mustard plaster applied to the abdomen and back. Recovery was complete in twenty-four hours. About one-third of the nut eaten consisted of a brownish, half- solid substance with a bitter taste, so that this was probably not a case of pure nutmeg poisoning. Case IV. — Reported by Howard.' Young married woman took about one ounce of finely powdered nutmeg in whiskey to bring on delayed menstruation. This was taken before retiring and she was awakened during the night by prseoordial anxiety and an intense burning sensation in the stomach. She soon sank into a semiconscious condi- tion, pulse 130, respiration oppressed, and face pale. The intellect was impaired, objects appearing multiplied and the head enlarged. The more severe symptoms were removed by coffee. A good deal of pain remained in the ovarian region for several days, however, and there was a profuse urination, the urine having the odor of nutmeg. The catamenial function was not restored. Case V. — Reported by Palmer." Woman took one and one-half nutmegs. Became drowsy in two hours and remained in this condition about an hour, the drowsiness amount- ing almost to stupor. This was followed by a condition of excitement, with sharp pain in the head, involuntary laughter, wild fancies, and incessant talking. Presently there was pain in the region of the heart, with cold extremities, and a sinking sensation. The face was pale, and the pulse weak and thready. These symptoms lasted more than an hour, during which two doses of ammonium bromide were given. The next morning the dose was repeated. The patient remained unusually nervous for several weeks. Case VI. — Reported by Pinnock.'" Healthy woman took one-half nutmeg, out finely, in a glass of hot ale. Symptoms came on in a few hours and consisted of nausea, sleepiness, paUor, coldness and prickling of the skin. The pupils were dilated, the respira- tion weak, and the pulse could not be felt. The heart was weak and fluttering. Stimu- lants and emetics were administered and the patient recovered. Cases VII and VIII. — Reported by Dodge." Boy, aged 8, ate two nutmegs. A number of hours later was found in semi-comatose condition. The coma gradually deep- ened in spite of the administration of ammonia, brandy, and atropine sulphate, and death occurred on the following morning. There was complete suppression of urine. Girl, aged 3, sister of the boy, ate one nutmeg, which was afterwards mostly thrown out by vomiting. She also became partly unconscious, but was normal on the following morning. GEORGE B. WALLACE. 353 Case IX. — Reported by Car\-ell.'^ Woman, aged 21, in third mouth of pregnancy, took five powdered nutmegs in one dose to produce abortion. Was seen three hours later with following symptoms: Pulse, 140; temperature, normal; delirious for a few minutes, calling out loudly, then sinking into a stupor from which she could be aroused to again become delirious. She complained of an iuunense weight pressing upon her. An emetic was given, but little nutmeg appeared in the \-omitus. Castor oil, potassium bromide, and ammoniated tincture of v£ilerian were also given. K,(^co^'(!ry was complete in about two days. Xo abortion occurred. Case X. — Reported by Gillespie." Woman took five grated nutmegs in warm water to produce abortion. Symptoms of poisoning came on in one hour. These were dizziness, vertigo, and inco-ordination ; the face was greatly flushed and swollen, the pupils contracted, the pulse 130, but of good strength ; there was marked nausea but no vomiting. Frontal headache was severe, but there \\'as no abdominal pain. Vomiting was induced by 30 grs. of zinc sulphate, and the symptoms were much relie^-ed for about an hour when sud- den circulatory failure and collapse came on. Spirits of ammonia were administered and gradual improvement and recovery followed. No narcotic effects were produced in this case. Abortion did not occvu\ Case XL — Reported by Air." Woman took one nutmeg in hot water to bring on delayed menstruation. In two hours had extreme thirst, giddiness, and prostration, with an intolerable restlessness, causing her to walk up and down the room almost incessantly, holding on to furniture for support. She also complained of a feeling of tightness in the chest. Emetics and coffee were given in treatment. The symptoms persisted through the night and passed awa}- the next morning. Case XII. — Reported by Tyler;'* Woman took one nutmeg steeped in gin to bring on menstruation. The symptoms were giddiness, faintness, and disturbed vision; the pulse was feeble, the skin cold and clammy, and the pupils were dilated. There was no abdominal pain. The treatment consisted in giving an emetic, which removed a large part of the nutmeg, and brandy. The recovery was rapid. Case XIII. — Reported by Alexander." A man took one nutmeg for diarrhoea. All the next day he felt stupid, giddy, and very drowsy. Strong coffee and castor oil caused the symptoms to pass away. Case XIV. — ^Reported by Devlin." Young woman ate one nutmeg about 7 p.m. She then went to bed and slept heavily till late next morning when she was aroused with some difficulty. She got up feeling very drowsy, had no appetite, and vomited several times. At 10 :30 a.m., when seen by Devlin, vomiting had stopped, face was pale, and she was unable to raise the eyelids. There was no giddiness, but a marked drowsiness and disinclination for exertion. Strong coffee and other stimulants were administered and recovery followed. Case XV. — Reported by Bentlif." Man took one grated nutmeg in milk as treat- ment for boils. This was taken at night. He awoke early in the morning with a feeling of giddiness. Was unable to stand, complained of great pain in the head, and could not distinguish surrounding objects. When examined by Bentlif, he could be aroused suffi- ciently to answer questions, complained of great thirst, his mouth and tongue feeling " parched, " and of numbness of the limbs. His pulse was 70, regular, respiration normal, pupils somewhat contracted. He was given calomel and castor oil, and on the following morning had quite recovered. Case XVI. — Reported by Sawyer." Boy, aged 3, said to have eaten five large nut- megs. Two or three hours later complained of feeling dizzy and soon fell asleep. Tem- perature and respiration were normal, pulse regular, but slightly slowed, muscles com- pletely relaxed, pupils widely dilated. All attempts to arouse him failed. There was no 354 ON NUTMEG POISONING. delirium at any stage of the poisoning. He remained in this condition about thirty hours and then recovered consciousness as if awakening from natural sleep. Cases XVII and XVIII. — Reported by Hammond.^" Woman took one and one-half nutmegs in hot water to produce abortion. In two hours had severe burning pain in stomach, followed by a dreamy, semi-conscious condition, which gave way to complete unconsciousness lasting six hours. She was kept awake by repeated violent shaking, Later, she vomited and was given whiskey. In ten hours was practically well. Woman took three nutmegs to produce abortion. In two hours had intense pain in the epigastric region and some nausea, but no vomiting. Later, she became entirely unconscious and could be aroused only with the greatest difficulty; respiration 9, pulse 100 and feeble, temperature 100° F. The eyes were turned upward and bulging, the pupils were equally dilated, the extremities were cold, and both nails and lips purple. There was relaxation of the sphincters of the bladder and rectum. Vomiting was induced by zinc sulphate and ipecac, and strychnine and brandy were freely administered. Con- sciousness slowly returned, and in twelve hours after the poisoning had occurred there was complete recovery. No abortion occurred in either case. Case XIX. — Reported by Reading.^' Woman, three months pregnant, took three nutmegs to produce abortion. Later, she vomited freely several times and then passed into a condition of low muttering delirium, with occasional silly laughter and with the hallucination that she had two heads. She could be aroused from this condition when sharply spoken to, but immediately relapsed into it. The pulse was 120, of good strength and volume, and the respiration was also somewhat accelerated. The treatment consisted in the administration of chloral, twenty grains being followed by improvement. There was no abortion. Case XX. — Reported by Hogdon.^^ Woman took two nutmegs and after some hours repeated the dose. The symptoms of poisoning came on a few hours after the second dose was taken. The pulse rate increased to 120, the pupils became dilated, and sleeplessness and anxiety were noted. There was no abdominal pain. On the following day the gen- eral condition was about normal, but on the third daj^ the pulse again became accelerated and there was some dilatation of the pupils. Complete recovery then followed. Case XXI. — Reported by Simpson.^ Woman took two grated nutmegs in a small amount of gin, to cause abortion. Patient was found lying in bed in a drowsy condition a,nd somewhat delirious, the delirium taking the form of confusion and mistaking one person for another. She complained of great tightness across the chest and of vertigo and faint- ness on attempting to stand. She had vomited several times. The pulse was 75 and rather feeble; pupils were normal. She was given stimulants and later depressants, and gradually recovered, the weakness and vertigo lasting for two days. There was no abortion in this case. Case A'A'//.— Reported by " H. E. C. F. ™ Man, aged 23, ate four nutmegs. Was seen six hours later when he complained of great drowsiness, thirst, burning sensation in the throat, and weakness and heaviness in the lower extremities. Pulse, 75, full and strong; respiration, 20. There was no pain or tenderness over the abdomen. Was given zinc sulphate, but no nutmeg was found in the vomitus. Caffeine citrate was also given, with strong coffee. The patient then slept for twelve hours and awoke recovered, save for some weakness in the lower extremities. Case XXIII. — Reported by Merritt.^* Woman, aged 27, had been suffering from monorrhagia for two weeks, to relieve which she made a decoction of two grated nutmegs to a pint of water. After drinking one-fourtli of this she passed into a deep sleep lasting about ten hours. On awakening she drank the remaining three-fourths of the decoction. GEORGE B. WALLACE. 355 She again became drowsy and was soon in a stupor from which she could 1)C arouscil with difficulty. She was finally aroused from this and brandy administered. Lalcr, she became slightly delirious; respiration was 42 per minute, with prolonged shallow expira- tion; pulse was about 140, weak and irregular, pupils slow to react to light, and extremi- ties cold and cyanotic. Strychnine sulphate was gi\-cn in repeated doses and recovery was rapid and uneventful. The menorrhagia ceascil on the night of the poisoning and menstruation was normal the following month. Case XXIV. Reported by Daschewski.^* .-V young woman took two nutmegs weigh- ing about 11 g. Giddiness, palpitation of the heart, general weakness and cramps in the extremities appeared in a few hours and terminated in sleep wliich lasted 4S hours. Daschewski estimates the \-olatile oil of two nutmegs at 0.5 g. Case A'A'T'. Reported by Watson." Man took a teaspoonful of mace and in IJ hours complained of pain in the head, nausea, giddiness, and fulness of the head, faintness, cold shi\-ering, and precordial anxiety. The feet were cold, the eyes injected, and there was some mental confusion. Complete reco^■ery after three days. Treatment, emetics and brandy. From an examination of these cases it would appear that nutmeg has a well-marked depressant action on the brain, as is evidenced by the giddiness, inco-ordination, drowsiness, and stupor so often seen. The dehrium and excite- ment sometimes described, point on the other hand to a stimulation, so that there may be a mixture of stinuilant and depressant action. The depression is much the more marked, however, especially after large doses, and in fact the stimulation may almost be considered an exception to the general action. The drug apparently has no abortifacient action, although it undoubtedly is very -widely used by the laity for this effect. There is but one fatal case recorded, although in a number of the others the poisoning was very severe and dangerous. That death has not occurred more often is probably due to the fact that the crude nut was taken in all cases, so that the separation and absorption of the oil was necessarily a slow and gradual process. Then again, vomiting occurred in a large number of the cases, a good part of the poison being thus removed. The rational treatment of the typical poisoning would consist of the removal of the poison from the stomach by means of emetics or the stomach-tube, and from the intestine by such non-irritant purgatives as the saline cathartics, and of such central nervous stimulants as caffeine or strychnine, when symp- toms of depression appeared. Chemistry. The nutmeg fruit contains starches, fats, and albuminous matter in varying proportions, and in addition from three to eight per cent of a volatile oil. After the oil has been extracted with ether from the nutmeg, the residue ad- ministered to animals produces no noticeable effect whatsoever, even when given in enormous doses, while the oil, on the other hand, given in small doses, causes the same symptoms which follow the administration of the whole nut- meg. The volatile oil then is the important constituent of the nutmeg, and 356 ON NUTMEG POISONING. is wholly responsible for the symptoms in the cases of poisoning mentioned above. It resembles other volatile oils in its physical properties, being a clear, sparkling fluid, insoluble in water, soluble in alcohol, ether, chloroform, etc., possessing a burning irritant taste, and an agreeable odor. It has been the subject of a number of chemical researches, but the results obtained have not been in complete accord. The more important results are as follows : Wright^' found the oil to consist chiefly of two terpenes, of the formula CjoHiB. In addition he obtained small amounts of cymol, C10H14, myristicol, CioHijO, a body of the formula C10H13O, and a non-volatile solid residue to which he gives the formula C4oH5e05. Gladstone^" found, in addition to terpenes, a substance C10H14O, which possesses practically the same boiling point as Wright's myristicol, and is probably identical with it. Semmler,^^ in a sample of oil obtained from Schimmel & Co., found only terpenes, distilling at about 50° C. at 8 mm. Hg. pressure. In this oil, he states, were no cymol, myristicol, or body of higher boihng point. In the oil of mace, however, which is obtained from the aril, or integument covering the nutmeg, and is apparently very similar to nutmeg oil in composition, he found a crystal- hne body which he calls myristicin, and to which he gives the formula C12H14O3. His myristicin was in that part of the oil distilling at from 148° to 158° C. at 10 mm. pressure and occurred here with a body giving phenol reactions. The phenol body he destroyed by treating the crude oil with metallic sodium in vacuo, the oil being kept cool during the early part of the reaction and then gradually heated. When the temperature reached 148° to 149°, the myris- ticin passed over as a clear oil, which, when cooled, formed a white crystalline mass. By adding bromine he was able to form a dibrommyristicin, also crystalline, and a number of other substitution compounds were similarly prepared. He gives the structural formula of myristicin as CeHj — QQjj^ C 4H7 In a later paper^^ Semmler speaks of myristicin as being obtained from both nutmeg and mace, so that he evidently obtained it from other nutmeg oils examined. Having determined that the physiological action of nutmeg depended entirely on its volatile oil, I have endeavored to ascertain whether any one of the constituents of this is wholly responsible for the symptoms produced. Accordingly the oil* was subjected to fractional distillation at 14 mm. Hg. pressure, and the action of the different distillates tested on animals. In par- * The oil used was obtained from Schimmel & Co. in 1899. GEORGE B, WALLACE. 357 ticular I have attempted to isolate the myristicin of SemmkM-. The oil was thus divided into different parts, boihng at about 50°, 68°, 110°, 122°, 135°, and 150° C. Each of these was redistilled at least twice, in order to make as complete a separation as possible. "When these wore administeretl to cats, which give the most typical symptoms with the oil, it was found that the two lowest distillates produced practically no symptoms, while the others showed a tox- icity increasing with the boihng point, and greatest in the highest distillate, 150°. Thus, 5 cc. of the highest distillate per kg. weight of animal, produced in a cat restlessness, salivation, inco-ordination, with subsequent depression and death in two days. A similar amount of a lower distillate, 110°, caused some inco-ordination, but not serious poisoning, and recovery was complete after a few hours. The other two distillates, 122° and 135°, brought on symp- toms more marked than the 110° distillate, but recovery always occurred. The poisonous principle is thus obviously not a terpene, but possesses a much higher boihng point. As it is extremely difficult to separate completely these higher boihng substances by distillation, it is probable that enough of the highest distillate (150°) remains with the lower ones to account for the symptoms produced by these. It seems possible, however, that these possess a certain toxicity, similar in kind to that of the highest distillate, but much less in degree. Were myristicin present in this oil, it would be found in that part boiling at about 150°. This distillate composes about twenty-five per cent of the total oil, while the lowest distillates, the terpenes, form about fifty per cent. The highest distillate then was redistilled several times until a constant boiling point, 149° at 14 mm. pressure, was obtained. A long series of combustions of this substance were performed of which the following may serve as exam- ples: I. 0.20915 g. gave C02.5255 g. and HjO.llSl g. corresponding to 68.52 per cent C. and 6.27 per cent H. II. 0.2096 g. gave C02.5298 g. and H2O.II955 g. corresponding to 68.88 per cent C. and 6.33 per cent H. III. 0.2633 g. gave CO2.66OI g. and H2O.I535 g. corresponding to 68.37 per cent C. and 6.47 per cent H. Average 68.59 per cent C and 6.36 per cent H. Semmler's formula for myristicin requires 69.9 per cent C and 6.8 per cent H, so that at first I was of the opinion that the poisonous principle was identi- cal with myristicin. The presence of a small amount of impurity would then explain the divergence in my analyses from those of Semmler. But every attempt to reach a closer approximation failed. Semmler found a phenol associated with myristicin, but no phenol reactions were obtained from my substance. Treatment with metallic sodiuin in vacuo also failed to change the combustion results. M3Tisticin crystallizes very readily, even to some 358 ON NUTMEG POISONING. extent from the oil when exposed to cold, but I could obtain no crystals, either from the various specimens of oil examined nor from the high distillate. Semm- ler also found a dibrommyristicin by saturating myristicin with bromine, but I found that when bromine was added to the high distillate, that con- siderably more of it was absorbed than was necessary to form a dibrom com- pound, and that the result was a dark brown product, from which nothing could be obtained on distillation except a few drops of clear fluid, which proved to be identical with the high distillate. Semmler found further, on distilling the crude oil with metaUic sodium in vacuo, that the myristicin crystallized out in the condenser, but on following his procedure exactly, I obtained no crystalline substance, but the usual fluid high distillate. Thus, though the boihng point and the organic analysis point to myristicin as the poisonous principle in nutmeg oil, the other facts lead me to believe that Semmler's myristicin and the phenol body were both absent in the specimens examined by me. The isolation of myristicin is so simple, according to its discoverer, that it seems quite impossible that I could have overlooked it. It is more probable that while it is present in some oils, it undergoes some modi- fications in others, leading to a poisonous principle which is closely related to it but is somewhat more highly oxidized, is non-crystalline, and fails to form definite bromine compounds. The further investigation of this question, which is purely a chemical one, I may leave to professional chemists. Pharmacological Action. That nutmeg acts as a depressant on the nervous system has long been recognized. Thus, the older writers on materia raedica — Murray,* Rumph, Pareira, Gmelin, and others — describe its narcotic properties, and it was at one time recommended as a soporific. Purkinje'^ made himself the subject of an experiment to determine its action, eating three nuts, which produced marked nervous depression amount- ing to stupor. In 1851 Mitscherlich^* conducted a number of experiments with the oil of nutmeg on rabbits. He found that 4 cc. of the oil given to a small rabbit by the stomach caused death in thirty hours. The symptoms were a decrease in the pulsations and strength of the heart, difficult breathing, diminished heat of external parts, and finally death without convulsions. With smaller doses there were frequent and powerful pulsations of the heart, slight acceleration of respiration, at first restlessness, later muscular weakness, little or no loss of sensibility. He states that the oil is a strong irritant and causes inflamma- tion of the stomach and jejunum. * Murray refers to an old legend in the following words : " Avi quoque paradisiacae in deliciis est, quae vero inde inebriatur et mortua decidit, quibus factis formicae adrepunt et • pedes eonsumunt." (Appar. Medicam. VI, p. 138, 1792.) GEORGE B. WAT,LACE. 359 Cad^ac and Meunier'^ found that when the oil of nutmeg is injected intra- venously it produces quite marked symptoms of poisoning, these consisting of inco-ordination and ataxia, tremor of the entire body, depression of the nervous system, and finally complete unconsciousness, with loss of reflexes. During the earlier stages of the poisoning there is a period of quite noticeable respiratory stimulation. The heart and circulation appear to be but little affected, there being only a slight weakening of the heart. In some experiments performed by Wood'" results similar to those of Cad6ac and ileunier were obtained. In most of my experiments the highest distillate (149° C. at 14 mm. pressure) has been used, since it acts in much smaller quantities and is more easily ad- ministered than the crude oil. If a frog is placed in a dish containing water to which a few drops of the distillate has been added, there is seen first a short stage of restlessness coming on in three or four minutes, during which the animal moves about continuously and tries to escape from the dish. The reflex activity is sUghtly increased at this time, but not markedly so. After this condition has lasted about five minutes the frog ceases its movements, sitting quietly most of the time with its head depressed. The stage of quietness passes into one of depression, which, finally, after an hour or so deepens into paralysis, with abolition of reflexes. If the frog is removed from the dish as soon as paralysis is evident, the depression gradually passes away, and after about twelve hours the frog has regained its normal condition. If allowed to remain in the dish, however, death invariably results. The following experiment illustrates these changes : Frog (Jtana temporaria). 3 P. M. — Placed in dish containing 15 cc. water to which five drops 149° distillate had been added. 3 :05 — Frog is moving and jumping about continuously, and attempting to escape from dish. 3 :10 — Is quiet, with only an occasional movement. 3:15 — Quite noticeably depressed. Can be handled without attempting to escape. Assumes normal position, however, when placed on back. Reflexes slightly weakened. 3 :30 — Depression greatly increased and frog remains on back when so placed, although making feeble attempts to regain normal position. Reflexes much weakened. 4 :00 — Complete loss of voluntary and reflex movements. Stimulation of sciatic nerve by electric current followed by practically normal response, showing that motor nerve ends and muscles are unaffected by the poison. If, instead of administering the oil as in the above experiment, it is injected into the abdominal lymph sac of the frog (two drops being sufficient) the same symptoms are induced, excepting that the restlessness is not as well developed. If a ligature is tied around one of the legs, excluding the sciatic 360 ON NUTMEG POISONING. nerve, the loss of reflex activity occurs as quickly here as in the leg to which the poison has access, showing that the paralysis is wholly central. The effects of the oil on the frog's heart are of but little importance when compared to those on the nervous system, for although an exposed heart is literally bathed in the oil, it still continues to beat for a number of hours. The changes in detail are as follows : Frog, with brain destroyed and heart exposed. Time. Beats per minute. 2 p. M 43 2 :10 44 1 drop of oil (149°) placed on heart. 2:16 44 2 20 44 2 drops applied. 2 :33 37 2 drops applied. 2 :40 34 2 drops applied. 2 :45 33 5 drops applied. 3:00 29 3:30 26 4:35 19 In this experiment twelve drops were applied directly to the heart without stopping it, whereas two drops are sufficient to completely paralyze the central nervous system. This lack of effect cannot be because the oil fails to pene- trate into the heart, for when applied in alcoholic solution or in an acacia emulsion it produces no greater effect. The action of the oil on the frog then consists of a weak stimulation followed by a depression and paralysis of the nervous system, with a slight cardiac depression. The short period of stimulation is probably due more to local irritation than to any central nervous action, since the oil possesses quite marked irritant properties. The depression and paralysis, on the other hand, are due wholly to a direct action on the central nervous systerfi. These changes" in the frog produced by oil of nutmeg are quite similar to those caused by other volatile oils. For example, Heffter^' found that safrol, the active constituent of oil of sassafras, produced the same slight weakening of the heart and depression and paralysis of the central nervous system, while the oil of pulegium has been shown by Lindemann ^' to possess a similar action. The last-named oil has, however, in addition a curare-like action on the motor nerve ends, not possessed by either of the other two. The action on mammalia is very similar in kind to that on the frog. In the rabbit there occurs a condition of nervous depression and inco-ordination which gradually deepens into complete unconsciousness, the heart, although somewhat weakened, being not seriously impaired until near the end. Tremors, as mentioned by Cadeac and Meunier, I have not found to occur constantly, being quite marked in some cases while not noticeable in others. In cases in which these were present, sudden starts, such as a normal animal makes if unexpectedly struck, occur at short intervals and seem to be the forerunners GEORGE B. WALLACE. 361 of the tremors. The following experiment shows the typical symptoms induced : Rabbit; weight, 1,500 g. 1 :30 P. Ji. — Administered by stomach tube 5 oc. of distillate at 149° in emulsion with acacia. 2 K)0 — Rabbit is more quiet and drowsy than normally, malting almost no spontaneous movements. Sudden starts occur at intervals of two or three minutes, such as would be seen if the animal were suddenly aroused from light sleep. 2 :15 — The movements are more clumsy and sluggish. .\ tremor is noticeable, especially of the head. The animal remains lying on its side when so placed. 2:30 — Depression is quite evident, although complete unconsciousness is not yet present. 3 :30 — ^Narcosis has deepened, the rabbit lying on its side. The corneal reflex is present, but not as active as normal ; the heart seems unaffected, but the rate of respiration has fallen from a normal of about sixty per minute to thirty-eight, and the breathing is somewhat labored. 7:30 — Complete unconsciousness; corneal reflex absent; respiration very slow and weak (twelve per minute) ; the heart seems now much weakened, the number of beats per minute having fallen from 150 to 72. The rabbit died about 9 p.ii. Of the mammalia the cat seems the most susceptible to the action of the oil, and doses which in the rabbit would produce no marked symptoms, cause in this animal fatal poisoning. If a sufficient dose of the oil (about 0.4 g. of the 149° distillate per kg.) is given to a cat by the stomach, symptoms of nervous stimulation appear within about ten minutes. These consist of a state of restlessness, the head being thrown excitedly from side to side, while the animal keeps moving about in the cage, and of a tremor, beginning in the head and later spreading over the whole body, resembling closely the tremor seen in carbolic acid poisoning. A profuse salivation is an invariable accompaniment. Within a short time a period of quiet replaces the excitement stage, the cat l}dng down with its head resting on the floor. The tremors persist but the salivation decreases. When the cat attempts to walk, inco-ordination is apparent, the animal staggering about and frequently losing its balance and falling to the floor. A mydriasis of varying degree is usually present at this time. The reflexes, although not absent, become much weaker than normally, and the cat pays little attention to painful stimulation, to which, in its natural condition, it would readily respond. Soon a condition of semi-consciousness sets in, during which the heart is not much affected, while the respiration is rather labored and feeble. In some cases the unconsciousness deepens, the respiration gradually becomes weaker, the heart shows signs of impairment, and death finally comes in from eight to twelve hours after the ingestion of the poison. As a rule, however, after the stage of unconsciousness has been reached and has persisted for a few hours, a gradual improvement in the cat's condition 362 ON NUTMEG POISONING. sets in, and in about fifteen hours after the oil has been taken the animal seems to have regained its natural condition, save for a state of abnormal quietness and disinclination to move about. This improvement is but tem- porary, however, for the cat becomes gradually weaker and more depressed, eating nothing and paying but little attention to its surroundings, until coma comes on, followed by death in from thirty-six to seventy-two hours after the oil was taken. Other symptoms occasionally seen are vomiting and purging which may come on at any stage of the poisoning. A blood pressure tracing shows that the oil has no greater effect on the mammalian circulation than on the amphi- bian. A cat was anesthetized with chloretone, a mercury manometer connected with the carotid artery, and the oil, in a fine emulsion with acacia, injected into the external jugular vein. The result is shown in the following: Time. Pressure. .0.1 CO. distillate of 149° injected. .0.2 cc. .0.2 cc. .0.5 cc. . respiration very weak. . respiration ceased at this point, while the heart continued to 3 :05 170 mm. Hg 3:20 170 3 ;25 160 3:35 160 3:45 Ill 4:00 80 beat for a short time afterwards. The fall in blood pressure appears, then, to be of no great importance in the poisoning, a fairly good pressure remaining till the breathing had stopped. At one stage in this experiment, immediately after the 0.5 cc. had been injected, there was noted an increase in the rate of respiration, which was, however, of short duration and not as well marked as the acceleration men- tioned by Cadeac and Meunier. If the oil is given by subcutaneous injection instead of by the stomach, the same symptoms are induced, save that salivation does not occur and the stage of excitement is not so well developed. If death is sufficiently delayed, an abscess usually forms at the point of injection, owing to the irritant action of the oil. Without entering into detail, it may be stated that in the dog, the same symptoms of inco-ordination and ataxia, drowsiness, tremor, and finally unconsciousness are seen. A condition of restlessness or excitement during the early poisoning is commonly but not invariably present, while, in addi- tion, vomiting and purging usually occur and are quite severe. In summing up the characteristic effects of the oil on mammaUa it is found that the typical symptoms produced are a short stimulation, followed by inco-ordination, muscular weakness, tremor, and unconciousness, with the heart and circulation not markedly affected and death resulting from paralysis of the respiratory center. The stimulation, as was stated in the case of the frog, is probably not so GEORGE B. WALLACE. 363 much of central origin as of a reflex natun-, due to the irritant effect at the point of appHcation; this irritant action also explains the salivation and the vomiting and purging. It is possible, ho\v(n-er, that then^ is some direct stimulation of the cerebrum, and the acei^kn'ation of the respiration would point to this action on the medulla. Thi- entire stimulation, however, is of very slight degree. The tremor of carbolic acid poisoning is commonly stated to be due to stimulation of some indefinite part of the mid-brain, and this explanation might be applied to the tremor of nutmeg poisoning, esjx'dally since there is no fall in temperature to account for it. It is not inconceivable, however, that the tremor is due not to stimulation of the mid-brain, but to a derangement of function of this part through a depressant action. The inco-ordination, muscular weakness, drowsiness, and stupor, are brought on through purely depressant effects on the central nerA'ous system. Pathological Changes. From the marked fatty degeneration, occurring after safrol,'' oleum pule- gii," thymol,*" oleum rosmarini," and other volatile oils, it might be ex- pected that the closelj^ related oleum myristicse would induce a similar tissue change. Were this the case it would explain the effect on the cat, where, after the more acute symptoms have passed off, a stage of apparent recovery is noticed, to be succeeded by a general depression and death. A post-mortem examination shows that some fatty degeneration is present in the liver and kidney, but in a slight degree only, and in nowise comparable to the phos- phorus-hke effects produced by the other oils just mentioned. Accompanying this fatty degeneration is a slight cloudy swelling. The heart, spleen, and lungs do not partake in these changes. The local action of the oil on the stomach and intestine is more prominent when the oil is given by the mouth. These organs often contain a bloody, mucus-like substance, and their walls show areas of severe inflammation and blood extravasations. Small areas of necrosis and ulceration may also be seen. The urine is somewhat diminished in amount and contains considerable quantities of albumen. After large amounts of the oil have been given, the urine contains also a substance, not sugar, which reduces FehUng's solution and turns the plane of polarized light to the left. This is probably glycuronic acid, since most of the volatile oils are excreted as glycuronates. Although these pathological changes are not in themselves sufficiently marked to be the cause of death in the more chronic poisoning, they undoubted- ly aid greatly in producing the general weakness and depression. Death in these cases is due then to the combined effects following the pathological changes and a direct depressant action of the oil on the central nervous system. In conclusion, I wish to express my gratitude to Professor Cushny, whose suggestions and assistance have been invaluable. 364 ON NUTMEG POISONING. REFERENCES. 1. Lobelius — Stirpium historia Observationes, 1576, p. 570. 2. Hughes. — Manual of Pharmacodynamics, 1880, p. 695. 3. Paullinus. — Nucis moschatse curiosa descriptio, Leipsig, 1704, quoted by Lewin, Nebenwirkungen der Arzneimittel, 1893, p. 728. 4. Cullen.— Treatise of Mat. Med., 1812, Vol. II, p. 145. 5. Barry.— St. Louis Clinical Record, 1879-80, p. 133. 6. Matthews.— Prakt. Artz., Wetzlar, 1880, p. 125. 7. Gaulke.— 76id., p. 217. 8. Howard.— Phil. Med. Times, 1880, p. 438. 9. Palmer.- Am. Joum. Pharm., Vol. LVII, p. 23. 10. Pinnock. — Australas. Med. Gaz., 1887, p. 274. 11. Dodge.— N. Y. Med. Record, 1887, p. 624. 12. CarveU.— Brit. Med. Joum., 1887, I, p. 1317. 13. Gillespie.— Phil. Med. Times, 1887, p. 726. 14. Air.— Brit. Med. Joum., 1887, I, p. 1201. 15. Alexander.— Zfeid., 1887, I, p. 1085. 16. Tyler.— /6td., 1887, I, p. 1201. 17. 'Delvin.—Ibid., 1887, I, p. 1201. 18. Bentlii.—Ibid., 1889, II, p. 1389. 19. Sawyer.— N. Y. Med. Journ., Vol. L, 1889, p. 354. 20. Hammond.— Med. News, 1891, Vol. LVIII, p. 580. 21. Reading.— Therap. Gaz., 1892, p. 585. 22. Hogdon.— Amer. Med. and Surg. Bull., 1892, p. 1492. 23. Simpson. — Lancet, London, 1895, 1, p. 150. 24. H. E. C. T.— Guy's Hosp. Gaz., 1898, p. 535. 25. Merritt.— Amer. Medicine, 1901, p. 684. 26. Manquat.*— Traits 4i6m. de Th^r., 1900, Vol. I, p. 622. 27. Watson.— Provincial Med. and Surg. Journ., 1848. Vol. XII, p. 37. 28. Daschewski. — Yuzhno russk. Med. Gaz., 1895, p. 191, cited from Pharm. Ztschr. f. Russland XXXIV, p. 264, 1895. 29. Wright.- Journ. Chem. Soc, 1873, p. 549. 30. Gladstone. — Journ. Chem. Soc, 1872, p. 1. 31. Semmler.— Ber. d. d. chem. Gesel., 1890, p. 1803. 32. Semmler.— /6id., 1891, p. 3818. 33. Purkinje.— Quoted by Vogt, Pharmakodynamik, 1832, Vol. II, p. 638. 34. Mitsoherlich. — Buchner's Rep. d. Pharm. XVI, 1851, p. 104. (Abstract) Amer. Journ. Pharm., 1851, p. 128. 35. Cad6ac et Meunier. — Journ. d. Med. Vet. Zootech. Lyon, 1890, I. XV, p. 5. 36. Wood. — Therapeutics; Its Principles and Practice, 1900, p. 615. 37. Heffter.— Arch. f. exp. Path. u. Pharm., Bd. XXXV, S. 342. 38. Lindemann.— /6id., Bd. XLII, S. 356. 39. Falk.— Therap. Monatsheft., Bd. IV, S. 448. 40. Husemann. — Arch. f. exp. Path. u. Pharm., Bd. IV, S. 280. 41. Schreiber. — Quoted by Lewin, Lehrb. der Toxikol., S. 289. 42. Van Bosch. — Quoted by Husemann Handb. der Arzneimittellehre, 1874, p. 590. 43. J. F. Gmelin. — Allgemeine Geschichte der Pflanzengifte, Zweite Auflage, Niirn- berg, 1803, p. 488. * A case of poisoning is quoted which is probably that given under reference 9. A CONTEIBUTION ON THE MORPHOLOGY OF SUDORIP- AEOUS AND ALLIED GLANDS, G. CARL HTJBEE., M.D., Junior Professor of Anaiomy and Director of Histological Laboratory, University of Michigan, and EDWARD WILLIAM ADAMSON. {From the Histological Laboratory, University of Michigan.) A number of investigators have in recent years made use of the Born wax- plate method of reconstruction in ascertaining the exact form and relation of certain anatomic structures in the adult too small to be isolated by the ordinary technical methods, and yet too large and too complicated to be fol- lowed in serial sections. We may mention the contributions by Miller on the form and structure of the terminal air passages and the relation of the blood- vessels thereto, the important work of Dr. Florence R. Sabin on the structure of the medulla oblongata, and the investigations of Maziarski, whose careful work has thrown new light on the much-debated question of the form and relationship of the secretory compartments of the salivary and other com- pound alveolar and tubular glands, and Johnston's careful study of the arrange- ment of the terminal arteries and blood-capillaries of the glomeruli of the kidney. The figures accompanying these contributions, illustrations of the models obtained in the course of the investigations, offer more convincing proof of the value of this method and its wide application in anatomic research than would be done by extended discussion. The literature bearing on the wax-plate-reconstruction method per se, and the steps of the method are so well Icnown that further reference to it is deemed unnecessary, especially since Bardeen has quite recently drawn attention to the value of the rnethod in anatomic and embryologic investigation and has described in detail the steps to be pursued. In our investigation we reconstructed sudoriparous glands from the foot, pubicandaxillaryregionsof the human adult, and, to show developmental stages, several glands from the plantar region of human embryos of various ages; also so-called modified sweat-glands from the circumanal and the axillary regions, ceruminous glands from the external auditory canal, and glands of Moll from the eyeUd. The tissues employed were fixed in formalin, dehy- drated, and carefully imbedded in paraffin. Attention need hardly be drawn 365 366 MORPHOLOiJY OF SUDORIPAROUS AND ALLIED GLANDS. to the fact that, since one of the requisites of wax-plate reconstruction is a perfect series of sections, too much consideration cannot be given to a thorough imbedding of the tissues to be sectioned. In our work, the cutting of serial sections was facilitated by the employment of a method now much used in this laboratory when cutting thin sections of tissues imbedded in paraffin, which on being cut are prone to become torn or distorted, or which, by reason of hard- ness, are difficult to cut: namely, imbedding tissues in relatively hard parafhn (Griibler's 58° C), and while cutting, keeping a layer of distilled water on the knife. By means of this method we were able to make uninterrupted series, numbering from 300 to 600 sections. In certain of our series the sec- tions measured 5 ii in thickness (sections of embryonic tissue and of the skin of the plantar surface of the foot of a full-grown individual). The sec- tions of the remaining series measured 10 ij- in thickness. After selecting the gland to be reconstructed in &ny one series, the various sections of the gland in question were sketched with the aid of the camera lucida. The magnification used was either 400 or 200 diameters, the former when using sections of 5 ii- thickness, the latter when employing sections of 10 ij- thickness, since in all our work we made use of wax plates of 2 mm. thickness, and it was necessary to maintain an equal ratio between the diameter of the magnification of the drawings and the sections, and the thickness of the plates used and the thick- ness of the sections. The sketches on the paper were then transferred to wax- plates, after which the parts representing the sections were cut out, piled in series, and fastened together by means of pins and nails, and finally fused by means of heated modeling tools. The figures accompanying this article are, in part, free-hand drawings, controlled by measurements, and in part copies of photographs. They all show a uniform magnification of 100 diameters (except Figs. 9 to 12, which show a magnification of 66 diameters). A study of the figures will, therefore, afford a ready means of comparison of the rela- tive size of the glands from different regions, the relative size of the secreting portion of the tubules of the glands studied, as also an exposition of the form of the glands investigated. The large number of illustrations given seemed necessary to a clear understanding of the subject under consideration. Sudoriparous Glands — (Sweat-Glands) . In a Latin dissertation, published in Breslau, in 1833, and in an article which appeared the following year in Miiller's Archiv., Wendt mentions the fact that Purkinje had observed in preparations of the skin hardened and cleared in " Liq. Kali carbon. " certain spiral threads (Spiralfaden) occurring in the epidermis. Wendt, in tissues prepared in the same manner, was able to confirm this observation, and extend it in so far that he was able to trace the spiral threads beyond the rete Malpighi into the cutis, where they had a varied form and course. Certain of the threads observed in the G. CARL HUBER AND EDWARD AMLLIAM ADAMSON. 367 cutis were perpendicular to the surface, others had a spiral course. The ends of the threads were found in the deeper layers of the cutis, somewhat enlarged, bent or otherwise formed. Wendt recognized a lumen in the thread and spiral and regarded these structures as the sweat-glands. It is evident from the description and figures given (especially Fig. 1, Plate I), that the observer saw only the ducts of the glands in question. Almost contemporary with the above observation, there are recorded by Breschctand Roussel de Vauz^me the results of their investigation on the skin, made on tissues treated with hot water or by maceration methods. These investigators observed the spirals in the epidermis, their extension into the cutis, and their termination in glands, which were, however, not fully described. The value of their observa- tions suffers by the fact that they described a special variety of glands, forming a mucous secretion which was thought to cornify to form the epidermis. Even though their work contains many errors, it would seem just to credit their publication ^dth containing the first account of the sweat-glands. An article published by Gurlt gives the first correct account of the shape of the sweat- glands. He states that these glands lie in the deeper part of the cutis, extend- ing into the subcutaneous adipose tissue. They are described by this observer as consisting of a tube repeatedly twisted ("einem vielfach gewrmdenen Schlauche"). He studied the glands in man, horse, sheep, pig, and dog, and made interesting comparisons of the shape of the glands as observed in the different animals studied. The figures found on Plates IX and X, accompany- ing his article are, we believe, the first giving a relatively correct conception of the shape and structure of the sweat-glands. Krause, in a most interesting and complete account of the structure and physiology of the skin, discusses at some length the shape, structure, distribution, and function of the sweat- glands. He speaks of them as consisting of a simple tube and differentiates between the secreting portion and the duct; he also discusses their cellular fining. Hassal, on the other hand, speaks of the sweat-glands as consisting of a tube, which anastomoses at variable distances, thus forming a meshwork of loops, aU emptying into a common duct. V. KoUiker is given priority in calUng attention to the layer of non-striated muscle cells found internal to the membrana propria of the secreting portion of the tubule of nearly all sweat-glands. Reference may, also, be made to the contributions of Horschel- mann and Heynold. The latter emphasized the fact that the duct extends for a considerable distance into the coil of the gland ("betheiligt sich der Gang auf eine lange Strecke an der Bildung des Drusenknauels"). He also drew attention to the cuticular lining of the duct. It is, however, not our purpose to enter at length on a discussion of the Hterature bearing on the structure and anatomy of the sweat-glands, since the majority of observers who have investigated these structures have concerned themselves mainly with the microscopic anatomy and only incidentally with their shape and the 368 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. arrangement of the tubules. The citations made were inserted by reason of their historic interest. In a sudoriparous gland there are recognized a coiled portion, comprising the secreting portion of the tubule and a portion of the excretory duct, and situated in the panniculus adiposus, just beneath the cutis, and a slightly wavy portion which passes through the cutis to reach the epidermis, and is recognized as excre- tory duct. Above the rate Malpighi, the duct loses its distinctive cellular lining, passing through the horny layer of the epidermis as a spiral canal of from one, two to twenty turns, to terminate on the free surface in a sweat pore. The entire gland, as is well known, consists of a single tube. Now and then, though very rarely, two glands empty into one diict. The coiled portion of the gland presents in sections a round or oval form. Single sections give, however, an inadequate idea of the arrangement of the tubule in the coil, and Fig. 1. — (A B.) Two A-iews of a model of the coiled portion of a sudoriparous gland from the plantar region of a man. even in serial sections it is diffictolt to follow the different portions of the re- peatedly cut tubule, through a sufficient number of sections to enable the observer to formidate a mental picture of an entire gland. Our reconstruction of these glands was undertaken with a view of obtaining certain data concern- ing the shape of the gland and the arrangement of the tubule in the coiled portion not fully determined by a study of sections. It was hoped that with the aid of models light might be thrown on the following points, which suggested themselves as of sufficient importance to deserve investigation : 1. To ascertain the length and the arrangement of the tubule in the coiled portion of the gland and to determine, if possible, from a study of several models, whether any particular arrangement might be regarded as funda- mental or typic. 2. To ascertain the relation of the end of the tubule to the duct. 3. To determine the extent of the participation of the excretory duct in the formation of the coil. 4. To ascertain the manner of the development of the coiled portion of the gland. G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 369 5. It occurred to us that such models would serve as a basis for more accurate illustrations of the coiled portions of the sweat-glands than the Utera- ture now affords. The figure so frequently copied from Todd-Bowman is obviously diagrammatic and incorrect, and the figure (Fig. 84, Band V. The Anatomy of the Skin, von Brunn.) found in von Bardeleben's Handbuch of Anatomy, one of the most recent contributions on the anatomy and structure of the skin, must be regarded as semi-diagrammatic. In that portion of the figure giving the coiled portion of the gland, the arrangement of the tubule cannot be regarded as normal, as the figure was drawn from an isola- tion preparation cleared in acetic acid. Our own results, as above stated, are based on a study of models obtained by reconstruction. In Fig. 1 are represented two views of a model of a "sweat-gland from the plantar region of an adult. This gland was regarded as typical and of average size. The group of tubules constitut- ing this gland, as seen in sections, was of oval shape, and was readily traced through a series of fifty-eight sections. Measurements made on the model give this gland a long diameter of 0.38 mm. V. Kolliker gives the average size of a sweat-gland as 0.3 to 0.4 mm. This model, as also those of the other sudo- riparous glands studied, shows conclu- sively that these glands consist of a repeatedly coiled or twisted tubule. The length of the tubule in the coiled portion of the gland shown in Fig. 1 is 4.25 mm., of which 1.25 mm. fall to the excretory duct and 3 mm., to the secretory tubule. The excretory duct, which is readily recognized by its epithelial lining, forms an essential portion of the coil of the gland; in certain sections of the series it was cut as often as six to eight times. The greater portion of the duct is situated inside of the coil (Fig. 1, A), and comprises eight distinct loops. The secreting portion of the tubule forms six primary loops, each of Fig. 2. ■Coiled portion of sweat-gland from pubic region of a woman. 370 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. which presents secondary undulations. Four of the primary loops are seen in B, Fig. 1. The transition from the excretory duct to the secreting portion of the tubule is quite abrupt, see Fig. 1, A. In this, as in the other models of the sweat-glands made, the end of the tubule is situated near the entrance of the excretory duct into the coil. This relation of the terminal part of the tubule is of interest in so far as it suggests the manner of the development of the coiled portion of these glands, as will be discussed more fully in considering the models obtained by reconstructing sweat-glands from embryos. In Fig. 2 we represent a gland from the hairy portion of the pubic region of a woman. This gland measures 0.9 mm. in its longest diameter, 0.5 mm. and 0.25 mm., respectively, in its other diameters. The gland was situated Fig. 3. — Coiled portion of sudoriparous gland from axillary region of a man. between the sebaceous glands of two neighboring hair follicles, and was rather sharply bounded by two relatively large trabeculse of fibrous connective tissue. The tubule comprising the portion of the gland reconstructed measures 10.4 mm., of which 4.4 mm. is duct and the remaining 6 mm. secreting portion. The duct passes along one side of the gland to its lower end, after which it forms an intricate series of loops, only a few of which are seen in the figure, and not all of which are seen in the model. The loops of the duct form a compact mass surrounded by the loops of the secreting portion of the tubule, except on the side shown in the figure. The secreting portion of the tubule forms primary loops, each possessing a number of undulations. One of the larger loops is directed toward the epidermis and for a distance is parallel to the duct, as is shown in the figure. Fig. 3 represents the coiled portion of a sudoriparous gland from the axillary region of an old man. It was situated in the immediate vicinity of one of the large axillary glands. The entire gland forms an irregular ovoid mass and measures 0.9 mm. in its long diameter. The tubule of the coil measures 9.9 mm., of which 4.4 mm. is duct and 5.5 mm. secreting portion. The duct, after entering the coil, forms numerous compact G. CARL RUBER AND EDWARD WILIJAM ADAMSON. 371 loops, in part surrounded by the loops of the secreting portion; tliis latter portion consists of six primary loops, each of which shows secondary foldings. From a study of the glands here reproducetl, and from several others inves- tigated, we were not able to recognize any typic arrangement of the loops forming the coiled portion of the sweat-glands, either as to the number of the primary loops or as to their disposition. It seems to us, however, that we are warranted in concluding that the shape of the coiletl portion of the gland, as also, to some extent at least, tl\e arrangement of the primary loops, is depend- ent on the relation and proximity of the gland to the neighboring structures, such as hair follicles and sebaceous glands, larger blood-vessels, larger tra- beculse of connective tissue, and also neighboring sweat-glands. As concerns the latter, it may be stated that it is now and then exceedingly difficult in a series of sections to trace with certainty a single gland, especially toward the begin- ning and end of a series of sections of any one gland. Loops of neighboring glands are often in close proximity and are apparently surrounded by a common laj-er of somewhat denser connective tissue, so that a separation of the tubules belonging to two or, now and then, even three contiguous glands can be made ^"ith certainty only by reconstruction. The close relation of the end of the tube to the place of entrance of the excretory duct into the coiled portion of the sweat-glands suggests that the coil of the gland is not formed by an onward growth of the end of the tubule during its development, but rather that the end of the gland, after reaching the stage of development when it is directed toward the epidermis (see Fig. 5), retains a relatively fixed position, and that the coil is developed by the forma- tion of primary and secondary folds, or knucldes, proximal to the first-formed loop. A comparison may here be drawn between the development of the sweat- glands and that of the urinif erous tubules ; in the latter, the end of the tubule is early retained in a relatively fixed position by the glomerulus of the Malpighian corpuscle. Sweat-glands from several fetuses and from a child a few days old were reconstructed to ascertain, if possible, the manner of the development of these glands. The material at our disposal was not large enough to give us a complete series of stages. The stages studied may, however, receive brief consideration. As is well known, the sudoriparous glands develop as solid buds from the germinal layer of the epidermis. According to v. Kolliker and Minot, they appear on the soles of the feet in the early part of the fifth month. During the fifth month, this epidermal ingrowth reaches the subdermal tissue and becomes folded upon itself, the club-shaped end now pointing toward the epidermis. In Fig. 4 are shown two developing glands from the skin of the plantar region of a fetus about five and one-half months old. The gland anlage, at this stage, as shown in the figure, is nearly straight, with a club- shaped enlargement at the end. In Fig. 5 are shown the sweat-gland anlagen from the same region of a fetus about six months old. They show three stages 372 MORPHOLOGY OP SUDORIPAROUS AND ALLIED GLANDS. in the development of the primary loop, or crook, and further, and especially in A and B of this figure, there may be observed just above the terminal loop, or crook, a deviation from the long axis, suggesting the formation of knuckles. A similar condition is seen in the gland anlagen shown by Minot in Fig. 314 (section of the sole of the foot of a fetus of the fifth month), of his work on Human Embryology. By the end of the eighth month the coiled portion of the gland is readily recognized, as may be seen from a study of Fig. 6. In the two glands reproduced in the figure, the tube of the' coil measures 0.75 mm. Fig. 4. — The anlagen of two sudo- riparous glands from the plan- tar region of a human fetus of five and one-half months. Fig. 5. — (A B G.) Three developing sudoriparous glands from the plantar region of a human fetus of six months. In both glands, the end of the tubule is seen near the end of the relatively straight portion of the glands (duct.) Two primary loops may be recognized in each gland, each loop having one or more secondary folds. The close relation of the end of the tubule to the duct is clearly shown in Fig. 7, repre- senting models of two sweat-glands from the plantar region of a child about one week old. The tubule of the coiled portion of A of this figure measures 1 mm. and of B, 1.2 mm. V. Kolhker gives, as the length of the tubule of the coil of sweat-glands from the heel of the new-born, 0.13 mm. to 0.15 mm.; this is obviously an error, probably due to faulty punctuation. The coiled portion of the gland shown in A presents three distinct loops, while in B, four rather irregular loops may be made out. The duct forms in the two glands an inconspicuous portion of the coil. In the series of developing glands studied by us, there is lacldng a stage showing conclusively the formation of loops after the development of the first loop, as shown in Fig. 5. The formation of additional loops is suggested in the gland anlagen shown in this figure, while G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 373 Fig. 6. — (A B.) Two developing sudorip- arous glands from the plantar region of a human fetus of the eighth month. The anlage of the coil and a portion of the duct are shown. Fig. 7. — (A B.) Two sudoriparous glands from the plantar region of a child about one weelc old. The coiled portion and a part of the duct are shown. in the glands shown in Figs. 6 and 7 these are well developed. The close relation of the end jaf the tubule to the duct and the configuration of the loops in the coiled portion of these glands are such that it seems probable that their formation may be attributed to a folding of the tube as it grows in length rather than to an onward growth of the end of the tubule in such directions as would result in the formation of loops. Circumanal Glands. The occurrence in the circumanal region of glands resembling sudoriparous glands, though differing from them morphologically, was first noted by Gay. In his communication on this subject, he states that in an area having the form of an elliptical ring, with a width of 1.25 cm. to 1.5 cm. and situated 1 cm. to 1.5 cm. from the anus, there were found glands of the type of sweat- glands, but differing from them in being much larger. They were designated by him as circumanal glands. The dimensions of these glands, as given by this observer, are about three times that of the ordinary sweat-glands, the tu- bules being proportionately large, now and then showing a lumen 0.1 mm. to 0.2 mm. in diameter. Gay describes their structure as similar to that of the sweat-glands found in other regions, and calls attention to the fact that they possess a well-developed musculature. These observations were confirmed by v. Kolliker and Horschelmann; Heynold; on the other hand, sees no reason why a special group should be made of these glands. Our own observations were made on tissues taken from a woman at autopsy. We are able to confirm Gay's observations to the extent of finding a glandular ring of about the same dimensions and location as given by him. In this area, we find four quite distinct types of sudoriparous glands : 1. Sudoriparous glands like those found in other regions. 2. Large sudoriparous glands, the circumanal glands of Gay. 374 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. 3. Branched sudoriparous glands, tubulo-alveolar glands. 4. Branched sudoriparous glands, with relatively straight duct, ending in a relatively large saccule or vesicle, from which arise secondary tubules and alveoli. The glands of the first type need no further consideration. They extend to the region of the internal sphincter, as has been stated by v. Kolliker. In a series of sections made from a piece of tissue about 4 mm. thick, taken from the circumanal region about 1 cm. from the anus and extending through the glandular ring, we find four glands of the type described by Gay ; Fig. 8. — (A B.) Two \-iews of a model of a relatively small circumanal gland of Ga3^ G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 375 two of these are within the series of sections made, the other two are only partly within the series, the one at its beginning, the other at its ending. In Fig. 8 are shown two views of one of these glands. The gland is of irregular oval shape and measures 0.9 mm. in its long diameter and 0.4 mm. in its short Fig. 9. — (A B.) Two views of a model of a relatively large circumanal gland of Gay. diameter. The gland consists of a tubule, arranged in the form of loops, and has a length of 6 mm. The greater portion of the coil of the gland is made up of the secreting portion of the tubule, which represents about the same diameter throughout, but is somewhat flattened and slightly larger in the last loop. The duct, which scarcely enters the coil, differing in this respect from other sweat-glands, is for a distance lined by cubical epithelium. For the greater 376 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. portion of its course through the cutis, the duct has a lining of low stratified pavement epithelium and a tortuous course. In the course of the secreting portion of the tubule, there are found a number of short, bud-like branches (five in the entire gland), four of which are clearly shown in Fig. 8, B. They possess a narrow lumen and present a structure similar to that of the parent tubule. They have, we believe, thus far escaped observation. Another of the circumanal glands reconstructed by us is shown in Fig. 9. This gland measures 1.5 mm. in its long diameter and 0.5 mm. in its short diameter, and consists of a repeatedly folded and twisted tubule, having a length of 9 mm., of which 1.5 mm. may be regarded as duct (in the coil of the gland). The duct, as shown in A, is comparatively straight, and runs to one side of the. coil. The secreting portion of the tubule consists alp3,ost throughout of irregularly spherical, oval, or cylindric vesicular enlargements, connected by tube seg- ments of much smaller diameter. The vesicular enlargements are readily recognized in sections, and are no doubt those measuring 0.1 mm. to 0.2 mm. mentioned by Gay, while the constricted segrrlents of the tubule appear by comparison as small tubules, often not larger than those found in ordinary- sweat-glands. Circumanal glands of this type may, we believe, be recognized in section by the presence of tubules showing this dissimilarity in size. The larger and smaller segments of the tubules are throughout lined by the same epithelium. In this gland, as in the one shown in Fig. 8, there are found small side branches. A reason for the existence of the vesicular enlargements has not occurred to us. Whether they are to be regarded as a means of ob- taining an increase of the secreting surface, or whether they serve the purpose of reservoirs in which the secretion may be temporarily stored, cannot be decided until more is known concerning their function and the nature of their secretion. We estimate that, on an average, ten of the circumanal glands of Gay are found in one square centimeter. . The type of glands above designated as branched tubulo-alveolar sudorip- arous glands are of especial interest. Little mention is made in the literature of the existence of branched sudoriparous glands. Mention is here and there made of two glands opening into one duct. Mention of branched s\yeat-glands is made by v. KoUiker, who expresses himself as foUows: "Many years ago I observed in the large axillary glands a division of the tubules, the branches dividing further and terminating separately after giving off a number of alveoli. Now and then anastomoses between branches were observed. Recently I have several times seen a similar branching of the tubules in the circumanal glands. " Fig. 193 of his Handbuch der Gewebelehre, page 252, showing branching in the tubules of a circumanal gland, confirms this statement. This figure resembles very closely several of a series of drawings made by us while recon- structing one of these glands. Horschelmann doubts the existence of branched sudoriparous glands. G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 377 Our best example of a gland of this type is reproduced in Fig. 10. This gland measured 1.25 mm. in its long diameter and 0.5 mm. in its short diameter. In it there may be recognized a parent tubule about 4 mm. long, of which about 0.75 mm. may be regarded as excretory duct, which shows no branching and has a comparatively straight course (in one of the glands studied, the duct was very tortuous, but not coiled). Its structure is similar to that of the duct Fig. 10.— Two views of a model of a branched tubulo-alveolar sudoriparous gland from the circumanal region. 378 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. of ordinary sweat-glands. The primary tubule of the secreting portion of the gland shown in Fig. 10 is lined throughout by secretory epithelium. It pre- sents a very tortuous course, forming four primary loops. Along its entire course, with the exception of one prominent loop, there arise from the tubule larger and smaller branches, some of which are mere alveolar expansions, while others form distinct secondary tubules, which in turn show alveolar protrusions or further division. The majority of the short branches end in not very well marked saccular expansions. Intermediary tubules, such as are Fig. 11. — Two views of a model of a sudoriparous gland with relatively straight duct ending in saccular enlargements, from which arise tubules and alveoli. found in the serous (and mucous) salivary glands, were not observed. Besides the branching and alveolar protrusions above noted, the primary tubule pre- sents a number of vesicular enlargements, united by segments of smaller diameter. That these glands are sudoriparous glands there seems no question, both from their location and also from the fact that they, in common with other sudoriparous glands, possess a single layer of non-striated muscle cells external to the epithelium. We estimate their number (in the circumanal region) to be about the same as that of the circumanal glands of Gay, ten to twelve in a square centimeter. The other type of sudoriparous glands observed in this region may be regarded as a modification of the branched tubulo-alveolar glands just described, presenting, however, certain well-defined characteristics, which are readily G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 379 recognized in the models, as also in the sections. In Fig. 11 \vc reproduce a relatively small gland of this type, which measures 0.6 mm. from the lower end of the duct to the lower end of the gland. In all the glands studied, the duct has a relatively straight course, passing obhquely or nearly perpen- dicularly through the cutis. In the gland shown in the figure, the duct was nearly parallel to the plane of the sections in the greater part of its course. The duct is lined by a low stratified pavement epithelium. The secreting portion of these glands does not present a distinct primary tubule, neither can one speak of a coiled portion. The duct ends in a relatively large vesicular enlarge- ment, which has in sections the appearance of a small cyst. This vesicular en- largement presents a number of constrictions giving it in sections a somewhat irregular outline. Its form varies ; in the gland reproduced, it is of irregular oval shape ; in other glands it is irregularly cyhndrical in shape and may be bent upon itself so as to be cut twice in certain of the sections of the series. From it there arise a number of alveolar protrusions and short tubules. In the gland sketched, one of these tubules again shows enlargements and gives rise to three secondary tubules, which undergo further division. The relatively large vesicle and also the alveoli and tubules arising therefrom, are lined throughout by a secretory epithehum, external to which there is found a not very clearly defined layer of non-striated muscle cells. The appearance in sections of glands of this type is very characteristic. Around or to one side of one or several rela- tively large tubules, with proportionately large lumen, there are observed a varying number of smaller tubules, in size about the same as the secret- ing portion of the ordinary sweat-glands. The branching of the smaller tubules, which occurs to a Umited extent, is not readily seen in sections. These glands are not so numerous as the other glands found in the circumanal region. So far as we have been able to ascertain, sudoriparous glands of this type have thus far not been described. Axillary Glands. In the axillary region there occur, as is well known, certain large sudorip- arous glands, known as axillary glands. Hassal states that these glands were first described by Horner and Robin. In text and hand books and in special articles the axillary glands are described as resembling the sudoriparous glands in general structure, differing from them in that they are much larger and possess proportionately larger tubules. As previously stated, v. Kolliker has observed a branching of the secretory tubules of these glands, also an anastomosis of these branches. This observer gives their size as varying between 1 mm. and 3 mm. in thickness and 2 mm. to 7 mm. in breadth. Hey- nold, who considers them quite fully, mentions a fact which we are able to con- firm, namely, that the secretory tubules of the axillary glands show numerous expansions and constrictions. He also calls attention to the fact that the 380 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. excretory ducts of these glands vary in shape and structure. In some instances, the large secreting tubule was observed to become abruptly narrower and to join a relatively short and small excretory duct, which is similar in structure to the excretory ducts of ordinary sweat-glands; again, the secretory tubules, after becoming somewhat constricted, may join a duct with large lumen and presenting many tortuous turns; and finally glands were observed in which the secreting tubule joined a duct with " enormously enlarged " lumen. These observations also we are able to confirm. Our own observations on the shape and general structure of the axillary glands are, as concerns their shape, based on the study of a relatively large gland, modeled by the plate method. These glands are so large and com- phcated, and the method so time-consuming, that no time remained for their further study. The gland shown in Fig. 12 measures 1.45 mm. in breadth and 0.6 mm. in thickness. In comparison with the measurements given by v. KoUiker (2 mm. to 7 mm. breadth, 1 mm. to 3 mm. thickness), the gland reconstructed by us would appear relatively small. We are not prepared to state that glands measuring 5 mm. to 7 mm. in breadth do not occur in the axillary region. We wish, however, to call attention to a possible source of error in making such measurements from sections, namely, that the prox- imity of the glands is such that it becomes exceedingly difficult to state where one begins and another ends. We must confess that the drawings from which our model was made contained not only the gland reconstructed, but also portions of'two other large axillary glands and a complete small sudoriparous gland, the tubules of all of which were so closely placed and disposed in such a manner that they appeared to belong to one gland, and it was only by recon- struction that we were able to separate the respective glands. The dimen- sions given by us are obtained from measurements made on the model. The gland shown in Fig. 12, giving only the secretory tubule, consists of a single tubule, which shows no branching and which measures approximately 30 mm. The figure may serve to show the general course of the tubule, although only a relatively small proportion of its loops are exposed. As stated by Heynold, and as clearly seen in the figure, the secretory tubule of these glands presents numerous constrictions and enlargements. The secretory tubule is hned throughout by the same epithelium. The duct of the gland shown in the figure is relativelj/ large and tortuous and passes obhquely across the face of the gland exposed in the figure. In certain regions of the axillary space, these glands are so numerous that they form an almost continuous layer of glandular tissue, situated at the junction of the cutis and the panniculus adiposus, some of the loops extending for some distance into the cutis, others reaching the deeper strata of the adipose tissue layer. Besides these large axillary glands, there are found in the axillary space numerous small sudorip- arous glands, which in structure and shape are like the ordinary sweat-glands G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 381 (see Fig. 3). We also observed a relatively small number (three in a piece of tissue having a surface area of a sq. cm. sectioned serially) of branched tubulo-alveolar sudoriparous glands, resembling tliose described for the cir- cumanal region (Fig. 10). Ciliary Glands, or Glands of Moll. The eyehds contain at the ciliary borders sudoriparous glands, known as ciliary glands, or glands of Moll, the ducts of which empty into the hair follicles of the cilia in conjunction with the sebaceous glands, or they may empty between the cilia. The secreting portion of these glands is found either between the follicles of two neighboring cilia, or between a cilium and the orbicularis palpebrarum or the muscxilus ciliaris Riolani. Moll, who first described these glands, gives the following account of them: ''We found the sweat-glands of this region (edges of eyelids) of peculiar form. They are very narrow, Fig. 12. — Large sudoriparous gland, or axillary gland. 382 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. measuring only tV i^^i™- i^ breadth and attain a length of f mm. to | mm. They consist, evidently, of a single tortuous tubule, which now and then forms a loop extending backwards (riickwarts) ; in other cases the tubule, as it extends from the depth to the surface, has a zigzag course. The secreting portion of the tubule joins a straight duct which terminates in a hair foUicle. " Waldeyer, in a most excellent article on the structure of the eyelid, speaks of the glands as having a "very wide, long, and little twisted terminal tubule." Stieda states that the sweat-glands lose their characteristic coil as the edge of the Fig. 13. — Relatively small sudoriparous gland from the edge of the upper eyelid, ciliary gland, or gland of Moll. Fig. 14. — Sudoriparous gland from edge of upper eyelid, ciliary gland, or gland of Moll. eyelids is reached. Here they are less twisted, have a relatively large tubule and empty by means of a straight and narrow duct into the mouths of hair follicles. This he confirms in a later communication. Sattler, who devotes an article to the consideration of the form and structure of the ciliary (Moll's) glands, states that they " are characterized by their relatively large and simple secreting tubules, which form no coils, but have a zigzag course or show S-shaped windings. Similar statements concerning the form and shape of these glands are found in the accounts given by Tartuferiand Schwalbeand generally in the special and general text-books in which they are considered. G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 383 Our own observations based on a stiuly of models of a nunilior of glands of Moll lead us to the conclusion that these glands arc not composed of a simple tubule, having a zigzag or slightly twisted course, but of a tubule which shows branching and distinct alveolar eiilargements, as may be seen from the figures presented. In Fig. 13 is shown a relatively simple gland of this type, measuring 0.5 mm. in its long diameter (exclusive of duct) and 0.35 mm. in its short diameter. The secreting portion of the gland consists of a relatively large primary tubule of slightly tortuous course, which may be regarded as an ex- tension of the duct, and which ends in a T-shaped division only partly shown in the figure; two other prom- inent branches, or tubular buds, are also observed. This gland, though not quite typic, is reproduced by reason of its simple form and since it shows so clearly a branching of . the primary tubule. In Fig. 14 is reproduced a gland, measuring 0.6 mm. in its long diameter and 0.5 mm. in its transverse diameter. This gland consists of an irregular tube folded upon itself and forming a simple coil. From the tube there arise two short tubular branches. This gland is composed of large vesicular enlargements united by short and relatively very narrow segments. The disposition of these is such that the enlargements in part cover up the constricted, uniting seg- ments; the entire gland, as may be seen from the figure, has the ap- pearance of an alveolar gland. In the sections containing this gland, ments presented the appearance of Fig. 15. — Large sudoriparous gland from edge of upper eyelid, ciliary gland, or gland of Moll. the cut vesicular or alveolar enlarge- small cysts rather than tubiiles, and woidd have been regarded as pathologic, did they not show a normal epithelial lining and join a duct with narrow though distinct lumen. The duct, which 384 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. shows a slight vesicular enlargement at the lower end, is lined throughout by a low stratified pavenaent epitheUum and terminates in the mouth of a hair folUcle. The secretory tubule, enlargements, and constricted portions show a lining ol secretory epitheUum. In Fig. 15 is shown a relatively large gland of Moll, measuring 0.75 mm. in length and 0.5 mm. in breadth, in which an alveolar type is more clearly brought out. The relatively short and straight duct has a diameter as great as the secreting portion of the tubule, especially at its lower end, where it presents a distinct vesicular enlargement, differing in this respect from the majority of ducts seen in the series of sections studied. The duct is lined by a low pavement epithelium. The greater part of the secreting portion of the gland may be said to be composed of a tubule with irregular spiral course, showing very marked vesicular or alveolar enlargements, united by very narrow segments, situated for the greater part in the interior of the gland. From the tubule arise kt irregiilar intervals three narrow tubules, each of which ends in a large alveolus. The short tubular segments uniting these alveoli to the parent or primary tubule, as also the true alveoli, are Hned by a secretory epithelium (columnar), the constricted portions of the tubule by a lower epithelium of the cubical variety, and very probably non-secretory. We regard the glands of MoU as tubulo-alveolar in character, the tubules, however, showing a branching only to a limited extent. From investigations on the microscopic structure of these glands, we are able to confirm, though not extend, the observations made by other investigators — Stieda, Tartuferi, Sattler, Schwalbe, and others. Ceruminous Glands. ^ ^ Attention ..was first drawn to the ceruminous glands by Stenon (Observa- tiones Anat.omicse-, 1656), and they were more fully described by Duverney (Traits de Torgane de I'ouie, 1683(?). They were long regarded as sebaceous glands. Pappenheim (Fror. N. Not., 1838) and Wagner (Icone's physiologica, 1838) first called attention to their tubtdar character. We have taken the Hberty of maldng these extractions from Alzheimer's recent article on the ceruminous glands, in which the literature is very fully considered. In the majority of text-books or special articles in which these glands are considered, they are spoken of as coil-glands (Knaueldriisen), and are likened in their shape and structure to the large axillary glands or to the glands of Moll (Tartuferi). Fig. 191 of V. Kolhker's Handbuch der Gewebelehre supplements his brief description of these structures. In this figure, which must be regarded as diagrammatic, so far as concerns the ceruminous glands, these glands are represented as consisting of a simple tubule arranged in the form of a com- pact coil. Alzheimer, who has written very fully on these glands, and whose article contains much accurate information, speaks of them as consisting of G. CARL HUBER AND EDWARD WILLIAM ADAMSON. ;585 a coiled portion and a duct which ends at the sui-racc in a funnel-shaped expan- sion. Schwalbe has given, we believe, the most accurate description of these glands. He states that the coils of the glands have au elongated form, with the long axis not perpendicular to the surfac(> but at an angle, and extendinc Fig. 16. — Ceruminous gland from the external auditory canal of a middle-aged man. into the subcutaneous tissue to the neighborhood of the cartilage. The glands are relatively large, now and then measuring 1.5 mm. in their longest diameter. The secreting tubules of these glands are relatively large, while their ducts are nearly straight and have a diameter of 12 ij.. Schwalbe further states that lie now and then saw a branching of the duct at an acute angle and " it would appear that in the body of the gland in the secreting portion of the tubule 386 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. such branching may now and then take place. " Fig. 33, page 177, Sin- nesorgane, das aussere Ohr, Schwalbe in Bardeleben's Handbuch der Anatomic des Menschen, Fiinfter Band, zweite Abth., also Fig. 164, page 436, of Schwalbe's Lehrbuch der Anatomie der Sinnesorgane, show the general struc- ture, shape, and arrangement of the tubules correctly. They also show branching of the duct and the secretory tubule. Hassal appears to have recognized a branching of the secretory tubules of the ceruminous glands, if we interpret the following statement taken from his work correctly: "The ceruminous glands are similar in structure to the other sweat-glands described inth£,t they consist of twisted tubules forming loops, which finally unite to form a single excretor}' duct. ' ' Huschke also is said to have seen branching in the tubules of these glands. Our own observations extend and amplify the more correct account of the shape of the ceruminous glands given by Schwalbe. In Fig. 16 we reproduce a model of a ceruminous gland from the external auditory canal of a middle-aged man. This gland measures 1.25 mm. in its long diameter, 0.9 mm. in its transverse diameter, and has a thickness of 0.5 mm. Only a short segment of the lower end of the duct is shown in the figure. The secretory tubule, which forms a continuation of the duct, pre- sents the form of a verj' irregular, tortuous tubule, showing vesicular enlarge- ments, united by narrower segments. Along its course at irregular intervals there arise short tubular protrusions, giving the whole tubule a very irregular form. Within the coil of the- gland, forming the portion of the gland shown in the figure, especially to the right, the primary contorted tubule divides and gives off branches, thus forming four relatively long tubules of very irregular form and much contorted, along the course of which there arise short tubular branches and alveolar protrusions. One of the branches terminates in the upper right portion of the figure, another at X, about the center of the figure, and a third forms the lower part of the figure and from its exposed position shows clearly the secondary branching of the tubules. From other glands, carefully studied and in part reconstructed, we are convinced that the ceruminous gland, more fully described and presented in the figure, may be regarded as typical, and that the secretory tubules of the ceruminous glands present a much more irregular form and show more frequent branching than the statements generally made would lead us to infer ; and further, from what has been said, it appears that there is not the similarity in form between the large axillary glands and the ceruminous glands which is expressed very generally in the literature, nor can a parallel be drawn between the glands of Moll and the ceruminous glands. We regard the ceruminous glands as branched tubulo-alveolar glands. In conclusion, it may be stated that our study of the sudoriparous and allied glands has enabled us to differentiate more sharply between the ordinary or typical sweat-glands and the modified sweat-glands than is permitted by a G. CARL HUBER AND EDWARD WILLIAM ADAMSON. 387 study of sections or of specimens prepared by maceration foUowt-d by teasing or crushing. The ordinary sweat-gland, as has long been known, consists of a single tube coiled at the lower end. Attention may again be drawn to the extent of the participation of the so-called duct in the formation of the coil of these glands. In the glands of this type modeled by us, on(--( juarter to two-fifths of the length of the tubule constituting the coil is duct, the remaining portion of the tubule being lined by secretory epithelium. Of the modified sweat- glands modeled by us— circumanal, ceruminous, ciliary, and axillary— all but the axillary (in the particular gland completely modeled by us) show more or less branching of the tubules, with saccular or alveolar enlargements at the ends of the branches or in the course of the tubules. In the axillary glands, the tubule is very much larger and longer and of more irregular shape than in the ordinary sweat-glands. Between the larger axillary glands are found smaller ones with branched tubules. The modified sweat-glands (with the exception of the large axillary glands) may be classed as branched tubulo- alveolar glands. In nearly all the glands of this type, the duct is relatively straight, showing a variable degree of tortuosity or looping, not forming, how- ever, distinct loops or coils, as observed in the ordinary sweat-glands. The duct in the modified sweat-glands is also relatively shorter than in the ordinary sweat-glands. From observations at hand, it would appear that the modified sweat-glands differ also in their mode of development from the ordinary sweat- glands. Alzheimer has dra^^-n attention to this point. He has shown that the ceruminous glands develop from the anlagen of the hair follicles of the external auditory canal and not directly from the germinal layer of the epider- mis, as is the case for the ordinary sweat-glands, and suggests a similar mode of development for the cihary glands and makes it probable that the modified sweat-glands of other regions develop in the same way, although the necessary data are, at present, lacking so far as the circumanal and axillary glands are concerned. It is also probable that the modified sweat-glands form a special secretion containing odoriferous constituents, or otherwise differing from the secretion of the ordinary sweat-glands. Attention has also been drawn to the difference in the length and the arrange- ment of the duct in the ordinary and the modified sweat-glands, and a question as to the reason for the existence of a long and coiled duct in the glands of the former type naturally suggests itself. A satisfactory answer is, at present, not at hand. That the perspiration is a secretion and not a filtration is gener- ally conceded. The experiments of Goltz showing the presence of sweat nerves in the sciatic of the cat and dog are well known ; Arnstein has traced secretory nerve fibers to the gland cells of the tubules of the sweat-glands; and Joseph has shown a structural change in the secretory cells of the sweat-gland when perspiration was induced by electrical stimulation or by drugs. There is no evidence to show that the cells of that portion of the duct of the ordinary 388 MORPHOLOGY OF SUDORIPAROUS AND ALLIED GLANDS. sweat-glands, participating in the formation of the coil, possess secretory- function, and the arrangement and shape of this portion of the duct are such that it does not appear to serve the purpose of a reservoir. It has occurred to us that in the recent work of Ciishny on diuresis and the permeabiUty of the renal cells there may be found an explanation of the existence of the long, coiled duct in the ordinary sudoriparous glands. This observer, in a series of carefully planned experiments, brings forth strong proof to show that with the fluid passing through the glomerular epithelium there are carried certain salts and urea, the salts and urea in the proportion in which they occur in the blood plasma, and that in passage through the uriniferous tubules a certain per cent of the fluids and certain salts are again absorbed, the salts in proportion to their diffusibility or their permeability of the renal cells. In his experiments, the chlorides were more readily absorbed than the less diffusible sulphates, phosphates, and urea. We may suggest as a hypothesis, which we cannot support by experimental data, that an analogy, from the physiological point of view, exists between the ordinary sweat-glands with long, coiled duct and the uriniferous tubules, and that a certain per cent of the fluid secreted in the secreting portion of the sweat tubule, possibly also some of the chlorides, is again absorbed in the coiled portion of the duct, while in the modified sweat- glands with relatively shorter and less coiled duct and with a special secretion, such absorption does not take place. LITERATURE CONSULTED. Alzheimer, A. — Ueber die Ohrenschmalzdriisen. Verhandl. d. Wiirzburger physik.- med. GeseUschaft. Neue Folge, Bd. XXII, 1888. Bardeen, R. D. — Bom's method of reconstruction by means of wax plates, as used in the Anatomical Laboratory of Johns Hopkins University. The Johns Hopkins Hospital Bulletin, Vol. XII, 1901. Breschet and Roussel de Vauzeme. — Recherches anatomiques et physiologiques sur les appareils t^gumentaires des animaux. Annales des Sciences naturelles. T. II, 1834. (Seen only in reviews.) Brunn, A. von. — Sinnesorgane in Handbuch der Anatomie des Menschen. (Von Bardeleben). V. Band, Erste Abth., 1897. Cushny, Arthur R. — On Diuresis and the Permeability of the Renal Cell. Journal of Physiology, Vol. XXVII, 1902. Gay, A. — Die Circumanaldriisen des Menschen. Sitzungsb. d. k. Akad. d. Wissensch. Math.-Naturw. CL, Bd. LXIII, 2 Abth. Wien, 1871. Gurlt. — Vergleichende Untersuchungen tiber die Haut des Menschen und der Haus- Saugetiere, besonders in Beziehung auf die Absonderungsorgane des Haut-talges und des Schweisses. Miiller's Archiv f. Anat. u. Phys., II Jahrgang, 1835. Hassal, A. H. — Mikroskopische Anatomie des menschlichen Korpers. (Uebersetzt von Otto Kohlschiitter), 1852. Reynold, H. — Ueber die Knaueldriisen des Menschen. Archiv f. Path. Anat., etc., Berl., Bd. LXI, 1874. G. CAUL HUBER AND EDWARD WILLIAM A DAMSON. 389 Horschelmann. — Anatomische Untcrsuchungcu iiber die Schweissdriiscii der Mcnschen. Diss. Dorpat., 1875. Johnston, W. B. — A Reconstruction ol' a Glomerulus of the Human Kidni>}'. The Johns Hopkins Hospital Bulletin, 1900. Joseph, Max. — Ueber Schweiss und Tulfulriiscn.sccrction. Archi\- fiir Anat. und Physiol., Physiol. Abth., 1891. Krause. — Haut in Wagner's Handworterbuch der Physiologie, Bd. II, 1844. V. Kelliker, A. — Handbuch der Gewebelehre des Menschcn, Knaueldriisen., p. 247, Bd. I, 1889. V. KoUiker, A. — Entwickelungsgeschiohte des Mensohen und der hbheren Tiere. (Development of glands of skin, p. 793.) 1S79. Maziarski, St. Ueber den Bau und die Einteilung der Driisen — Anat. Hefte, Bd. XVIII, Heft I, 1901. Miller, W. S. — Das Lungenlappchen, seine Blut-und Lymphgefasse. Arohiv f. Anat. und Entwickelungsgesch., Leipz., 1900. Minot, C. S. — Human Embryology. (Glands of the Skin, p. 563) 1892. Moll, J. A. — Bemerkungen liber den Bau der Augenlider des Mensohen. A. von Graefe's Archiv f. Qphthalm., Bd. Ill, Abth. 2, 1857. Sabin, Florence R. — Model of the MeduUa, Pons, and Mid-Brain of a New-Born Babe. Contributions to the Science of Medicine dedicated by his pupils to Professor Welch, 1900. Sattler, H. — Beitrag zur Kenntniss der modificirten (Moll'sohen) Schweissdriisen des Lidrandes. Archiv f. mik. Anat., Vol. XIII, 1877. Schwalbe, G. — Lehrbuch der Anatomie der Sinnesorgane, 1887. (Glands of Moll, p. 243. Ceruminous glands, p. 436.) Schwalbe, G. — Sinnesorgane. Das aussere Ohr, Handbuch der Anatomie des Menschen. (V. Bardeleben) V. Band, Zweite Abth., 1898. Stieda. — L'eber den Bau der Augenlidbindehaut des Menschen. Archiv f. mik. Anat., Bd. Ill, 1867. Stieda. — Ueber die Caruncula lacrymalis des Menschen. Archiv f. mik. Anat., Bd. XXXVI, 1890. Tartuferi, F. — Le glandule di MoU studiate neUe palpebre del I'uomo e degli altre mamiferi e comparate aUe tubolari cutane. Archiv per le scienzi mediche., Vol. IV., 1879. Waldeyer, W. — Mikroskopische Anatomie der Cornea, Sklera, Lider und Conjunctiva in Graefe und Saemisch' Handbuch der gesamten Augenheilkunde, Vol. I, 1874. AVendt, A. — Ueber die MenschUche Epidermis. Miiller's Archiv f . Anat. und Physiol., 1834. ON THE PEEPAEATION AND USE OF COLLODIUM SACS. CHAE.LES S. GORSLINE, A.B., M.D. (From the Hygienic Laboratory of the University of Michigan.) The idea of growing bacteria within the body and inside of a permeable sac or tube, thus protecting the germs from the action of phagocytes, led to the introduction of collodium capsules. The earliest attempt in this direction was that of ilorpurgo and Tirelli.^ These workers endeavored to cxiltivate the tubercle bacillus in celloidin capsules, one slipped over the other. These were placed under the skin or in the peritoneal cavity of rabbits. In 1896, Metchnikoff , Roux, and Salimbeni,^ made the first really valuable contribution on- the usefulness of the sac method. They showed that the cholera toxin can diffuse through the walls of a collodium sac when placed in the peritoneal cavity of a guinea-pig. With sacs having a capacity of three to four cubic centimeters the diffusion was such as to almost always produce a fatal intoxication in the experimental animals. Since then the French bacteriolo- gists have used the sac method to increase the virulence of various bacteria. Thus Vincent^ employed the procedure in his studies upon the transformation of the saprophytic B. megaterium and B. mesentericus vulgatus into patho- genic varieties. In a similar research on the destruction of the spores of the Hay bacillus in the body, Podbelsky^ made use of a novel dialysing membrane. Owing probably to the difficulty of making collodium sacs he replaced these with tubes made out of the lining membrane of reeds. Nocard and Roux,^ by means of the collodium sac were enabled to suc- cessfully cultivate the ultra-microscopic microbe of pleuro-pneumonia of cattle. In the same year Nocard" was able by this method to increase the virulence of the human bacillus tuberculosis to such an extent that it killed chicken. In other words the tubercle bacillus from mammals was thus trans- formed into the aviary variety. The method of preparing the collodium sacs is not detailed by any of the foregoing investigators. Only the merest outline is given of the procedure. '■ Archives Ital. de Biologie, Vol. XVIII, p. 187 ; Ref . in Centralbl. f . Bacteriol., Bd. XIII, p. 74, 1893. ^ Aimales de I'lnstitut Pasteur, Vol. X, p. 261. 3 Annales de I'lnstitut Pasteur, Vol. XII, p. 787, 1898. * Annales de I'lnstitut Pasteur, Vol. XII, p. 431, 1898. = Annales de I'lnstitut Pasteur, Vol. XII, p. 240, 1898. " Annales de I'lnstitut Pasteur, Vol. XII, p. 564, 1898. 390 CHARLES S. GORSLINE. 391 The method, however, as carried out in the Pasteur Institute, was fully described by Novy.^ In place of the heavy perforated glass tube with which th(> Fi-ench workers surrounded the sac to prevent this fi'om bursting, Novy strengthened it by placing a similarly perforated test-tube within the sac. Incidentally, it may be said, that this safeguard is not necessary except when very large sacs are used or when the sac is to remain within the animal for many months. It is wholly unnecessary to go into the full description of the method since this can be found in the text-book referred to. This paper is concerned with an improved way of preparing the sac itself, and more esiDecially in detaching the same from the glass tube on which it is molded. The procedure as followed heretofore was to roll a glass tube of suitable diameter, with rounded end, in the coUodimn until the desired thickness was obtained, after which the tube was dried in the air and eventually placed in water. The most difficult part of the process was to remove the sac from the tube. This was done by everting the edge and then stripping off the sac as one would a tightly fitting glove. Unfortunately, the walls would frequently tear, and as a result the whole pro- cess woxild have to be repeated. It may be said that a really perfect sac was the exception rather than the rule. A considerable amount of time was thus consumed in the preparation of a single sac. To avoid this an extremely simple way of detaching the sac was devised. Before describing this modifica- tion it may be well to refer to several methods of preparing sacs which have appeared in the past two years. McCrae,^ following Pmdden, employed gelatin capsules as a frame-work on which to deposit the collodium. The capsules were attached to the end of a glass tube and after the sac was molded the gelatin was removed by means of hot water. In a more recent paper by Grubbs and Francis,^ a somewhat similar procedure is described. The authors make use of the Novy perforated tube, the openings of which they close with gelatin. After depositing the collodium on the outside the gelatin is removed by means of hot water. The McCrae procedure has been slightly modified by Harris.* The following method enables one to prepare sacs of any desired diameter or length. It is just as easy to make a sac one or two inches in diameter and twenty inches in length as the small one which is used for insertion in the peri- toneal cavity of an animal. This is in itself a distinct advantage since, on account of the permeabihty of the walls the larger sacs may be used for dia- lysing purposes in place of the parchment tubes. Experiments on this point will be given presently. The collodium may be of the U. S. P. strength, but it is advisable to allow ' Laboratory Work in Bacteriology, 1898, p. 496. Uoum. Exp. Med., Vol. VI, p. 635, 1901. ' Bulletin No. 7, of the Hygienic Laboratory of the U. S. Marine Hospital Service, 1902. *Muir and Ritchie's Manual of Bacteriology, Amer. Ed., 1903, p. 67. 392 PREPARATION AND USE OF COLLODIUM SACS. it to thicken by spontaneous evaporation. The roll-tubes are the same as those emploj'ed by Novy/ except that the rounded end is provided with an opening about two milhmeters wide. Obviously a test-tube perforated at the end and lengthened by means of a cork and tube may be used. The procedure is as follows : The perforated end of the roll-tube is first closed by a thin film of thick coUodium. This may be done by bringing the tube down vertically till it just touches the coUodium. Particular care must be taken to prevent the collodium from entering inside the tube for if this does take place it will render the subsequent operation of detaching the sac very difficult. The best way of sealing the opening is to touch it several times with a cork which has been moistened with collodium, or by passing a brush over the surface. The film is now allowed to dry in the air for about a minute. If not thick enough or insufficiently dried it is likely to be dissolved in the next step. The glass cylinder or bottle containing the collodium is now inclined as much as possible without spilling the contents. The tube is then inserted in a horizontal position and applied to the surface of the liquid, but is not immersed. It is then rotated so that about three-quarters of the circumference of the tube with its coating of collodium is exposed to the air. If the liquid is thin it may be necessary to withdraw the tube from the cylinder and rotate it in the air for a minute or two so as to get a firm layer. The tube is again inserted, care being taken to avoid the formation of air bubbles on contact, and rotated so as to form another layer of collodium. This operation may be repeated once more if necessary. It is preferable, as already stated, to use a thickened collodium in which case the first rotation will give a sufficient coating on the tube. When the desired thickness of collodium has been deposited on the tube the latter is withdrawn and rotated in the air for a minute or two until the reagent has set. This point is judged best by touching the thickest part of the collo- dium, which should not adhere to the finger or be too soft. As soon as the collodium has properly set the tube is immersed in distilled water for a few moments. If the laj^er has not been sufficiently dried it will become opaque- white on contact with the water and such sacs are brittle and not suitable for dialysing purposes. The next step is to remove the sac. This, the most difficult part of the old method, is now a very easy procedure. The tube is filled with water which, by steady blowing, is forced through the opening, and following the line of least resistance, it creeps up between the sac and the tube. While still blowing, the sac is gently twisted back and forth until the water has separated it from the tube on all sides. The neck of the sac is then cut or torn loose, when the sac will slide off with ease. ^ Locus citus, p. 497. CHARLES S. GORSLINE. 393 The detached sac is noAV placed in distilled water, after which it may l)e fastened to a constricted tube, according to the directions gi\'en by Novy. With x'ery little care sacs can be jirepared so transparent as to he scarcely seen when placed in water. The method not only permits the easy preparation of sacs for the ordinary peritoneal work, but also enables one to make these of any desired size for dialysis. At Professor No\'y's suggestion a number of experiments were made with large sacs with reference to their dialysing power. These are very interesting and instnictive and can be used for the demonstration of osmotic change. In no other way is it possible to obtain so perfect a view of the interchange which takes place. A sac was attached to a piece of tubing by means of a rubber band. Dis- tilled water was placed inside and the sac thus equipped was suspended in a one per cent solution of sodium chloride. At the end of fifteen seconds a portion of the liquid was removed from the inside by the aid of a pipette and tested with silver nitrate. A positive reaction was obtained showing that dialysis was already taking place. A normal solution of sodium chloride was placed in another sac which was then lowered into a deci-normal silver nitrate solution. Almost immediately a beautiful frost of silver chloride appeared upon the surface of the sac. This gradually increased in intensity till droplets of the precipitate began to fall. Sliding gently down the side, each droplet left a fine trail of white silver chloride. If the solutions or their strengths are reversed the precipitate will form within the sac and the liquid on the outside will remain clear for some time. With equimolecular solutions the precipitate will form in the walls of the tube. Another very interesting demonstration can be made by placing a strong ferric solution inside the sac which is then immersed in water to which a few drops of potassium sulphocyanide have been added. A red stream appears on the outside of the sac. A like result can be obtained with a solution of an alkali and some phenol phthalein. The dialysis of a saturated solution of magnesium sulphate is particularly instructive. For this purpose a large sac about fifteen inches long is prepared, filled with the solution, and immersed in distilled water. In a few minutes two currents are observed — one upward within the sac, the other downward on the outside. The current on the outside is marked by little syrupy drop- lets of magnesium sulphate which can be seen to form along the sac, and falling downward, they form a plainly visible syrupy stream which flows to the bottom of the jar. Inside, the current is marked by similar droplets of water which, on account of their low density, rise to the surface. If, at the beginning of the experiment, the two liquids are placed at the same level, it will soon be seen that the liquid on the inside rises as a result of osmotic 394 PREPARATION AND USE OF COLLODIUM SACS. pressure. In one experiment where a sac about three-fourths of an inch in diameter and about twenty-two inches long, attached to a tube of hke diameter, was used, the Hquid rose to a height of nineteen inches. To appreciate the rapidity with which dialysis takes place it may be said that in an experiment similar to the one just mentioned, 200 c.c. of a saturated solution of magnesium sulphate were placed in the sac, which was then im- mersed in running tap-water. At the end of four hours and twenty minutes, the liquid in the sac gave the same reaction with barium chloride as did the running water. It is evident from this that the coUodium sac can be used to advantage where rapid dialysis is desired. In view of the rapid dialysis which takes place it was probable that colloidal substances, such as pepton, albumose, starch, dextrin, albumin, and enzymes would pass through the coUodium membrane. Some experiments were made with these substances and positive tests were obtained in less than twenty-four hours when the dialysis was carried on at 35° C. With a five per cent solution of albumin a good test was obtained with nitric acid and heat within six hours. The other substances were employed in strengths of one-half to one per cent. THE PHYLOGENY OF THE FOEEAEM FLEXORS. J. PLAYFAIR McMURBICH, Ph. D., Professor of Anatoiiii/, Unii'crsiti/ of Michigaii. Notwithstanding the vohiminous literature of descriptive myology, com- paratively little has been accomplished to«'ard the determination of the exact homologies of the limb muscles throughout the vertebrate phylum. Something has been done in the way of elucidating a fundamental plan for the mammalian muscles, especially through the efforts of Ruge, Cun- ningham, ^Yindle, Leche, and von Bardeleben, to mention only some of the more recent authors, but even mth regard to this group there are still gaps to be filled out and the earlier stages in the , phylogeny still require study, notwithstanding Eisler's very important contribution to that side of the story. That author (1895) has made a careful study of the limb muscles of the urode- lous amphibia, taking Menopoma as a type, and has attempted to reduce the mammalian condition to a modification of what obtains in that group. Hav- ing omitted to consider the reptilia, however, Eisler has missed some important points bearing on the question, and I propose in the following pages to give the results of observations made on both amphibia and reptilia, and hope to demonstrate a detailed homology of the arm muscles in these groups and then to extend the homologies to the mammalian muscles. My attention was primarily directed to the subject through some study which I had made of the perforated flexors of the hand and foot. It has been a general custom to regard these muscles as equivalent, and to assume that the primary condition, so far as mammalia are concerned, is repre- sented in the arm, and that there has been a secondary recession of the muscle into the foot in the lower Hmb (cf . Wiedersheim, 1893) . On looking into the matter it seemed that the evidence which could be brought forward in support of such a theory was decidedly scant, and I determined to test it by a phylo- genetic study, beginning with an attempt to trace the evolution of the flexor sublimis of the arm. This muscle, as a distinct element, being, however, confined to the mammalia, it was evident that in order to obtain a correct appreciation of its significance and a basis for its comparison with the flexor brevis of the foot, it would be necessary to discover what structures, if any, represented the subhmis in the lower vertebrates. Thus the investigation broadened to include a determination of the phylogeny of the entire flexor- pronator mass of the forearm and it is to the results of this portion of the problem that the present paper will be devoted. I hope to consider at some 395 396 THE PHYLOGENY OF THE FOREARM FLEXORS. future date the muscles of the leg in a similar manner and so return to the question of the equivalency of the muscles in the two limbs. A few words are necessary regarding the forms studied and the methods employed. My first intention was to approach the question from the embryo- logical side, and to study the development of the forearm muscles in embryos of Amblystoma tigrinum, Anolis sagrcei,^ the rabbit and man. I soon dis- covered, however, that this method would not yield the desired results, for in the mammalian embryos the forearm muscles, when first distinctly recog- nizable, have practically the adult arrangement. The same result has been obtained by Lewis (1902) in his admirable study of the development of the arm in man, and it would seem that there is a very extensive condensation in the ontogenetic development of the limb muscles in the mammalian. It is prob- able that the entire phylogenetic history of the forearm muscles of man, for in- instance, is condensed into the stages during which the muscles are represented by an undifferentiated mesodermic blastema and that, therefore, anomalies of reversion are referable to the possibilities, dependent on past history, latent in this blastema. The embryological method being then excluded, it was necessary to have recourse to comparative anatomy. Careful dissection revealed much that was of importance, but far more valuable results were obtained from the study of serial sections. From these the topographic relations of the various muscles and their nerve supply could be determined with certainty, and the pictures presented were so much more perfect and striking that I finally relied on the sections rather than on dissections, employing the latter mainly for confirmation. As types of the urodelous amphibia I studied by both dissections and sections Amblystoma tigrinum and by sections only Plethodon erythronotum. Of the reptilia I studied Phrynosoma cornutum, Liolepisma laterale, Callisaurus draconoides and Chrysemys picta, and of mammalia I examined Didelphys vir- giniana (the material of which I owe to the kindness of Dr. C. F. W. McClure, of Princeton University), the cat, the mouse, and man, employing for my serial sections advanced embryos of these forms instead of adult individu- als, simply as a matter of convenience in preparation. I made use of von Ebner's decalcifying solution, embedded in paraffin, cut to a thickness of 20 fi and stained on the slide either with .picrolithium carmine or with Delafield's haematoxylin followed by van Gieson's picrofuchsin, this latter method giving excellent differentiation of the various tissues. ' For material of this form I am indebted to the kindness of my friend, Dr. Henry Orr, of Tulane University. J. PLAYFAIB, McMURRICH. 397 I. The Forearm Flexors of the Urodelous Amphibia and Lacertilia. It is well known that the flexor muacles of the forearm of the urodele amphibia may be regarded as consisting of three layers. The most super- ficial layer consists of muscles arising from the internal condyle of the humerus and extending longitudinally to be inserted either into the carpus or into a strong palmar aponeu- rosis; the middle layer is made up for the most part of oblique muscles arising from the ulna and passing distally and radially to be inserted into the palmar aponeurosis, one muscle only, the ulno-carpahs, having an almost longitudinal direction and being inserted into the carpus; while the third layer consists of a sheet extending obhquely across between the ulna and radius. The superficial layer is divided into three or four muscles; (1) the palmaris superficialis (Fig. 1, PS), which occupies the median por- tion of the layer and inserts into the palmar aponeurosis; (2), the flexor carpi ulnaris (F. C. U.) ; (3), the flexor antihrachii ulnaris {epi- trochleo-anconeus) , Eisler, which inserts into the ulna, and is more or less perfectly differentiated in different forms; and (4), the /Zexor Fig. 1. . Transverse section through the lower half of the forearm of Amhlystoma tigrinum. FCR, flexor carpi radialis; FCU, flexor carpi ulnaris; PP'-PP'", first to third por- tions of palmaris profundus; PQ, pronator quadratus; PS, palmaris superficialis; R, radius; rp, ramus profundus; rswi, ramus superficialis medialis; rsu, ramus super- ficialis ulnaris; U, ulna; UC, ulno-carpalis. carpi radialis (F. C. R.). The obhque muscles of the middle layer are divided by the ulno-carpalis into an ulnar and a radial portion, the latter being again more or less dis- tinctly divided into two portions, so that altogether the layer is composed of four muscles. The most ulnar of these and therefore the most superficial may be termed the palmaris profundus III (Eisler) (P. P. Ill) ; it arises from the ventral surface of the lower part of the ulna and is inserted into the under (dorsal) surface of the palmar aponeurosis. To the radial side of it and separating it at its origin from the palmaris profundus II is the ulno-carpalis 398 THE PHYLOGENY OF THE FOREARM FLEXORS. (U. C), which, arising from the iilna, descends almost longitudinally to be inserted into the distal row of carpal bones. More radially lies the palmaris projundiis II (P. P. II) which resembles closely the palmaris profun- dus III, arising from the radial side of the lower part of the ulna and inserting into the dorsal surface of the palmar aponeurosis toward its radial edge; and, finally, most radial of all, is the palma- ris profundus I (P. P. I), which arises from the lower part of the ulna and also from the carpus and may be traced distally and radially to an insertion into the aponeurosis and the base of metacarpale II. As has been already stated the distinction between portions I and II is not always quite evident and there is also a close rela- tionship between I and the muscle of the third layer, the pronator quadratus (P. Q.), both being sup- plied by the same nerve; portion II has, however, a different nerve supply, receiving branches from the same stem which supplies portion III. The relations of these muscles as seen in sections may be perceived from Fig. 1, which represents a transverse section through the lower half of the antibrachium of Amblystoma tigrinum. Turning now to the lacertilia one finds a condition which seems at first sight far removed from that obtaining in the amphibia. There is a greater amount of longitudinal division of the muscle layers and a diminution in the amount of the oblique musculature in the middle layer, as well as a tendency for it to associate itself more or less closely with the superficial layer. Taking the condition found in Phrynosoma cornutum as typical, the arrange- ment of the muscles at about the middle of the forearm is as shown in Fig. 2. Starting from the ulnar side there is first the flexor carpi ulnaris (F. C. U. and rsu Fig. 2. Transverse section through the middle of the forearm of Phrynosoma cornutum. FOR, flexor carpi radialis; FCU and FCU', lateral and medial portions of the flexor carpi ulnaris ; PP II, III, deep portions of the palmaris communis ; PQ, pronator quadratus; PS, superficial portions of the palmaris communis; PT, pronator teres; R, radius; rp, ramus profundus; rsm, ramus super- ficialis medialis; rsu, ramus superfieialis ulnaris; U, ulna. J. PLAYFAIR McMURRICH. 399 F. C. U.'),, consisting of two tlistinct slips; tnu'od distally these fuse to form a single tendon, which inserts into tlic- ulnar sitl(> of the carpus, while proxi- mally they separate mor6 and more, their origins from the internal condyle being separated by the epitrochleo-anconeus, whose ins(U'tion into the ulna lies above the level of the section figured. The median portion of the arm is occupied by a strong mass which forms the flexor digitonim profundus (auct.), although it would be better to use the term flexor communis digitorum em- ployed by Stannius, or, better still, pal- maris communiSj if we are to regard it as a single muscle. In reality it consists of five distinct portions, only four of which are seen in Fig. 2. Two of these four (P. S.) are superficial, occupying the interval between the more median head of the flexor cai-pi ulnaris and the flexor carpi radialis (F. C. R.), and the other two (P. P. II and III) are deeper, one resting immediately upon the ventral surface of the ulna, while the other lies ventral to the pronator quadratus (P. Q.) and the radius. The fifth portion (Fig. 3, P. P. I) is short and is an oblique muscle arising from the ventral surface of the ulnar side of the carpus. All five portions insert distally into the palmar aponeurosis, which, in the majority of the forms studied, contains a strong volar carti- lage. Proximally the superficial portions take origin from the internal condyle, while the deeper portions arise from the ulna, or, in the case of the fifth portion, from the ulnar side of the carpus. On the radial side of the arm are two muscles, a more superficial flexor carpi radialis (F. C. R.) and a deeper pronator radii teres (P. T.), both of which arise from the internal condyle of the humerus, the former having its insertion into the radial side of the carpus and into the base of metacarpale I, while the latter inserts into the lower part of the radial side of the radius. Finally, constituting the deepest layer, there is a pronator quadratus (P. Q.), extending between the radius and ulna, and in the proximal part Fig. 3. . Transverse section through the wrist of Liolepisma laterale. FOR, flexor carpi radialis; F. C. U., flexor carpi ulnaris; In, intermedium; PPI, flrst part of deep portion of the palmaris communis ; RL, radiale ; rp, ramus profundus; rsu, ramus super- flcialis ulnaris; UL, ulnar; vc, volar . cartilage. 400 THE PHYLOGENY OF THE FOREARM FLEXORS. of the arm a 'pronator accessorius (Mivart), which arises from the internal con- dyle and is inserted into the radius. Before proceeding to a comparison of the individual muscles of the am- phibia and reptilia,a description of the nerves of the fore arm in the two groups will be necessary, for they present a remarkable similarity in their arrange- ment and will serve as guides in the determination of some of the more obscure homologies. In the amphibia the flexor muscles of the forearm are supplied by a large trunk which enters from the brachium towards the radial side and constitutes what has been termed the N. brachialis longus inferior. It passes obliquely inwards between the flexor carpi radialis and the radius and soon divides into a superficial and a deep branch. The ramus profundus (Fig. 1, rp) passes behind the pronator quadratus, which it supplies, and descends the arm in that position to the lower edge of the muscle, where it comes to lie immediately below, i. e., dorsal to the palmaris profundus I, to which it sends fibers. The superficial branch passes toward the median line of the arm ventral to the pronator quadratus, and divides into two branches after giving a twig to the flexor carpi ulnaris and to the epitrochleo-anconeus. The two branches may be termed the ramus ulnaris (rsu) and the r. medialis (rsm) . The former gives off a second branch to the flexor carpi ulnaris and passes obliquely across the ventral surface of the ulno-carpalis, which it supplies, and then descends the arm upon the ulnar side of that muscle. In Amblystoma it lies in the lower part of the forearm between the ulno-carpalis and the palmaris profundus III, and as it nears the carpus, it bends ulnarwards between the latter muscle and the ulna and comes to lie superficially upon the ulnar side of the arm. In Plethodon, however, in which the origin of the palmaris profundus III does not extend so high upon the arm, the nerve begins to bend ulnarward before it reaches the origin of the muscle, and consequently possesses somewhat different relations to it than it does in Amblystoma. The medial ramus breaks up into a number of branches which ramify in the substance of the palmaris superficialis, one, however, descending a short dis- tance to give off a branch to the palmaris profundus II and also to the palmaris profundus III. The supply of the various amphibian muscles may be tabulated, then, as follows : Epitrochleo-anconeus Flexor carpi ulnaris [■ R. superficialis ulnaris. Ulno-carpalis Palmaris superficialis Palmaris profundus II [■ R. superficialis medialis. Palmaris profundus III Palmaris profundus I Pronator quadratus [• R. profundus. Flexor carpi radialis J. PLAYFAIR McMURRICH. 401 In the reptilia the main nerve stem for the flexor mu.scles of the forearm enters from above upon the radial side, and as in the amphibia, may be termed the N. brachialis longais inferior. It e;uiy divides into two stems, a ramus profundus and a R. superficiahs, whoso general relations are practically iden- tical with those found in the amphibia. The R. profundus (Fig. 2 rp) bcuils mesially and dorsally, curving around the radius, and comes to lie dorsal to the pronator quadratus, in which position it descends the arm. It supplies the pronator quadratus and also the pronator accessorius ami thc^ flexor carpi radialis, and at the lower border of the quadratus it passi's ventrally so as to lie upon the ventral surface of the carpus (Fig. 3, rp), gi\dng off a branch to the oblique portion of the palmaris communis. The R. superficiahs divides, as in the amphibia, into a R. mediaUs (Fig. 2, rsm) and a R. ulnaris (rsu). The latter passes obhquely across the arm between the superficial and deep layers of the palmaris communis, reaching the ulna at the lower edge of the insertion of the epitrochleo-anconeus, which muscle it supphes, also sending twigs to the lateral head of the flexor carpi ulnaris. It then continues down the arm, lying to the ulnar side of the deep portions of the palmaris communis and so passes into the manus. The R. medialis follows at first the course of the ulnaris until it reaches approximately the median line of the arm, when it gives off branches to the more medium head of the flexor carpi ulnaris. Early in its course it gives a branch to the pronator radii teres. It passes down the arm between the superficial and deep portions of the palmaris communis, both of which it supplies and in which it is finally lost. Tabulating the muscles according to their nerve supply the arrangement is as follows : Epitrochleo-anconeus ) r, c • i- i Flexor carpi ulnaris (lateral head) \ ^- suPf^rficialis ulnaris. Palmaris communis (superficial portions) 1 Palmaris communis (deep portions II and III) o c • i „, j- r Flexor carpi ulnaris (median head) f^" s^^Perfioialis mediahs. Pronator radii teres J Palmaris communis (oblique portion) ^ Pronator quadratus U profundus. Pronator accessorms f F^"'""'-'^ Flexor tiarpi radialis J We are now in a position to make a comparison of the individual amphibian and reptihan muscles. On the ulnar side the epitrochleo-anconeus has become more distinctly separated from the flexor carpi ulnaris in the reptilia, and with the latter muscle a portion of the palmaris superficiahs has associated itself to form the medial head, while the rest of the palmaris superficiahs, represented by the superficial portions of the palmaris communis, shows a tendency to divide into two portions. ' ^ The palmares profundi II and III are represented by the deep portions 402 THE PHYLOGENY OF THE FOREARM FLEXORS. of the palmaris communis shown in Fig. 2. They have, however, undergone a very important modification by the extension of their origin proximally upon the bones of the forearm, so that they have acquired a more longitudi- nal direction, a condition which is associated with a reduction of the width of the volar cartilage into which they are inserted as conpared with the amphibian palmar aponeurosis. The proximal extension has occurred chiefly in connec- tion with portion II of the muscle and with it there has been a certain amount of extension of its origin radialwards. The palmaris profundus I has retained its original oblique direction and also its primary origin from the lower end of the ulna and the ulnar carpal bones, and has thereby been brought into some- what different relations to the other portions of the profundus than obtained in the amphibia. In a section through the distal part of the forearm of Ambly- stoma the profundus I appears as the most radial of the profundus muscles, while in the reptilia it seems to be the most ulnar. The identification of the muscle in the two groups rests mainly on its nerve supply, and if this be accepted as a sufficient criterion, an explanation is to be sought for the apparent difference in its position. I believe that this can be found in the change in the direction of the second and third portions of the profundus and the migration of their origins proximally, the profundus I being thereby permitted to occupy. exclu- sively the lower part of the ulna and the ulnar side of the carpus, arid since its insertion in the reptilia is into the dorsal surface of the volar cartilage, while the other portions of the profundus insert into its proximal border, there is no obstacle in the way of a conversion of the arrangement seen in the amphi- bia into what occurs in the reptilia. One muscle of the amphibian forearm I have not been able to recognize in the reptifia. This is the ulno-carpalis. The ramus- ulnaris of the super- ficial branch of the inferior brachial nerve passes across the ventral surface of this muscle and descends the arm upon its ulnar surface and in the reptilia the corresponding nerve has the same relations to the second part of the palmaris profundus, using that designation for the portion of the palmaris communis which has been identified with the amphibian profundus II. Arguing from this topographic relation, it seems possible that the muscle has been incorporated in the reptilian profundus II. Such a condition would, however, necessitate a decided alteration of the insertion of the ulno-carpalis, which must have shifted from the carpus to the palmar aponeurosis; and further- more, I find no branches of the ramus ulnaris, which supplies the amphibian muscle, entering the substance of the reptilian palmaris profundus. While I hesitate to express a conviction that the muscle is unrepresented in the reptiha, the weight of evidence seems to me to point that way. The flexor carpi radialis seems to be equivalent in the two groups, while the pronator radii teres seems to correspond to the radial portion of the amphib- ian palmaris superficialis. In Liolepisma the nerve which passes to the flexor J. PLAYFAIR McMURRICH. 403 carpi radialis arises from the N. brachialis inferior longus before its division into the deep and superficial rami, but it is more ni>arly associated with that portion of the nerve which becomes the R. profunilus and I have therefore associated it with that ramus. The branch to the pronator teres, on the other hand, was the first branch from tlie R. suporficialis medialis and there is, accordingly, good reason for belie\-ing that the pronator teres and the flexor carpi ulnaris are quite distinct structures. The reptihan pronator accessorius is supplied, like the pronator quadratus, from the R. profundus and I see no reason for doubting the conclusion of Fiirbringer (1870), that it represents the upper portion of the amphibian quadratus. The homologies of the amphibian and reptilian muscles as described above may be tabulated thus : Amphibia. Ulno-oarpalis Epitrochleo-anconeus Flexor carpi ulnaris Reptilia. ? Epitrochleo-anconeus Flexor carpi ulnaris (lateral headl Palmaris superficialis ( Flexor carpi ulnaris (medial head) •J Palmaris communis (superficial portions) ( Pronator radii teres Palmaris profundus III Palmaris profundus II j Palmaris communis ( (longitudinal deep portions) Palmaris profundus I ( Palmaris communis ( (oblique deep portions) Pronator quadratus Flexor carpi radialis \ Pronator quadratus ( Pronator accessorius Flexor carpi radialis II. The Forearm Flexors of the Mammalia. In the amphibia and reptiha it is evident that the forearm muscles proper end at the wrist joint, their action upon the digits being through their insertion into the palmar aponeurosis, from which the palmar muscles arise. In the mammaha it is customary to regard the long digital flexors as extending from their antibrachial origins to the phalanges, and in comparing them on this basis with the corresponding muscles of the lower groups, it is necessary to assume that there has been either an extension of the origin of certain palmar muscles proximally, or a shifting of the insertion of antibrachial muscles dis- tally, or, perhaps, a combination of both these processes. My results show that such a way of regarding the long flexors is erroneous and that if we are to obtain correct homologies we must compare only the antibrachial portions of the mammahan flexors with antibrachial muscles of the amphibia and rep- tiha, the palmar portions being comparable to palmar structures, tendons or muscles. My reason for this conclusion will be given in a subsequent section of this paper, but in what follows here attention will be directed solely to the strictly antibrachial portions of the mammalian flexors. 404 THE PHYLOGENY OF THE FOREARM FLEXORS. I regret greatly that I have not been able to include a monotreme in the material studied, for, to judge from the descriptions to which I have access, they present most interesting resemblances to the conditions obtaining in the reptilia. The tendency toward an indistinctness in the separation of the superficial and deep layers of the forearm flexors seen in the reptilia is appar- ently carried further in the monotremes, there being recognizable in them, as distinct muscles, only a flexor carpi radialis, a pronator radii teres, a flexor communis digitorum, an epitrochleo- anconeus, and a flexor carpi ulnaris, the last in Echidna being united with the flexor communis to about the middle of the forearm. Dissections have failed so far to reveal any divis- ion of the flexor communis into con- stituent elements such as may be recognized in other mammals, and it would be interesting to determine whether or not such a division could be recognized in sections. Lacking information on this inportant point I must perforce take, as my start- ing-point for a consideration of the mammalian muscles, a condition in which a differentiation of the flexor communis has occurred, a condition a little in advance of what is found in the monotremes and yet a little below what is found in such a mam- mal as the opossum, in that it fails to show any differentiation of the antibrachial portion of the flexor sublimis. I take such a condition for comparison with the lower forms Fig. 4. Transverse section through the forearm of a hypothetical mammal. ai, anterior interosseous nerve ; C, cen- tralis; FOR, flexor carpi radialis; FCU' and FCU^ lateral and medial portions of the flexor carpi ulnaris; m, median nerve; PL, palmaris longus; PQ, pronator quadratus; PT, pronator teres; R, radius; Ra, radialis ; RC, condylo-radialis ; U, ulna ; u, ulnar nerve; UC, condylo-ulnaris ; U, ulnaris. rather than one in which the forearm portion of the sublimis is differentiated, because this muscle is peculiar to the mammalian series and possesses within that series a somewhat complicated development, which may more conveniently be considered later on. The arrangement of the muscles in the somewhat hypothetical condition may be supposed to be as follows. Superficially upon the ulnar side of the forearm is the flexor carpi ulnaris (Fig. 4, F. C. U.) arising by two heads, one from the internal condyle of the humerus and the other from the olecranon process, and inserting below into the ulnar side of the carpus. In close prox- J. PLAYFAIR McMURRICH. 405 imity to this muscle is the epitrochleo-anconeus, also arising from the condyle of the humerus and inserting into thc^ olecranon process. Upon the radial side there is a flexor carpi radiahs (F. C. R.), extending from the internal condyle to the base of one or more of the radial metacarpals, and a pronator radii teres, again from the condyle and inserting a varying dis- tance down the outer surface of the radius. The median portion of the arm is occupied by (1) a palmaris longus (P. L.), extending from the internal condyle to the palmar aponeurosis, and (2) a large mass, composed of several more or less distinct portions and which may be termed the flexor communis digitorum. The constitution of this muscle has been admirably elucidated by Windle (1890), and I propose to follow closely his description of it, based as it is upon a profound and critical analysis of its various components. JMy observations have confirmed his for the most part, the only modification which I shall make being the omission for the present of a sublimis component. I do this because, as I hope to show later, the subhmis is far from being an equivalent muscle throughout the mammalian series, a Adew which differs fundamentally from that apparently held by Windle. Omitting the sublimis, then, as a distinct component, there are recogniza- ble in the flexor communis digitorum five components. Three of these, named by Windle the condylo-radialis (Fig. 4, R. C), the condylo-ulnaris (U. C.) and the centralis (C), have their origin from the internal condyle of the humerus ; the fourth and fifth components, the radialis (Ra) and the ulnaris (Ul), on the other hand arise from the bones from which they derive their names. All the five components unite in a common tendon. Finally, as one of the mammalian muscles there is to be mentioned the pronator quadratus, which extends across between the distal two-thirds of the radius and ulna, its upper border occasionally, however, reaching almost the proximal ends of the bones. If, now, we attempt to employ the nerve supply of these muscles as a guide to their homologies in the lower vertebrates, we are at once met with the difficulty that the arrangement of the nerves in the mammalian is very different from what was characteristic for the lower forms. In place of the single nervus brachiahs inferior longus entering the forearm to supply all its flexor muscles, there are two such nerves, the ulnar and the median. In a general way the ulnar corresponds to the R. superficialis ulnaris of the lower forms and the median to both the R. profundus and the R. superficialis radialis, but the homology cannot be carried into detail. Indeed, from what I have observed in the forms I have studied, I am inclined to beUeve that the median and ulnar nerves are not perfectly equivalent throughout the mammalian series, the ulnar, for instance, in one case containing fibres which in another case are included in the median. That such may be the case has already been pointed 406 THE PHYLOGENY OF THE FOREARM FLEXORS. out by von Bardeleben (1891), who finds in the mammalia a plexus formation between the median and ulnar in the proximal part of the forearm, and Kohl- brugge (1897),' arguing from the differences he finds in the nerve-supply of apparently homologous forearm muscles in different mammals, goes so far as to maintain that the median and ulnar nerves are not to be regarded as definite and invariably equivalent nerves, but merely as paths which may conduct elements of different origins. I may say that in the arm of the human embryo I employed in the present study, a strong branch was given off from the median at the level of the branch- ing of the brachial artery, and following the course of the ulnar artery, it passed obliquely inward between the sublimis and profundus muscles to join and become completely incorporated in the ulnar nerve. Krause and Telgmann (1868) mentioned this condition as of occasional occurrence in man and state that it is almost constant in apes. But while we cannot employ the nerve supply as a certain basis for the homologies of the mammalian muscles, yet it may yield accessory evidence if we can determine the general plan of the rearrangement of the nerve fibers, which has taken place. I believe the rearrangement may be pictured as being along the following lines : The separation of the N. brachialis inferior longus into its forearm branches has, in the intermediate forms between the reptilia (or amphibia) and mam- malia, receded up the arm until in the mammalia it occurs practically at the brachial plexus. During the recession a very considerable change in the relative position of the R. profundus has occurred, since it has left its deep situation and come forward to join the greater part of the R. superficialis radialis, forming a stem, the median, which lies ventral to the deeper muscles and has on the whole a radial position. That portion of the original profundus, however, which supplies forearm muscles remains more or' less distinct from the median and forms the anterior interosseous nerve which passes to its destination anterior to the pronator quadratus. The ramus superficial medialis, which, even in the lower forms, splits into numerous branches soon after its entrance into the forearm, probably associates itself partly with the R. profundus to form the median and partly joins the R. superficialis ulnaris to form the ulnar, and it is in the relative amounts of it which enters into the composition of each of these nerves that variation occurs in the mammalia. Bearing in mind, then, that we probably have in the mammalian anterior interosseous nerve the representative of the portion of the R. profundus which ' I regret that I have not been able to consult this paper. The statement made con- cerning it is based on the review of the paper by von Bardeleben in the Ergebnisse fur Anatomie und Entwicklungsgeschichte, Bd. IX, 1899. J. PLAYFAIR McMURRICH. 407 is supplied to the forearm, the portion of that nerve destined for the hand being induded in the main stem of the median, and considering also the topo- graphic relations of the deep portions of the flexor communis digitorum sup- plied bjf it, it seems that we are justified in identifying these portions of the communis with the portion of the reptilian palmaris profundus which is sup- plied by the R. profundus; in other words, the radialis and ulnaris portions of the flexor communis are together almost equivalent to portion I of the palmaris profundus, but not entirely so, since as a rule the ulnar portion of the ulnaris also receives some tmgs from the ulnar nerve, and the portions of the muscle so suppUed probably represent another portion of the deep muscles of the lower forms, but exactly which, it is difficult to state with certainty. There seem to be two possibihties worthy of consideration; either (1) the twigs for the ulnar nerve, which enter the muscle, represent a portion of the R. ulnaris, in which case it is necessary to turn to the amphibia to find a homo- logue for the muscle fibers in the ulno-carpalis, or (2) the twigs represent a portion of the R. superficials medialis which has associated itself with the R. ulnaris, and in this case the muscle fibres would represent either the second or third portion, or both these portions, of the palmaris profundus. I am not prepared to say which of these possibilities is correct, although I am much more inclined to favor the second than the first. However that may be, I feel confident that in the radialis and ulnaris por- tions of the flexor communis digitorum we have the representatives of the palmaris profundus and that these muscles are the only representatives of the profundus. The remaining portions of the flexor communis, together with the palmaris longus, represent the palmaris superficialis, and it is inter- esting to note that the mammalian muscles have the same relations to the elbow joint, so far as their general origin is concerned, as those of their reptilian prototypes; there has been, in other words, no skipping across the joint of the profundus group of muscles. For the remaining muscles the homologies are less complicated. The flexor carpi radiahs is equivalent throughout all the forms under consideration ; the pronator radii teres, which in the majority of mammals lacks the cojonoid head, seems undoxibtedly equivalent to the reptilian muscle of the same name; this is also true for the epitrochleo-anconeus, which, it may be remarked, is, as Leche (1898) has suggested, a member of the flexor group and not one of the extensor series with which it has usually been classified. The flexor carpi ulnaris presents a slight difficulty, it being a question whether it represents the compound muscle so named in the reptilia, or merely the lateral head of that muscle. The double origin of the mammalian muscle, which is usual, seems to indicate that it is the equivalent of the entire reptilian muscle, in which case we have further evidence that a portion of the R. superficialis medi- alis is included in the mammalian ulnar nerve. 408 THE PHYLOGENY OF THE FOREARM FLEXORS. Tabulating the homologies stated above we get the following results : Amphibia. Ulno-carpalis Epitrochleo-anconeus Flexor carpi ulnaris Palmaris superficialis Palmaris profundus III Palmaris profundus II Palmaris profundus I Pronator quadratus Flexor carpi radialis Reptilia. ? Epitrochleo-anconeus Flexor carpi ulnaris (lateral head) Flexor carpi ulnaris, (medial head) -{ Palmaris superficialis Pronator radii teres Palmaris profundus II, III Palmaris profundus I Pronator quadratus Pronator accessorius Flexor carpi radialis ii Mammalia. ? , Epitrochleo-anconeus Flexor carpi ulnaris (lateral head) Flexor carpi ulnaris (medial head) r Palmaris longus J Portio condylo-radialis I Portio condylo-ulnaris (^Portio centralis Pronator radii teres Portio \dnaris Portio radialis Pronator quadratus Flexor carpi radialis The results which I have recorded above differ materially from those obtained by Eisler (1895) from the comparison of the mammalian muscles with those of an amphibian. His homologies may be stated briefly as fol- lows : The superficial palmar of the amphibia is represented by the palmaris longus, having become very much reduced in size correlatively with a marked increase in the size of the deep palmars. Of these the palmaris profundus II, gradually extending its origin proximally and radially, becomes transformed into the flexores digitorum profundus and longus pollicis; the palmaris pro- fundus III similarly migrates proximally upon the ulna and eventually, pass- ing over the elbow joint, reaches the internal condyle of the humerus and be- comes the flexor digitorum sublimis; while the profundus I is normally unrepre- sented in the mammalian forearm, but occasionally appears as the anomalous radio-carpeus of Fano (the flexor carpi radialis.brevis seu profundus of Wood). It seems to me that these results are open to criticism along three general lines. In the first place, the omission of all consideration of the reptilia has placed Eisler at a disadvantage in having no bridge over the enormous gap which undoubtedly exists between the urodelous amphibia and the mammalia. Even if we accept an amphibian ancestry for the mammaUa, it seems probable that the ancestors were much more reptilian in character than are any of the existing urodeles, and furthermore, not only must the mammalian musculature be referred back to the amphibian, but so must the reptilian. Accordingly we may expect to find in the reptilian muscles, if not direct evidence of the phylogeny of the mammahan conditions, at all events indications of the lines along which it proceeded, and it seems to me, this expectation has been fully borne out by the results described in the preceding pages. There is certainly much more general similarity in the arrangement of the reptilian and mamma- lian forearm musculature than in that of the amphibia and mammalia. J. PLAYFAIR McMURRICH. 409 In the second place, Eisler has failed to take into consideration th(> evidence derived from the nerve supply of the amphibian musculature. It may not be possible as j^et to institute a certain homology between the amphibian and mammahan forearm nerves, but I believe that I have shown a sufficient general equivalency' to warrant the acceptance of the nerve supply as important corroborative evidence. The identification, therefore, of the palmaris pro- fundus II with its nerve supply from the R. superficialis medialis with the mam- mahan flexor profundus supplied by fibres which represent the R. profundus, seems very doubtful, unless the evidence from other sources is more than ordinarily convincing, and that it is so has not, I beheve, been demon- strated. Thirdly, the homologies proposed by Eisler demand a very considerable modification in the topographic relationship of the muscles. A muscle, the profundus III for example, which is clearly a portion of the deep layer in the amphibia, becomes, in the mammalia, the superficial flexor sublimis, altering its topographic relations to the principal nerves of the arm. Such an alteration is of course possible, btxt its probability is greatly diminished if an homology can be found which does not demand it, and I have shown that there is such an homology. Indeed, the superficial and deep layers of the amphibian forearm musculature are clearly recognizable in both the reptilia and the mammalia, and there seems no reason for manufacturing homologies which require their confusion. Furthermore, it seems to me that an homology which demands an extensive migration of muscle masses across joints should be viewed with suspicion, and such a migration is demanded by Eisler's identification of the palmaris profundus III with the flexor sublimis. With the enormous reduction which he supposes to have occurred in the palmaris superficialis, room is afforded upon the internal condyle for such a migration, but, as has just been indicated and as will be shown later, there is evidence to show that this reduc- tion has not occurred. Hence, independently of the a priori objections to a migration of a profundus muscle across a joint, there is, in the present case, an additional objection on the ground that the muscle would have found the territory for which it was striving already pre-empted. Finally, a word concerning the identification of the palmaris profundus I with the anomalous flexor radialis brevis. I have had an opportunity for studying this muscle in a subject dissected last winter in the Anatomical Laboratory of this University, and from its general relations I should be strongly inchned to regard it as a portion of the flexor carpi radialis, though I cannot exclude the possibility of its derivation from the pronator quadratus. In either event, however, I agree fully with Le Double (1897) in assigning it to the group of progressive anomalies: "II est la consequence du morcellement plus complet de la masse flexopronatrice, et non un 'remnant' de cette masse pour me servir d'une expression du professeur Humphry. " 410 THE PHYLOGENY OF THE FOREARM FLEXORS. III. The Antibrachial Flexors in Man and the Evolution of the Flexor Sublimis. The flexor muscles of the forearm in man present certain departures from the condition which has been considered fundamental for the mammalia, the more important of these departures concerning the pronator radii teres and the flexor communis digitorum. The peculiarity in the pronator consists in its possession of a coronoid head in addition to the condylar one, the median nerve passing into the forearm between the two heads. This condition, so far as I am aware, occurs only in man and in the anthropoid apes, and in these forms it is associated with a marked reduction in the size of the pronator quadratus. There seems to be no doubt but that Macalister (1868) was right in regarding the deep head as something quite distinct from the pronator teres proper, and I believe we may go further than Macalister when he says in his earlier paper that it is to be regarded as 'the germ of a superior transverse muscle, the upper equivalent and co-ordinate of the pronator quadratus below." In its highest degree of development in the mammalia this latter muscle occupies the entire length of the forearm, and in Perameles and some species of Halmaturus, in the dog and the hyeena, its proximal portions are united with the pronator teres (Leche). In man and the anthropoids, as Macalister points out in his later paper (1869), we seem to have other instances of a similar fusion, in association with which there has been, however, a degener- ation of a considerable portion of the quadratus, only its proximal and distal portions persisting. In the case of the flexor digitorum communis the modifications are much more complicated. The most striking peculiarly of the human flexor is its separation into a large flexor sublimis seu perforatus and a flexor profundus seu perforans, and, furthermore, the separation of the profundus into the profundus of anthropotomy and the flexor longus poUicis. That the occurrence of a flexor longus pollicis is due to a differentiation of a portion of the profundus, to be more precise of a portion or aU of the portio radialis, seems beyond question. It is a muscle which has not infrequently been described as absent in the lower forms, or in other cases, its absence has been accounted for by a fusion with the profundus. To my mind neither of these expressions fits the case; the latter one imphes that it is an independent typical constituent of the mammalian fiexors which in certain cases has dis- appeared by fusion with the neighboring muscle, while the former implies that it is unrepresented. The occurrence of the muscle in a comparatively small number of the mammalia, e. g., in certain carnivores, Hylobates and man, indicates by no means indistinctly its secondary nature, and it certainly seems improbable that it could appear sporadically, as it does, without having some representative in the arms of forms nearly alUed to those which possess J. PLAYFAIR McMURRICH. 411 it. If it be a separated portion of the profundus, then it has a representative throughout the entire mammalian series, probably oven in forms which lack a pollex, for the relation of the profundus is not primarily to the individual digits but to a common tendon, a point which will be elucidated in the succeeding part of this paper. Independently of the decided difference in the views of Eisler and myself as to the homologue of this muscle in the amphibia, it seems to me that Eisler is wide of the mark in attempting to disco\-er an indication of its independent existence in the lower forms, as he does in the partial separation of the pro- fundus II into two portions. Its independence from the mammalian profundus is too recent phylogenetically to warrant a hope of an absolute identi- fication of it in the amphibia. Its separation occurs only within the mamma- Han phylum, and indeed only in certain of the more highly specialized mem- bers of that phylum. I can see no reason for supposing that the occurrence of the muscle in the dog and the hysena has any phylogenetic relation to its occurrence in man; it seems rather to have been developed, i. e., separated from the profundus independentlj' in the two cases. The palmaris longus is a muscle which may well be regarded as typical of themanimaha, though its absence in the monotremes, that is to say, its lack of separation from the flexor communis, implies that its differentiation has occurred within the limits of the phylum. The available evidence seems to point to its having been the first separation from the common flexor, and its distinctness from the other components does not seem to he equal in different forms. In other words, it is doubtful if the muscle is an absolutely equivalent structure throughout the mammahan series, but this, as well as the question as to the nature of the palmar fascia to which it is attached, can be more satisfactorily discussed later. The subhmis, like the palmaris longus, has been differentiated from the flexor communis digitorum within the limits of the mammalian phylum and is not an equivalent muscle throughout the group, since it contains a greater portion of the flexor communis in man and the higher forms than it does in the lower. It is hardly necessary to remark that the identification of the subhmis with the fiexor brevis digitorum (perforatus) of the reptiUa, which has so frequently been made, is incorrect. The comparison of the subhmis in different mammals must rest upon the recognition of its relations to Windle's five portions of the flexor communis and these relations are as yet unknown in the majority of the mammalia. I shall, accordingly, first describe what I have found in the forms which I have studied, namely in the opossum, the cat, the mouse, and man, and employ the table contained in Windle's paper only after I have established the prob- able line of differentiation. In a section through the upper part of the arm of an opossum (Fig. 5) 412 THE PHYLOGENY OF THE FOREARM FLEXORS. the five portions of the flexor communis are clearly recognizable, the condylo- ulnaris (C. U.), lying beneath the medial head of the flexor carpi ulnaris (F. C. U.^) and the palmaris longus (PL) and ventral to the ulnaris (Ul), the condylo-radiahs (CR) lying to the radial side of the palmaris and the condylo-ulnaris, while between it and the radialis is the slender centralis (C), which corresponds to the "slender little spindle of muscle, quite distinct from the rest," described by Coues (1872), whose identiflcation of it with the flexor longus pol- licis is manifestly erroneous. Tracing the various portions down the arm, it is found that the condylo-ulnaris decreases in size rather rapidly, its fibers passing into a flat tendon which lies on the surface of the muscle in contact with the ulnaris. A portion of the muscle, repre- sented approximately by the portion which is shaded in Fig. 5, may, however, be traced on- ward to the wrist where it passes into a tendon lying to the ulnar side of and superficial to the large tendon which is formed by the fusion of the main condylo-ulnar tendon with the other four portions of the flexor communis. Later the superficial condylo-ulnar tendon divides into three slips, which pass to the second, third, and fourth digits and are three of the tendons of the flexor digitorum sublimis. On tracing the palmaris longus distally it is found to develop upon its deep surface a slender tendon, represented by the shaded portion in Fig. 5, and toward the wrist the rest of the muscle passes into a flat tendon which is lost in the palmar fascia. The slendor tendon can be traced onward into the hand below, i. e., dorsal to the flat tendon and the palmar fascia, and verging toward the ulnar side of the hand, it becomes the sublimis tendon to the mini- mus. This origin of a portion of the sublimis from the palmaris differs from Fig. 5. Transverse section through the forearm of the Opossum, ai, anterior interosseous nerve; C, centralis; CR, condylo-radialis ; CU, condylo-ulnaris; FOR, flexor carpi radialis; FCU' and FCU^, lateral and medial portions of the flexor carpi ulnaris; m, median nerve; PL, palmaris longus; R, radius; Ra, radialis; U, ulna; u, ulnar nerve; UL, ulnaris. The shaded areas represent the flexor sublimis digitorum. J. PLAYFAIR McMURRICH. 413 the account given by Coues, the only doscrijition of the myology of the opossum I have been able to consult; tliis author, as well as Windlc, (k-rives all four tendons from the condylo-ulnaris. The origin of tlio minimal tendon from the palmaris was found both in my sections and dissection, and as will be seen, is in harmony with what occurs in other forms in which the sublimis is in a low state of differentiation. In the cat the condylo- ulnaris is not distinctly sep- arated from the \ilnaris in the upper part of the arm, and it contains a tendon im- bedded in its substance which is continuous with a tendon on the ventral surface of the ulnaris. Apparently a portion of the condylo-ulnaris inserts into the ulnaris tendon and this unites ^\ith the other four portions to form the profundus tendons as in the opossum, but the rest of the muscle, the shaded portion in Fig. 6, can be clearly seen to divide near the wrist into a smaller radial and a larger ulnar portion which remain distinct from the ulnaris and in each of which a tendon develops. The tendon belonging to the larger portion eventually divides into two slips which form the sublimis tendons for the third and fourth digits, the tendon for the smaller portion of the muscle passing to the second digit. The fourth tendon, that for the fifth digit, is formed by a slip from the pal- maris longus, the relations of the subhmis being very similar to what obtained in the opossum, though the amount contributed to it by the condylo-ulnaris is somewhat greater in the cat. In the mouse the conditions are somewhat different, however. As usual the five portions of the flexor mass and the palmaris longus can be recognized (Fig. 7) and tracing them downward it can be seen that the ulnaris, radiaUs, centraHs, and condylo-radialis all unite together to form a profundus tendon. But the condylo-ulnaris remains quite separate from the rest and at the wrist Fig. 6. Transverse section through the forearm of an embryo cat of 7 cm. Lettering the same as in the preceding figure. 414 THE PHYLOGENY OF THE FOREARM FLEXORS. divides into three portions which, becoming tendons, pass as the perforated tendons to the second, third, and fourth digits, there being in this form, or at all events in the single individual I studied, no sublimis tendon to the fifth digit and hence no contribution to the subhmis from the palmaris longus. ^ These three forms afford a very definite clue to the relations of the subhmis to the flexor com- munis. It is principally associ- ated with the condylo-ulnaris, the portion to the fifth digit, however, being derived from the palmaris longus. In the opossum but a small portion of the condylo- ulnaris is devoted to the forma- tion of the sublimis, the contri- bution is distinctly greater in the cat and the entire muscleis taken up into it in the mouse. Having then some indication of the line which the differentia- tion of the sublimis follows, we may now turn to the table given by Windle (1890) and inquire whether it reveals any further differentiation along the same line. And first of all we may consider his account of the ar- rangement in the rat. The condylo-ulnaris is stated to be absent in this form, while the other four portions of the flexor mass are recognizable, and the sublimis is indicated as being an independent muscle. This may with propriety be interpreted, on the basis of what I have found in the mouse, that the condylo-ulnaris has been completely taken up into the sublimis, and applying the same inter- pretation to other forms tabulated by Windle, we find that the same condition obtains in the majority of the rodentia. Proceeding to higher forms we find that in CebTis capucinus the condylo-ulnaris is again wanting and that, fur- thermore, the subhmis is closely associated with the condylo-radialis, that is to say, the sublimis not only includes the whole of the condylo-ulnaris but also receives a contribution from the condylo-radialis. The same condition occurs also in Cynocephalus maimon, with the addition that in this form the centralis has also disappeared, having, I imagine, been taken up into the sublimis. This disappearance of the centralis is also noted for several other time a longitudinal division of the super- ficial layer of the palmar aponexu-osis occurs, a strong slip of it being continued distally upon the palmar surface of each superficiahs shp (Fig. 10), while beneath each superficialis slip a thickening appears in the deep layer of the aponeurosis. :More distally the lateral portions of each superficialis slip separate (Fig. 10, F. S.') and pass dorsally to fuse with the corresponding flexor brevis profundus (F. P.), while th(> remaining median portion later on divides into shps, which may be traced distally to their insertion into either side of a strong fibro-cartilaginous nodule occurring at the metacarpo-phal- angeal joint. In the mean time the deep layer of the palmar aponeurosis has divided longitudinally into shps or tendons, which are the continuations of the thick- enings already mentioned as occurring in it, and one of these tendons lies im- mediately below the median portion of each flexor superficialis slip passing to the digits under consideration (Fig. 10, pt). When the final division of the muscle slips occurs, the tendons derived from the deep layer of the aponeurosis pass ventrally between the two terminal slips of the muscle and unite with the superficial tendons, passing on ^^ith them to be inserted into the base of the terminal phalanges. The flexores medii, compared with the superficials, are small muscles. As they are traced distally that for the fourth digit (Fig. 9, F. B. M.'') lies over the palmar surface of the fourth metacarpal, being separated from it by the corresponding flexor profundus (F. B. P.'"), while that for the third digit (F. B. il.'"); niuch smaller than the other, lies rather to the ulnar side of its metacarpal. The muscle of the fourth digit divides longitudinally into three shps (Fig. 10, F. M., F. M.^ and l), that upon the ulnar side uniting with the subjacent flexor profundus IV, while the median and radial slips insert into the metacarpo-phalangeal fibro-cartilage, the median one entering into close relationship Tsith the underlying median portion of the flexor profundus (F. P.'). The muscle of the third digit, owing to its more ulnar position with reference to the axis of the digit, lacks a radial slip, dividing into only two portions (F. M. & I), the more radial of which corresponds to the median part of the foiirth muscle and_ like it unites somewhat closely with the underhdng portion of the flexor profundus, while the ulnar portion inserts independently into the ulnar side of the metacarpo-phalangeal fibro-cartilage. The points to which attention needs to be especially directed for our pres- ent purpose are (1) the splitting of the palmar aponeurosis into two layers by the origin of the flexores breves superficiales, (2) the formation of tendons by the deep layer of the aponeurosis, which, after the division of the flexores superficiales into their terminal slips, pass up between them to join the tendons 420 THE PHYLOGENY OF THE FOREARM FLEXORS. from the superficial layer of the aponeurosis, and (3) the origin of the flexores breves raedii from the under surface of the deep layer of the aponeurosis. Turning now to the reptilia one is at once struck by the fact that there is no strong aponeurotic layer covering the surface of the flexor brevis super- ficialis (flexor sublimis seu perforatus Auct.) (Fig. 11, F. B. S.). On the other hand, a very strong aponeurosis (vc), frequently partly transformed into cartilage, is present beneath the flexor superficiaUs, giving origin to this muscle from its palmar surface and receiving the insertion of the forearm muscles as described on a preceding page. We may for convenience confiae our attention mainly to the muscles associated with the three middle digits, for the same reason that led us to dis- regard the lateral digits in the amphibia. Traced distally the central portion of the superficialis sheet divides into three portions (Fig. 11, F. B. S.), which pass to the three digits we are considering, and underneath each por- tion there is a strong tendon which is a distal continuation of the volar car- tilage. Shortly before reaching the metacarpo-phalangeal joint each por- tion of the superficialis splits into two slips, which separate so as to lie one on each side of the strong tendon just mentioned and gradually fade out into the fascia covering that tendon. The muscles which correspond to the amphibian flexores breves medii reach a much greater development than in the lower group and are arranged in two distinct layers, the superficial one (Fig. 11, Z) lying immediately beneath the volar cartilage, from which it takes origin, while the deeper one (pi) is in relation with the underlying mtacarpal bones. This latter layer does not concern us at present and will be left for consideration on another occasion. The superficial layer, when traced distally, divides into four portions which pass to the II-V digits, there being no portion for the poUex. Each portion lies beneath the corresponding portion of the flexor superficiaUs, being separated from it by the strong tendon derived from the volar cartilage. More distally each of the portions corresponding to digits II-IV divides into two shps which Fig. U. Transverse section through the palm of Liolepisma laterale. F.B.S., flexor brevis digitonim superficialis; I, lumbricalis, i. e., superficial layer of the flexor brevis medius; pi, palmar adductors, i. e., deep layer of the flexor brevis medius; vc, volar cartilage. J. PLAYFAIR McMURRICH. 421 come to lie on either side of the corresponding strong tendon and are finally- inserted into opposite sides of the base of the metacarpo-phalangeal fibrocarti- lage of the digit to which they belong. In the case of the fifth digit the conditions are slightly different, in that the superficial sheet of the flexor mediiis does not extend laterally beyond its radial border, and hence, when the division of the sheet into separate slips takes place, that for the minimus lies upon the radial side of its digit and does not divide into two terminal slips as do the others, but inserts entirely into the radial side of the arthrodial fibro- cartilage. The arrangement of these muscle-slips is shown diagrammatically in Fig. 12. If now we proceed to compare these arrangements with those seen in the amphibia we arrive at the following conclusions. The portion of the super- ficial layer of the amphibian palmar aponeurosis which covers the flexor brevis superficialis, has disappeared in the rep- tilia or is represented in the flexor, if one prefers to state it that way. The more proximal portions of the aponeurosis, however, are represented by the volar cartilage and the strong tendons which are continued distally toward the fingers from the volar cartilage are, in their proximal portions, the representatives of the tendons formed from the deep layer of the amphibian aponeurosis. Beyond the point of the bifurcation of the shps of the flexor superficialis these tendons in the amphibia fuse with the tendons from the superficial layer of the aponeu- rosis, and it is probable that in the reptiha the tendons from the same point are equivalent to those of the amphibia. The annexed diagram (Fig. 13) will give, I trust, a sufficiently clear idea of the arrangement in the two groups, the portion of the amphibian aponeurosis which has disappeared in the reptilia being indicated by the stipphng. From the reptiUan arrangement as interpreted above to the mammalian the passage is easy. The tendons which are continued distally from the volar Fig. 12. Partly diagrammatic repre- sentation of the arrangement of the lumbricales (Q, in Ldolepisma later- ale. F.B.S., flexor brevis super- ficialis; ph, phalanx; pt, profundus tendon. Fig. 13. Diagram showing the mode of formation of the profundus tendon. F.B.S., flexor brevis superficialis; F.B.M., flexor brevis medius; P.A., palmar aponeurosis. The stippled portion of the aponeurosis disappears in the reptilia. 422 THE PHYLOGENY OF THE FOREARM FLEXORS. cartilage to the digits clearly correspond to the mammalian profundus tendons and the superficial laj'er of the flexor brevis medius is, I believe, equivalent to the mammalian lumbricales. The deep layer of the medius, however, entering as it does into relation with the flexores breves profundi and the metacarpals, is probably represented in the mammalia by the palmar adductors, an homology which I hope to consider in detail in a later paper. It is interesting to compare the arrangement of the lumbricales in Echidna as described by WestUng (1889)' with that which I have found in the superficial layer of the flexor brevis medius of the reptilia. The reptilian equivalents of the sublimis tendons are indicated, I believe, by the condition found in the monotremes and in certain marsupials. In Ornithorhynchus a muscle has been described as the flexor digitorum subhmis (Smith, Westling), which has essentially the same relations as the flexor brevis superficiahs of the reptilia, and in Thylacinus and Phascogale (Cunningham, 1882), this muscle is represented by four minute tendons which arise from the strong tendon of the flexor communis digitorum. The communis tendon I take to be practically the homologue of the reptilian aponeurosis in which the volar cartilage is developed and the small tendons which arise from its sur- face are therefore equivalent to four of the slips of the flexor brevis superficiahs of the reptilia, which have undergone, as so frequently happens, transformation into connective tissue (see von Bardeleben and Bland Sutton). The identification of the tendon of the flexor communis with the reptihan palmar aponeurosis is not, however, quite exact, for there exists in the mam- malia a palmar fascia which covers the sublimis tendons and receives the inser- tion of the palmaris longus. This muscle is a portion of the condylar flexor mass of the forearm and is, as has alreadj'' been seen, closely related to the subhmis, containing, in the lower mammals, elements which in higher forms are included in that muscle. This being the case it must be supposed that the original insertion of the palmaris was with the rest of the flexor mass into the palmar aponeurosis, and that with the separation of the palmaris there has also been a separation of a palmar layer of the aponeurosis to form the mammalian palmar fascia. The relations of the superflcial thenar and hypothenar muscles to the fascia support this view of its origin, since these muscles are persisting portions of the flexor brevis digitorum superficiahs. The correct equivalent, accord- ingly, of the reptilian palmar aponeurosis in the mammalia is the tendon of the flexor communis -plus the palmar fascia, but it should be pointed out that there is a strong probability that the distal portions of the mammahan fascia may represent the portion of the amphibian palmar aponeurosis which has disap- peared in the reptilia I have studied. The phylogenetic history of the mammalian long flexors, which has been ' I have not been able to consult this paper, but the figure which bears on this point is reproduced by Leche in the Mammalia of Bronn's Thierreich. ,1. PLAYFAIR McMURRICH. 423 traced in the preceding pages, n\:iy be briefl}- stated as follows : In the pri- mary condition the entire flexor mass of the forearm terminates at the wrist, a certain portion of it inserting into the bones of the forearm and carpus and the rest into a strong palmar aponeurosis. From the latter two sets of muscles take origin, (1) from its substance the flexors breves superficiales, and (2) from its deep surface the flexors breves metlii. By the mode of origin of the first of these the palmar aponeurosis is divided distally into two layers, a more super- ficial one which is prolonged distally into strong tendons which insert into the bones of the terminal phalanges, and a deep one also prolongeil into tendons which pass between the terminal slips of the flexores superficiales to unite ■nith the superficial tendons. This is the amphibian stage. In the second or reptilian stage the portion of the superficial layer of the palmar aponeurosis which covers the flexores breves superficiales disappears and the action of the forearm flexors wMch insert into the aponeurosis is distributed to the digits entirely through the tendons of the deep layer, which, together with the persisting terminal portions of the superficial tendons, may be recognized as the equivalent of the mammalian profundus tendons. The portions of the two layers of the forearm flexors which act on the aponeurosis fuse more or less completely, the flexores breves superficiales retain their amphibian relations while the flexores medii divide into two layers, the more superficial of which represents the lumbrical muscles of the mammalia. In the last or mammalian stage the flexores breves superficiales become transformed more or less completely into the tendons of the flexor sublimis, and as the scale is ascended, a gradually increasing amount of the superflcial portion of the flexor communis separates to become continuous with these tendons, until, in man, the entire condylar portion of the muscle, except so much as is represented by the palmaris longus, is taken up into the flexor sublimis. In the cases of the first and fifth digits some departures from the processes outUned above occur, but these may be more conveniently discussed in con- nection with the historj' of the other hand muscles in a later paper. The results recorded above as to the relations of the subhmis tendons to the forearm muscles agree in general with those arrived at by Eisler, but I have succeeded, I believe, in tracing with greater exactness the processes by which the final arrangement has been acquired. Eisler has failed to per- ceive the true relations of the profundus tendons to the amphibian palmar aponeurosis, relations which it would be diflEicult to discover without the aid of sections. He has, however, recognized the relations of the flexores breves superficiales to the subhmis and the probable end to end union of the two mus- cles, an arrangement which, as he points out, throws clear fight on many of the anomalies occurring in connection with the human sublimis. 424 THE PHYLOGENY OF THE FOREARM FLEXORS. REFERENCES. 1881. K. Bardeleben. — Muskel und Fascie. Jenaische Zeitschr. fiir Naturwiss., XV, 1881. 1891. K. Bardeleben. — Ueber Innervierung, Entstehung und Homologie der distalen Gliedmassenmuskelu bei den Saugetieren. Verhandl. Anat. Gesellsch., V, 1891. 1872. E. Coues. — The Osteology and Myology of Didelphys Virginiana. Mem. Boston Soe. Nat. Hist., II, 1872. 1882. D. J. Cunningham. — Report on some points in the Anatomy of the Thylacine (^Thylacinus cynocephalus), Cuscus (Phalangista maculata), and Phascogale {Phascogak calura). Reports of the Scientific Results of the Voyage of H. M. S. Challenger. Zool., V, 1882. 1895. P. Eisler. — Die Homologie der Extremitaten. Abhandl. Natur. GeseUseh. zu Halle, XIX, 1895. 1870. M. Fiirbringer. — Die Knochen und Muskeln der schlangenahnlichen Sauriern. Leipzig, 1870. 1897. J. H. F. Kohlbrugge. — Muskeln und periphere Nerven der Primaten. Eine vergleichende-anatomische und anthropologische Untersuchung. Verhandl. Akad. Wetens Amsterdam (2), V, 1897. 1868. W. Krause and J. Telgmann. — Die Nerven- Varietaten beim Menschen. Leipzig, 1868. 1898. W. Leche. — Mammalia. In Bronn's Klassen und Ordnungen des Thierreichs, Bd. VI, Abth. V, I, 1898. 1897. A. F. Le Double. — Traits des variations du systSme musculaire de I'homme. Paris, 1897. 1868. A. Macalister. — On the nature of the Coronoid portion of the Pronator radii teres. Journ. of Anat. and Phys., II, 1868. 1869. A. Macalister. — On the arrangement of the pronator muscles in the limbs of vertebrate animals. Journ. of Anat. and Phys., Ill, 1869. 1897. J. Bland Sutton. — Ligaments, their Nature and Morphology. London, 1897. 1884. C. Westling. — Beitrage zur Kenntniss des peripherischen Nerven system. II, Omithorhynchus. Bihang. Svensk. Vet. Akad. Handl., Stockholm, IX, 1884. 1889. C. Westling. — Anatomische Untersuchengen iiber Echidna. Bihang, Svensk. Vet. Akad. Handl., Stockholm, XV, 1889. 1893. R. Wiedersheim.; — Der Bau des Menschen als Zeugniss fiir seine Vergangenheit. Freiburg, 1893. 1890. C. A. Windle. — The flexors of the digits of the hand. I. The muscular masses in the forearm. Journ. of Anat. and Phys., XXIV, 1890. CLINICAL OBSEEVATIONS ON CONGENITAL AND ACQUIRED TRANSPOSITION OF THE VISCERA. JAMES RAE ARNEILL, A.B., M.D., Instructor in Clinical Medicine, University of Michigan. Cases of sitiis viscerum inversus have, until recent years, been discovered, studied, and reported chief!}' by anatomists and pathologists. As a result of a correspondence with a large number of medical men of experience and promi- nence, it woidd seem to me that the pendulum has swung about of late years, and that now a much larger percentage of these cases is discovered by the clinician. This fact is a natural result of the much more frequent and careful physical exaniiimtions that are now being made. However, even with this improvement, all physicians know that a very small percentage of even sick people, submit to careful physical examination, and that the percentage of the entire population who are thus examined is extremely small. . Situs \'iscerum inversus is of great interest to the embryologist and anato- mist. To the internist it presents very instructive problems in physical diagnosis. Gruber made a remarkable collection of all cases found in literature up to the year 1865. Of the seventy-nine cases which he reports, only five or six were discovered during life. " This fact seems rather remarkable, since his report covers more than two centuries. In 1643 Petrus Servius reported the first case as occurring in Rome. Kiichenmeister, up to the year 1888, collected 149 cases. The majority of these were discovered in the anatomical laboratory, and on the post-mortem table. Pic, in 1895, increased the number of reported cases to 190. Cases of partial situs inversus are not so common as those of complete transposition. Lochte, up to the year 1894, collected thirteen cases of this incomplete variety. In more than half verj- poor descriptions were given. An attempt to collect all of the cases of situs viscerum inversus is extremely difficult and unsatisfactory, because they are reported in the literature under numerous headings. Perhaps something hke 300 of these cases have been reported. Since the spring of 1897 I have seen in hospital and private practice six cases of transposition of the viscera. It is a rather remarkable fact that three of these rare cases were seen within the short period of six months, and the last two cases within a period of one month, an experience which I doubt being equaled even in the largest hospitals of the country. I am also familiar 425 426 TRANSPOSITION OF THE VISCERA. with three other cases of situs inversus — discovered in the living subject by members of the Internal Medical staff, two by Dr. Warthin, and one by Dr. Cowie. The case which I saw first was with Dr. Chadbourne, in June, 1897, and has been reported by him, and perhaps by a number of others, for the man was a professional freak, who went the rounds of medical schools exhibiting his ana- tomical abnormaUty for a com- pensation. This same pa- tient, Schuppel by name, again presented himself at the hospital in July, 1902, and was demonstrated to my summer school section. He showed a complete transposition of the viscera, so far as could be demonstrated by clinical means. Inflation of the stomach, and colon, showed these viscera to be trans- posed. (See Figs. 1 and 2.) The seven cases besides Schuppel which I present in this article are bona fide pa- tients, two farmers, a dental student, a literary student, a housewife, a business man, and a laborer, who learned for the fiirst time of their peculiar abnormality. My experience inclined me to believe that this condition was much more frequently detected in the living patient than one would infer from the literature. Acting upon this belief, nearly two years Fig. 1. Photograph of Schuppel. — The upper curved line on the patient's right indicates the area of absolute cardiac dullness. The X just below this indicate? the apex beat. The curved line on the patient's right siae indicates the splenic dullness. The straight line on the patient's left side indicates the upper border of the liver dull- ness; the small x below indicates the lower border. The lowest curved line indicates the greater curvature of the stomach. ago I sent letters to a number of prominent clinicians, asking them the number of cases of congenital trans- position of the viscera they had encountered in their work. Letters were also sent to a number of anatomists. The following interesting statements were received : JAMES RAE ARNEILL. 427 Dr. Wm. Osier, Johns Hopkins Hospital: " Since- the opening of the hospital we have seen, I should say, at least six cases. " Dr. Eshner, Philadelphia Polyclinic: " I beg to state that I cannot recall having seen a case of transposition of the viscera. " Dr. J. C. N\'ilson, Jefferson Mcdii-al College, Philadelphia: "I ha\-c encountered three demonstrable cases of transiiosition of the \iscera, all adult niuli's, " Fig. 2. Radiograph of Schuppel. — Same case as shown in Fig. 1. Dark shadow is heart on right side, with apex pointing to right. Dr. R. H. Fitz, Harvard Medical College : " I do not recall the number of living cases seen; I have no recollection of having examined, after death, a case of dextrocardia. " Dr. Warren Coleman, New York: "I do not recall having seen a case with transposi- tion of the viscera. " Dr. I. N. Danforth, Chicago : " I never saw such a case in my life. " Dr. Allen Jones, Buffalo : " I have not seen a case of dextrocardia. " Dr. Frank Billings, Rush Medical College, Chicago : " I have seen three cases of con- genital transposition of the heart and abdominal viscera in my private and hospital work; one discovered in a babe about two weeks old, one in a young man between twenty and thirty, and one in a man of sixty-six. The last suffered from gall-stones and was operated by Dr. Fenger of Chicago. AU three were unrecognized until the time when I first saw 428 TRANSPOSITION OF THE VISCERA. each case. Besides these three I have seen a case in a young man "who traveled about exhibiting himself in medical schools. " Dr. John H. Musser, University of Pennsylvania, Philadelphia : " I have never seen any cases of congenital dextrocardia, except those belonging to and reported by other physi- cians, as Dr. J. Dutton Steele and Dr. Joseph Sailer. " Dr. J. M. Anders, Medico-Chirurgical College, Philadelphia : " Three cases, but they have been seen while visiting various institutions here and elsewhere. " Dr. R. C. Cabot, Harvard Medical College : " I have seen two cases of transposition of the Adscera. They were hospital cases in the service of Dr. Shattuck. " Dr. H. B. Newberry, Denver: "I have seen two cases, one a child of two or three, also reported by Dr. J. A. HaU and seen in consultation with him. A second in a young adult whom I was examining for life insurance. The abdominal viscera were displaced in both cases as shown by the liver. " Dr. James Tyson, University of Pennsylvania, Philadelphia : " I recall only one well- remembered instance of transposition of the viscera occurring in my clinical experience. I remember seeing one other case years ago in a dissecting-room. " Dr. James Munson, Superintendent Michigan Insane Asylum at Traverse City: "While a member of the staff of the Pontiac Michigan Asylum, I saw a case in which the diagnosis of complete transposition of the viscera was not made 'intra vitam', but at autopsy. During the last year of his life he had Addison's disease. I reooUect that previous to his death the diagnosis of an enormously enlarged spleen had been made, and that the dis- placement of the heart had also been recognized, but my recollection is that it was thought to be due to dilatation. " Dr. Robert Preble, Chicago Medical College : " I have seen four cases of transposition of the viscera, three clinical, one post-mortem. Two of the clinical were stock cases in Vienna, the third is now a student at North Western. In addition to these, there is another case in which the examination is incomplete, discovered while examining heart for anaes- thesia. " A. Jacobi, College of Physicians and Surgeons, New York : " Only two that I remember. " Dr. Atkinson, Baltimore: "Two cases of transposition of viscera appearing at my clinic, and who traveled widely exhibiting themselves. " Dr. C. G. Jennings, Detroit Medical College : " I have never met such a case. " Dr. I. W. Blackburn, National Hospital for the Insane, Washington, reported a very interesting case of transposition, which came to autopsy.' With his permission an excel- lent photograph of this case is presented in this article. (See Fig. 3.) Dr. C. B. Nancrede, Ann Arbor: "One case, seen in the anatomical laboratory of the University of Pennsylvania, of which I helped to prepare a wax model. This case was reported by Dr. N. Hickman. " Dr. Irving Haynes, Cornell Medical College: "Although I have superintended the dis- section of about 2,500 subjects I have never seen a case of transposition of the viscera. " Dr. Frederick Gerrish, Bowdoin Medical College, Portland, Maine: "I have never encountered a case of transposition of the viscera. " Dr. D. K. Shute, Washington, D. C. : " It has never been my good fortune to see any cases of transposition of the viscera. " Dr. S. Yutzy, Ann Arbor: " I have seen one case of transposition in ten years' service in the anatomical department. During this period about 750 subjects have been dissected. The ease was a middle-aged man. " Members of the staffs at the four Michigan State Asylums inform me that the Pontiac case reported by Dr. Munson is the only case to be found on their records. ' Washington Medical Annals, Vol. I, No. 2, 1902. JAMES RAE ARNEILL. 429 Since writing this article, Dr (retired), has reported to mc viscera, in a prominent officer U. S. A., whicli was first recog- nized in 1867, wlien he was examined for admission to West Point. The above writers are all men of large experience. Five well-known internists and four professors of anatomy- replied that they had never seen a case of transposition. Of the thirty-five cases noted in the above letters, all save six were discovered dur- ing life. In a few instances (travehng freaks) the same cases have probabh' been ex- amined and reported by several of the writers. The majority, however, were discovered for the first time, and did not make a practice of exhibiting themselves. In this country, at least, cases of congenital transposi- tion are at present much more frequently discovered during life than after death. Note the contrast between Gruber's re- port and my own. He collected seventy-nine cases, of which only five or six were discovered during life. In my collection, which simply covers the cases reported in the preceding letters, together with those with which I am personally famihar, there are forty-four cases, of which thirty-eight were dis- covered during life. P. R. Brown, major and surgeon U. S. A. a case of complete transposition of the Fig. 3. Dr. Blackburn's Case. Photograph. — Trans- position of viscera. Tei-ms transposed in description: a, Right common carotid artery; b, innominate vein; c, left common carotid artery; d, innominate artery; e, left lung (?); /, right lung (?); 3, pulmonary artery; h, aorta; i, right ventricle of heart; k, appendix of right auricle; I, diaphragm; m, right lobe of liver; n, suspensory ligament of liver; o, left lobe of liver; p, gall-bladder; q, cardiac end of stomach and situation of spleen ; r, pyloric end of stomach; s, gastrocolic omentum; t, trans- verse colon dragged downward; u, situation of cecum; v sigmoid flexure; w, omentum. 430 TRANSPOSITION OF THE VISCERA. Report op Unpublished Cases. Case I. — Wm. B. consulted me June 13, 1900, complaining of symptoms of indigestion. He is a farmer thirty-two years of age ; is tall, slender, and poorly nourished. On examin- ing his lungs by percussion, tympany was encountered in the sixth intercostal space, right nipple line — there being no liver dullness. On the left side dullness was encountered on the sixth rib, nipple line, and eighth rib, mid-axiUary line. Litten's phenomenon was present on both sides. Auscultation of both lungs showed good vesicular breathing. Heart. — I was much surprised to find the apex beat in the fifth intercostal space one inch inside of right nipple line, easily seen and felt after exercise. Percussion. — Showed tympany in the normal area for heart dullness. On the right side, however, there was an area of dullness encountered on the fifth rib sternal line, and extending about one inch to the right; no dullness to the left. Abdomen. — Liver dullness was encountered on the left side, nipple line, at the lower border of the sixth rib. It extended in this line down to margin of ribs; while in the middle line to one-third of the distance from ensiform to navel. Just to right of middle line liver dullness merged with heart dullness. The liver was not felt. Spleen. — In right mid-axillary line dullness was encountered at the upper border of the ninth rib and extends down to the tenth intercostal space. It was about the size of a silver dollar. The spleen and kidneys were not palpable. Stomach was distended, but unsatisfactorily. Testicles. — The right was lower than the left. Case II. — A. G. S. dental student, twenty-one years old, was sent to me by director Pitzpatrick of the gymnasium, who suspected some cardiac anomaly. There was no history of disease of the lungs, pleura or heart. He was never a good distance runner because of shortness of breath. Por years he has thought that his heart was on the right side because he could feel it beating there. This patient was in good health, but did not look robust. Examination of lungs was negative. Heart. — Apex beat was felt in fifth intercostal space just inside of right nipple; dullness began on fourth rib, right sternal line, and extended out to right para-sternal line. There was no heart dullness on the left or under the sternum. Auscultation was negative except that at times the rhythm was somewhat irregular. The second sound in right second intercostal space was louder than that in left. The sounds were heard best to the right, but there were weak heart sounds heard in the location of normal apex beat. Abdomen. — Li^-er dullness was encountered on the left side; upper border, left sternal line, seventh costal cartilage, nipple line, seventh rib, mid-axillary line eighth intercostal space. This dullness extended down to margin of ribs. Spleen. — In outlining the lower border of the right lung, I was surprised to find no liver dullness. On percussing in the right mid-axillary line I found a small area of dullness beginning at the ninth rib, and extending down about two inches and forward to the anterior axillarj' line. Stomach. — When distended with acid and soda, showed a dislocation; the greater curvature was about three inches below navel; the lesser curvature one inch above. The fundus of the stomach seemed to be on the right side, there being here a broader area of distention than on the left. Colon. — When distended with air, filled first on the left side; extending about two inches above the navel. There was very little distention on the right side. Testicles. — The left hung lower than the right. Spinal Column. — No curvature. X-Ray Examination. — With the fluoroscope I obtained a good view of the heart beat- ing to the right of the sternum. JAMES RAE ARNEILL. 431 Case III.— I. C, merchant, age sixty, patient at the University Hospital. Tliis man had a chronic bronchitis and chronic rheunintism. He never suspected any dislocation of his organs. He was emaciated. Percussion and auscultation of lungs revealed signs of emphysema and bronchitis, but of no condition which coula have caused a dislocation of his heart to the right. Heart. — Apex beat could not be seen or felt on left side, l)ut in fifth intercostal space, right side, nipple line, a faint pulsation was seen and felt. Percussion. — Along left sternal line tliere was good resonance down to the sixth rib, where liver dullness was encountered. On the right side absolute heart dullness was encountered on the fourth rib, sternal line, and extended to the right on a level with the fourth rib to the para-sternal line, with fifth rib to within one inch of the nipple. Auscultation. — -Over apex on right side the sounds were strong; no murmurs were detected. In the second intercostal space, right, the second sound was much accentuated; in the second intercostal space, left, the second sound was very weak. Abdomen. — I^nfortunately a complete examination was not made. Case IV. — E. W., male, age twenty, a student in the University of Michigan. The young man came to the out-patient department of the medical clinic, because of symptoms of an acute bronchitis. In making a routine physical examination. Dr. Prentiss Cleaves discovered the anomalous position of the organs. I was asked to examine the case by Dr. Cleaves and made the following notes : The patient, though not robust, was fairly well developed mentally and physically, and was greatly surprised when told that his organs were transposed. He is left-handed ; the left testicle hangs lower than the right, and he keeps the testicles on the left side. Examination of Lungs. — It was negative except for signs of an acute bronchitis. Heart. — The apex beat was distinctly seen and felt in the fifth intercostal space, right para-sternal line. It was also felt a little further out and likewise in the fourth intercostal space. There was a rather forcible shock in this region. A faint pulsation was seen and felt in the fourth intercostal space just to the left of the sternum. There was a dis- tinct heart shock on palpation in this region. Percussion. — There was resonance along the left sternal line as far as sixth costal car- tilage where relative dullness was encountered. In the mid-sternal line dullness was en- countered opposite the sixth rib. To the right of sternum dullness was encountered in the fifth intercostal space. It extended out to the right about one and one-quarter inch; on the left it ran into li^er dullness. Auscultation. — The heart sounds were heard best to the right of sternum. The first sound at the apex was accentuated and murmurish. The sounds at the base were rather weak, both seconds about alike in strength, perhaps a little stronger in the left second intercostal space. Ahdomen. — It was symmetrical, and a trifle below the level of the ribs. Liver. — There was no dullness on the right side in the normal liver region, but instead it was found on the left side. The upper border was seventh costal cartilage, para-sternal line, seventh rib, nipple line, ninth intercostal space mid-axillary line. The lower border corresponded ■^^•^th the margin of the ribs. Spleen. — Dtillness was very small, about the size of a silver dollar, in region of tenth rib, mid-axillary line. Neither liver nor spleen could be felt. Stomach. — The gastro diaphane was introduced and the light was distinctly seen in the right hypochondrium about the edge of ribs, in the region of the ninth rib. Inflation of stomach showed greater fullness on the right side. Sigmoid. — The electric light was not sat-isfaotorily seen. On inflation, judging by 432 TRANSPOSITION OF THE VISCERA. auscultation, the air seemed to enter the right side first, suggesting a right-sided location of the sigmoid. Fluoroscopic Examination. — Showed the pulsating heart on the right side, with the apex pointing to the right. Case V. — Mrs. A. P., age twenty-seven, came to the University Hospital because of epileptic fits. Through the courtesy of Dr. Herdman I was able to make a complete examination of this patient. She knew of her peculiar anomaly, having been informed of it by her family physician. She was rather proud of her interesting anatomy, and took keen pleasure in the quandary into which students were thrown on examining her chest. She was of a neurotic temperament and had frequent fits. A pelvic examination was made by Dr. Peterson and salpingitis and extreme anteflexion of the uterus were found. The patient was poorly nourished; weight, 105 pounds; height, about average. The hair was quite gray. ' , Thorax. — Small, slight funnel chest, fair expansion. The examination of the lungs was negative. Heart. — The apex beat was neither seen nor felt on the left side, but was distinctly seen and felt on the right side, in the fifth intercostal space about one-half inch inside the nipple, with patient sitting up. Percussion. — The heart dullness began on the fourth rib, right sternal line, and ex- tended to the right, almost to the para-sternal line. There was no dullness in the normal heart area, but instead the note was replaced by resonance and tympany, down to the fifth rib, where liver dullness was encountered. Auscultation. — To the left of sternum the sounds were very weak; to the right they were normal in strength and accentuation, except that the second sound in the second intercostal space, right, was much stronger than that in the left. Abdomen. — The liver dullness was encountered on the left side, with normal upper and lower boundaries. - -. Spleen. — There was a small area of dull tympany in the right mid-axillary line, about the ninth rib. The left kidney descended on deep breathing so that almost the entire organ could be felt. The right kidney could not be felt. Colon. — On distention with air through the rectal tube, the patient stated that she could feel the air going in on the left side, but auscultation with the stethoscope showed, with each pressure of the bulb of the syringe, louder sounds on the right side. Case VI. — I am indebted to Dr. Cowie for the following report of this case : Mr. S., traveling salesman, aged fifty. His symptoms pointed to a chronic gastritis. An inter- esting point in his history was the fact that he was examined for life insurance five times without having this anomaly discovered. Examination of the lungs was negative. Heart. — Apex beat not seen or felt in normal position. There was no dullness to the left of sternum. Apex beat was plainly seen and felt in the fifth intercostal space, just inside right nipple. Auscultation showed heart sounds loudest to right of sternum. Abdomen. — In place of liver dullness in right hypochondrium there was a tympanitic note. In this region succession sounds were brought out by deep palpation. In the left hypochondrium liver dullness was encountered, upper border nipple line, seventh rib. This extended down to the margin of the ribs, and as far forward as the sternum. Stomach. — On distention with air it was seen to lie in the right hypochondrium, par- tially in the epigastrium near the median line. It was not dislocated downwards. Kidneys were not palpable. Testicles. — -The right was much lower than the left. The genitals were kept on the right side. JAMES RAE ARNEILL. 433 Case VII.— The notes on this case are very brief. Patient was colored, male, twenty odd years of age, was examined in the out-patient department by Dr. Warthin. The condition of congenital dextrocardia was satisfactorily determined, but so much enthusiasm and interest were manifested in his case, that he refused to return for a complete examina- tion of the abdominal viscera, fearing that he was to be subjected to experimentation. Acquired dextrocardia is a fairly common condition, resulting from contracting disease of the right lung or pleura, which pulls the heart over under the sternum, or even beyond. Again, it may be due to large collections of fluid (serous or purulent) in the left pleural cavity, which push the heart to varying degrees to the right. Inflammatory adhesions may fix the heart in this position, even after the cause has been removed by absorption or operation. A less common cause is the presence of a new gro^Hh involv- ing the left lung and pleura. A very striking case of this latter variety came under my observation last spring in Dr. Dock's clinic. Fig. 4 illustrates well the conditions which were present. The patient was a small woman, forty- five years of age. She was emaciated, weighing only ninety pounds. Three years ago she weighed about 115 pounds. On inspection of the thorax, the left side was a trifle fijUer than the right. On deep breathing expansion was practically limited to the right side. Percussion showed absolute duUness over the entire left chest, front and back, except a very small area at the apex, which gave a tym- panitic note. The dullness also extended across the sternum in front to the right para- sternal line above the fourth rib, while below this it reached almost to the nipple line. This dullness and that of the liver were continuous. As the figure well shows, I obtained resonance over only a very small part of the thorax in front. Behind, the right lung gave a hjrper-resonant note over its entire extent down to four fingers below the angle of the scapula. Litten's phenomenon was very distinct on the right, the shadow line moving from the eighth rib down three inches in mid-axillary line. On the left side it was absent. Auscultation. — Over the dull area, front and axilla, breath sounds were practically Fig. 4. Acquired Dextrocardia. — New growth in mediastinum. The entire area between the dark lines was absolutely dull. O =apex beat of heart near right nipple. Finger tips show marked clubbing. 434 TRANSPOSITION OF THE VISCERA. absent; behind there was extremely weak vesicular breathing, perhaps conducted from the right side by the spinal column. Over the resonant area of the right lung there was exaggerated vesicular breathing. The voice sounds were well transmitted over the right resonant area, while practically absent over the left dull area. Vocal fremitus was very strong over right, but absent over left chest. Heart. — Apex beat was felt in fourth intercostal space nipple line, right side, most dis- tinctly one inch inside of nipple line. There was, also, a distinct pulsation in second and especially in third intercostal space, right, extending almost to nipple line. Heart duUness was continuous with the large area of duUness to the left and below; it extended to within an inch of the right nipple on a line with the fourth rib. Auscultation. — Over the dislocated apex the first and second sounds were moderately strong. Third intercostal space, right para-sternal, first sound was murmurish, second was accentuated; second intercostal space, left, sounds were weak. Pulse was small, of low tension and regular. After a consideration of these signs, I concluded that there was an enor- mous collection of fluid in the left pleural cavity, which caused a dislocation of the heart to the right. A large exploratory needle was used twice by the medical interne, Dr. Pray, with no results, beyond withdrawing a small amount of blood. Three days later I also used exploratory aspiration twice, but with negative results. As a result of these findings, I came to the conclusion that the cause of the dullness and dislocated organs was a new growth involving the left pleura, lung, and perhaps the mediastinum. Another possibility which must be considered is that of an old empyema, with greatly thickened pleura, and a large collection of thick purulent material. However, it seems most likely that we would have been able to demonstrate pus had it been present. A striking physical sign was the remarkable clubbing of the fingers, which is shown in the figure. The patient came to the hospital complaining of great weakness, a bad cough, and shortness of breath on exertion. The present illness began four years ago with a cold. She was then trou- bled ■\^^th shortness of breath on bending over, with pains in the left side and shoiilder. The cough became much more troublesome during the past two or three years. It was always dry until a few weeks ago, when the patient caught cold; since then she has raised some sputum. During the past few weeks she has had very little pain in the left side, but there is a feeling as if the lungs were compressed. During the patient's stay in the hospital her temperature seldom reached 100°. She came to the hospital with the diagnosis of consumption. Exami- nation of the sputum for tubercle bacilli while there was always negative. This patient died in January, 1902. Her attending physician, Dr. Fay, of Carlton, Mich., wrote me that autopsy revealed an enormous mediastinal tumor, about the shape and size of a medium-sized waterrnelon. JAMES RAF, ARNEILL. 435 Another striking case of acquired dextrocardia is well illustrated in Fig. 5. The change in the position of the heart is secontlary to an eni]\yema of the left chest of long standing, with extreme retraction of the (>utire left side;. The left side of the chest, with the exception of the apex, does not (>xpand. The right side expands fairly well. Heart. — Forcible pulsation is seen to the right of the sternum in the second, third, fourth, and fifth intercostal spaces, strongest in the fourth. The apex beat is in the fifth intercostal space, nip- ple line. AuscuUaMon. — At the apex both first and second sounds are loud and clear cut. In the second intercostal space, right, the second sound is accentuated. In the sec- ond intercostal space, left, both first and second sovmds are murmurish. In this area splash- ing sounds could be heard with each pulsation of the heart during forced inspiration, due in all probability to the sudden move- ment of the purulent fluid and air in the empyemic cavity (open pneumothorax) . The patient was taken sick in December, 1900. Two months later, he noticed the heart beating on the right side, in about the same posi- tion which it occupies at present. In December, 1901, Dr. Nancrede resected six ribs. Marked retraction of the left chest followed, but the position of the heart was not altered by the operation. Fig. 5. Acquired Dextrocardia. — Empyema with resection of ribs. Heart displaced to right en masse. X = areas of cardiac pulsation. Theories Explaining the Development of Congenital Transposition OF THE Viscera. These are chiefly of interest to the embryologist and anatomist, and will hardly be understood except by those who have devoted special study to embryology. The following may be mentioned : Von Baer explained transposition by the turning of the embryo in the opposite direction; that is, the embryo normally lies on the left side of the umbihcal vesicle, but if it lies on the right side, then we have transposi- 436 TRANSPOSITION OF THE VISCERA. tion. According to him, this occurs at the beginning of the developmental period. Forster considered situs inversus a malformation, in which the transpo- sition of the anlagen takes place in the first embryonal formation. In the double monster, the foetus of the right side shows a complete transposition, ■ while the foetus on the left side shows a normal situs. Fig. 6. Radiograph of same case as shown in Fig. 5. Heart displaced en masse, show- ing difference between congenital dextrocardia, as seen in Fig. 2, and acquired dex- trocardia. Rindfleisch believed that the spiral turning of the blood column is respon- sible for the displacement of the heart. Normally it flows from left to right, but in situs inversus an opposite direction must obtain. The asymmetry of the heart is made responsible for all asymmetry in the animal body. Virchow emphasized the influence of the umbilical cord. In situs inversus, it is wound spirally to the right ; in the situs solitus to the left. Kiichenmeister thought that the location of the fertilized disc at the sur- face of the egg is the essential thing. The normal situs in single birth proba- JAMES RAE ARNEILL. 437 bly depends upon growth of the germ from below upward, instead of from above downward. He says that from this it must be self-evident that the turning of the embryo has been inverted. This must also affect the later spleen side and the side of the arterial heart. Concerning the congenital partial situs viscerum, solito inversus, which shows itself either in the chest or belly, but not in both places at the same time, he believes that the growth on the whole follows the type for the situs inversus. The rarer partial situs is an inhibition formation which grows according to the type of the normally pro- jected embryo. Martinotti in the transA'crsus of the single born, emphasized the condition of the vena omphalo-mesenterica, first mentioned by Dareste. The direction which the heart loop takes, depends upon the dissimilar growth of the two halves of the vascular area. Under normal conditions a dissimilar formation of the two halves exists. The left omphalo-mesenteric vein is more devel- oped than the right; the right gradually disappears. The heart reacts in a very sensitive waj' toward the cause of situs transversus. Marchand held that the loop formation of the vena omphalo-mesenterica about the intestines under normal conditions prevents the intestines from sUpping toward the right. Hence turning to the right takes place if the loop formation is absent. He considers the left-sided persistent vena omphalo- mesenterica the cause of the right position of the stomach. In a more recent contribution, Marchand did not believe that the development of the vena omphalo-mesenterica can have an influence upon the rotation of the stomach. Lochte advanced the view that the growth of the organ considered in the sense of situs sohtus, is associated with a persistence of left-sided omphalo- mesenterica and umbilical veins, while those of situs transversus totalis, on the other hand, are associated with corresponding right-sided veins. To the cHnician, transposition of the viscera presents many interesting problems in differential diagnosis. The displacement of the heart to the right, makes it necessary to examine the lungs and pleura carefully, in order to exclude acquired displacements. The discovery of an enlarged area of dullness in the left hypochrondrium suggests a number of possibilities ; it is most hkely an enlarged spleen, either of leukaemia, malaria, spleno-megaly, etc. This point is illustrated by actual cases in practice. In Munson's case, referred to above, the diagnosis of an enormously enlarged spleen had been made; and the displacement of the heart was thought to be due to dilatation. In the normal patient it is a very common experience to find an entire absence of liver dullness in the right hypochrondrium. It is also common for the apex beat to be neither seen nor felt on the left side, especially if the patient is quiet and in the horizontal position. Heart dullness is also fre- quently absent. So it is easy to understand how these cases of transposition are often overlooked. 438 TRANSPOSITION OF THE VISCERA. It is possible to mistake an aneurism of the arch for a dislocated heart. This fact was recently brought to my attention. Gruber refers to the f ollo^^dng errors in diagnosis : In one case of trans- position, a pain in the right hypochrondrium led to the diagnosis of a chronic inflammation of the liver. In another case, a soldier was wounded in a duel in the right hypochrondrium; from the position of the wound, and the vomit- ing of green fluid, it was thought that the Uver had been penetrated. In a third case, in the Wiirzburg clinic, the transposed liver was diagnosed as a spleen tumor. In a fourth case, one of a cancer of the pylorus in a trans- posed stomach, the hard tumor felt deep in the left hypochrondrium as though to belong to the stomach or the pancreas. Gruber arrived at a number of interesting conclusions from a study of 79 cases. Concerning the sex, there were 49 men, 19 women, and 11 in which sex was not mentioned. These individuals lived as long as those with nor- mally placed organs. Five of the 19 women lived to an age between 70 and 84. The women were normally fruitful. One gave birth to 12 children. An:iong the 79, 4 died an unnatural death, and only 4 were extremely mal- formed. There was transposition of both chest and abdominal organs in 71 ; of the abdominal organs alone in 8. In the first kind the transposition was complete ; in the latter incomplete. Lungs were transposed in 35 of 71 cases ; the right had two lobes, and the left three. In two cases they were not trans- posed; in 2 both lungs had 2 lobes; in 1 the right had 1 lobe and the left 2 lobes. Spine: Curvature of the dorsal portion is mentioned in only 11 cases. In 7 of these it was to the left, in 4 to the right, as normally. We cannot draw the conclusion that persons with transposition are more likely to be left- handed than those with normally located viscera. The position of the tes- ticle was mentioned only 7 times. In 4 the right was lower, in 1 the left. In 1 the left had not descended. The lower position of the right testicle is unimportant as a sign of situs inversus. In only 9 cases were there notes on the position of the kidneys. In 7 the left was lower, in 2 the right. In 32 cases in which the vessels arising from the arch of the aorta are mentioned, these were transposed 29 to 30 times. H. Steinhaiiser mentions the fact that in the operation of oesophagotomy, it is well to know that the oesophagus lies to the right of the trachea in per- sons with transposition. In situs partialis, the transposition of the abdominal organs may be very irregular. In one case the stomach and duodenum were normally located, while the other organs were transposed. In another case the hver alone was transposed. In 1888, a case of pure dextrocardia, with congenital pulmonary stenosis, without malposition of the viscera in general, was shown to the Vienna Medical Society by Dr. Gruss. In discussing the case. Von Bamberger concurred in JAMES RAE ARNEILL. 439 the diagnosis, and remarked that Professor Schvotter had stated that no single case of pure dextrocardia had ever been proved, whereas all anatomists of experience, for example, Rokitansky, Friedberg, Forster, etc., had mentioned such cases, and he himself had seen two. The above quotation emphasizes the fact that partial situs is a much rarer condition than complete inversus. If the transposition is located in the abdominal cavity, it will most likely be overlooked in the physical examination. ON THE SUPEEIOE SPHINCTEE OF THE RECTUM. ROBERT C. BOURlcAND, A.B., BI.D., Late Instructor of Anatomy, Department of Medicine and Surgery, University of Michigan. (From the Anatomical Laboratory, University of Michigan.) Ever since Houston first described certain valvular projections upon the internal surface of the rectum, there has been a great deal of discussion con- cerning these structures, and many varying opinions have been enunciated as to their existence and character. At one extreme, stand those who deny absolutely the presence of any normally appearing projections, while at the other, are those who claim that three or four valves are constantly present. Within these boundaries is the statement that one valve (the most prominent one mentioned by Houston), is to be found in every normal rectum, while others are occasionally to be found. It is to this statement that the writer subscribes, basing his ideas on a number of observations, both on the cadaver and the living subject. Turning to Houston's account, we find the following description: "The valves exist equally in the young and in the aged, male and female, but some variations in number and position are seen. Three is the average number, though sometimes four, and sometimes only two, are present. The position of the largest and most regular valve is about three inches from the anus, opposite to the base of the bladder. The fold of next frequence is placed at the upper end of the rectum. The third in order occupies a position midway between these two, and the fourth, when present, is attached to the side of the gut, about one inch above the anus." Since the appearance of Houston's observations, the principal writers on the subject of rectal valves agree that there is one valve or projection con- stantly present. This is situated from six to eight cm. above the anus, oppo- site the base of the bladder, and characterized by a thickened mass of circular muscle fibers within it. It is with this most prominent valve, and its nature, that this paper has to do. A brief mention of the more important of the early viewb concerning this structure may not be out of place. Nelaton and Velpeau describe a structure which they call the sphincter swperieur, situated three or four inches above the anus. Hyrtl finds a marked constriction and narrowing of the lumen of the intestine, caused by a thickened mass of circular muscular fibers, pbced 440 ROBERT C. BOURLAND. 441 some three or foiir inches from the amis, and termed by him the sphincter ani tertius. Kohlrausch describes the flica transversalis recti, a structure similar in location and consistency to the pphincter sup6rieur. Baur, in 1863, made what is perhaps the most important contribution to the hterature of the subject which has yet appeared. In brief, his results are as follows: One principal vah-c, valve a, corresponding in position to the principal valve of Houston, partially encircles the gut in a spiral way, the right half of the valve being somewhat higher than the left. In twenty-one cases, ranging in age from embryos of four and five months, to middle-aged adults, the right half of valve a is situated 7.7 cm. and the left half approxi- mately 5 cm. above the anus. The right half of the valve is much the ?tronger and better developed of the two. In every case the valve was found to con- sist of a duphcature of mucous membrane, having at itt base a strongly devel- oped mass of circular muscle fibers. Next in importance to valve a is valve b, situated higher in the rectum. Besides these, other irregular and inconstant transverse folds may be found below valve a . The most important of Baur's results is his conclusion th-^t valve a, the sphincter ani tertius, the sphincter superieur, and the plica transversalis recti are one and the same structure, a statement which brought order out of con- fusion, and which has simplified greatly the discussion and investigation of the rectal vslves. Since the time of Baur, Bouisson has contributed an important paper to the subject under discussion. He states that there are transverse ridges in the rectum, which disappear when the gut is dilated, but that in addition to these, there are, in certain individuals, distinct valve-like processes high in the rectum which do not disappear, and which may be regarded as the imperfect expression of a third sphincter. Toldt finds a very strongly developed valve about 10 cm. above the anus, near the sacro-coccygeal junction. This is constant, and is situated on the right and anterior wall of the rectum, in the majority of cases, and is known as the plica transversalis recti. At its base, the circular muscle coat is thick- ened, forming the musculiis sphincter ani tertius. In recent years the most important papers on this subject have been those of Cooke and Martin. To the latter belongs the credit of having collected the greater portion of the literature bearing on the subject. Both Cooke and Martin agree with Houston in the salient features of their investigations concerning the possible number, constancy, and location of the valves. Thus far I have endeavored to show, by citing previous investigations, that, beyond all question, there is normally in every human rectum at least A -'■ "yJ>^ "-^: 442 ON THE SUPERIOR SPHINCTER OF THE RECTUTVI. one constHnt and ineffaceable transverse fold, • attached to the inner wall of the gut, from 5 to 8 cm. pbove the anus. I have never failed to find this structure. As a rule, it is located upon the posterior and right lateral wall of the rectum, being seen occasionally ante- riorly and to the left. In size the valve is exceedingly variable. In some individuals it may be merely a pmall transverse ridge, involving at its attach- ment from 2 to 4 cm. of the rectal wall, while in others it exists as a large semilunar fold, 2 cm. in breadth, standing well across the lumen of the gut, and involving at its base ( i/l.f ^, _ from one-third to one-half the cir- cumference of the rectum. In some cases a condition is seen corresponding to that described by ,'?>;I3'^'' ■:. .•^; ■:;■■■ V. .;■-■.■; ■ s'^'- " Baur, namely, a spirally arranged valve. This is brought about by the presence of a second valve, opposite to and somewhat below the principal : _, >"r,->- .. -~ - ■ -J- one, the two overlapping and sim- t'- • ulating a diaphragm. Histologically, the valve is seen Fig. 1. — Transverse section through rectal ^ ■ , r r i i r , , . 1 • J. • -ti,- to consist 01 a told oi mucous mem- valve showing sphincter superior within the valve, m, mucosa; mm, hyper- brane inclosing muscularis mucosse, a trophied muscularis mucosse; sm, sub- thickened submucosa — and as a rule, mucosa; ss, sphincter superior; Im, a thickened band of circular muscle longitudinal muscular layer of the ^^^^^^ ^here is, without exception, a distinct thickening of the circular muscular coat, either in the valve or at the base of the valve, or again m the immediate vicinity of the valve. That is to say, in some rare instances the thickening of the circular muscular layer m^y not correspond exactly to the location of the fold of mucous membrane, but may be situated a slight distance from it. In the fcEtal rectum the valve is seen to contain a duplicature of the cir- cular muscular layer, the entire coat, thickened at this point, lying within the valve. (See Fig. 2.) In the adult, the doubling of the circular muscular layer is not marked, but there is a very definite thickening of this layer. In all of the specimens examined microscopically, the muscularis mucosae was well developed, and in some cases was markedly hypertrophied, this over- growth being especially prominent in those cases in which the thickened mass of muscle was not actually within the valve. In the specimen shown in Fig. 4, the muscle mass was situated somewhat superior to the fold of mucosa, while the muscularis mucosae in the inferior wall of the fold was greatly thickened. ROBERT C. BOURT.AND. 443 rm suggesting that in this condition the absence from the fold of the thickened muscular layer is in some measure compensated for by the hypertrophied muscularis mucosae. The significance of this A-alve, ct)iistant in appearance, irregular in size, and showing a definite histological structure, is a matter of speculation. Com- parative anatomy throws little light on this subject. Baur, after investigations embracing a large munber of mammals, -,';■/,. ■• states that no mammal save man has ■^^-^i^"'''"■Z:■:^^ the plica transversalis recti. The most prominent features of the structure, and those upon wliich an h3rpothesis may be reasonably based, seem to be the fold of mucosa and the thickened mass of muscle. Of these, the most character- istic is the latter, and in the mind of the ^vriter, it is to this aggregation of mus- cular fibers that the question of the sig- nificance of the valve must be referred. The fold of mucosa is a structure of sec- ondary importance. Folds to the number of three or four may be found in the rectmn in varying locations, but they are apt to disappear entirely on dilata- tion of the viscus, and in the opinion of the writer, their morphological signifi- cance is Httle or none. In the majority of cases, the thick- ened muscular tissue is well within the valve proper, and forms the greater part of its substance, or it lies in the b?se of the valve. The rare instances in which the thickened muscular layer does not correspond exactly to the location of the fold of mucosa, may with pro- priety be considered variations from the normal, such variations as may be expected in a structure which is essentially irregular and in all probability rudimentary. H3rrtl ascribes to the structure the properties of a sphincter, saying that in diseased conditions necessitating the division of both the internal and external sphincter ani, the faeces are retained by the sphincter ani tertius, as he denominates the principal valve. Baur also thought that the valve plays a part in the withholding of faecal matter, stating that it seemed very possible for the sphincter tertius to hypertrophy and take the place of the internal sphincter when necessary, as in vicarious changes in other organs. Fig. 2. — Transverse section through rectal valve showing approach to the fcetal condition in the sphincter superior, m, mucosa; mm, muscu- laris mucosae; sm, hypertrophied submucosa; ss, sphincter superior; Im,, longitudinal muscular layer of the rectum. 444 ON THE SUPERIOR SPHINCTER OP THE RECTUM. From the fact that there is in every rectum examined a definite thickening of the circular muscular layer at one point, and at no other point, and that this thickened mass of muscle partially encircles the intestine, while in con- (Wfijv*- ~WiM Fig. 3. — Transverse section through rectal valve showing the sphincter superior at the base of the valve, m, mucosa; mm, muscularis mucosse; sm, submucosa; ss sphincter superior ; Im, longitudinal muscular layer of the rectum. '0^^^mx'4.i:it_^ %1*feW*^ junction with a similar structure upon the opposite wall, it may completely surround the gut in a spiral way, I am led to believe that in this irregular and peculiar structure we have, as Bouisson says, "the imperfect expression of a third sphincter." While I would not ascribe to the structure as it usually exists any marked sphincter action, it is certain that in a hypertrophic state it may exercise this function, both by means of the thickened muscular ring, and the thickened, fibrous submucous tissue. So that, in spite of its many irregularities and variations in size, one must recognize here the existence of a concrete structure, which may with propriety be accorded the name of sphincter superior, first given it by Nelaton and Velpeau. Doubtless a more detailed research upon the recta of lower forms will serve to throw new light upon the significance of the structure, and go far towards determining its real character, whether retrograde or a product of progression. Fig. 4.^-Transverse section through a rectal valve showing the sphincter superior above the valve and marked hypertrophy of the mus- cularis mucosae in the inferior wall of the valve. M, mucosa; mm, hypertrophied muscularis mucosie; sm, submucosa; ss, sphincter supe- rior; Im, longitudinal muscular layer of the rectum. ROBERT C. BOURLAND. 445 REFERENCES TO LITERATURE. C. B. Ball. — The Anal \':il\'es; Their Origin ;ind Pathogenic Significance. Mathew's Medical Quarterly, Louisville, 1894, A'ol. I, p. 191. H. Baur. — Ueber die Falten des Mastdarms. Beitrage zur Anatoraie und Physiologic. Giessen., 1863, Vol. Ill, p. 1. E. F. Bouisson. — Des vices de conformation dc I'anus et du rectum. Paris, 1851. J. R. Chadwick. — The Functions of the Anal Sphini'ti'r.><, so-called, and the Act of Defecation. Trans. Amer. Gynecol. Soc, 187S, Vol. II, p. 43. A. B. Cooke. — A Study of the Rectal A'alves — Experimental and Clinical. Philadel- phia Med. Journ., 1900, Vol. V, p. 964. . J. Henle. — Handbuch der Systenuitischen Anatomic des Mcnschen. J. Houston. — Observations on the Mucous Membrane of the Rectum. Dublin Hos- pital Reports, Vol. V, 1830, p. 158. J. Hyrtl. — Lehrbuch der Anatomic des Menschen. Wien., 1889. O. Kohlrausch. — Zur Anatomic und Physiologic der Beckenorgane. Leipzig, 1854. T. C. Martin. — New Evidence that the Rectal Valve is an Anatomical Fact. Mathew's Medical Quarterly, 1896, Vol. Ill, p. 310. J. M. Mathews. — Report on Diseases of the Rectum. American Pract. and News, Louisville, 1886, Vol. II, p. 26. Rosswinkler. — Beitrage zur Anatomic der Harnrohre und Harnblase des Mannes. Wien. med. Wochenschr., 1852, p. 435. R. A. Vance. — Rudimentary Structures in the Human Rectum. Medical and Surgi- cal Reporter, Vol. XXXVIII, 1878, p. 203. PAEOTITIS FOLLOWING ABDOMINAL SECTION. W. H. MOEIiEY, Ph.B., M-D., Assistant in Gynecology and Obstetrics, University of Michigan. Sequelae and complications incident to surgical operations upon the abdomi- nal and pelvic viscera are of interest alike to the specialist and to the general practitioner. Inflammation and swelling of the parotid gland, as a post-opera- tive complication of abdominal and pelvic surgery, are of sufficient rarity to demand the careful recording of every case. The relation of the parotid gland to the viscera of the abdomen and pelvis is one of the many unexplained prob- lems that exist to-day. A careful review of the anatomy of the parotid gland fails to explain why these glands, some days after the removal of the ovaries, become congested, swell, and sometimes suppurate. Many authors, among, them Stephen Paget, Bumm, and Goodell, believe the connection to be a nervous one, which ner- vous connection is made through the great sympathetic system. These authorities cite the complication of orchitis following parotitis in the male as an argument in favor of the sympathetic relation between the testis and the parotid, and state further that a similar condition no doubt exists in the female; but the position of the ovary and the reluctance of women to submit to an examination for a slight pain in the ovarian region, are sufficient reasons for many cases passing unnoticed. Some investigators state that secondary parotitis is caused by toxins being carried from the seat of the abdominal operation to the parotid gland by way of the lymph and blood channels. But the advocates of this theory fail to explain why these toxic agents exert a selective action for the neck organs and why pyemia and septicemia do not accompany every case of secondary parotitis. To state that there is a con- nection through the sympathetic nervous system between the parotid glands and the pelvic and abdominal viscera is simply one way of masking our igno- rance, but until some one suggests a more plausible theory, we shall be com- pelled to accept this in lieu of anything better. Before considering and reviewing the cases of secondary parotitis reported in the Uterature, I shall report briefly and succinctly a case of this kind follow- ing a salpingo-ovariectomy which occurred in Dr. Peterson's service at the Hospital of the University of Michigan. Miss E. G., aged twenty-one, was admitted to the gynecologic service of the University Hospital, October 9, 1901. Her family history was negative. Menstruation appeared at the age of ten, and has been regular both as to time and duration up to two years ago. 446 W. H. MORLEY. 447 when the present, illness began. She has al\v:i,ys In-en troubled with ocMistipation and fre- quent burning micturition, and she states that she has hiid " malaria" nearly e\Try sum- mer for a number of years. The present trouble began two years ago with painful menstruation and a dull heavy pain in the pelvis. She has had a gradually increasing le\iforrhcal discharge with a history of bladder trouble, together with se\ere peh-ic inflammation following the taking of a hea\-y cold last June. Two years ago the increasi- auscultation of the heart sounds. Bonders^ investi- gated the relatiA-e diuation of systole and diastole with various pulse rates, using the stethoscope in the usual way upon the human subject, recording the time of the somids by tapping a key connected with an electric signal which, along '(\-ith a chronograph, marked on a moving surface. He used mus- cular work as a means of increasing the pulse rate, but all his records were taken after the work ceased. Hering,* working upon dogs, studied the functions of the cardiac nerves. He recorded the respiratory moA-ements by a modification of Marey's method of air transmission, and connected his stethoscope with the rubber bulbs which served as receiving instruments. He was able to count the pulse in this way during both rest and work. Graphic records 'of the pulse date from the invention of the sphygmograph by Vierordt in the early fifties.^ His instrument was soon improved upon, and the principle was extended by Marey and others to the taking of records from other arteries than the radial, thus multiplying the possibilities of the method; but it was many years before the effects of muscvdar work received more than occasional and incidental attention. In 1863 Marey" published records of the pulse taken during inspiratory and expiratory efforts, incidental to an investigation of the effect of respiratory movements on the circulation. Thurston,' Edgren,* and Chapman,' using the graphic method, made extensive studies of the changes in systole and diastole accompanying changes in pulse rate. Here muscular work was used as a convenient means of quickening the pulse, but apparently was not of itself of any interest to the investiga- tors. Within the last decade the sphygmograph has been used in at least six ' St. Petersb. Med. Wochenschr., 1901, p. 79. ' Deutsches Arch. f. Klin. Med., Bd. LXXI, p. 539. ' Dublin Journ. Med. Sc, Vol. XLV, p. 225. * Arch. f. d. ges. Physiol., Bd. LX, p. 429. ' See Landois: Arterienpiols ; Berlin, 1872. ' Physiologie Medicale der Dang., Paris, 1863, p. 296. ' Journ. Anat. and Physiol., Vol. X, p. 494. « Skandin. Arch. f. Physiol., Bd. I, p. 67. » Brit. Med. Journ., 1894, I, p. 511. 464 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. different studies of the effects of muscular work. The work of Kolb/ extend- ing over several j^ears and including many kinds of athletic and gymnastic exercises, was published in England in 1892. Two years later Herman Christ^ published a study of the pulse before and after work done upon Dr. Jacquet's stair apparatus, using normal individuals and convalescents as subjects. The work by Christ was continued by Staehelin.' Egger^ studied the effect of pain, using the effect of Hght work as a control. Babbitt' made studies simi- lar to those of Kolb. Zuntz and Schumburg^ made an exhaustive study of the physiology of marching, including pulse records. In all these investiga- tions the ordinar}' clinical instruments were employed and apphed as usual to the radial artery. The studies were limited to the periods before and after the work. The plethysmograph has been used in studying the pulse by Athanasiu and Carvallo' and by Binet and Courtier.' In addition to the work done upon man, much valuable information has been secured upon this subject through work done upon animals by Marey," Kaufmann,^" Johansson," Jacob, '^ MacWilliam," Hunt," and others. These investigators made use of a mercury manometer connected with a cannula inserted in an artery. The superiority of the graphic method in a study of this kind over any method involving the observation of a series of phenomena by a single observer is obvious. A graphic record includes more than can be grasped by any observer, no matter how well trained. The records being preserved and read repeatedly, the chance of error is greatly reduced. Time relations can be worked out on a graphic record with a precision which cannot be approached by direct observation. But the sph3^gmograph and the plethysmograph, the two instruments hitherto used in studies of the pulse, are so sensitive to any movement or jar that the most essential part of the problem — the period dur- ing the work — ^must be given up if the work is at all vigorous. Not wishing to ' Physiology of Sport, London, 1892. 2 Deutsches Arch. f. Klin., Bd. LIII, p. 102. = Deutsches Arch. f. klin. Med., Bd., LIX, p. 79; LXVIII, p. 147. < Arch. f. Psychiat., Bd. XXXI, p. 274. * Am. Physical Education Review, 1901, p. 240. ' Physiologie des Marsches; Berlin, 1901. ' Loc. cit., p. 351. " L'ann^e Psyohologique, 1896, p. 30. ' Loc. cit. '» Arch, de Physiol., 1892, pp. 279 and 495. " Skandin. Arch, de Physiol., Bd. V, p. 20. " Arch. f. Anat. and Physiol., 1893, p. 305. '^ Proc. Roy. Soc, London, 1893, p. 464. " Am. Journ. Physiol., 1899, II, p. 395. WILBUR P. BOWEN. 465 be subject to such limitations, the writer mad(- a series of trials of different forms of appliances, and finally decided upon th(> apparatus described in the next section. Methods Employed in this Research. A graphic record of the pulse is taken from the carotid artery by means of two tambours connected by a tube. The receiving tambour, made from a tin box-cover, is 5 cm. in diameter and 1.5 cm. deep. This tambour is without the usual membrane, the open side being pressed against the skin of the neck over the artery by a U-shaped spring, one end of which is connected with the tambour by a ball and socket joint, while the other end terminates in a rounded block which bears against the opposite side of the neck. To prevent the edge of the tambour from cutting the sldn, a wire rim was put on so as to give a bearing surface 2 mm. wide. The connecting tube, which leaves the tambour from the side, is of light rubber near the neck, where lightness and flexibility are required; but as much of it as possible is of thick-walled rubber or glass. A tube of between 3 and 5 mm. bore gives best results. At some point in the tube a glass T-tube is inserted, and the side tube is fitted with a short piece of rubber tubing. The side branch is left open while the apparatus is put in place and adjusted, and is closed with a spring clip when a record is desired. One point in favor of this apparatus is its convenience of application; it requires but an instant to put the receiving tambour in place or to remove it. The essential point in which it differs from the similar instruments used by Marey and Edgrenis the appHcation of the tambour directly to the skin, without any intervening button or membrane. This greatly reduces the disturbing influence of movements of the body, but it does not entirely prevent them. It is therefore necessary to choose kinds of work in which the neck can be held comparatively still. The recording tambour is 1.5 cm. in diameter and is covered with a thin rubber membrane, which transmits its movements to the recording surface by a Ught celluloid lever. Three electro-magnetic signals are used. One, connected with a pendulum or fork, records the time; another, connected with the apparatus upon which work is done, records each movement or revolution; the third is used to indi- cate the time relations of two records taken at once, as described on page 489. The recording tambour and signals are mounted on a steel rod one foot in length, whose ends turn freely in sockets. The two sockets are clamped to two other rods, which in turn are clamped to a rigid support in such positions that the first rod turns in a vertical plane. The support is fastened to the table at a proper distance from the kymograph, and a horizontal rod is clamped to the pivoted rod so that it will strike the support just as the writing points reach the paper. A light rubber band holds the apparatus in this position 466 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. while a record is being taken, and a weight suspended by a cord passing over a pulley turns the ^^Titing points away when desired. The tambour recording the pulse is so placed that the axis of its writing lever is on a level ^^vith the writing point of the signal which records time, and the lever of the tambour is adjusted to write shghtly above the time record. With this arrangement the correct time relations of the pulse curve can be found by simply describing arcs cutting the two records, the radius being the length of the writing lever, and the center on the base line of the time record. This plan avoids the construction of a perpendicular for eafih interval measured, and thus saves a vast amount of time and lessens the chance of errors. The records are taken on papers 13 cm. wide and from three to five meters in length, which pass around two drmns ; one of these is the drum of a heavy kymograph, and the other a smaller one made of a section of brass tubing 5 cm. in diameter. The smaller drum is mounted on a rod whose beveled ends turn in sockets, and these are clamped to two horizontal rods, which are fastened to a rigid iron supporting standard. One leg of the latter is pivoted, allowing sufficient adjustment for tightening the paper. When the time is recorded in seconds, a continuous record can be taken by the use of one meter of paper for each four or five minutes; when a fork curve showing fiftieths of a second is recorded, a meter of paper is required for each thirty seconds. The table supporting the recording apparatus is fastened to the wall of the building, to avoid vibrations of the floor. The apparatus upon which work is done is placed on the floor and brought as near this table as possible, to avoid long connecting tubes.- In the experiments made thus far, heavy and moderate work has been done upon a lathe and a bicycle, while a telegraphic key has been used in studying the effects of the lightest work. The lathe is worked by a treadle moving through an arc of 15 cm. and carries a 55 pound fly-wheel 23 inches in diameter. A wooden lever rests in a polished groove in the rim of the fly-wheel, and a weight which can be moved along the lever provides a resistance which can be varied from 5 to 25 kg. The record is taken while the subject stands beside the lathe and rests more or less of his weight upon it by his arms, to prevent oscillations of his body, while he works the treadle with one foot. The movements are such that oscillations of the body are troublesome; and although several successful experiments have been made upon it, the lathe has been replaced in later work by the bicycle. The frame of a bicycle is fastened rigidly upright, the wheels removed, and the chain carried beyond the rear of the frame to a sprocket mounted on a shaft which carries a heavy fly-wheel and a broad band-wheel with a polished s\irface. While searching for such material in the scrap-room of a WILBUR P. BOWEN. 467 machine shop an old grindstone was found which auswcM-cd the reciuiremcnts without other labor or expense than the addition of the sprock(>t. The stone weighs about 125 pounds, is -well centered on a shaft which also carries a suitable band-wheel and is mounted on a hea\'>- iron frame. A picH'c of leather belting is fastened at one end, passed half A\-ay round th(> band-wheel, and attached to a weight-holder at the other end, making it possible to vary the resistance at the pedals from 4.5 kg. up to any desired amount by the addition of weights. The handle bars are set firmh' in one position, and an adjustable wooden support helps to hold the rider's shoulders steady. The momentum of the fly-wheel carries the pedals past the " dead point, " so that the sensation produced wliile running the machine is quite like that experi(>nced in riding a bicycle on a smooth road. By using toe-clips and taking care to pedal evenly, one can work upon this machine almost to the limit of his strength and record a pulse curve in which the pulse beats caii be counted readily and a large per cent of them analyzed. (See Figs. 1 and 7.) The telegraph key is operated with the forefinger, and has a range of move- ment of 1 mm. and a resistance which varies between 40 and 4,000 grams. The subject is seated before the table upon which the key is fastened, his arm resting on the table, to avoid the necessity of any movement or effort other than that required to move the key. All the experiments were made first with the wTiter as the subject; then they were repeated upon others as many times as circumstances permitted. The writer's age is thirty-eight, height and weight medium, health good; ac- customed to a large amount of exercise from 1890-1900, but having much less of it at present. The other subjects have been men in good health, but not in athletic training, mostly medical students between the ages of twenty and twenty-five. The Change in Pulse Rate when Muscular Work Begins. 1. The Latent Period. — With the beginning of muscular work the pulse quickens with surprising promptness and rapidity. To find how soon the acceleration begins, a series of experiments were made in which the work con- sisted in tapping the key as rapidly as possible, the records being taken with rapid drums and timed with a fork. The general character of the records is shown by Fig. 2 ; the results are given in Table I. As may be seen by a glance at the table, the four heart cycles immediately preceding the beginning of the work show changes in duration such as are usually seen in the pulse of a resting subject; changes that in the main are small, the length of the cycle alternately increasing and decreasing with a rhythm which differs in the different cases, and which therefore cannot be due to the work nor to any constant condition of the experiments. 468 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. Table I. Duration op Consecutive Heart Cycles in Hundredths of a Second. Subject. Before Work. Work b'gins During Work. Work ends. After Work. W. P. B ... 76, 76, 78, 78 74 70,66,64,62,64, 60, 62, 62, 62, 64 64 64, 66, 68, 74 78, 80, 77, 80 82 73,69,70,70,67. 67, 66, 66, 65, 63 66 67, 74, 72, 73 " 76, 74, 75, 78 77 72, 68, 66, 70 68, 70, 69 66 70, 72, 74 o.M. c. :;: 88, 90, 96, 100 98 90,82,74,74,74. 74, 74, 72, 70, 76 74 80, 86, 86, 84 E. A. B. ... 76, 78, 78, 80 80 76,64,60,60,64. 60, 66, 64, 66, 64 64 68, 74, 76 G. R 84, 84, 88, 88, 90 80, 78,76,74,68. 70, 76, 76, 74, 74 74 72, 78, 86, 84 H. H. P .. 92, 94, 88, 90 90 84,82,76,74,70. 70, 64, 62, 60, 58 58 62, 74, 72, 76 R. C. R. . . . 94, 88, 90, 90 92 86,82,74,72, 72. 70, 76, 78, 74, 74 74 88, 90, 96, 90 C. J. L. ... 90, 94, 96, 98 , 90 86, 84, 70, 70, 70. 70, 80, 84, 76, 82 76 , 76, 82, 88, 84 W. P. B. . . 76, 72, 72, 76 78 74,70,66,64,62. 64, 60, 62, 62, 64 64 66, 68, 74, 76 74, 76, 74, 72 74 70, 66, 62, 62, 64. 62, 64, 62, 62, 64 64 68, 66, 70, 72 77, 81 83 80, 77, 74, 70, 71 . 70, 71, 69, 69, 67 70 79, 85, 89, 92 " 77, 82, 85, 83 80 79, 70, 65, 64 65, 65 73 81 " 89, 92, 93, 89 83 77, 74,75,73,76. 73,76 78 81, 93, 89 " 85, 89 92 83, 74, 74, 72 76, 78, 77, 79, 82, 83 85 86, 84 " 72, 72, 72, 72 71 66,66,66,64,64. .. 68, 70. 72, 93, 66 89 66 92 64,62,64,64,62. 88, 77, 76 B. p. N. '.': 92, 84, 82, 84, 80, 78, 80, 96 84 90 80 84 88 80 90 80 80 86,84 78,76 78, 80, 80 78,70 76, 74. 75 The causes of this variation in the length of the heart cycle in the resting subject have been investigated at great length, Traube/ Hering/ Sommer- brodt,' Fredericq/ and Gley'' being prominent among earlier investigators, while Binet and Courtier" and Lombard and Pillsbury' have studied the subject more recently. As the result of all this work the minor changes in the duration of the cycle during rest are believed to be due to respiratory, vasomotor, and psychic changes. These changes appear to continue without interruption into the cycle in which the work begins, for this cycle has the same average length as the pre- ceding one, while as to the phase of the change there are eight cases with an increased length, ten cases with a decrease, and five cases with no change. Some of these cases of decrease in the length of the cycle may be due to the work, or partly so ; but in view of the fact that on the whole the changes are substantially the same as in the cycles preceding, it is not safe to assume that the work has had any effect at all up to this time. With the next cycle after the work begins there is a marked change. The cycle is shorter than the one before in every case; the amount of the change is in general much greater than the preceding changes; and the decrease con- ' Centralbl., f. d. Med. Wissensch., 1865, p. 881. 2 Sitzungsb. d. k. Akad. d. Wissensch., Wien, 1869, XL. ^ Deutsches Arch. f. klin. Med., Bd. XXIII, p. 542. * Arch, de Biol., 1882, III, p. 55. ^ Arch, de physiol. norm, et pathol., 1881, p. 742. « L'ann^e psychologique, 1896, p. 42. ' Amer. Journ. Physiol., Vol. Ill, p. 201. WILBUR P. BOWBN. 469 t2 ■¥ £ jg ^ CD a-S Rf &1 fl<&< 470 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. tinues without interruption for several cycles following. Such changes must without question be attributed to the work. The number of these experiments is too small to serve as the basis of an exact determination of the latent period, but the change is so prompt and so constant in response to the work that it may be used in arriving at a rough approximation. It must be observed first of all that it is not possible to begin the work in each case at a chosen point in a cj^cle ; the beats follow each other too quickly. In this series the records show that the work began in the first quarter of the cycle in five cases, in the second quarter in five cases, in the third in seven cases, and in the last quarter in six cases. It will be shown in a later section of this paper that the acceleration of the pulse at the beginning of muscular work is wholly due to a shortening of the diastole ; in other words, we have no evidence from the pulse curve of any response on the part of the heart until the end of the next cycle after the work begins. The delay, there- fore, which occurs between the beginning of the work and the first evidence of quickened heart action shown by the pulse curve varied in the twenty-three cases between one and two heart cycles as its extreme limits. The variation in the length of the delay is evidently due to variation in the time of beginning the work. For example, the two cases in which the work began earliest in the cycle show a delay of 1.49 seconds, while the six cases in which it began in the last quarter of the cycle show delays varying between 0.85 and 1.11 seconds; the delay being longer in the first instance simply because the work began farther away from the firSt point on the pulse curve at which a change in pulse rate can be observed. It follows that the latent period of the heart when influenced by muscular work cannot be far from the shortest delays observed, or about the length of one cardiac cycle, which A-aried in the present cases from 0.64 to 0.90 seconds. It is interesting to notice in this connection that Bonders^ found a latent period of the dog's heart, when influenced through the vagus nerve, varying from .6 to .8 of a cardiac cycle. By cutting the vagus in the dog Hunt^ ob- tained a shortening of the cycle in which the cutting occurred; he also obtained by stimulation of a sensory nerve a shortening of the next cycle after the work began, and by stimulation of the accelerator nerves a shortening which took place after a delay of several seconds. 2. The Manner in which the Pulse Rate Changes. — Another important point brought out by the results in Table I is the rapidity with which the pulse rate changes at the very start. A glance at the table is sufficient to observe this in a general way, but since the minor changes in rate that have been mentioned continue into the period of work and cause small irregu- larities, we get a more definite picture of the change from averages. The last ' See Tigerstedt's Physiologie des Kreislaufes, p. 239. 2 Loc. cit., pp. 436, 441, 438. WILBUR P. BOWEN. 471 eight cases in the table were taken to study hitent period only, and do not include enough cycles for the present purpose; excluding these and taking the average duration of consecutive cycles from the remaining fifteen cases we have: before the work begins, 82.3, 83.0, 83.4, 84,8; at tlie time of beginning, 84.2; and during work, 78. 6, 73.7, 69.4, etc. The average duration of the cycles preceding the work and including the one in which it began show only minute variations, but vnth the commencement of the work we have a shorten- ing amounting to 17.5 per cent within the time of three cycles; and of the three the change is greatest in case of the first one. In order to study the manner of this change through the entire period of acceleration, records were taken during work lasting from one to six minutes. In these experiments the records were taken on slowly moving drums and timed in seconds. The results are given in Tables II, III, and IV with suffi- cient fulness for present purposes. In compiling the tables the number of pulse beats for each period of ten seconds was first determined from the rec- ords; each of these numbers was then multiplied by six, giving the pulse rate per minute for the consecutive periods, as shown in the tables. In each experiment we fuid immediately after the beginning of the work a rapid rise in pulse rate, which will be called the primary rise. This rise, after continuing for a period varying from ten seconds to two minutes, gives place to an approximately uniform portion, which will be spoken of as a plateau. For the sake of clearness it is assumed that the primary lise continues until it merges into a plateau lasting at least thirty seconds. Any further prolonged rise wiU be called a secondary rise; any brief rising and falling of the rate will be called minor changes. Table II. Pulse Ratks per Minute for Consecutive Periods op Ten Seconds, Showing the Effect of Tapping a Key Rapidly for Periods Varying phom One to Six Minutes. (Seven Subjects.) Number of Experiment 10 11 12 13 f Before Work . { I During Work After Work 66.0 68.4 66.0 76.8 84.0 88.8 89.4 90.0 92.1 91.2 67.8 66.0 67.8 66.6 7.3.0 7.5.0 78.0 96.0 102 105 102 99.0 103 102 68.0 66.0 67.8 66.6 69.0 63.0 67.2 85.8 94.8 99.0 96.0 97.8 87.0 92.4 82.8 76.2 72.0 73.2 75.6 78.0 75.0 86.4 87.0 90.0 90.0 90.0 93.6 92.4 81.0 78.0 78.0 64.8 64.8 66,0 78.0 84.0 86.4 90.0 90.0 85.2 86.0 86.0 60.0 60.0 60.0 79.2 78.0 76.2 90.0 90.0 81.6 78.0 78.0 79.2 60.0 66.6 67.0 84,0 97.2 102 102 102 97.2 94.8 81.0 73.2 73.0 69.0 58.8 60.0 63.0 69,0 80.4 83.4 81.6 88,2 96.0 96.0 73.2 66,0 60,0 61,2 63,4 52,2 64.0 66.0 72.0 81.0 84.0 84.0 84.0 84.0 66.0 64.2 64.2 58.8 66. Oi 65.4 69.0 70.2 67.2 72.0 78.0 78.6 84.0 84.0 84.0 76.8 72.0 70.8 81.0 81,0 81,0 80,4 83.4 80.4 78.6 72.0 67.2 67.2 66.6 67,8 66,6 65,4 72,0 81,0 82,2 84,0 83.4 83.4 78.0 72.0 69.6 66,0 72.0 72.0 72.0 78.6 85.8 88.2 94.2 94,2 88,6 90,0 84,0 80,4 81,0 80,0 72,6 75.0 72.0 87.6 91.2 90.0 92.4 90.0 90.0 90.0 75.0 77.4 72.0 472 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. Table III. Pulse Rates per Minute fob Consecutive Periods of Ten Seconds, Showing the Effect of Five Minutes of Lathe Work. Work Done by One Subject. Experiment Number. 1 2 3 4 f 69.0 66.0 60.0 64.0 69. C 66. C 58. S 67.8 [ 72.0 67.2 58.8 69.0 f 81.6 84.0 78.0 * 90.0 87.0 82.8 82.2 90.0 84.0 86.4 88.8 90.0 87.0 87.0 90.0 90.0 84.0 84.0 91.8 90.0 84.0 84.0 93.0 90.0 87.0 86.4 93.0 90.0 82.8 87.0 93.0 90.0 88.2 84.0 90.0 88.8 91.2 90.0 94.2 88.8 94.8 90.0 94.8 [ 88.8 90.0 90.0 90.0 r 84.0 75.0 75.0 81.6 75.0 72.0 67.8 78.0 72.0 71.2 64.8 72.0 72.0 67.8 62.4 72.0 72.0 69.0 62.4 66.0 72.0 66.0 60.0 64.8 8 9 10 Before Work . During Work After Work . 69.0 70.8 66.0 60.0 64.8 60.0 87.0 * 96.0 93, 98.4 99. 96.0 100. 99.0 106. 99.0 114 99.0 117 100 126 100 126 105 102 105 129 132 132 84.0 105 78.0 96. 72.0 93. 69.0 90. 69.0 87. 66.0 84. 72.0 72.0 72.0 90.0 102 111 114 120 126 135 141 144 144 144 144 126 120 114 114 112.. 105 73.2 75.0 78.0 72.6 75.0 71.4 73.8 76.2 68.4 100 112. 126 132 138 144 144 156 150 156 150 144 126 123 111 109.2 108 104 109.8 120 132 138 141 142.2 144 150 150 153 121.8 111.6 106.8 102 96.0 96.0 108 124.2 134.4 144 150 156 156 156 163.8 163.8 156 138 121.8 115.2 112.8 111 112.2 * Pulse curve not legible. Table IV. Pulse Rates per Minute for Consecutive Periods of Ten Seconds, Showing the Effect of Five Minutes of Bicycle Work. Work by One Subject, Except in 10. Experiment Number. 1 2 3 4 5 6 7 8 9 10 11 61.2 56.4 56.4 76.2 70.8 71.4 79.8 76.2 79.2 72.0 66.0 Sitting in Chair 61.2 56.4 56.4 69.6 68.4 71.4 77.4 76. S 78.0 75.0 68.0 64.8 58.8 56.4 70.8 70.8 73.2 79.8 73.2 77.4 74.2 67.2 70.8 61.2 61.6 73.2 70.2 78.6 82.2 79.2 81.0 80.4 78.0 Upon Bicycle \ 66.0 60.6 61.2 72.0 70.8 76.8 82.2 79.2 82.2 80.4 78.6 72.0 66.0 64.2 74.4 73.2 75.6 83.4 83.4 82.8 76.8 80.4 78.0 80.4 78.0 84.0 96.0 99.0 102 97.8 102 93,0 inn, 2 80.4 79.8 87.0 96.0 100.8 108 106.2 108 120 100.8 112.4 79.2 82.8 90.0 96.6 103.2 115.8 109.8 117.6 132 105 116.4 73.2 82.2 93.0 99.6 107.4 123 112.8 124.8 142.2 108.0 118.8 76. S 76.2 90.0 102 109.8 128.4 114 130.8 150 112.2 121.8 76.0 78.0 89.4 100.8 111.6 129 114 136.8 144 116.4 126 76.2 80.4 96.0 97.8 114 130.8 120 139.2 144 123.6 132 During Work .... 75.0 80.4 82.8 82.2 96.0 95.4 97.2 96.0 115.2 115.2 132 132.4 117 117 142.8 144.6 156 162 126 126 132 132 82.2 84.0 93.0 102 115.2 136.8 120 147 168 126 132 77.4 84.0 97.2 98.4 114 136.8 118.2 148.2 168 126 132 78.0 78.0 93.6 97.2 111 132 121.2 150 180 126 138 73.8 80.4 94.8 91.2 113.4 147 123 156 180 142.4 144 73.2 80.4 96.0 94.2 110.2 144 120 156 180 144 144 , 78.0 80.4 96.0 94.8 109.8 144 122.4 159 180 141 144 73.2 73.2 86.4 90.0 102 135 111 160 174 132 138 67.8 72.0 79.2 81.0 94.8 129.6 105- 138 162 132 126 After Work < 64.8 65.4 67.2 64.8 75.6 72.0 79.2 73.2 90.0 87.6 120 110.4 98.4 102 126 120 156 151.6 126 114 117 114 66.0 63.6 68.4 72.0 82.8 108.2 102 117 147.6 104.4 108 70.8 64.8 67.8 72.0 78.0 102 97.2 120 142.2 99.0 110 WILBUR P. BOWEN. 473 It is desirable to study tlie aA-erage of the pulse rates given in a series of cases, when permissible, to throw out the disturbing minor changes. Taking an average of the fourteen cases of Table II during the primary rise we have as follows : Average rate per minute . . . Increase over preceding pe- riod Bef(»re work. Last 1 sec 68.6 During Woik. 1st. 10 sec, 80.5 11.9 2cl. lOaeo. 86.5 6.0 3cl. 10 sec. S9.4 2.9 4th. 10 sec. 90.7 1.3 5th. 10 sec. 91.3 0.6 It is plainly the rule here for the acceleration to be most rapid at the start and for the amount of change in each successive period to become less and less until the primary rise is finished. Plotting a curve to represent the average Pulse Rate i 1 1 1 — •- -. 1.- — .1..., \ / y' i / \ 60 •— «•— ■ "*". ^•- FiG. 3. Cvure of pulse rate, showing the effect of tapping a Icey rapidly. Plotted from the average of the fourteen cases of Table II. Arrows indicate the time of beginning and stopping work. pulse rate of the fourteen cases, laying off units of time as abscissse, and pulse rates per minute as ordinates, we obtain the curve shown in Fig. 3. In the experiments of Tables III and IV there was such wide variation in the speed and intensity of work that an average is hardly practicable; but the results given in these tables agree with the above rule in every case, except as minor changes cause apparent exceptions in some instances in the later stages of the acceleration, when the changes due to the work have become small. Choosing Experiment 8 of Table IV as a typical example and plotting its ciirve we obtain Fig. 4. Figs. 3 and 4 picture clearly the character of the primary rise : very rapid at first, then gradually subsiding, and finally merging into a plateau. There are some small irregularities in these tables which result from the manner of reading the curves. It rarely happens that the line separating the 474 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. periods of ten seconds coincides exactly with the beginning of a primary pulse wave, and consequently fractions of cycles are of common occurrence in the results. It was not considered necessary, in view of present purposes, to meas- ure the fractions of cycles accurately, but instead they have been estimated care- fully by the eye with the aid of a hand lens. It may be observed that in Table III the fractions are not estimated so closely as in II and IV. This is because in the lathe work oscillations of the body caused slight defects in portions of the record, making it impossible to determine the fractions as accurately as I^ulse Rate —•'' — — ^— "^e"^ ^•- \ ,•' ,.' ^•' -.■' \ / i \ ' i / \ i / 1 1 i \ A ^_ 1 ''.^' 7' / 1 \J — ^ ji (10 T i 1 1 1 T Fig. 4. Curve of pulse rate, showing the effect of five minutes of vigorous work on a bicycle. Plotted from results of Experiment 8 of Table IV. M, mount ; arrows indi- cate time of beginning and stopping work. in other work. The irregularities in Table II during the period of the plateau are no doubt due, in part, to changes in the speed and force of the work. All of these changes, however, are so small in comparison with those resulting from the work that they can have practically no effect on the general form of the primary rise, as illustrated by Figs. 3 and 4. The superiority of the graphic method for studies of this kind is well illus- trated by the completeness and accuracy with which these results show the sudden and rapid primary acceleration of the pulse under the influence of work. Griinbaum and Amson"- have been over this ground with great care, using the most perfect method of direct observation of the pulse ever employed; yet, since they could not determine the duration of individual beats, they could not find the exact manner of the change nor the latent period ; and since they could not count uninterruptedly, they missed critical periods. In fact, their ' Loc. cit., pp. .557-586. WILBUR P. BOWEN. 475 protocols show that thi\\- missed in evci\\' case the few seconds during which the change is most rapid. This will accomit for some minor differences between their curves and mine. 3. The Caiiscs of the Primary fijsc— The rate of the heart can be influenced hj two groups of causes; those acting directly upon the heart, and those act- ing indirecth- through the nerxous mechanism A\'hich controls it. Among in- fluences acting directh- upon the heart may be mentioned heat and waste pro- ducts carried from the working muscles in the blood stream; probably heat and waste products formed in the heart as the result of its own increased activity ; and circulatory conditions determining the supply of blood to the heart and resistance to the discharge of blood by it. The nervous mechanism which controls the heart is influenced b>- a similar set of conditions : heat and waste products brought by the blood, changes in blood pressure, and above all by nervous impulses coming from all parts of the body. When we compare the latent period of the heart and the rapidity of the primary acceleration of the pulse with the time required for these various causes to become effect! A-e, we are compelled to conclude that the only influ- ence acting quicldy enough to account for the first change in pulse rate when muscular work begiiis is that of nervous impulses arising in the motor centers of the cerebral cortex and in the muscles, and acting upon the inhibitory center. In some cases of heavy work what we have called the primary rise lasts long enough for all of these causes to become effective; in some cases of light work, the plateau is reached so soon as to exclude all of them except the effect of nervous impulses acting on the heart centers. Consider first the causes acting directly upon the heart. The effect of waste products in the blood has been studied by Johansson^ and by Athanasiu and Carvallo,^ who foimd that these were capable of quickening the heart beat. Neither of these authors gives the length of the latent period of this effect, but both say that it is long. The products of metabolism, before they can act upon the heart, must traverse half of the systemic circulation, the whole extent of the pulmonary circulation, and enter the coronary arteries from- the aorta. According to the more recent estimates of the circulation time in man this must require from thirty-six to forty-five seconds,^ if we make no allowance for the diffusion of the waste products from the tissues into the blood. All investigators agree that the acceleration of the heart caused by waste products comes on slowly and increases gradually as work continues. It is certain, therefore, that this cause cannot account for the first change in pulse rate. ' Loc. cit., pp. 60-62. ^Loc. cit., p. 561. ' Stewart: Manual of Physiology, Phila., 1900, p. 124; Richet: Diotionaire de Physiol- ogie, see Circulation. 476 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. The heat developed in the working muscles, like the waste products, must travel in the heart in the blood stream before it can influence the rate of the beat; it cannot, therefore, become effective in causing acceleration much sooner than the waste products. Anything which causes more rapid filling of the heart thereby tends to quicken the beat, but there is no conceivable waj' in which light work can bring this about the instant the work begins, unless it should be through a compression of the abdomen causing an emptying of the abdominal veins. An act involving violent effort often causes strong contraction of the abdomi- nal muscles and consequent compression of the abdominal viscera; but this phenomena is not marked in light work, and cannot account for the first accelera- tion, which comes so promptly, even in the lightest work. Several tests were made of the effect of compressing the abdomen strongly by means of a strap, but the acceleration of the pulse produced was very slight and lasted only a few seconds. Marey^ and Kaufmann,^ who worked upon the horse, lay great stress upon the dilation of peripheral vessels and the resulting fall of blood pressure as the cause of the increase in heart rate during work. Work cannot bring about the first beginning of dilation of vessels in less than three seconds,^ and accord- ing to MacWilliam^ it requires about thirty seconds to cause any increase in the heart rate after a considerable fall of blood pressure has taken place. Even if we assume that the results obtained by Marey and Kaufmann hold good in case of man, we must still look elsewhere for the cause of the first acceleration at the beginning of the muscular work. ' Contrary to the theory of Marey and Kaufmann, most recent observers* agree that muscular work causes a rise of blood pressure in man. It is not at all probable that a rise of blood pressure would cause an increased pulse rate; but in the absence of positive proof it may be well to consider the time in- volved. As stated by Tigerstedt,^ the height of the blood pressure at any time depends upon: (1) the resistance of the peripheral vessels; (2) the mass of blood in circulation; and (3) the energy of the heart's action. A constriction of peripheral vessels by action of the vasomotor center requires at least 1.5 ' Loc. cit. 2 Loc. cit., pp. 279, 497. " Francois-Franck : Travaux du Laboratoire de Marey, 1876, p. 39; and Bowditch and Warren: Joum. Physiol., 1886, AT:I, p. 440. * Loc. cit., p. 470. ^ Edgecumbe and Bain: Joum. Physiol., Vol. XXIV, p. 48; Maximowitsch and Rieder: Deutsohes Arch. f. klin. Med., Bd. XL VI, p. 329; McCurdy: Am. Joum. Physiol., Vol. V, p. 95; Kornfeld: Wiener Medicinische Blatter, 1899, p. 631; Masing: Deutsohes Arch, f. klin. Med., 1902, Bd. LXXIV, p. 253. « Loc. cit., p. 332. WILBUR P. BOWEN. 477 seconds.* The work cannot instantly incroasp the mass of blood in circulation except through effort great enough to compress the abdominal viscera and drive blood from there into the general circulation; a result not produced in hght work, as previouslj^ stated. It is \'or>- improbable that increased pres- sure can have anything to do with the first increase in pulse rate accompany- ing light work. i\Iost -observers report that changes in arterial pressure have little or no immediate effect on the rate of the pulse.^ A cause of increased heart rate which I have never seen mentioned as such is the formation of heat and waste products in the heart itself, as the result of its owia increased activity. This cannot of course take place at the be- ginning of the acceleration, but in consideration of the very rapid rate that we find accompanying work, it may well be considered as an influence of con- siderable importance later. Causes influencing the heart through its nervous mechanism show a marked difference in effect depending upon which of the cardiac nerves is involved. Withdrawal of the inhibitor}- influence of the vagi increases the pulse rate rapidly after a short latent period; increase of accelerator influence increases the pulse rate slowly after a long latent period.^ It follows that the first acceleration accompanying work must arise through diminished vagus action. As to the origin of the nervous influences which act upon the cardiac cen- ters to cause the acceleration during work, several views have been advanced. Hering* concluded from his experiments upon the dog that an increased ampli- tude of respiratory movements gives rise in the lung to impulses which inhibit the vagus center. Examination of the respiratory curves, which accompanied the pulse curves in nearly all my experiments, show that no increase in the amplitude of respiratory movements occurs at the beginning of work; on the contrary, the breathing movements are always diminished at first, and some- times aU but suppressed. Several experiments were made to test the effects upon the pulse rate of such changes in respiratory movements as occur in work. These changes in respiration alone caused only slight effects on the pulse rate, and the usual effects of the work were not in any apparent degree affected, either by volimtary suppression of the breathing or by maintaining it at nor- mal depth and rate during the work. The resiilts of the tests showed plainly that the cause assigned by Hering cannot account for the first acceleration. Johansson^ compared the effects of voluntary and reflex movements, and concluded that the impulses causing the change in pulse rate originate in the motor centers of the cerebrum, and that the impulses arising in the muscles ' Bowditch and Warren : Loc. cit. ■ ^ Tigerstedt : Loc. cit., p. 295. 'Hunt:Loc. cit., p. 436. * Hering: Loc. cit., p. 491. = Loc. cit., p. 63. 478 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. are not iiiiportant. i\IacWilliam^ practically agrees with Johansson. Athan- asiu and Carvallo^ concluded that the impulses coming from the muscles are all important, and that influences from the cerebral centers have no effect on the pulse rate. The latter conclusion is based on experiments made upon paralytics and upon animals poisoned with chloralose,^ and it seems to me that such experiments cannot be conclusive ; for in both cases the cerebral centers are impaired, and would not be expected to discharge motor im- pulses of normal intensity. Jacob^ and Egger' have shown that sensory stimuli can increase the pulse rate, and Hunt" has shown also that impulses from either the brain or a peripheral nerve can cause the acceleration. The general conclusion to be drawn from all this evidence is that in all probability impulses from both the voluntary brain centers and from the con- tracting muscles influence the regulating mechanism of the heart and thus in- increase its rate of beat. The latent period of the nerves involved is short enough to make this possible; moreover, the same conclusion is warranted for other reasons. We know that impulses are discharged from the cortical centers to initiate each voluntary movement,' and that impulses arise in the muscles as a result of each contraction.' It is also kno^vn that such impulses spread to parts of the nervous system entirely outside of their direct paths, and influence other nervous mechanisms, even those which have nothing to do with the work in hand. For example, clenching the hand increases the force of the knee jerkf practice in clenching one hand increases the strength of the other hand.'" The diffuse spread of impulses which always accompanies vigorous work undoubtedly reaches the cardiac centers and is responsible for the first and most rapid acceleration of the pulse. The continuation in some cases of an unbroken primary acceleration for two minutes or more (as in the case plotted in Fig. 4) gives ample time for all of the causes of increased heart action to become effective. Some of the influences discussed above as possible causes probably do not aid in the effect at all; reasons will be given for thinking that some of them do not act until later. Some idea of the relative importance of nervous influences in causing the change may be obtained from the fact that in the average pulse rate ' Loc. cit., p. 478. ' Loc. cit., pp. 554, 567. ' Richet : Dictionaire de physiologie. See chloralose. ■* Loc. cit. ' Loc. cit. « Loc. cit., pp. 439 and 446. ' Schafer: Text-Book of Physiology, Edinburgh and London, 1900, Vol. II, p. 722. ' ' Sherrington : Ibid., p. 1002. ' Lombard : Am. Journ. Psychology, Vol. I, p. 5. '» Wissbr and Richardson : Psychological Review, 1900, Vol. VII, p. 29. WILBUR P. BOWEX. 479 showiin Fig. 3, fifty -two per cent of the entire rise is accomplished in the first ten seconds and ninet>--one per cent in the- first thirty seconds. In the heavier work of Table IV, all the ^\-ay from t^\'enty-two to seventy-five per cent of the rise occurs within ten seconds, and from fifty per cent to one hundred per cent \^ithin thirty seconds. An increase in pulse rate from sixt}- or seventy up to one hundred or more, which is frequently obtained by tapping a telegraph key as rapidly as possible, seems at first thought scarcely credible as the result of such an insignificant amount of work; and one is Ukelyto conclude that it is not due to the work at all to am- large extent, but rather to the psychic conditions brought about by the experiment. In A-iew of the fact that the cause of the change in pulse rate is chiefly or wholly a nervous influence acting on the vagus center, it is entirely correct to speak of it as a psychic influence; but the idea that such a change in pulse rate arises from psychic changes independent of the work can be set aside at once. It was at one time the theory of the writer that the acceleration is due to the condition of nervous tension induced by the effort to work rapidly, and that it can be produced as well by an effort of the will, especially if one imagines the work as being done. It required only a few tests of this matter to prove the fallacy of this view. Seated at the table mth the finger on the key, the wi-iter tried again and again, but in vain, to bring about the same acceleration by imagining himself engaged in a desperate trial of speed in tapping against an imaginary opponent, with aU the excitement commonly attending an athletic contest. No amount of nervous tension that could be assumed was able to produce one-sixth as much increase in pulse rate as was instantly secured when the finger began to do its work upon the key. The psychic influence, if we choose to call it such, which causes the acceleration, is inseparably associated with muscular work; and the discharge of motor impulses is an essential part of the process causing the increase in pulse rate. Changes in Pulse Rate feom the Completion of the Primary Rise to THE End of the Work. 1. Minor Changes. — These changes are not seen with any regularity when the pulse is counted in intervals as long as ten seconds. They are seen, how- ever, in many instances, and may be found in nearly all cases by estimating the length of individual cycles. They are not so apparent during the most rapid part of the primary rise, for obvious reasons. As before stated (page 10), the minor changes in pulse rate are usually attributed to respiratory, vasomotor, and psychic changes. Since these influences are present and active during work, as well as when the subject is at rest, it is reasonable to assume that the minor changes during work also result from these causes. No attempt 480 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. has been made to study the minor changes in this research, except to recog- nize their presence and thus avoid confusing them with the effects of work. 2. The General Course of the Curve of Pulse Rate. — The main facts on this point, such as the existence of a plateau and a secondary rise, can be readily seen in Tables II, III, and IV. The duration and height of the plateau and the secondary rise, with their relations to the conditions of work, are shown for the bicycle work in Table V. In this group of experiments a plateau seems to be present in every case; the same is true of the lathe work, although in some cases it occurred during the portion omitted from the table. The experiments of Table II are not valid here, for the work was not • perfectly uniform in speed and force. It may be observed in Table V that the first five cases of bicycle work gave no secondary rise; the remaining six cases gave a secondary rise continuing to the end of the work. The questions then arise : Would a secondary rise occur in any or all of the first five cases if the work was continued longer ; and would the secondary rise in the last six be followed by a plateau on continuing the work for a sufficient time? Table V. Changes in Pulse Rate During Bicycle Work, as Related to the Conditions op THE Work; Compiled from the Same Records as Table IV. Conditions of work. Effects upon heart rate. t. s Primary rise. Plateau. Secondary rise. X S fl fl S . V 0) oj ■e-s el o . II o 8 c 03 a I'm 3 c o i. o la ■rj t-i 2 0) o 1 n ^1 a o g 3 R 1^ la 1" 'z rt Pi ^■" Min. Sec. s Min. Sec. gH WILBUR P. BOWEN. 491 recording the pulse waves in correct time. The natural vibration period of the tambours was found to be very short— between 0.03 and 0.04 seconds— and their amphtude so small that no indication of any interference with the much greater pulse waves could be detected. Moreover, pulse waves varying in length from 0.08 to 0.45 seconds were recorded readily, so that there seems to be no occasion to discredit the apparatus. Table X. Changes in Duration op Systole and Diastole, Expressed in Hundredths of a Second. Showing the Effect op Vigorous Work on a Bicycle. Two Subjects. Number of experiment. Before work began. Soon after beginning work. Near time of maximal pulse rate. Soon after work ceased. Several minutes later. S. D. S. D. S. D. S. D. S. D. 1 2 3 -I 5 6 7 25.2 26.0 25.0 24. S 26.8 27.0 25.5 55.4 64.7 58.7 47.2 64.0 66.0 69.5 26.5 28.0 24.8 25.4 26.0 28.0 27.5 33.2 34.0 26.0 28.4 34.0 42.0 40.5 26.4 25.2 20.4 22.0 22.0 20,0 21.0 36.0 31.0 20.9 20.0 28.0 25.0 28.0 24.4 25.0 19.0 22.4 22.0 20.5 22.0 49.4 32.0 22.5 30.0 40.0 33.0 42.0 24.6 26,0 22.0 23.8 25.0 25.0 25.0 53.3 56.5 43.0 38,4 63,0 62.0 57.0 Table X gives results obtained in the manner just described during the progi'ess of several experiments upon the bicycle. Here the greatest change in systole is nearly twenty-five per cent, and the greatest change in diastole sixty- 200 tha, sec. ^- — •% ^' ■^ \ , ' ^ — -~ — __ s N ti ■«• \ / \ ■" "■^" Ml ^ ■■ Fig. 8. Changes in duration of .systole and diastole due to heavy and rapid work upon a bicycle. S, curve of systole; D, curve of diastole. Arrows indicate the time of the beginning and cessation of work. 492 THE PULSE RATE, AS MODIFIED BY MUSCULAR WORK. six per cent. In the curves of Fig. 7, which were taken during an experiment involving more violent work than any in this table, the systole shows a change of nearly thirty per cent and the diastole about eighty per cent. The changes in duration of systole and diastole foimd in this experiment are plotted in the curves of Fig. 8. This figure shows a shortening of diastole until it is less Temp. Fahr. _ ^ ■— — ~~.' -^» ^ . — V ^ — — " 1 * -'' "— . ■^-v \ / •* X ,'•' / \ /'■ A — s -s, — , -.. \, ^ "**, t -V — ■ — .' t \ Fig. 9. Changes in duration of systole and diastole due to moderate bicycle work of fifty-two minutes' duration. ' of Pennsylvania, upon the nature and distribution of the new tissue in cirrhosis of the liver. His conclusions were that the white fibrous tissue was increased in all forms of cirrhosis of the liver accompanied by an increase of yellow elastic fibers which arose frona the pre-existing fibrils of Glisson's capsule and the adventitia of the vessels; that both of these connective tissue elements penetrated in some cases the lobule in all forms of cirrhosis, but to a lesser degree in hypertrophic, and that with the present methods of technique it was impossible to differentiate the reticulum from the white fibrous tissue of the liver lobule. Pearce (1901) studied chronic congestion of the lungs and found that the increase in density of the lungs was due to an increase of the yellow elastic tissue. This increase was progressive, and appeared first in the alveolar walls, but in older cases the connective tissue septa and the pleura showed the change. He thought that the newly formed elastic tissue arose from the pre- existing fibrils in the alveolar walls, and not from those in the vessel walls. In the same year Hedinger pubUshed the results of his observations on the gro^rth of sarcoma in the intima of veins and arteries in "struma sarcoma- toides. " In the periphery of the tumors he found many newly formed fibers of yellow elastic tissue, but in the center of the gro-wth only few fibers were present. In 1902 there appeared a number of articles which contained records of observations on the new formation of fibers of elastic tissue. , The first of these was pubhshed by Inouye, who reported the results of his study of the amoimt . of yellow elastic tissue in twenty carcinomata of the stomach. He foimd that the yeUow elastic tissue fibers were pressed apart and destroyed, but occasionally he found newly formed fibers which arose from the pre-existing elastic tissue elements. The second article was by Meinel, who examined a carcinoma of the stomach and found a varying amount of newly formed yellow elastic tissue, the largest number of fibers being pres- ent in the region of the fundus and the smallest number in the pyloric region of the growth. In the same year Oliver published the results of his study of yellow elastic tissue in twenty-two cases of cirrhosis of the liver. He found that the yellow elastic tissue was most abundant in atrophic cirrhosis, where the fibers were coarser, formed larger bundles, grew between the lobules, and seemed to constitute the greater part of the newly formed tissue. In hyper- trophic cirrhosis he found that the elastic tissue did not make up a very large part of the new tissue; that the yellow elastic fibers were finer, formed smaller bundles, and extended along the capillaries into the lobules. In syphilitic 498 TUMORS OF THE SKIN. cirrhosis the greater part of the new tissue was not yellow elastic tissue, although distinct fibers were fairly abundant in the interlobular tissue, the external capsule, and about healed gummata. As to the origin of these newly formed fibers, he found that they arose principally from the walls of the vessels, but that some arose from the walls of the bile-ducts. This brief review of the literature shows that the conclusions reached by the various authors have varied; and with the hope that new points might be added towards the solution of these questions, this research was under- taken. The material employed was fixed either in alcohol, formalin, bichlorid of mercury, Mijller's fluid, or picric acid; hardened; embedded in paraffin; sectioned and the sections stained on the cover-glass. After removal of the paraffin the sections were placed in a saturated aqueous solution of bichlorid of mercury for several hours, washed in water, stained in Weigert's elastic tissue stain, followed by Mallory's connective tissue combination. The object of this long procedure was, that a comparative study might be made of the different elements of the connective tissue, and of their relations. Since Mallory's combination stains only tissues hardened in solutions of bichlorid of mercury, the sections were placed in a saturated, aqueous solution of bichlorid of mercury, and the results obtained were the same as after fixation with this solution. The material used came from the following sources: 16 papillomata (con- dylomata), 8 being from the vulva, 1 from the anus, 3 from the vagina, 3 from the penis and 1 from the scalp; 2 hemangioma simplex; 1 hemangioma and 6 Ijonphangioma simplex hypertrophicum; 2 healed hemorrhoids; and 3 cases of elephantiasis, 1 of the labia, 1 of the thigh, and 1 of the scrotum. Papillomata. Vulva. — As has been stated, Kromayer found that yellow elastic fibrillae were present in the oldest papillomata only. I am able to confirm this to a certain extent, since sections taken from a young papillomata, of from four to six weeks growth, show no yellow elastic tissue; but in somewhat older papillomata, where sections can be made through the main stem of the growth, it is possible to trace fibrillse up this stem for a very short distance. Some of the fibrillse seem to end here abruptly, but few fine fibrillse are present in some of the nearest connective tissue papillae of the growth, and can be traced back to the old fibrillse of the skin. In sections of still older papillomata, a fine network of yellow elastic fibrillse can be made out in some of the subepithelial connective tissue papillse some distance from the stem, but very few of which are arranged around the blood-vessels. Anus. — In the papilloma from the anus, numerous subepithehal papillse FREDEHmv A. BALDWIN. 499 show the presence of ycUow ehxstic fibrilhe arvangetl Hke those of the older papillomata from the vulva, but prc^seut in larger minib(n\s. Vagina. — In two of the papillomata from the x'agina, large fibrils, which arise from the parent fibrils of the parent submiicosa, run from some distance up the stem and are associatcxi with fibrilhe which arise from the walls of the blood-vessels. At a distance of about one-half of a millimeter up the stem, these decrease greatly in size antl number, and farther up the stem, a few fine fibrillte are present '\\-hich end in delicate threads in certain of the subepithelial connective tissue papillte. In the other papilloma, more of the connective tissue papillffi contain these fine fibrils, which are more delicate than those of the other papillomata, and the stem does not contain as many large fibers. These end-fibril Ise in all of the three growths studied have no relation with the vessel walls. Penis. — In one papilloma of this organ, the connective tissue of the stem contains large fibrils of yellow elastic tissue which branch and rebranch to form in the stem a network of fine fibrillffi, but many of these fibrillsB pass into the subepithelial papillse and have no connection with the vessel walls. The second of these growths shows a small number of fibrils in the stem, but more of the subepithelial papillse contain the fine fibrillse than the other papil- loma of the penis. The fibrils and fibrillse resemble those of the other tumor, both in their relation to each other and to the vessels. Scalp. — ^The lower layers of this tumor contain a very dense network of yellow elastic tissue. From this network, fibrils arise which extend up into the center of each large subepithelial papilla. These fibrils branch and rebranch to form another network nearer the epithelium, not so dense as the first. From this latter network fine fibrillse extend upwards and anastomose under the epithelirmi to form a dense layer of yellow elastic tissue. Hemangioma Simplex. Everywhere the connective tissue forming the walls of the blood spaces contains many fine and a few coarse fibrils of yellow elastic tissue, which form a network of varying destiny in different places. All the fibrils and fibrillse have their origin from the coarse fibrils found in the base of the tumor, which arise from the deep fibrils of the dermis. Hemangioma Simplex Hyperteophicum. The one specimen at my disposal shows at three places a bundle of heavy fibrils of yeUow elastic tissue surrounded by connective tissue, and projecting towards an upper layer of the dermis which separates the tumor mass from the epideriTiis. From these coarse fibrils a few fine fibrillse run into the tra- beculse of connective tissue of the tumor mass proper. All the larger vessels of 500 TUMORS OF THE SKIN. the tumor contain yellow elastic fibers which give off few fibrillae, but these run only a short distance from the vessel wall and end by gradually diminishing in size. Lymphangioma Simplex Hyperteophicum. All the tumors of this variety here considered projected above the surface, and one of them was very papillary. Sections from this tumor show fine fibril] se of yellow elastic tissue in the connective tissue of each one of the smaller nodules. These fibrillse form a fine network, and in some places they are in close relation with the fibrils which arise from the older vessels, while others seem to be independent of the vessels, and some can be traced back to the stem of the tumor where they arise from coarse fibrils coming from the dermis. Sections from the other tumors show the same conditions, but to a lesser degree. Healed Hemorrhoid. The two specimens of healed hemorrhoids consisted of tags of connective tissue, containing dilated vessels and covered with squamous epithelium. Throughout their interior large fibrils of yellow elastic tissue are present, but they have no definite arrangement nor relation with the other elements of the connective tissue. Most of the yellow elastic tissue present is in the form of these coarse fibrils; in some places however they give off fine fibrillse. In one of these tags they form small, circumscribed, closely packed networks. The vessels of these tumors contain-but little yellow elastic tissue which is present as fine fibrillse and extends only a short distance from the vessel wall. Elephantiasis. Labia. — Sections of tissue removed from this case show a moderate amount of yellow elastic tissue. Each of the subepithelial papillse contains a few fibrillse of elastic tissue, which form a loose network; in the deeper portions of the tissue, larger fibrils are present, some of which run parallel to the bundles of white fibrous tissue, while others do not ; in still deeper portions of the tissue areas are seen which under low power give the appearance of almost a solid mass of bluish black color, but under higher magnification such areas are seen to consist of dense, circumscribed, interlacing networks of yellow elastic fibrillse. The latter have no definite relation to each other or to the blood- vessels, although some are found a short distance from the vessels, but have no connection with the fibrillse of the vessel wall. Thigh. — ^The skin of the thigh of this case was very rough and nodular, and one of these nodules served as material for sections. The center of the FREDERICK A. BALDWIN. 501 stem which connects the nodule to the skin contains :i nc^tworlc of fine fibrillffi of yellow elastic tissue, which is intimat(4y bhnuleil with a mass of white fibrous tissue, but the fibers of the \'ariefies of connecti\'e tiss\ie are in no place parallel. Under the low power, the network appears as a bluish haze, in which are ob- served few scattered fibrils; under the high power this network presents the appearance of a dense mesh-work of fibrillte, having no definite relation with the fibers of white fibrous tissue with which they are intermingled. Many of the fibrillje present the appearance of a string of fine granules connected by a fine thread. Certain of the fibrillffi of this mesh-work branch and run perpendicularly toward the epithelial surface. These fibrillse branch but little and form a loose net-work. In no place in the section are the fibrillse in connection with the fibrils in the walls of the blood-vessels. Scrotum. — The structural changes observed in material obtained from this case are seen, more particularly, in the papillary layer of the dermis. The subepithelial papillse are beginning to lengthen, due to an increase of connective tissue, which is composed of loosely arranged bundles of wavy fibers, and the overlying epidermis is slightly nodular. The yellow elastic tissue is seen in the deeper layers as coarse fibers, surrounding the involuntary muscle, into the perimysium and epimj'sium of which small fibrillse are sent. Besides these fibrillse, the blood-vessels have their own yellow elastic tissue. There is present a dense network of fibrillse in the dermis under the area showing the increase of connective tissue. The fibrillse of this network form a dense, irregular mesh-work, which sends branches into the more superficial layer. These branches extend as far as the basement membrane of the epidermis, forming a loose network, which presents no definite relation to the fibers of the white fibrous tissue, and have no connection with the blood-vessels. The study of these cases show, beyond a doubt, that a new formation of yellow elastic tissue takes place in papillomata, angiomata, healed hemorrhoids, and elephantiasis, since, in many sections, the fibrillse of yellow elastic con- nective tissue are found in greater number than that in the normal skin from which the growth arose, and in all sections the fibrillse have different arrange- ment and relations from those of the normal. Certain of the newly formed fibrillse, no doubt, arise from the yellow elastic tissue of the vessel walls, but by far the greater number arise from the pre-existing fibers of the parent dermis, which, by longitudinal growth, form the new fibrillse. That most of these new fibrillse do not arise from those of the vessel walls is shown con- clusively in the subepithelial papillaj where the yellow elastic tissue is present at some distance from the vessel walls, which in these areas consist of a single layer of endothehum without any yellow elastic tissue and also in the lower layers, where the greater number have no relation to the vessels. 502 TUMORS OF THE SKIN. Retictjlar TlSStTE. In his original article, Mall described reticulum as a form of connective tissue having some characteristics of yellow elastic and others of white fibrous tissues, basing his conclusions upon digestive experiments. His results and conclusions have not been generally accepted, and Flexner expresses the general sentiment when he says that with the present technical methods, retic- ular and white fibrous tissues cannot be separated. When Mallory published the formula of his stain for reticular connective tissue, he thought that it was a specific for reticulum, but later this was found not to be the case, since it stains both white fibrous and reticular tissues a similar shade of blue, and conse- quently there is no generally accepted technique for the isolation and staining of reticulum, and so the conclusions of this paper are based upon the digestive experiments and results of staining with Mallory 's connective tissue stain. For purposes of comparison, sections were made from pieces of skin taken from different regions of the body and stained according to the method given above or digested after the Spatholz method — the results obtained being very similar in both cases. Sections of normal skin which include the fat, show throughout this layer fine bluish, almost homogeneous fibrillse which separate the fat cells. The coarse fibrillae are increased in the layer above the fat and form here a network of branched, almost homogeneous fibrillse, which holds together the white and yellow elastic tissues, and is collected around the blood-vessels in a denser net- work, and surrounds the sweat-glands where it condenses to form a thick layer which serves as a basement membrane for the epithelium. Above the dense layer of the dermis which contains the white fibrous tissue these fibrillsa form a fine feltwork, which gives off branches to form- the entire structure of the connective tissue papillse upon which the epidermis rests, and condense to form a basement membrane for the epithehum, similar to that formed in the sweat-glands. Papillomata. — The frame work of very young papillomata consists entirely of a loose network of fine, almost homogeneous, fibrillse, which branch and anastomose and is more compact under the epithehum where, in some places, it appears as a fine, homogeneous fine upon which the epithehal cells rest. The capillaries run in this network, which condenses somewhat to form a supporting membrane for the endothehum. In somewhat older papillomata, bundles of white fibrous tissue are seen in the stem of the growth, but the fine network remains and occupies the same relations as in the young papillomata. In older growths, yellow elastic tissue appears. Angioma Simplex. — Continuous with a network of fine, branching, homo- geneous fibrillar in the underlying dermis is a similar network in the tumor. It consists of fine fibrillae, which branch and anastomose to form the dehcate FREDERICK A. BALDWIN. 503 framework which binds together the white fibrous and yellow elastic tissues and vessels and condenses to form a delicate, almost homogeneous, line upon which rest the endothelial cells lining the blood or lymph spaces. Angioma Simplex Hi/pcrirophicum.—'Wha.t is said of this tissue in the angi- oma simplex is true of that found in angioma simplex hypertrophicum, and in addition, it ma>' be said that small fibrillEB arise fronr the network outside the small masses of hypertrophic endothelium, run between thenrand separate the individual cells. Elephantiasis: Labia. — Sections taken from this specimen consist of a loose feltwork of fine, homogeneous branching fibrillae which form a framework for the whole section. It holds together the different elements, and is slightly condensed imder the rete Malphigi and around the sebaceous glands to form a basement membrane for the epithelial cells. Thigh. — Sections from the skin of the thigh of this case show the same general arrangement of this fine network as in the sections from the labia, but more of these fibrillse are present than in those of the labia. Scrotum. — The upper la}'er of the dermis, the layer showing the patho- logical changes, is composed almost entirely of fine homogeneous branched fibrillse arranged in a feltwork which holds in its meshes few strands of white fibrous and yellow elastic tissues. In conclusion, it may be said that our present technique does not differ- entiate the so-caUed reticular tissue from white fibrous tissue, but these methods reveal the presence of a tissue closely resembling, if not identical with, the reticular tissue of ilall in all papillomata, angiomata, both simplex and hyper- trophicum, and in elephantiasis. LITERATURE. Flexner. — Nature and Distribution of the New Tissues in Cirrhosis of the Liver. Uni- versity Medical Magazine, 1900-01, Vol. XIII, p. 613. Hamilton. — A Peculiar Form of Fibrosarcoma of the Brain. Journ. Experiment. Med., 1899, Vol. lY, p. 597. Hamilton. — On the Presence of New Elastic Fibers in Tumors. Journ. Experiment. Med., 1900, Vol. V, p. 131. Hedinger. — Ueber Intima-Sarcomatose von Venen und Arterien in sarcomatosen Strumen. .^chiv f. Path. Anat., 1901, Bd. CLXIV, p. 199. Hohenemser. — Ueber das Vorkommen von elastischen Fasem bei cirrhotischen Prozessen der Leber und Niere. Archiv f. Path. Anat., 1895, Vol. CXL, p. 192. Inouye. — Ueber das Verhalten des elastischen Gewebes bei Magen Carcinon. Archiv f. Path. Anat., 1902, Vol. CLXIX, p. 278. Jores. — Zur Kenntniss der Regeneration und Neubildung elastischen Gewebes, Beitrage z. Path. Anat., 1900, Vol. XXVII, p. 381. Kromayer. — Elastische Fasem, ihre Regeneration und Widerstandsfahigkeit. Mo- natschr. fiir Prakt. Dermatol., 1894, Vol. XIX, p. 117. 504 TUMORS OF THE SKIN. Mall. — Reticulated Tissue and its Relation to the Connective Tissue Fibers. Johns Hopkins Hospital Reports, 1896, Vol. I, p. 171. Meinel. — Ein Fall von Karzinom des Magens mit starker Entwicklung des elastisohen Gewebes und iiber das Verhalten dieses Grewebes im Magen bei verschiedenem Alter. Miinchener Med. Wochensohr., 1902, Vol. XLIX, p. 359. Melnikow-Raswedenkow. — Histologische Untersuchungen iiber des elastischen Gewe- be im normalen und im pathologisohen veranderten Organen. Beitrage z. Path. Anat., 1899, Vol. XXVI, p. 546. Oliver. — Elastic Tissue in Cirrhosis of the Liver. Transactions of the Chicago Patho- logical Society, 1902, Vol. V, p. 96. Pearce. — The Increase of Elastic Tissue in Chronic Passive Congestion of the Lungs. Joum. Med. Research, 1901, Vol. I, p. 258. White. — Distribution of Connective Tissue in New Growths. Johns Hopkins Hospital Bulletin, 1900, Vol. XI, p. 209. ON TEIPHENYLMBTHYL. BY M. OOIffBERG, Sc.D., Junior Professor of Organic Chemistry, University of Michigan. The conception of valence which underhes our theories of the constitution of organic substances has been the guiding light in the remarkable develop- ment of organic chemistry during the second half of the last century. By a strange coincidence it so happens that nearly all of the few elements which enter into the composition of organic substances possess what has been called a constant valence. It is assumed that hydrogen is always monovalent; oxygen divalent; carbon tetravalent; while nitrogen is either trivalent or pentavalent, the two states of valence being easily told apart, according to the function of the element in the given compound. With each of these ele- ments always possessing the same valence, the problem of assigning to a given body the correct constitution from comparatively few experimental facts becomes much simplified. In many instances the empirical composition and the molecular weight furnish in themselves sufficient data for a structural formula of a compound. Not so simple would the case be if we had to deal with carbon as an element possessing a variable valence. The tens of thou- sands of organic substances, those collected from the animal and vegetable kingdom as well as those prepared in the laboratory, have, in a remarkable way, corroborated the h}rpothesis first proposed by Kekule, that carbon is always tetravalent. So strikingly uniform has the accumulated evidence been to that effect, that the tetravalence of carbon has in itself become the guiding principle in the marvelous development of synthetic organic chemistry. Reac- tions are predicted to take place in this or in another way, because we firmly believe that the valence of carbon will, under these new conditions, still remain the same. It is, therefore, no wonder that when Nef , in a series of important papers, calls attention to the fact that carbon monoxide is not the only substance wherein carbon is divalent, chemists are rather slow to accept his interpretations. Nef has shown that the behavior of many organic substances can best be explained on the assumption that they contain a divalent carbon. But as their behavior is also not entirely inconsistent with a structure wherein carbon is tetravalent, the latter view, even if not the most plausible, will ap- peal to most chemists. In view of all this, it was not without considerable hesitation that the writer has published his views as to the constitution of triphenylmethyl. The view propounded in regard to this interesting substance has since then received strong support from different sources. The hypothesis 505 506 ON TRIPHENYLMBTHYL. that carbon acts, at least in this particular instance, as trivalent has proved useftol in explaining properties and reactions of substances which hitherto had no explanation at all. In the following pages is given a brief historical account of the work during the last three years. Those interested in the experimental evidence and in the details of the work are referred to the original papers.' I. Preparation of Triphbnylchlormethane. C6H5-7C — CI. CeH/ On the one hand, the readiness with which triphenylmethane is foimed under a variet}^ of conditions, and on the other hand, the apparent impossi- bility of obtaining the corresponding tetraphenylmethane by reactions which a 'priori lead us to expect its formation led Victor Meyer to the conclusion that the cause of this is to be found in stereochemical arrangements. That is, the three phenyl groups may take up so much space around the central carbon atom that such a complicated fourth group as another phenyl cannot become attached to the same carbon atom. It was with a view to test this theory that I undertook the preparation of tetraphenylmethane. After several months' work, about half a gram of a hydrocarbon was obtained, which, accord- ing to its composition, its moleciilar weight, its tetranitro derivative, etc., was judged to be the looked-for tetraphenylmethane.^ The existence of this body at all made it then desirable to review all the old methods that have ever been tried for the preparation of tetraphenylmethane. By none of these methods, however, could I obtain any indication of the formation of tetraphenylmethane. One of the inethods most fiequently mentioned in the literature consisted in the action of aluminium chloride upon a mixture of carbon tetrachloride and benzene, the expectation being that the reaction would proceed as follows : 4C6H, -^CCl, = CCCeHj), -h4HCl. That Friedel and Crafts were mistaken in reporting the formation of tetraphenylmethane by this method was shown by E. and O. Fischer. E. and » Journ. Amer. Chem. Soc, Vol. XXII, pp. 752, 757; XXIII, p. 496; XXIV, p. 597. Amer. Chem. Journ., XXV, p. 317; XXIX, p. 364. Ber. d. Chem. Ges., Bd. XXXIII, p. 3144, 3150; XXXIV, p. 2726; XXXV, p. 1822, 2397, 3914; XXXVI, p. 376, 1088. 2 Ber. d. Chem. Ges., Bd. XXX, p. 2043; XXXVI, p. 1088; Journ. Chem. Soc, Vol. XX, p. 773. UUmann and Miinzhuber, in a paper which has just been published (Ber. d. Chem. Ges., Bd. XXXVT, p. 404), describe the formation of tetraphenylmethane by a new method which leaves no doubt as to the correct constitution of this hydrocarbon. The hydrocarbon so obtained is found by UUmann and Miinzhuber to be identical in every way -with the one described by myself in the papers referred to. M. GOMBERG. 507 0. Fischer/ and a great nuiuy others, report that by distiUing the products of the above reaction, just as is done in the preparation of triphenylmethane from benzene and chloroform by the analogous method, 3CeHo +CHCI3 = CH(C„HJ3 +3HC1, the same hydrocarbon, namely, tripheuyhnethane, is the main product also when carbon tetrachloride is employed. Now, it appeared to me that any tetraphenylmethane, even if formed, would probably be decomposed by the appHcation of such high temperature as 300° to 500° C. So instead of dis- tilling the products of the reaction I subjected them to fractional crystalliza- tion. It was soon found that what is really formed here is not triphenyl- methane, but triphenylchlormcthane, its formation taking place according to the followdng equation : 3QHe + CCI4 = C(CeH5),Cl + 3HC1. Conditions were then established which furnished almost a quantitative yield of this extremely important substance. Previously to this, this halogen derivative was always prepared by the action of phosphorus pentachloride upon triphenylcarbinol, and the latter in its turn was obtained from the cor- responding hydrocarbon by oxidation : (C,H5)3CH + = (CeH,).,C - OH, (CeH5)3C- OH + PCl5 = (QHJ3C.CI +POCI3 +HC1. From the description in the literature I am inclined to think that E. and 0. Fischer were the only ones that had triphenylchlormethane in the pure state. By the new method, which is here outlined, large amounts of this halo- gen derivative can be easily prepared. The crude product is almost pure, and by one recrystallization can be obtained absolutely pure. II. Teiphenylmethylpeeoxide. As tetraphenylmethane could not be obtained in large quantities, it was decided to test the validity of certain reactions applied to that compound, on a substance as nearly analogous to it in composition and structure as possi- ble. Such a substance should be hexaphenyl ethane. It seemed that with the possibility of obtaining any desired amounts of triphenylchlormethane the preparation of hexaphenylethane should not present any difficulty. I proceeded according to the well-established methods for such condensations. The triphenylhalogenmethanes were subjected to the action of various metals, on the supposition that the metal wonld remove the halogen, allowing in this way the two radicals to unite with each other: (C6H5)3 C. CI , r,TvT _ (C6H5)3 C , p|^ p| (CA)3C.Cl+2Na-(c,H,)3C+2^"^^- > Ami. Chem. (Liebig), Bd. CXCIV, p. 254. 508 ON TRIPHENYLMETHYL. Accordingly, triphenylbrommethane in benzene was treated with metallic sodium, but without success. The chloro compound gave no better results. Molecular silver was substituted for sodium. After several hours' boiling a white crystalline body began to separate, and on filtering the hot benzene solution a considerable amount of the same substance separated on cooling. It was recrystallized from benzene, gave a constant melting-point, 185° C, and contained no halogen. In its high melting-point and in its only slight solubility in the usual organic solvents it resembled closely tetraphenyl- methane, and this new body was taken for hexaphenylethane. An elementary analysis gave, however, the following results : Carbon 93.83 87.93 Hydrogen 6.17 6.04 The low per cent of carbon found was rather surprising. It was explained on the assiunption that this was, perhaps, an instance of a hydrocarbon which is not easily burned. The next combustion was, therefore, carried on in an atmosphere of oxygen from the very beginning. The substance was mixed in the tube with copper oxide; a very high heat was applied towards the end of the combustion. The results were as follows : Carbon : 87 .74 Hydrogen 6 . 46 An entirely new lot of the material was then prepared. Ten grams of triphenylehlormethane and ten grams of silver gave, after several hours' boiling, four grams of the same hydrocarbon. This was recrystallized twice from benzene and twice from chloroform. It was perfectly free from halogen, was snow-white, and melted at 185°-186° C. The combustion was again made in an atmosphere of oxygen, lead chromate being used instead of copper oxide. Carbon 87.77 Hydrogen 6.23 The next analysis was made in a bayonet tube, the tube being filled with fine copper oxide for about four-fifths of its length, to insure combustion of any methane gas which would perhaps otherwise escape. The combustion was carried on slowly, and a very high heat was used towards the end. Carbon 88.23 Hydrogen 6.34 Several new lots of the same substance were made both from the triphenyl- brommethane and triphenylehlormethane, and purified by successive recrys- tallization from benzene, chloroform, acetic ether, and carbon disulphide. They all gave the same results, entirely concordant with each other. Com- bustions were then made in a porcelain tube; applying the direct heat of the fui'nace; also by the moist method with chromic acid in concentrated sulphuric M. GOMBERG. 509 acid/ but with no better result. Another combustion was then made of this body, by the copper oxide method; this combustion was made on a sample of material each crystal of Avhich was separately picked out and was found to be perfect, but with no different results. I therefore came to the conclusion that the body under consideration was not a simple hydrocarbon, but an oxygtni derivative. The results of all the analyses established be>ond doubt that the composition of this substance was CioHisO. It was hariUy to be expected that the triphenylmethyl group in the triphenylchlormethane would be destroyed by such mild treatment as the action of sih-er upon the latter at ordinary temperature. Consequently, the substance should be represented by the formula, (C6H5)3C— O,— which, in accordance with our notions of structural chemistry, was probably (CeH5)3C-0 0-C(CeH3)3. In other words, this substance was the peroxide of the radical, (CeH5)3C. The oxygen could come from either of two sources : first, the molecular silver may have contained some oxide of the metal; second, the atmospheric oxygen may act upon the hydrocarbon. As only the molecular silver and not the finely powdered crystaUine metal appeared to act in this case, a very pure sample of the former was prepared. The moist sih-er, as obtained by reduction of the chloride with zinc, was di- gested for a day with dilute sulphuric acid. It was then thoroughly washed by decantation, digested for several hours with ammonium hydroxide, again washed with water, then with alcohol, absolute alcohol, ether, and finally with benzene. This sample of silver gave, however, results not differing from those previously obtained. To make it more certain that the oxygen did not come from the silver, I substituted other metals for it. Mercury and zinc were found to act equally well, if not better, and the yield of the peroxide obtained by either of these two metals was even greater than in the case of silver. Another great advantage in the use of zinc and mercury is that the reaction takes place at ordinary temperature. I next proved that it is really the oxygen from the atmosphere which oxidizes the hydrocarbon. By working in an atmosphere of carbon dioxide no such insoluble peroxide is produced, even on weeks', and months' treatment of triphenylchlormethane in benzene with silver, zinc, or mercury. The moment, however, the solution is exposed to air, the formation of the peroxide begins instantly. After a series of experiments I settled upon zinc as the best reagent with which to carry on this reaction. Ordinary granulated zinc, zinc strips, zinc dust — all act upon the halogen compound. In all f\u1;her expeiiments zinc in the form of fine turnings was employed. The suc- cessful preparation of the unsaturated hydrocarbon requires the adsolute 'Fritzsche.— Ann. Chem. (Liebig), Bd. CCXCIV, p. 79. 510 OX TRIPHENYLMETHYL. exclusion of oxygen from the apparatus. Corks are to be avoided, and even rubber stoppers exposed to the action of benzene vapors become after a while porous. Carefully ground glass connections weie emploj^ed in all this work. If the assumption that this oxygen derivative was a peroxide of the con- stitution 0-CfCeH, /C0H5 0-CfCA was actually correct, what could the substance in the solution, prior to its oxidation, be, if not simply — (C6H5)3^C— , i. e. triphenylmethyl? This conclusion, while not absolutely certain, seemed at least quite prob- able. It remained to prove, however, how correct was the assumption in regard to the constitution of the oxygen compound itself. The constitution of the peroxide was established as follows : 1. The composition of the substance led to the empirical formula, CigHijO. 2. Careful molecular weight determinations established that the rational formula of this substance must be represented by CjgHgoOj. 3. The same oxygen derivative was made by an entirely different method, namely, by the action of sodium peroxide upon triphenylchlormethane. Theie could be no doubt that in this case the reaction proceeded according to the equation, 2(CeH5)3 C.Cl -hNa-0-O-Na = (C6H5)3C - - O- CCCeHs)^ -h2NaCl. As this oxygen derivative was in every way identical with the one obtained by the oxidation of the unsaturated hydrocarbon on exposure to air, the con- clusion seemed justified that the substance obtained by the latter method was actually the peroxide. 4. By the action of concentrated sulphuric acid, the oxygen derivative was converted almost quantitatively into triphenylcarbinol, resembling in this way peroxides of metals. NaA+H20 = 2NaOH-f-0. [(CeH,)3C] A +H,0 = 2(CeH,)3COH + 0. Air these proofs taken together made it evident that we had to deal here with a peroxide. Hence, the conclusion that the unsaturated hydrocarbon, pre- vious to its oxidation, was triphenylmethyl, gp,ined in probability. Of course, the establishment of the constitution of the peroxide did not lead to the above conclusion with a certainty, because even if it be true that the peroxide has the constitution assigned to it, it does not follow that it could not be formed from M. GOMBERG. 511 some other substance than triphenylniethyl. Foi' example, it might have been formed according to the following equation: 2(C„H,),=C = CeH,+H20 + = (C„H,)3C-0-0-C(C„I-l5)3. It may be said right here that the oxidation of the unsaturated hydro- cai-bon takes place even if ever\- possible precaution to exclude moisture has been insured. Therefore, this particular phase of the question may be left out from further consideration. III. Action of Halogens upon the Unsaturated Hydrocarbon. The solution of the h>'drocarbon obtained by the action of zinc upon a benzene solution of triphenylchloi methane is sensitive not only toward oxygen but equally so to^va^d halogens. The latter are absorbed by the hydrocarbon just as readily as ox>-gen. I hoped that by the action of bromine or chlorine the original triphenylhalogenmethane will be obtained. (CeH5)3=C- +Br = (CeH)3CBr. But the reaction was found to be rather complicated, giving rise not only to addition, but also to substitution. The absorption of iodine was more promising. Mter many trials I have obtained from twenty grams of tri- phenylchlormethane something like one gram of an exceedingly unstable iodine derivative, which contained 33.2 per cent of iodine, while triphenyliodo- methane should contain 34.3 per cent. This substance also agreed in such few reactions as could be studied on so small a sample in my posesssion with the reactions of triphenylbiommethane and triphenylchlormethane. These results, unsatisfactory and inconclusive as they were at that time, made it, however, appear more probable than before that we had here to deal actually with the radical triphenylmethyl. The question arose, whether the peroxide or the iodine compound owe their formation to one distinct individual substance, or whether they might not have been produced by the union of two or more different hydrocarbons? It was, therefore, decided to isolate, if possible, the hydro- carbon itself in as pure a state as possible, and subject it, hy itself, to the action of oxygen and to that of halogens. lY. Isolation of Triphenylmethyl and of its Ester and Ether Deriva- tives. The Ester and Ether Derivatives. — Preliminary experiments upon the resi- due obtained by evaporating the benzene solution of the unsaturated hydro- carbon indicated that the hydrocarbon was only slightly soluble in acetic ester and in ethyl ether. In the hope that the hydrocarbon might at once be obtained in a crystalline state, experiments were carried out using the ethyl ether and the acetic ester instead of benzene as solvents in the reaction of 512 ON TRIPHENYLMETHYL. metals upon triphenylchlormethane. It was soon found that the reaction would go on smoothly only when the solvents are perfectly free from traces of moisture and alcohol. The least trace of either of these two substances in any of the reagents caused an evolution of hydrogen. This was explained as follows: triphenylchlormethane is very readily attacked by water or by alcohol with the production of hydrochloric acid. The acid so formed attacks the metal : r (CeHOsC.Cl +HOH = (C,H5)3C.OH +HC1. l(CA)3C.Cl+HOC3H,=(CeH3)3C.OCA+HCl. Zn + 2HCl = ZnCl3+H2. The acetic ester had to be purified very carefully by distilling it first over potassium carbonate to remove any acetic acid present; second, it had to be distilled over phosphorus pentoxide at least twice in order to remove water and alcohol. The phenomenon observed by the use of this pure solvent is very striking. When triphenylchlormethane is dissolved in the pure acetic ester and a strip of pure zinc is introduced the solution turns yellow at once. In a few minutes crystalline specks appear on the metal, which grow larger and larger, and the whole metal soon becomes coated with a deposit of large, beautiful, perfectly transparent crystals. These crystals show all the proper- ties of the unsaturated hj'drocarbon. They absorb iodine when dissolved in benzene; exposed to air they form the peroxide. An elementary analysis of the crystals showed, however, that this substance was not the pure hydro- carbon, triphenylmethj'l, but that it carried from four to five per cent of something other than carbon and hydrogen. It was thought that perhaps this lack of carbon could be accounted for either by partial oxidation or bj- the presence of some zinc derivative, which was mixed with or adhered to the crystals. It was therefore necessary to construct an apparatus which would insure the purity of the product, and in which the unsaturated hydrocarbon could be protected from the least exposure to air from the very beginning to the end of the operation. The following is a description of the apparatus (Fig. 1) which answers all these requirements. Twenty grams of triphenylchloi methane are dissolved in 120 c.c. benzene in flask A. Ten grams of zinc turnings are introduced, and a stream of dry carbon dioxide is rapidly passed through the solution until all the air in the flask has been displaced. The stop-cocks a and h are then turned off and the solution is allowed to stand several days. The zinc attacks the triphenyl- chlormethane at once, giving rise to triphenylmethyl and zinc chloride. The zinc chloride so formed does not separate as such, but unites with a portion of the undecomposed triphenylchlormethane, forming an insoluble, sirupy mass, which settles to the bottom of the flask, as it is completely insoluble in benzene. The reaction can, therefore, be represented as follows : M. GOMBERG. 513 ( 2(CeH,)3C.Cl +Zn= 2(C„HJ3C +ZnCl, I ZnCl^ + (CbH,)3C.C1. = (CoHJsCCl.ZnOlj, or, 3(C„H,)3C.C1 +Zn=2(C„H,),C+(C„HJ3CCl.ZnCl2. It is evident from the above equations that only two-thirds of the tri- phenylchlormethane employed in this reaction are reduced to triphenyl- methyl. "When the reaction is completed (after five to ten days) the flask is connected with the apparatus B, as is shown in the figure. The apparatus B Fig. 1. is exhausted by means of a pump and then filled with dry carbon dioxide; this process is repeated several times in order to drive out all the air from the flask. The flask B is then once more exhausted, and the solution from A is siphoned off into the bulb B by turning the double-way stop-cock C so as to establish the connection between A and B. About 20 c.c. of benzene are then allowed to run into flask A through the tube D in order to wash the zinc, and also siphoned off into B. Flask A is then disconnected from B. The benzene is distilled off under diminished pressure at a temperature of 25° C. and the vapors are condensed by surrounding the suction flask with ice. The heating of the benzene solution is done by surrounding the lower half of bulb B with several coils of rubber tubing, through which steam is continuously passing. After all the benzene has been distilled off, hot acetic ester is added to the solid residue through the funnel E, and the residue is dissolved in that solvent by the help of gentle heat. In allowing the solution to cool, large 514 ON TRIPHENYLMETHYL. transparent colorless crystals separate. The distilling flask B is now discon- nected from the condenser, and is connected with a suction flask by means of the rubber stopper under the stop-cock G. On opening the stop-cock G just a little, the solution can be sucked out from B, leaving the crystals behind. The crystals are then washed by introducing the pure solvent through the fuimel E. After the crystals have been washed several times, the pump is attached to the tube H, and a slow stream of carbon dioxide is gradually let in through the stop-cock G. By alternately exhausting the flask B and filling it with dry carbon dioxide, it is possible to dry the crystals perfectly in about an hour. They are removed from the flask by cutting off the funnel at the point (c). The crystals which were obtained in this way when analyzed proved to be not triphenylmethyl itself, but a combination of the unsaturated hydro- carbon with acetic ester. The composition of this substance is [(CbH5)3C]2. CH3CO2C2H5. This was proved not only by the analysis of the substance itself, but also by the fact that on heating it to a temperature of 70° C. the substance lost in weight the calculated amount (15.5 per cent), and the con- densed vapor was identified as acetic ester. The same operation with ethyl ether instead of acetic ester as a solvent, gave the corresponding ethyl ether derivative of the composition, [(CeH5)3C]2. (C2H,)20. It was soon foimd that other esters and ethers unite in similar proportions withthetriphenylmethyl. The study of these ether derivatives has been delayed because of the difficulty encountered in obtaining the ethers themselves in large quantities. The higher ethers, especially the mixed ethers, are pre- pared rather with difficulty whether we employ the method of Krafft* — the sulphonic acid reaction — or the method of Nef,^ which consists in the action of alkalies upon the sulphates of organic radicals. It is, however, quite cer- tain from the results so far obtained that triphenylmethyl unites with many ethers and esters in the proportion of two molecules of the hydrocarbon to one of the oxygen derivative. I suggested that the constitution of these com- pounds is probably as follows : (aH5)3C/ \C2H, ' (CeH,)3C^ ^COCHg' At the time when this suggestion was first made very little support in favor of this view was to be found in the chemical literature. But the classical work of Baeyer and Villiger upon the oxonium bases, which was pubUshed since then, established beyond any doubt that oxygen can act as a tetravalent element in a great many instances. This subject will be studied further, be- ' Ber. d. Chem. Ges., Bd. XXVI, p. 653. 2823 2829. 2 Ann. Chem. (Liebig), Bd. CCCXVIII, p. 1. J[. GOMBERG. 515 cause it is believed that the compounds here desoi ibed are the only instances we have where oxygen is linked by the four valences to carbon atoms. The question naturaly arises, if oxygen can be linked by all its vakuaces to carbon, may not tlie actual constitution of carbon monoxide be represented by the formula C = 0, instead of by 0=0? The ester and ether compounds of tripheuylmethyl are colorless crystalline substances. In their beha\-ior they resemble the tripheuylmethyl itself in every respect. They dissolve in different solvents with a yellow color; the solutions absorb iodine, and also give on exposure to air the insoluble tri- phenyl peroxide. This behavior indicates, therefore, that these bodies disso- ciate, when dissolved, into tripheuylmethyl and the ester or ether. Isolation of Triphenijl methyl. —li instead of acetic ester or ethyl ether we employ acetone as the solvent for the dry residue obtained on evaporating the benzene solution in the flask B, then the tripheuylmethyl itself can be obtained in the pure ciystaUine state. Apparently the hydrocarbon does not imite with acetone. Perhaps higher ketones may give rise to such com- binations. Tripheuylmethyl obtained in this way was found to be quite pure. It was analyzed and gave the following results : Calculated for Found. C.,H.,. I. II. C 93.74 93.27 93.28 H 6.26 6.69 6.22 Triphen3dmethyl when dry is a fairly stable substance. It may be exposed to air for a short time without danger of oxidation. It is white when freshly prepared, but soon turns pale yellow. It is soluble in benzene, carbon disul- phide, only slightly soluble in acetone, and hardly at all in petroleum ether. While colorless itself, the solutions of tripheuylmethyl are always yellow, no matter what solvent be employed. The solutions of the hydrocarbon when exposed to air instantly become covered on the surface with a crystalline crust of the peroxide; in a short time the whole of the tripheuylmethyl is converted into the oxidized product, and the supernatant liquid becomes colorless. As the acetic ester derivative of tripheuylmethyl is obtained in purer crystals and in larger yield than the hydrocarbon itself, this compound was used in all the following experiments. It was found that it showed all the reactions of tripheuylmethyl itself, and, as has been mentioned, when in solu- tion probably breaks down into the unsaturated hydrocarbon and the ester. This part of the problem was, therefore, successfully solved. The hydro- carbon could be obtained in pure crystalline state. The question, do the per- oxide and the triphenyliodomethane owe their formation to one distinct individual or not, could now be put to the test of experiment. 516 ON TRIPHENYLMETHYL. V. Triphenyliodomethane and Its Reactions. That the peroxide could be formed from the pure hydrocarbon was very easily settled. It was found that when the hydrocarbon, dissolved in benzene, is exposed to air, it gives from eighty-five to eighty-seven per cent of the calculated amount of the peroxide. The remaining portion is oxidized to an oil, which contains a little more oxygen than the peroxide itself. But the study of the iodo derivative presented a great many difficulties. The constitution of this had to be studied at first indirectly. The extreme instability of this halogen compound made it advisable to employ for most reactions only the solutions of it, without attempting to isolate the dry iodo compound. The study of the reactions proceeded along the following lines: first, titration; second, reactions with amines; third, formation of perhalides; fourth, formation of double salts; fifth, the separation of the iodo compound itself. Titration. — Accoiding to the equations : (CeH,),C+I=(aH5)3CI, [(CeH,)3C]2.C,H30,.C,H,+I = 2(CeH,)3CI+C,H302.CA, the hydrocarbon should take up 52.2 per cent, and the ester compound 44.2 per cent, respectively, of their weight of iodine. Weighed quantities of these substances were placed in a flask filled with carbon dioxide, and titrated with a standard solution of iodine in benzene. It was found from a large number of experiments that when about eighty per cent of the theoretical amount of iodine has been taken up, apparently an eqmlibrium is established. After that, the iodine is not absorbed, and it imparts to the solution its character- istic color. This is not so very strange. We know of a number of unsaturated bodies, such as crotonic and cinnamic acids, allyl alcohol,' etc., that do not take up the full quantity of iodine necessary for their complete saturation. Reactions with Amines. — If the substance formed by the addition of iodine is actually (C|jH5)3C— I, then it must show all the reactions which are char- acteristic of the analogous substances, (CeH5)3C — 01 and (CeIl5)3C — Br. The chloride and the bromide are very reactive bodies, resembling more acid chlo- rides, fike acetyl chloride, than the alkyl halides. They unite almost quan- titatively with different amines, giving rise to well-characterized derivatives, such as triphenylmethyl amine, the anilid compound, etc. The iodo compound was subjected to this test. The hydrocarbon was titrated and the required amount of the amine added at once to the solution. The amines were formed at once, in just as pure state and with just as good a yield as from triphenylchlormethane or triphenylbrommethane. The following amines were prepared by this reaction from the iodo derivative : ' Lewkowitsch — Chemical Analysis of Fats and Oils, p. 176. M. GOMBERG. 517 Triphenylamidomethane (C„H5)gC.NH„. Triphenylmethylethylamine (C„H,)3C.NHCJ2H5. Tnphenylmethylpropylamine (CaHJ^CNHCsH,. Triphenylmethylamylamine (C„H5)sC.NHC„H,,. Triphenylmethylaniline (C„H5)3C.NH.C„H.. Tiiphenylmethyl-ortho-toluidine (C„H„)8C.NHC,Hj. Ti-iphenylmethyl-meta-toluidine (CoHs^aC.NHC.H,. Tiiphenylmethyl-para-toluidine (C„HJ,C.NHC,H,. Perlmlides. — Another reaction of tripheuylbrommethane was discovered by myself some years ago.' I fomid tliat by the addition of iodine to a benzene solution of the bromide, a crystalline periodide of the composition (CeH5)3CBr.Ig, is formed, with a yield of about eighty per cent. Bromine gives an analogous perbromide, (CeH5)sC Br.Brj. Now, if in the titration of triphenylmethyl with iodine, a lai'ge excess of iodine in benzene or in carbon disulphide be added after the apparent end-reaction has been reached, a crystalline periodide of the composition (C(jH5)3CI.l5 is precipitated, the yield being about eighty per cent of the calculated amoimt. These two periodides, the one from the bromide and the one from the iodide, resemble each other in every respect. Both are dark blue crystalline compounds, both are equally stable, can be dried in a vacuum over sulphuric acid, and are not affected by dry air. On direct titration with sodium thiosulphate each gives up only five of the six atoms of halogen. Both periodides behave exactly alike, if certain precautions be observed, towards water, alcohol, metals, etc. A perfect analogy in the constitution of these two periodides cannot be doubted. If one has the constitution (CeH5)3^C— Br.Ij, the other must have the constitution (CeH5)3^C — I.I5, and hence, from this point of view, the original halides that gave rise to these perhalides must also have an analogous constitution. Double Salts. — -It has already been mentioned that zinc chloride gives a double salt with triphenylchlormethane. All attempts to obtain this salt in , crystalline form, like Norris^ did in the case of the compound of triphenyl- chlormethane with aluminium chloride, have failed. But it was found that other metallic halides, like stannic chloride, antimonic chloride, and others, give crystalline bodies with triphenylchlormethane. Tritolylchlormethane,^ which was prepared similarly to triphenylchlormethane, is especially prone to give such combinations with a variety of metalhe halides, such as mercuric chlor- ide, stannic chloride, ferric chloride, zinc chloride, etc. All these compounds are crystalUne and possess an intense color, from yellow to red. That the color is not due to the metal is evident from the fact that the zinc double salt is of an intense red color. This peculiar property of forming double salts is shared by tripheuylbrommethane in an equal measure. Monophenylhalogenmethane and diphenylhalogenmethanes possess also this remarkable tendency to unite ' Joum. Amer. Chem. Soc, Vol. XX, p. 790; XXIV, p. 612. 2 Amer. Chem. Joum., Vol. XXV, p. 54. ' Gomberg and Voedisch, Joum. Amer. Chem. Soc, Vol. XXIII, p. 177. 518 ON TRIPHENYLMETHYL. with halides of metals. In all eases where analyses have been made, the composition of these substances was found to be RjC.Cl+MeClx. Triphenyliodomethane was now in its turn subjected to the same test. It was found that it also exhibited the same tendency as the corresponding chloride and bromide, to give rise to double salts. The compound with zinc iodide was in this instance obtained in crystalline form. Triphenyliodomethane, (CeH5)3CI. — In all these three reactions, (1) reaction with amines, (2) formation of periodide, (3) formation of double salts, the iodo compound agreed in every respect with the triphenylbrommethane and tri- phenylchlormethane whose constitution has been well established. But as all these experiments were performed on samples of the iodo compound which had been prepared only in solution, there was still attached to these proofs a certain element of uncertainty. It therefore seemed advisable to isolate, if possible, the iodo derivative in dry crystalline state, and to perform the above reactions with the pure triphenyliodomethane itself. The method which has at last given me good results is as follows: Tiiphenylmethyl is dissolved, or merely suspended, in warm petroleum ether and titrated care- fully to the end-reaction with an iodine solution, also in petroleum ether. On the gradual addition of the iodine the insoluble triphenylmethyl changes to the more soluble triphenyliodomethane. On cooling in an atmosphere of carbon dioxide, the iodide separates in large yellow crystals. The petrolemn ether is siphoned off, the residue is washed with more petroleum ether, and dried in a stream of carbon dioxide. The analysis of the product so obtained gave perfect results for the composition (CeH5)3CI. The substance gave every one of the reactions which are characteristic of triphenylchlormethane and triphenylbrommethane. With water, the carbinol is formed; with alcohol, if certain precautions be observed, the ethoxy derivative is produced; with metals, we get the same unsaturated hydrocarbon, triphenylmethyl, as is obtained from triphenylchlormethane and triphenylbrommethane.' In addition to this, all the reactions described above as to the formation of substituted amines, the perhalides, and the double salts, were given also by this dry crystalline iodide. Triphenyliodomethane is, however, far less stable than the corresponding bromide and chloride. It cannot be exposed to air, especially when in solu- tion, without complete breaking down with liberation of iodine. Summary of the Reactions of Triphenyliodomethane. — The experimental work detailed in this chapter can now be summarized as follows : M. GOMBERG. 519 The iodo-oompound formed on the addition of iodine to the unsaturated hydrocarbon. 1. 2. Reacts with water and with alcohol, and gives the same products as the bromide and the chloride. With silver chloride, triphenylclilormethane is formed. 3. Gives with ammonia the same amine, almost quantitatively. 4. Gives with amines also the same sub- stituted bases; in some cases quantita- tively. 5. Forms with iodine, with about the same yield, a similar pentiodide, (G„H,)3CI.I,. 6. Exactly the same unsaturated hydro- carbon is produced by the action of metals upon the iodo-compound. 7. Forms similar double salts. Triphenylohlormethane and tri- phenylbrommethane. 1. Empirical composition : Ci,H,,A C,„H,,Br. 2. React with water readily and gi\-o triphenylcarbinol. With alcohol the ether is formed. With silver chloride the bro- mide gives triphenylcUormethano. 3. With ammonia the amine, (C|iH5)3 CNHj, is produced. 4. React with substituted amines and give the corresponding bases (CoH5)3CNHR. 5. The bromide gives rise to a perbro- mide and also to a periodide: (C,H,)3CBr.Br5; (C^H^jCBr.!,. 6. AVhen acted upon by metals the chlo- ride and the bromide give an unsatvirated hydrocarbon, which, on exposure to air, forms triphenylmethylperoxide, (CeH3)3C-0-0-C(CeH,)3. 7. Both unite ^ith halides of metals, forming combinations of the nature of dou- ble salts. The facts described fully justify, I believe, the following conclusions : 1. The iodo-compound is triphenyliodomethane, and its constitution must be entireh- analogous to that of the corresponding bromide and chloride. CeH,^C-I. 2. All the iodine added in the titration of the hydrocarbon is used up in the formation of this iodomethane; in other words, all the iodine is taken up by one carbon atom. If these facts are to be interpreted in terms of the valence theory it must be admitted, that in case the hydrocarbon here described is really imsaturated, then we are dealing with a case of unsaturation which is entirely distinct in its nature from what is generally designated by this term. The unsaturation of one carbon atom in a molecule always implies a simultaneous imsaturation of some other atom in the same molecule. In case of a hydrocarbon this second atom must, evidently, also be a carbon atom. We are accustomed to represent graphically such a state of unsaturation by a "double linking." In becoming saturated, as by the addition of halogens, each of the two atoms performs a similar function, each taking up one atom of halogen : >C = C<-f-2I =>C-C< I 520 ON TRIPHENYLMETHYL. I am not aware of any exceptions to this general rule. But in the case of the unsaturated hydrocarbon under consideration all the iodine, as I have shown, goes to one carbon atom. Hence, in the language of the valence theory, the unsaturation must be limited in this instance to one carbon atom in the mole- cule. The hydrocarbon must therefore, from this point of view, be represented by the constitution CeHj— C. VI. Salt-Like Nature of the Triphenylhalogenmethanes. Let us now examine a little closer the nature of these leactions, irrespective whether they are of the chloride, bromide, or iodide. That the halides should react with the amines is nothing strange. Alkyl halides in general give the same reaction. Acyl chlorides show it even more readily. It has been cus- tomary to look upon triphenylhalogenmethanes more as acyl chlorides than as alkyl chlorides. Indeed, the whole behavior of triphenylchlormethane re- minds one of acetyl chloride. It is very reactive, very sensitive towards water, alcohol, etc. But on what basis can we explain the fact that triphenylhalo- genmethanes form periodides? This property, as a rule, belongs to salts of the halide acids, hydrochloric, hydrobromic, and especially hydriodic acid. Some inorganic salts, like potassium iodide, mercuric iodide, etc., give rise to such perhalides. Salts of organic nitrogen bases, as substituted ammonium salts, aniline, and its homologues, pyridine, quinoline, and all alkaloids, also diazoniiun salts,^ all give perhalides of that nature, B.X.Xj (B standing for the radical of the base, X for the halogen). Not only the nitrogen bases, but also the iodonium salts^ give perhalides, as (C6H5)2l — I = l2; and even the latest salts, those of oxygen, such as Werner's' carboxonium salts, do that. Shall we then say that, by analogy, triphenylhalogenmethanes are also salts? Again, what classes of bodies form double salts? Here once more we can go through the same list. Inorganic salts, salts of organic nitrogen bases, phenylhydrazonimn salts, diazonium salts, carboxonium salts, aU give rise to double salts with metals. Shall we then say that in the light of this general property of salts, the triphenylhalogenmethanes are also salts? Such a com- parison of the triphenylhalogenmethanes with salts of different types suggested that it should be possible to prepare a variety of salts of the radical triphenyl- methyl. Norris^ observed that when triphenylchlormethane is dissolved in con- ' Hantzsch, Ber. d. Chem. Ges., Bd. XXVIII, p. 2754. 2 Hartmann u. V. Meyer, Ber. d. Chem. Ges., Bd. XXVII, p. 1594. « Ber. d. Chem. Ges., Bd. XXXIV, p. 3306. ' Amer. Chem. Journ., Vol. XXV, p. 119. M. GOMBERG. 521 centrated sulphuric acid, hydrochloric acid is given off. From this and other experimental facts he concluded^ that metals also probably split off from tri- phenylchlormethane not chlorine, but hydrocMoric acid. Kehrmann^ took the same view. Kehrmann also suggested that the yellow color which is formed when triphenylchlormethane is dissolved in sulphuric acid is due to a desmotropic form of the halide. H^ ^Cl This explanation did not seem to me plausible. It seemed to me that the reaction was much simpler, and is analogous to the reaction which takes place when benzyl chloride or benzotrichloride are treated with sulphuric acid, as for instance : CeHsCH^Cl + H^SO, = CeH^CH^ - - SO^OH + HCl . (CeH5)3C.Cl+H2SO, = (CeH5)3C-0-S020H+HCl. In other words, an acid salt of sulphuric acid is formed. On the addition of water this salt would break down, like so many other organic acid sulphates do, into the hydroxide and sulphuric acid: (CeH,)3C - OSO2OH +H,0 = (QHOjC- OH +H2SO,. Indeed, this reaction proved to be one of the best methods of preparing the carbinol in pure state. That Kehrmann 's interpretation does not explain the facts can be easily shown by the following experiment. A weighed quantity of triphenylchlormethane is placed in the small flask A (Fig. 2, p. 531), and through a separating funnel concentrated sulphmic acid is added. The hydro- chloric acid formed is now driven out by means of dry carbon dioxide and re- ceived in a solution of silver nitrate. Quantitative experiments have shown that every bit of the hydrochloric acid can be driven out in this way from the sul- phuric acid solution, and yet the yellow color of the solution is not diminished in the least. Hence, this color cannot be due to the Kehrmann 's quinone chlo- ' In a publication of recent date Norris(Amer. Chem. Journ., Vol. XXIX, p. 131) states that the experimental evidence which led him to this conclusion is, as I showed, faulty. ' Ber. d. Chem. Ges., Bd. XXXIV, p. 3818. 522 ON TRIPHENYLMETHYL. ride. Of course, it is possible that a corresponding quinone sulphate may have been formed, (C.H.).C=0<"^^ In order to test this view, triphenylchlormethane was treated with silver sulphate, in the hope of obtaining the normal triphenylmethyl sulphate [(C6H5)3C]2S04, which, according to my view, should possess about the same yellow color as the acid sulphate. On the addition of silver sulphate to a ben- zene solution of triphenylchlormethane the sulphate of the metal at once becomes coated with a yellow layer, and the solution also turns yellow. Still more pronounced is the reaction with triphenylbrommethane, suggesting beyond doubt the formation of the sulphate of triphenylmethyl.^ Other silver salts, as the nitrate, phosphate, ferricyanide, all give rise to yellow compounds when allowed to act upon triphenylchlormethane. Silver chro- mate forms readily the crystalline triphenylmethyl chromate. While, with exception of the chromate and the sulphate, the other salts have not been isolated and analyzed, their existence can hardly be doubted. Baeyer and Villiger^ from their study of a trianisylmethane reached a similar conclusion; namely, that the trianisylmethyl radical,, (CH30.C6H4)3C— , like the triphenylmethyl radical, is capable of forming salts. They designate such salts as carbonium salts. I was also able to show that the triphenylmethyl radical is actually basic, because like the ammonium radical it forms a hydroxide which is basic. When the carbi- nol is dissolved in benzene, and hydrochloric gas passed into the solution, it changes with the first few bubbles of the acid into triphenylchlormethane, and the reaction can be made quantitative : (CeH5)3C- OH +HC1= (CeH,)3C.Cl H-H^O. VII. Triphenylchlormethane and Triphenyis wei'e increased in size and could be distinctly felt as hard, cord-like bands, similar to the superior rectus of the right e>-e. As before, no involvement of the cehular tissue of the orbit was demonstrable. On April 15 he left the hospital. The enucleated eye was fixed and hard(-ned in Miiller's solution and em- bedded in celloidin. Section confirmed the conditions noted clinically. The cornea was thin and sho^^■ed a marked keratitis. Iritis was well developed and cyclitis was present. No pathological changes could be made out in the choroid or retina. There was beginning calcification of the lens — secondary cataract. .•Ul the inflammatory changes were of more recent occurrence than those observed in the muscles, thus proving that the myositis was the primary condition. In considering this case the question immediatel}' arises regarding the nature of the myositis: Was it a primary idiopathic condition, arising in the muscle itself, or a secondary process dependent upon some local or general condition? It will be noticed that, in the beginning, especially in the left eye. where the course could be watched, the symptoms were almost classical for tenonitis, viz. : Swelling and edema of the upper eye lid confined to the upper and retrotarsal part, slight exophthalmos, pain on the slightest movement, limitation of movement, and chemosis beginning over the insertion of one of the recti muscles, and gradually extending over the whole surface. Whether such a comparatively rare condition existed primarily cannot be definitely stated, although such a supposition might seem warranted with the rheumatic diathesis present. AH the sj^mptoms above mentioned, however, could be produced by a primary inflammation of the muscles. That this was the true condition seems more probable, for two reasons : First, in the right eye the only muscle involved was the superior rectus; if arising from direct extension of inflammation from Tenon's capsule, we should expect all the muscles passing through the latter to be involved equally; secondly, myositis resulting from an idiopathic tenonitis has never been seen in any of the reported cages of this affection. No other local condition, such as osteomyelitis, periostits or cellulitis was present to account for the condition. Among the general conditions of which myositis may be a local mani- festation are to be mentioned syphilis and the various other infectious diseases, metastatic infection following trauma or abscess formation, and rheu- matism. As there is no history of infectious diseases or other signs of their presence, their influence on this case can be excluded. Likewise trauma and abscess formation can be eliminated. Rheumatism, however, was present, developing simultaneously with the trouble in each eye. The latter is a con- dition which is so poorly rmderstood etiologically and anatomically that its bearing on such a case as this must needs be considered very indefinite. " In the articular form changes characteristic of inflammation have been seen 546 THE EXTRAOCULAR MUSCLES. in the neighboring muscles, their fibers being swollen and granular or in a state of vitreous degeneration. In the muscular form there is probably a slight manifestation of a similar tendency, but the affected muscles do not exhibit any visible changes, and when swelling, redness, heat, and tenderness are apparent, the muscle is probably invaded by some other form of inflam- mation." (Lyman.) In this case the phenomena observed correspond closely with the require- ments laid down by Scriba for primary myositis. The swollen muscles retained accurately their shape and were absolutely distinct from their sur- roundings. The eye itself was free from inflammation until a secondary process developed through the cornea as a result of the exophthalmos. For this reason, therefore, as well as by a process of exclusion, this appears to be a case of primary idiopathic myositis. The recognition of primary myositis dates back to Froriep and Virchow, about fifty years ago. According to the latter, the pathological changes in this affection may arise in the muscle-cell itself, or in the interstitial con- nective tissue, or in both, it being often difficult to determine from whence the process primarily originated. In the parenchymatous form the muscle fibers undergo a number of retrograde changes — ^loss of strise, cloudy swelling, fatty infiltration and degeneration, and finally destruction. When the process is essentially interstitial, it may lead to the formation of small-celled infiltra- tion; when more extensive, to pus, or to the formation of granulation tissue or fat tissue, tendon or bone, which substances forcing themselves between the muscle fibers result in a purulent, fibrous, or ossifying myositis, as the case may be. Of the latter two, nothing further need be said, as they are foreign to the subject. The acute inflammatory form has been described under the various names of polymyositis or monomyositis, dermatomyositis, when skin eruptions are marked, and neuromyositis. The histological changes in the muscles are practically identical in all these conditions. Clinically the neuromyositis can be differentiated by the preceding motor, sensory, and trophic symptoms characteristic of a nervous affection. Idiopathic inflammatory myositis was first described clinically by Unver- richt, Hepp, and Wagner, simultaneously, in 1887. The symptoms, first laid down, as based on these cases, of fever, pain, and swelling of the muscles of the extremities, edema and rashes of the overlying skin, the process pro- gressing upward to involvement of the muscles of respiration and deglutition, death resulting from suffocation or inhalation pneumonia, have been modified by the descriptions of subsequent less severe cases, so that at the present time it is quite impossible to present any classical set of symptoms. They vary, according to the severity of the infection and the ntimber of the muscles involved. EtiologicaUy, idiopathic myositis is closely related to acute periostitis. JOHN ELWIN GLEASON. 547 spontaneous acute osteomyelitis, and acute rheumatic arthritis. Similar causes predisposing, there is in each of these conditions an acute inflammation which in the different diseases localizes itself in the different tissues. The nature of the etiological factor in myositis is unknown. There is a general belief that it is an infectious process. Pfeiffer contends that the cause is an animal parasite on account of a similar condition due to protozoa seen in lower animals. No bacteria have been found in a typical case. Their absence gives support to the theory of an intoxication. Senator and Kell report cases which followed the eating of stale crabs and diseased fish. Senator also suggests that it may be the manifestation of an autointoxication. All that can be said is that probably more than one etiological factor must be considered, and in a given case, sometimes one sometimes another may be operative. Although formerly regarded as a necessarily fatal disease, the prognosis can now be said to depend upon the severity of the onset, the number and location of the muscles involved, and the height of the fever. Each case therefore is a law unto itself. The treatment is, of course, symptomatic, as the cause is unknown. Nu- merous remedies, among which may be mentioned the salicylates, potassium iodide, antipyrin and phenacetin, and locally poultices, carbolic acid com- presses, and ointments of various kinds have been employed, with but little success in shortening the disease. In Laquer's case a deep incision extending to the bone was the only procedure which would bring the condition under control. Likewise in this case, relief of tension by removal of the eye seemed to effect a cure. This case adds nothing to our knowledge of the etiology or course of the disease. Its local symptoms are the same as seen in other cases modified by position. In its recurrence it resembles Laquer's and Herzog's cases, in which the biceps of the left arm and the right vastus internus were involved respectively. In the former the condition developed seven times in fourteen years, and was followed the last time by an acute attack of articular rheuma- tism. The author suggests that the intermittent character is due to the breaking forth from time to time of a specific infectious virus similar to that supposed to cause rheimiatism, a theory which is supported by this case. It was formerly thought that the muscles of the extremities and those of respiration and deglutition were the only ones involved in idiopathic myosi- tis, the remaining muscles being spared. In subsequent cases, those of the face, tongue, and even of the heart have been found to be attacked; and in 1891 StriimpeU reported a case where the patient, without known cause, was seized with headache, nausea, and vomiting, and developed pain and sweUing in the muscles of the extremities, difficulty in masticating, swallowing, and speaking, and prior to death, ptosis and ocular paralysis. The micro- 548 THE EXTRAOCULAR MUSCLES. scopical examination showed the cord and nerves to be normal. The muscles presented an interstitial myositis with marked parenchymatous 'changes. The absence of involvement of the extraocular muscles has been advanced as one of the points differentiating idiopathic myositis from trichinosis. This case, together with that of Striimpell, would seem to demonstrate that such a distinction does not exist. Although by their position, the extraocular muscles are not liable to secondary inflammation in articular rheumatism, it seems probable, as Wright has recently emphasized, that they are liable to be the seat of a muscular rheumatism, and there seems to be no reason why they should be immune to other diseases of muscles. This case then can be said to add to our clinical knowledge of this rare condition by forcibty illustrating its possible malignancy by position, and by completing the list of muscles presenting this condition, so that it now may be truthfully said that any muscle of the body may be the seat of an idiopathic myositis. LITERATURE. Unverricht.— Zeitung fur Klin. Med. 1887, Bd. XXVIII, Deutsche Med. Woch., 1891. Wagner. — Deutsche Arch, fur Klin, Med., 1887. Hepp.— Berl. Klin. Woch., 1887. Senator.— Zeit. fur Klin. Med., 1888; Deutsche Med. Woch., 1893. Scriba.— Deutsche Zeit. fur Chir., Bd. XXII, p. 497. Walther.— Deutsche Zeit. fur Chir., Bd. XXV, pp. 260-286. Striimpell. — Zeit. fur Nervenheilkunde, 1891. Jacoby. — Jour, of Nervous and Mental Diseases, Nov. 1898. Bertelsman. — Miinchen Med. Woch., Bd. XLV, p. 1020. Kell.— Jour, of Am. Med. Assn., Bd. XXVI, pp. 967-970. Herzog. — Deutsche Med. Woch., 1898, p. 605. Herrick. — The Am. Jour, of Med. Sciences, 1896. ON THE CULTIVATION OF TEYPANOSOMA LEWISI. WARD J. McNEAL, A.B., As.'iinlant in Bacteriology, FREDERICK G. NOVY, Sc.D., M.D., Professor of Bacteriology. (From the Hygienic Laboratory, University of Michigan.) I. Trypanosomatic Infections. The past few years have witnessed marked progress in our knowledge of the protozoal diseases of man and of lower animals. The old observation of the presence of the Plasmodium in malaria has been supplemented and rounded out by the brilliant researches bearing upon the transmission of this disease by the Anopheles genus of mosquitoes. The etiology of Texas fever has been established by the discovery of the piroplasma or pyrosoma in the blood of infected cattle, and the role of ticks in the spread of the infection, as is well known, was recognized even prior to the work done upon mosquitoes. Similar diseases have been noted in dogs, and even in man. The presence of amoeba in large numbers in dysentery and liver abscesses of the tropics, and even in occasional cases of dysentery originating in our own latitude, have led to the belief in the causative relation of these organisms to such diseases. While the careful studies of recent years have shown that epidemic dysentery is largely due to bacterial infection, and in this way have cast doubt upon the part played by the amoeba, yet the fact remains that in certain types of the disease the evidence points to this organism as an etiological factor. The positive proof, however, is lacking, and some uncertainty will exist until new methods establish, one way or another, the exact relationship. At present we may look upon the Amoeba coli as a direct and immediate cause, or as a means of conveying the possible bacterial agent, or as a mere secondary invader, which finds the soil, prepared by another agent, suitable for its propa- gation. The sporozoa, represented by the plasmodium, and the rhizopoda, by the amoeba, are not the only classes of protozoa which possess a medical interest. A third class, the flagellated protozoa, and among these more especially the trypanosomes, claim careful consideration, inasmuch as the investigations of the past few years have established beyond a doubt that wide and destruc- 549 550 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. tive epidemics among cattle, horses, and related animals are due to these organisms. The first representative of this group was described by Gruby,* in 1843, who found it in the blood of frogs. Other observers had undoubtedly seen this organism before this date. In succeeding years similar parasites were found in the blood of diverse animals, so that at the present time' we may say that all vertebrates may harbor in their blood, for a greater or less length of time, some species of the trypanosomata. Usually, however, the infection is not injurious to the host, for of the large number of trypanosomes known, only a few can be said to be pathogenic. Our immediate interest obviously centers about the trypanosomes present in the circulation of warm-blooded animals. As pointed out by Doflein, trypanosomes were seen in the blood of rats and hamsters as early as 1845, but these early observations were forgotten and overlooked. Lewis, in 1877, noted the presence of these parasites in the blood of rats in Calcutta. In 1881 V. Wittich, in Germany, rediscovered, as we might say, in the hamster a like parasite, previously seen by others. This observation was corroborated in the same year by Koch, who regarded it not as a Spirillum or spirochaete, as others had done, but as a flagellated organism, probably identical with that seen by Lewis in rat blood. Kunstler, in 1883, noted a similar organism in the blood of a guinea-pig. In 1891 Jolyet and de Nabias reported a like parasite in a rabbit. The parasites in birds and fish were studied by Dani- lewsky. It may be of interest to note that Durham,^ in his report of the yellow fever expedition to Para, mentions having seen trypanosomes in abundance in the blood sucked up by a Stegomya fasciata from a confined bat. Crookshank, in 1886, carefully described the rat trypanosome of Lewis, and pointed out the presence of an undulating membrane and of a flagellum. The same organism was studied by Carter (1887), Danilewsky (1888), Chalach- nikow (1888), Lingard (1893), Rabinowitsch and Kempner (1899), Wasie- lewski and Senn^ (1900), Jiirgens* (1902), and lately by Laveran and Mesnil (1900-1903). The last four investigations are of particular importance with reference to the morphology and life-history of the rat trypanosome. In 1880 Griffith Evans, in India, observed trypanosomes, or spirochetes as they were first regarded, in the blood of horses, mules, dogs, cattle, and ' A very complete bibliography of the literature bearing on trypanosomes, prepared by Hassall, -ndll be found in the " Emergency Report on Surra, " by Salmon and Stiles, Bulletin No. 42 (1902), Bureau of Animal Industry, United States Department of Agricul- ture, Washington, D. C. Only new references will be given in this article. ^ Thompson Yates Laboratories Report, Vol. IV, part II, p. 563, 1902. Journ. Trop. Med., V, p. 270, 363, 1902. = Zeitschr. f. Hyg., XXSIII, p. 444, 1900. * Arch, f . Hyg., XLII, p. 265, 1902. WARD J. McNEAL AND FREDERICK G. NOVY. 551 camels affected with a disease known locally as Surra. It was not, however, until 1893 that this disease received a thorough study. In that year Lingard published his report on Surra, which work attracted special attention to this affection, and in a way led to all the subsequent work upon trypanosomatic diseases. In 1901 the disease was recognized in the Philippines by Kinyoun and Smith, and by Slee. In 1894 and subsequently, Bruce investigated the tsetse-fly disease, or Nagana, of Zululand, and demonstrated not only that trypanosomes were present, but also the part played by the tsetse-fly in the transmission of the parasites. , In the course of his travels in East Africa, in 1898, Koch studied this same disease, and was led to believe that it was identical with the Surra of India. As a rule, German writers, following Koch's example, do not dis- tinguish between the two diseases. There is unquestionably a very great similarity in the symptoms of these two diseases, and in the morphological and pathogenic properties of the trypanosomes present, but it is by no means estabHshed that the two are identical. It is interesting to note that Martini' has met with a case of the disease in the BerUn Zoological Garden, in .a pony from West Africa, and it is highly probable that similar importations, though not recognized, have occurred elsewhere. The same may be said to be true of human trypanosomiasis, which wiU be mentioned later. In Algeria and other Mediterranean countries a disease known as Dourine, or "Maladie du Coit, " is known to occur exclusively among breeding equines. In the blood of a staUion thus affected Rouget, in 1896, found a trypanosome which he was able to keep ahve by transmission through a series of rabbits for two and a half years. Recently Buffard and Schneider^ have questioned the correctness of the diagnosis in the case of Rouget's stallion, and for other reasons beUeved that the organism studied by him was that of Nagana. Rou- get,^ however, has been able to answer these criticisms. While there is a tendency to admit the identity of Surra and Nagana, there can be no doubt, notwithstanding the analogy of symptoms, that the trypanosome of Dourine is different from that of Surra or Nagana. Insects seem to play no part in the transmission, which occurs exclusively through coition. The disease was studied by Buffard and Schneider in 1900, and in the following year by Nocard, who demonstrated that dogs immunized to the Dourine were killed by the Nagana trypanosome. This result is in accord with the studies of Laveran and Mesnil (1901), who pointed out marked morphological and developmental differences between the two parasites. There can, therefore, be no doubt but that Dourine is specifically different from Nagana. ' Zeitschr. f. Hyg., XLII, p. 341, 1903. 2 Reoueil M6d. Y6Uv., (8 S6r.) IX, p. 721, Dec. IS, 1902. ' Ibid., X, p. 82, Feb. 15, 1903. 552 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. A similar disease of the horse, known as " Mai de Caderas, " has been noted in South America, notably in Brazil, Paraguay, Uruguay, Argentine, Chih, and Bolivia, and it would not be surprising, in view of the endemicity of trypanosomatic infections along the Mediterranean littoral and in South America, if a like disease was found in our southern states and in Mexico. The presence of a trypanosome in Caderas was first recognized by Elmassian^ in 1901, but the most extended description of the disease and of the parasite was given by Voges in the following year. Laveran and MesniP have compared the trypan- osomes of Nagana and Caderas and have reached the conclusion that the two diseases are due to distinct species of parasites. There are constant morphological differences in the organisms, and animals immunized to Nagana are subject to Caderas. These facts, together with slight differences in the symptoms, show that the two diseases are distinct. Early in 1902 Theiler, a veterinarian in Pretoria, Transvaal, found in cattle a trypanosome which was clearly different from that of Nagana. While the latter is pathogenic to a large variety of animals, the disease in question is peculiar to cattle. In the region where the disease occurs it is called Gal- ziekte, referring to a disease of the bile. The findings of Theiler have been reported by Bruce^ and by Laveran and Mesnil.* The latter workers have described the parasite from preparations sent to them by Theiler. It appears that the trypanosome of this disease ranges in size from 30 to 65 ,« in length, and is therefore the largest species known. It is interesting to note that cattle may be infected with this parasite and with Texas fever at the same time. Another specimen of blood sent by Theiler to Laveran showed a trypano- some different from the above, which, moreover, was unlike any other known species. Laveran designated it as Trypanosoma transvaaliense. The animal which supplied it had a triple infection — trypanosome, Texas fever, and rinder-pest. Nothing is known as to the pathogenicity of this organism. It would seem that the anti-rinder-pest inoculations with defibrinated blood have helped to spread the trypanosomatic and other infections among cattle. The extremely destructive trypanosomatic diseases thus briefly touched upon prevail among domestic and wild animals in the tropical and subtropical countries. For some time it seemed as if man was wholly free from this class of infections. In South Africa the tsetse-fly bites men, as well as animals, but while the bite in the latter is followed by Nagana, in the former it is without effect, showing that man is immune to the trypanosome of that disease. We • Ann. de I'lnst. Pasteur, XVII, p. 241, 1903. ' Compt. Rend. Acad. Sc, CXXXV, p. 838, Nov. 17, 1902. 3 The Lancet, March 8, 1902, p. 664. ' Compt. Rend. Acad. Sc, CXXXIV, p. 512; CXXXV, p. 717, 1902. WARD J. McNEAL AND FREDERKMv G. NOW. 553 know now, however, that while man is probably iinmune to the diseases mentioned, he is nevertheless subject to invasion by a certain trypanosome. Nepveu' reported in 1891 and in 1898 some observations upon the blood of malarial patients made in Algeria in 1888. While the claim is made that a trypanosome was found at that time it can hardly !)(> said that enough work was done to merit unreser\'ed acceptance of such a statement. The fii-st authentic case of trypanosomatic infection of man was reported by Button^ in 1901. The patient was an Englishman in the government service in Gambia. Forde,' the attending physician, had previously noted in the blood of this man what he called "small, worm -like, extremely active bodies." The illness began in May, 1901, and ended fatally in January, 1903. On the return of the patient to England a further study of the case and of the parasite was made by Annett.'' The chronic character of the infection in man was reproduced in animals. In the case of the white rats, some appeared refractory, others died in two to three months; in monkeys, it was fatal in about two months, while dogs seemed to be unaffected. The parasite appears to be smaller than the other pathogenic trypanosomes, and there can be no doubt but that it represents a distinct species. As pointed out by Annett. the chronic character of the disease, the scantiness of parasites, their apparent absence for long periods from the peripheral blood, and the occasional rise in temperature, may easily lead to a diagnosis of malaria. In fact, malarial parasites and trypanosomes may exist in the same blood, as was found by Button," when, on his return, he examined 115 films obtained from native children in Gambia. One of these showed malarial parasites and trypanosomes. The recognition of two cases in man led the Liverpool School of Tropical Medicine to send an expedition to Senegambia with the purpose of looking for other cases of human trypanosomatic infection. Button and Todd," in their preliminary account of the work, add to the two instances mentioned another case of infection in a white man and four cases out of 220 natives examined. To these seven unquestionable cases of human trypanosomiasis others are to be added. Manson,' after seeing Button's first case, was able to recognize the disease in a woman, the wife of a Congo missionary. The diagnosis was confirmed by Baniels, who succeeded in finding the trypanosomes in the blood, although daily ' Mem. de la Soc. Biol., Ill, p. 39, 1891; V, (10 S6r.) p. 1172, 1898. 2 Thompson Yates Laboratories Rep. Vol. IV, Part II, p. 455, 1902; Brit. Med. Journ., Dee., 1901, Jan. 4 and 11, and Sept. 20, 1902. =■ Journ. Trop. Med., V, p. 261, 330, 361, 1902. * Brit. Med. Journ., Feb. 7, p. 304; April 18, p. 927, 1903. ' Locus citus, p. 467. " Brit. Med. Journ., Feb. 7, 190.3, p. 304. 'Journ. Trop. Med., V, p. 330, 1902. 554 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. examinations made during the first two weeks were negative. More recently, Manson' has reported another case in a white woman, also from the Congo. The parasite was recognized in her blood by Broeden, of Leopoldville. Two other cases, occurring on the Congo, are mentioned by Manson. Kermorgant^ Ukewise reports the detection of trypanosomes in the blood of a Congo official. Brumpt^ has added another case also met with on the Congo river. Up to the present time then, seven cases of human trypanosomiasis have been found on the Gambia, and six on the Congo river. It is obvious that this type of infection is by no means rare in tropical Africa, and it is quite likely that trypanosomatic diseases of man will be found in other parts of the world. It wiU be seen from this brief outline of our present knowledge of the pathogenic trypanosomes that there are several distinct diseases caused by these organisms. For the present it may be well to keep Surra and Nagana distinct, for while there is no positive proof of their being different, yet there is some reason to believe this to be the case, and there is no real evidence of identity other than the absence of differentiating characteristics. New methods of work, such as cultivation and immunity experiments, will be needed to throw further light upon this point. Up to date the following extremely fatal diseases then are known : 1. Surra; occurring in India, Burmah, China, the PhiUppines, and possibly in Africa; affecting horses, mules, camels, buffaloes, etc. Cause, Trypanosoma evansi. 2. Nagana; prevailing in South Africa, and attacking horses, cattle, and wild animals. Cause, Trypanosoma brucei. 3. Dourine; a disease of equines in Algeria and other Mediterranean coun- tries; apparently transmitted solely through coition. Cause, Trypanosoma equiperdum. 4. Caderas; a disease of South America, affecting horses. Cause Trypano- soma equinum. 5. Galziekte; a cattle disease of South Africa. Cause, Trypanosoma theileri. 6. Human trypanosomiasis, met with in West Africa. Cause, Trypanosoma gambiense. II. The Cultivation op Animal Organisms. The remarkable development of bacteriology during the past quarter of a century is due to the introduction of methods of cultivation whereby pure growths of the organisms under examination can be readily obtained, and thus their relation to disease and to other changes can be demonstrated as that ^Ibid., VI, p. 85, 1903; Brit. Med. Journ., p. 720, March 28, 1903. ^BuU. de I'Acad. de M6d., XL VIII, (3 Ser.) p. 574, December 16, 1902. ^Md., XLIX, (3 Ser.) p. 368, March 17, 1903. WARD J. McNEAL AND FREDERICK G. NOVY. 555 of cause and effect. Moreover, as is well known, by means of such pure cultures it has become possible to disco^^er means for the prevention and even cure of many diseases. There are still some bacterial diseases where the organism present resists all methods of cultivation, and yet the constant presence of such an organism justifies, by analogy, our belief in its causal relation to the disease. Never- theless the absolute demonstration of such relation will not be accomplished until the organism in question is isolated and grown in a pure state, and until the disease is reproduced by means of such pure cultures. The recognition of protossoal diseases has led numerous workers to apply to the study of these diseases the culture methods which have been so fruitful in bacteriology. These attempts, however, have been more or less unsuccess- ful, and it has seemed as if the cultivation of animal parasites was beyond the reach of the investigator. Under these conditions it has come to be assumed that the constant presence of a given animal parasite in a disease established its causal relation, and, \indoubtedly, in many diseases, such as malaria, Texas fever, trypanosomiasis, and like affections, this is absolutely correct. On the other hand, in the so-called amoebic diseases grave doubts may be raised as to the part played by the organism. Such questions can be answered, positively or negatively, when it becomes possible to imitate the natural con- ditions and obtain pure growths of the protozoon apart from the diseased organism. KartuUs^ was perhaps the first to attempt the cultivation of supposed pathogenic protozoa. In the course of his work upon tropical dysentery and abscess of the liver, he endeavored to obtain cultures of Amoeba coli upon beef-tea, liquid blood serum, Uquid gelatin, and the Uke; but in this he was unsuccessful. On employing straw infusions previously heated and then inoculated with amcfibic material, he obtained better results. He secured mixed cultures of amoebae and bacteria, but although he claimed that the organism cultivated was Amoeba coli, this was doubted by Kruse and Pas- quale,^ by Casagrandi and Barbagallo.^* The use of Hquid media by Bosco and Perroncito,^ Miller* and others, led to similar results; that is to say, they were able to obtain a multiplication of amoebse in the presence of bac- teria. Most of the recent efforts at cultivating the amoeba have been made with solid media and there can be no doubt that these have been successful. The ' Ceutralbl. f. Bakt., IX, p. 365, 1891. ' Zeitschr. f. Hyg., XIV, p. 1, 1894. ' Centralbl. f. Bakt., XXI, p. 579, 1897. * R. Acad, di Med. di Torino., Nov. 29, 1895; Jan. 10, 1896. * Centralbl. f. Bakt., XVI, p. 273, 1894; Welch's Contributions to the Science of Medi- cine, p. 511, 1900. 556 OX THE CULTIVATION OF TRYPANOSOMA LEWISI. Amceba coli, however, has not been grown artificially. Such is the opinion of Behla,' and Doflein- even goes so far as to hold that most of the amoeba cul- tures obtained thus far are not real amcebfe, but rather amoeboid stages of myxomycetes. Celli and Fiocca^ (1894) may be said to be the first to employ with success solid media, in which Fucus crispus was used in place of agar. Their amoebae, however, were invariably accompanied by the various acciden- tal bacteria which chanced to be present in the material used. This was also the case with Tischutkin* and Schubert,^ who obtained mixed ^gar cultures. Beyerinck,'' in 1896, went a step further. He succeeded in obtaining growths from garden soil of two amoebae, one of which. Amoeba zymophila, he cultivated for more than a year on agar inoculated with a definite organism — Saccharomyces apiculatus. He expressed the belief that amoebae required solid food in the shape of bacteria or yeasts, and that without these organisms they could not develop. Gorini' succeeded' in growing Amoeba zymophila in the presence of S. apiculatus on potato. Frosch,* in 1897, likewise attempted to secure pure cultures of an amceba obtained from garden soil. Inasmuch as the amoeba was associated with a sporeless bacillus, he sought for a means of separating the two organisms. This he was able to do by placing old cultures, in which the amoebae had as- sumed the encysted condition, into a twenty per cent soda solution for seventy- two to seventy-four hours. The bacteria were thus destroyed while the cysts were unaffected. He was able to obtain no development of these cysts in the pure condition, or in the presence of dead bacteria, or of their products. On furnishing definite species of living bacteria he obtained good growths and for that reason he held that living bacteria were essential to the amceba. Zau- bitzer" employing Frosch's method obtained like results. Schardinger'" (1896) also obtained pure amoebae free from bacteria, but was equally unable to secure their multiplication without the addition of living bacteria. He used hay or straw infusion agar, which was slanted and the water of condensation was inoculated with the mixed amoeba material. A streak was then made along the middle of the agar and the amoeba followed the bacteria which developed on the line of inoculation. 1 Die Amoben, Berlin, 1898, p. 49. ^ Die Protozoen als Parasiten und Krankheitserreger, Jena, 1901, p. 36. 3 Centralbl. f. Bakt., XV, p. 471; XVI, p. 329; XIX, p. 536, 1896. ^ Centralbl. f. Bakt., II Abtheil., Ill, p. 183, 1897. = Hyg. Rundschau, VII, p. 73, 1897. « Centralbl. f. Bakt., XIX, p. 257, 1896; XXI, p. 101, 1897. ' Centralbl. f. Bakt., XIX, p. 785, 1896. 8 Centralbl. f. Bakt., XXI, p. 926, 1897. 'Ibid., XXX, p. 311, 1901. '» Ibid., XIX, p. 538, 1896; XXII, p. 3, 1897. WARD J. McNEAL AND FREDERK^K G. NOW. 557 Tsujitani' (1898) by means of Schardinsor's method was able to grow amoebae in association with pure cultures of various kinds of bacteria, nuch as cholera, typhoid, coli, green pus, etc. He found that the ama'bae would grow on cultures of these germs even after they had been killed by heat. Tsujitani, therefore, may be said to be the first, if not the only one, who has suc- ceeded in obtaining pure cultures of amoeba\ Anthrax bacilli and yeasts were not suitable as food. Although the anacebte strayed from the line of inocula- tion where the bacteria were growing, and in this way were separated from the latter, it was not possible to induce these pure amoebae to grow without solid bacterial food. Mouton,^ in his recent work (1902) on the enzymes produced by the amoeba, prepared mass cultures of these organisms, associated, however, with B. coli. These pure mixed cultures were obtained by making streaks on Petri plates containing an agar which was very poor in nutritive substances. The amoebae multiplied at the expense of the accidental bacteria present and eventually wandered from the streaks. By placing B. coh near by, the wandering amoebae reached these organisms, and by repeating this procedure eventually cultures were obtained wluch contained only amoebae and B. coli. Mass cultures in Roux flasks were then prepared with ease. Apart from the amoeba, very little work has been done in the cultivation of protozoa. Ogata^ (1893) by means of a capillary tube was able to separate motile protozoa from associated bacteria. In this way he obtained pure cultures of Polytoma uvella, and also isolated Paramecium aurelia. A pure culture of Protomonas spirogyrae, which belongs to the mycetozoa, was obtained by Schardinger.'' In biological laboratories, as is well known, certain protozoa are cultivated so to speak, by keeping these imder natural conditions. This is notably the case with the interesting Paramecium, the multiplication of which can be fol- lowed through many generations. This takes place by division and unless the organism is rejuvenated by conjugation with paramecia from a new source, it will eventually die out. Calkins,^ however, has recently shown that conjuga- tion is not necessary to this rejuvenescence, since he has been able to secure over five hundred generations in the course of fifteen months by the mere addi- tion of chemical substances which exert a stimulating action upon the organ- ism. Mechanical agitation had a like effect. It is worthy of note that Calkins observed periods of depression in his Paramecium, at which times the organisms » Cetitralbl. f. Bakt., XXIV, p. 666, 1898. 2 Annal. de I'lnst. Pasteur, XVI, p. 457, 1902. Theses presentees &, la Faculty des Sciences de Paris, 1902. ' Centralbl. f. Bakt., XIV, p. 165, 1893. * Locus citus. ' Arch. f. Entwickelungsmeehanik der Organismen, XV, ^. 139, 1902. 558 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. were smaller, failed to divide as frequently as before, and the transplantations often died out. This depression was overcome by a change in the medium used, or by agitation. Schardinger in his work on the amoeba has called attention to the need of varying the medium on which the organism has grown. In our own work we have repeatedly seen a decided improvement in the culture when the trans- plantation was- made to a medium which possessed a different reaction from that which had been used for some time. The brief account which we have given of the attempts made to secure cultures of protozoa serves to show how much is still to be done. There is unquestionably an enormous difference between the lower plants and animals in the ease with which they adapt themselves to changed conditions. While bacteria obtain their nourishment by osmosis, the amoeba and other animal forms seem to require solid food in the shape of other organisms. It does not follow, however, that all protozoa make use of such food. It seemed to us that the rat trypanosomes were particularly well adapted for the purpose of determining whether or not it is possible to cultivate protozoa without the presence of bacteria, and hence upon purely liquid food material. In the Hving body the trypanosomes undoubtedly live upon the soluble con- stituents of the blood. It is true that these organisms are exquisite parasites, and for that reason the task which we set ourselves appeared to be all the more difficult. The one advantage enjoyed at the outstart is the fact that the trypanosomes are present in the blood in pure culture and hence all experi- ments can be made with material free from bacteria. In view of the fact that trypanosomes are clearly pathogenic organisms, it seemed all the more desira- ble to secure cultures of these animal parasites. With such cultures it would then be possible to establish fully the causal relation of these organisms to the disease. Moreover, a method which would permit the successful cultivation of the rat trypanosomes would in all probability be applicable to the isolation of the parasites of Surra, Nagana, Caderas, Dourine, and the like. The arti- ficial cultivation of the trypanosomes which cause these diseases is obviously of far-reaching importance from the standpoint of the prevention and possible cure of these affections. Some of the problems which we have taken up have been satisfactorily solved. It will be shown that rat trypanosomes can be successfully cultivated in vitro. The cultures wiU develop in fluid media at room temperature, or at that of the incubator. Moreover, the inoculation of such artificial cultures reproduces the disease in rats. The results of our work along these fines is given in the following pages. While the manner of multiplication of the trypanosomes in these cultures has been studied to some extent, our observations are by no means complete, and for that reason will be taken up in a subsequent paper. WARD J. McNEAL AND FREDERICK G. NOVY. 559 III. Distribution and Viability of Trypanosomes. The trypanosome infection of wild rats may be said to have a worldwide distribution. It is by no moans certain, however, that Trypanosoma lewisi is the only species which is harbored by rats. Lingard has shown that the blood of rats in India may communicate Surra to horses, and for that reason he failed to distinguish between Trypanosoma leivisi and Trypanosoma evansi. We know that these two organisms are distinct species. In India, therefore, the rat at large may harbor at least two distinct trypanosomes. The fact that the rat is susceptible to laboratory inoculation with the parasites of Surra, Dourine, Nagana, Caderas, and of man, indicates the possibility of finding several species of trypanosomes in the blood of wild rats. Rats infected with Trypanosoma lewisi rarely die. In fact, it is the excep- tion to observe any very marked ill effects in either the white or gray rat after inoculation with the parasites. Occasionally the animal is so sick that it offers little or no resistance in process of handling. In the course of our work we have had occasion to inoculate about 150 to 200 rats, white and gray, and of these, nine died. There was an entire absence of bacterial infection in three of these cases and there can be no doubt but that death was due to the trypanosomes present. Of the latter, one died on the fifth day and two died on about the thirteenth day after inoculation, and their blood contained enor- mous numbers of the parasites. Considerable enlargement and softening of the spleen was found. The last two rats to die (B and C) received injections of cultures XXa and XX&, which see. Of the other observers, Jiirgens' is the only one to our knowledge who has noted fatal results in rats. Of the forty-seven young white rats which he inoculated, sixteen died: two of these on the fourth day, one on the twenty- fifth day, and the others in the second week. The animals sickened in a few days after inoculation, became dyspnoeic, developed edema in the hind legs, and showed subcutaneous hemorrhages. Streaks from the peritoneum and heart-blood remained sterile. Assuming the absence of anaerobic bacteria, for the presence of which he did not test, it would seem that Jiirgens either worked with a new species of rat trypanosome, or with a Trypanosoma lewisi, the virulence of which had in some way been increased. The first trypanosome which we found did not produce fatal results, whereas the parasite from another animal, obtained from the same place, gave us the results mentioned above. This difference could easily be due to a slight variation in the virulence of one and the same organism. During the past three years we have examined 107 wild rats in Ann Arbor. Of these five, or 4.67 per cent, were found to be infected. It is significant that ' Arch. f. Hyg., XLII, p. 276, 1902. 560 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. these five infected animals came from the same barn. Had the examination been limited to that one location it is quite likely that a much larger per cent of infected animals would have been seemingly observed. This fact should be borne in mind in the interpretation of the figures given by other workers. On the other hand, when the infected animals recover they become immune to subsequent inoculation and to this may be due the fact that some rats are refractory. Lewis found twentj'-nine per cent of the rats examined by him in Calcutta to be infected. Carter, at Bombay, obtained twelve per cent, while Lingard, presumably at Bombay, examined 3,105 rats and found 34.91 per cent of these to be infected with trypanosomes. Crookshank, in London, met with twenty-five per cent of infections; Rabinowitsch and Kempner, in Berlin, found 41.8 per cent. In Paris they were reported as not rare by Blanchard, while Laveran and Mesnil found only two infected rats out of forty-three examined. Buard,^ at Bordeaux, examined fifteen rats and found all of these to be infected. According to Calmette infected rats are common at Lille. Koch, at Dar-es-Salaam, in Africa, found ten infected rats out of twenty-four examined. Button found trypanosomes to be frequently present in the rats at Bathurst, Senegambia. "\'oges, in South America, in connection with his work on " Mai de Caderas, " examined numerous rats with negative results. Rouget, at Constantine in Algeria, likewise failed to find trypanosomes among the sewer rats. To our knowledge they have been met with in one instance in Detroit, and they have also been observed in the rats in Manila. The studies of Rabinowitsch and Kempner have shown that the infection is spread among rats by fleas and lice. We have had occasion to observe living trypanosomes in the blood in the stomach of lice. It is possible that other blood-sucking insects are instrumental in spreading the disease. The observa- tion of Button regarding the presence of trypanosomes in the blood of a bat and sucked up by the Stegomya fasciata has been referred to. As in the case of rats, it is possible for other trypanosomes to exist in the blood of wild animals without causing a fatal disease. Thus, according to Bruce, the Nagana parasite exists in the blood of wild animals without inducing severe infection, whereas the blood of these animals when sucked up by the tsetse-fly (Glossina morsitans) and transmitted through its bite to horses and cattle produces the fatal disease, Nagana. The similar role of insects in the transmission of Surra, Galziekte, and Caderas, is well recognized. Whether these insects cause the disease by inoculating the original fully developed trypanosomes which they have sucked up with the blood, or whether these trypanosomes give rise in the body of the insect to as yet unknown develop- mental forms, remains to be proven. ' Compt. Rend. Soc. Biol., LIV, p. 877, July 11, 1902. WARD J. McNEAL AND FREDERICK G. NOVY. 561 Numerous observations lia^-c been made upon the viability of trypanosomes in litro in the blood of the host. Attempts have, indeed, been made at culti- vating these organisms, but with the exception of that of Jurgens, they have been unsuccessful. Danilewsky observed living trypanosomes in blood kept in a pipette, or even under a cover-glass, for eight to nine days, at ordinary temperature. Involution forms per-sisted for twelve days. He also noted that the young parasites adapted themselves to artificial conditions, that is to say, they Uved longer than the old ones. Although this worker is usually credited with having cultivated the trypanosomes in capillaries this is not substantiated by any definite statement of his own. According to Rouget, Chalachnikow is credited with growing the parasite on dog serum. Six days after inoculation he ob- served, besides the ordinary forms, new ones showing different degrees of development. Rabinowitsch and Kempner, repeating Danilewsky' s observation, were unable to obtain any development of trypanosomes either in the hanging-drop or in capillaries. They found, however, that the parasite retained its vitality and virulence for a week when the blood was kept at room-temperature, or even when kept in the incubator. Wasielewski and Senn were hkewise unable to secure development of the rat trypanosomes. Laveran and Mesnil found that the rat trypanosomes died out in defibri- nated blood, or in blood diluted with salt solution, in four to five days, while, if placed in an ice-box at +5° to -1-7° they remained alive for fifty-two days. In the latter case the blood was still virulent although the trypanosomes could not be detected. They appeared more slowly (seven days), in the blood of the inoculated animal, than under ordinary conditions. They were unable to observe any multiplication in the hanging-drop at room-temperature, or in the tubes of blood kept in the ice-box. If bacteria chanced to be present the trypanosomes died earlier than would have been otherwise the case. The same workers note that fish trypanosomes can live for several days in vitro, and in this they confirm Berg who found the Trypanosoma remaki alive in blood at 12° for six days. Mitrophanow also kept this species alive for three to four days. Jolyet and de Nabias are said to have kept a rabbit trypanosome ahve in blood preparations for five days at room-temperature. Jijrgens' noted that the rat trypanosomes could be kept alive in micro- scopical preparations for seven days, at —5° to —8°. In a freezing mixture at — 17° they died out in two hours and the blood failed to infect animals. Hang- ing-drops and capillary tubes kept at +5° to -|-10° for thirty-two and fifty- three days infected rats after a period of incubation of seven days. At room- temperature they died out more rapidly (one to one and one-half weeks), and at 37° in two to four days. As he himself points out the rapid death at 37° is ' Arch. f. Hyg., XLII, p. 279, 1902. 562 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. due to the development of bacteria, but he notes that in aseptic preparations death also results in two to three days. Jiirgens was able to obtain evidence of the multiplication of trypanosomes in the hanging-drop. Such preparations made with trypanosome blood, drawn three to four days after inoculation and containing young trypanosomes, placed at 37° showed on the following day developmental forms which were not present in the beginning. This result was not obtained when the preparation was kept at room-temperature or when blood containing old tryp- anosomes was used. This observation with young trypanosomes is in accord with that of Danilewsky. Our own work, it will be seen, shows that the trypanosomes can grow at room-temperature and at that of the incubator. It is of interest to note that according to Kanthack, Durham, and Bland- ford the Nagana trypanosomes may be alive in the blood for two to three days, and even for five to six days, whereas in the dead body they die out within twenty-four hours. Plimmer and Bradford kept Nagana blood from a dog, in the presence of oxygen, for three days and although no trypanosomes could be detected they found it to be infectious. In hanging-drops they found it alive for five to six days. Bruce has also been able to preserve the virulence of Nagana blood in vitro for four days. According to Laveran and Mesnil, Nagana blood mixed with physiological salt solution can be kept in a virulent condition for three days at room-temperature. The results, however, were not constant. Undiluted blood, they found could become inert in forty-eight hours, whereas, if diluted with horse serum the parasites could live for three days. When kept in the ice-box at -|-5° to +7° the Nagana parasites died out the Same as when kept at room-temperature, in three to five days. They also observed the interesting fact that when the blood was subjected to —50° to — 55° for one-half to two hours and then melted that the Nagana trypanosomes were still alive. This observation shows that these organisms are fully as resistant as the rat trypanosomes, which, according to Jiirgens, died out at — 17° within two hours. The trypanosomes of Caderas, as in the case of Surra and Nagana, disap- pear from the blood after the death of an animal within twenty-four hours. Voges, as a rule, found that these trypanosomes, in the hanging-drop, lost their motion in from ten to fifteen minutes and that they completely: disap- peared in twenty-four to forty-eight hours. In three to four days the blood was no longer infectious. In one instance, however, he was able to infect a horse with blood which was fourteen days old. As to the trypanosome of Dourine, we have the observation of Rouget, that in fresh blood preparations it was motile for eighteen, but never after twenty-four hours. On attempting to cultivate it in the serum of rabbits and dogs he noted that a few trypanosomes were alive for twenty-four hours at 37°. WARD J. McNEAL AND FREDERICK G. NOVY. 563 It may not be out of place, in connection with the viability of these animal parasites, to mention that Laveran and Mesnil' state that Texas fever blood has been transported on ice from Argentine to France without loss of virulence. In view of the above facts we are convinced that Nagana, Surra, and other trypanosomes are not much weaker than the rat trypanosomes, if at all, and that, for that reason, it will be possible, as with the latter, to secure cultures in vitro. In our preliminary experunents on the cultivation of rat trypanosomes a large variety of media was tested. It was soon found that trypanosomes could be kept alive for several months when the infected blood was added to rat, rabbit, or Guinea-pig blood. In these trials every effort was directed to prevent bacterial infection, for, in common with other observers, we have found that bacteria soon destroy the trypanosomes, probably by altering the mediiun, and more especially the hemoglobin. The fact that the trypanosomes lived in the rat blood when diluted with other blood for a considerable length of time, even at ordinary room-tempera- ture, whereas, in undiluted blood they died out quite early, indicated that possibly multipHcation was taking place. The next step in our procedure was to add sterile, defibrinated blood to ordinary nutrient agar of varying reaction. For this purpose the agar was melted and then cooled to about 50°, when it was mixed with the defibrinated blood, usually three to one, after which the tube was slanted and allowed to solidify. The medium thus pre- pared is bright red and contains unaltered hemoglobin. The latter appears to be essential for we have repeatedly observed that when such blood agar was kept for some days, especially in the incubator, and as a result the hemoglobin had undergone cleavage, that it was no longer a useful medium. The blood was usually drawn from the carotid into sterile pipettes.^ A glass rod was passed through the cotton plug before sterilizing and by means of this the blood was readily defibrinated. The transfer of the defibrinated blood to the melted agar was made with sterile Pasteur bulb pipettes. In order to secure from the rat the largest possible yield of trypanosome blood without bacterial contamination, the best procedure is to draw it directly from the heart into a sterile pipette. ■ With reasonable care blood agar can be thus prepared without the least danger of bacterial contamination. As soon as the blood agar has solidified the tube is placed upright so as to allow as much water of condensation to collect on the bottom as possible. This fluid was then inoculated with the otherwise sterile trypanosome blood. When the inoculated tubes are to be kept at room-temperature for weeks » Compt. Rend. Soc. Biol., LIT, p. 816, 1900. ' Novy: Laboratory Work in Bacteriology, p. 460. 564 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. and months it is necessary to prevent desiccation. This can be readily done, and best by means of seaUng-wax. Rubber caps are not as useful. In the culture experiments at incubator temperature a different procedure was found necessary. The tightly rolled cotton plugs were first cut short, then charred in a flame and pushed inside of the tube. A number of these were then placed in a large desiccator, or preferably in a Novy anaerobic plate apparatus on the bottom of which was some cotton well soaked in mer- curic chloride. In this way an abundance of moisture and of oxygen was provided for, and at the same time desiccation of the tubes was prevented. By wiping down the outside of the tubes with the mercury solution practically all danger of mould contamination is done away with. It will be noted from the experiments which will be described that cultures at room-temperature retain their vitality for weeks and months. In one case, where the amount of fluid was originally large, the trypanosomes devel- oped slowly, and as a result persisted longer, being still ahve at the end of 306 days. On the other hand, in the incubator, the development of the trypano- somes proceeds more rapidly, reaches its maximum in about eight to twelve days, after which they die, which result usually takes place in about three weeks. The early death of the trypanosomes in the incubator, as compared with that of those developed in the room, is undoubtedly due to the altera- tions which take place in the medium, since the bright red of the hemoglobin soon disappears and gives way to the dark brown hematin color. Whereas in the hving rat the trypanosomes occur singly or in very small rosettes, our cultures have been characterized by the formation of enormously large rosettes. While Laveran and Mesnil, as well as Jiirgens, speak of agglu- tination rosettes containing thirty to fifty trypanosomes, those met with in the culture tube easily contain several thousands of individuals. The correct- ness of this statement wiU be readily seen when we say that we have repeatedly met with rosettes so large as to completely fill the field of a No. 7 Leitz objec- tive. In our opinion these interesting formations begin with the agglutination of a few individuals, which subsequently multiply by division, in situ, and thus enormous rosettes eventually are produced. As the rosettes increase in size the trypanosomes in the center, evidently cut off from their food-supply, begin to degenerate and give rise to round bodies. The cells on the periphery retain their motion the longest. Even- tually the round bodies degenerate into a mass of granules. IV. Cultivation at Room-Temperature. The source of the material was a white rat which had received an intra- peritoneal injection of trypanosome blood five weeks before, on April 11th, 1902. On May 16th, the rat was bled from the jugular vein and the blood was received directly in a sterile pipette. Four drops of the blood, which was very rich in WARD J. McNEAL AND FREDERICK G. NOVY. 565 the parasites, was planted in the water of condensation in a blood agar tube. The latter was ordinary nutrient agar, which was melted, cooled to 50°, and about one-third its A-olume of defibrinated rabbit blood added. After mixing, the contents were all(n\-oil to solidify in an inclined position. A small amount of condensation water accumulated at the bottom of the tube, and to this the trypanosome blood was added. Generation^ /.—An examination of the fluid was made a few hours after the inoculation. Only about t«-o or three trypanosomes were found in each field of the microscope. Se^-en days later, on May 23tl, numerous fair-sized rosettes were found. On the thirteenth day, May 29th, the rosettes were large and numerous. On the twentieth day, June 5th, the trypanosomes were very motile, and the rosettes, both large and small, were very numerous. On September 25th no motile forms were present. Round, granular forms were found which frequent attempts at cultivation, in this and other trials, have shown to be dead forms. Generation II. — One loopful of the culture in tube 1 was transferred on June 5th to the condensation water of a rat blood agar tube. An examination made the same day showed but eight trypanosomes in the cover-glass prepara- tion. Five days later, on the 10th, numerous actively motile rosettes were fovmd. On the seventeenth day, June 22d, numerous large and small rosettes, as well as single motile forms were present and the whole gave evidence of being a rich culture. On September 25th, only the granular, round bodies, referred to above, were present. Generation III. — A one per cent nutrient agar, to which was added an equal volume of defibrinated rabbit blood was used. The water of condensa- tion was inoculated on June 22d with a loopful of the growth from the second generation. No immediate examination was made. On the eighteenth day, July 10th, numerous large and small rosettes were found. On September 26th the culture contained masses of numerous small round granular bodies, to which, in many instances, were attached motile trypanosomes. Evidently these masses are the result of the degeneration of the rosette forms. A parallel culture with rat blood agar inoculated at the same time as the above showed on January 6th, nearly six months later, an abundance of large rosettes actively motile. Generation IV. — For this an agar containing the defibrinated blood of a wild rat was used. It was inoculated on the 12th of July from the preceding. 'We use the term " generation " in the same sense as employed in bacteriological work. It refers not to the division of a single cell, but to the entire growth formed in a tube prior to the next transplantation. It is to be noted that, while only one tube is described under each generation, as a matter of fact, a dozen or more tubes were inoc- ulated each time. These additional cultures showed little or no difference from the ones described. 566 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. No examination of this tube was made until September 25th (seventy-fifth day), when granular degenerating forms with attached trypanosomes were found. Evidently this culture had passed its maximum development and was undergoing degeneration. On October 27th actively motile parasites were still present, though not numerous. Generation V. — Defibrinated rabbit blood agar three weeks old was em- ployed. This was inoculated with one loopful of the fluid from the fourth generation on October 27th. Ten days later, November 7th, no trypanosomes could be detected. On the sixteenth day, November 13th, several small rosettes were found. This culture was again examined on November 20th, December 2d, and January 1st, but no marked increase in trypanosomes was noted. The agar used for this culture was decidedly alkaline, and several attempts at transplantation to alkaline blood agar gave negative results. On the other hand, when inoculated into freshly prepared, slightly acid rabbit blood agar, growth was obtained. Generation VI. — This was made on December 4th, and as just mentioned, a slightly acid agar, containing defibrinated rabbit blood was used. Seven days later, December 11th, an examination gave negative results. On the fifteenth day, December 19th, a few small rosettes and some actively motile single trypanosomes were found. On January 1st the culture showed evidence of degeneration, and on January 5th no parasites could be detected. Generation VII. — ^A freshly prepared acid agar containing defibrinated rabbit blood was inoculated on December 19th. The water of condensation was quite considerable. Examinations made on January 1st and on January 5th were negative. On January 28th (fortieth day) a few rosettes of four to twelve cells were found. On March 10th (eighty-first day) the fluid was extremely rich in trypanosomes. The preparations showed masses of trypano- somes which were actively motile and perfectly clear. No granular forms or degenerations were present. On April 22d (124th day) no organisms could be detected. Generation VIII. — This was made on March 11th from the preceding. An acid rabbit blood agar about a month old, but kept sealed, was used. On the thirteenth day, March 24th, it was very rich in rosettes, and in motile single forms. Motionless, round bodies were also present in rather large number. On the twenty-third day, April 3d, the same appearance was met with. On the forty-second day, April 22d, several actively motile rosettes were still present. Generation IX. — This was started from the preceding on March 24th. A blood agar, sUghtly acid in reaction, and about two months old (capped), was used. Very little water of condensation was present. On April 3d the exami- nation was negative. On April 22d (twenty-ninth day) fairly numerous, small, but actively motile rosettes were found. WARD J. McNEAL AND FREDERICK G. NOVY. 567 A duplicate of this generation was inoculated on April 3d. The slightly acid blood agar was but six days old. On April 22d it showed many small, actively motile rosettes. Generation A'.— This was planted on April 23d on a nutrient agar, one per cent, to which no alkali was added. Defibrinated blood from a rabbit was used as before. On the sixth day one good-sized active rosette and several single forms were found. On May 13th (twentieth day) rosettes, large and small and fairly numerous, were present. On May 17th about one-half a c.c. of bouillon was added to the Xth Gen- eration tube, and the liquid was then injected into the peritoneal cavity of a ■white rat. On May 21st about seven trypanosomes were found in each field of the blood taken from the rat. Developmental forms were present. On May 22d the blood showed about 75 parasites per field. On May 23d about 150 to 200 trypanosomes were found. The rat was evidently very sick; the respira- tion was greatly increased and fever was present. Another Xth Generation, from a different series, was inoculated into a white rat, also on May 17th. Four days later the blood showed about three trypanosomes in each field. On the fifth day there were about 40 per field. On the sixth day about 100 were present, together with developmental forms. Generation XL — This was started on May 13th from the preceding culture. The medium was similar to that used for the latter. On May 17th one rosette, consisting of about a thousand ceUs, was- found. On May 19th several small rosettes were present. It will be seen from the above that we have grown rat trypanosomes, at room-temperature, through eleven generations in the test-tube, for an entire year. There seems to be no reason why these cultures cannot be kept up indefinitely. After cultivation for a year, in vitro, the trypanosomes are as virulent, if not more so, than in the beginning. This is seen in the short period of incubation (four days) shown by the two rats inoculated with the Xth Generation. The trypanosome blood which was employed to start the next series was of interesting origin. On June 5, 1902, some of the heart blood of an infected rat was transferred to about 3 c.c. of defibrinated rabbit blood. The rat blood was very rich in the parasites, an approximate count showing about 150 trypanosomes per field of the microscope. An examination of the rabbit blood immediately after inoculation showed about five trypanosomes in each field. Five days later only about two trypanosomes were present in each field, and these were single and motile. The capped tube was kept in the dark at room-temperature. On September 25th, 112 days after inoculation, large motile rosettes were found to be present. From a single experiment like this it is not possible to assert that multiplication actually took place, since the rosette formations can be considered as mere agglutinations of the original 568 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. trypanosomes. On the following day (September 26th) 0.3 c.c. of the blood was injected into the peritoneal cavity of a white rat. A wild rat was also injected at the same time with a like amount. The latter showed the first trypanosomes in its blood — on the sixth day — while in the white rat they appeared on the seventh. The nimiber of parasites gradually increased and were single lontil the tenth day. On that day agglutination rosettes appeared in the blood of each rat and persisted for about two days, after which the rosettes disappeared, and only single, actively motile trypanosomes were to be found. This appearance of rosettes on the tenth to the twelfth day is almost a constant phenomenon when fully developed trypanosomes are used for inoculation. It will be seen from the above that trypanosomes can be kept, and perhaps grown, in a single tube for 113 days, and that such material is as infectious as fresh blood. In a similar experiment to that just given the second generation of a room culture was added to an inclined rat blood agar to which some rat blood was added so as to furnish a liquid medium. This inoculation was made on June 22d. On July 10th, an examination failed to show any trypanosomes. The tube was then set aside and again examined on September 26th, when many small rosettes, as weU as motile single individuals were found. On October 27th, the rosettes were numerous and actively motile. On January 6th, 198 days after inoculation, very large rosettes were present, and the culture was clearly very rich in the organisms. No examination was then made until April 4th, when no trypanosomes could be detected. It is evident, therefore, that trypanosomes can be kept in a suitable medium for^ over six and a half months. This experiment and similar ones show that when trypanosomes are inoculated into a relatively large volume of the medium the growth is slow and the vitality of the culture persists for a considerable length of time. On the other hand, when the material is planted into a small volume of fluid the maximum growth is reached in a shorter time, and the culture dies out in less time than when an abundance of fluid is used. The longest viabihty which we have observed was in an experiment similar to the preceding. The medium was prepared by adding to two parts of agar kept at 50°, one part of rat blood, and one part of a solution of glycocoU and sodium asparaginate. The latter solution contained one per cent of each of the amido acids. The tube was then inclined and allowed to soUdify, after which some defibrinated rabbit blood was added to supply the culture fluid proper. This tube was inoculated on July 12th with one loopful of a second generation culture. The first examination on July 27th showed a few actively motile rosettes, together with a few single organisms. On September 25th numerous small active rosettes were present. On January 7th large masses of round granular bodies with a few motile organisms on the WARD J. McNEAL AND FREDERICK G. NOVY. 569 edges were observed. On April 4th, 266 days after inoculation, the large clumps of round bodies, obviously degeneration forms, were still present, as well as a few motile forms. On April 26th (288th day) the tube showed many bunches of granular round bodies, some of which were very large (20 // in diameter). Many small rosettes of actively motile spindle forms, slightly granular, were present. On May 14th (306th day) many motile forms were still present. It is quite certain that active multiplication occurred in this case. Whether the addition of aniido acids to the medium influenced favor- ably the viabihty of the organism we are not able to say. The results obtained in other cultures where amido acids were used seemed to indicate an increased growth, but further experiments are needed along this line before a definite conclusion can be reached. V. Cultivation at Incubator Temperature. The trypanosome blood which was employed to start this series was obtained from the heart of the white rat already referred to, which received, on Septem- ber 26th, 0.3 c.c. of the fluid from a room culture (first generation) 113 days old. The blood was drawn from the heart of this rat on December 4th. It was fairly rich in trypanosomes, since sixty to seventy of the organisms could be coxmted in each field of the microscope. Generation I. — A drop of the above blood was added to a freshly prepared, sHghtly acid agar containing defibrinated rabbit blood. The inoculated tube was set aside at 35°. In this and subsequent trials, in order to prevent desicca- tion, the tubes were set in a Novy anaerobic jar, on the bottom of which was placed some cotton and a solution of mercuric chloride. The penetration of moulds was prevented by using compact cotton plugs in the tubes, which were charred in the flame before the tubes were set aside. It is advisable to moisten the labels, and to wipe the outside of the tubes with the mercury solution. With these precautions no difficulty is experienced in preventing mould con- taminations. The above culture, planted on December 4th, was examined seven days later, and was found to be exceedingly rich in trypanosomes, which were chiefly single. Some rosettes were present. On the fourteenth day rosettes predominated. On the 23d (nineteenth day), no motile organisms were pres- ent. Numerous round granular bodies, probably degenerations, were in evidence. Generation II. — A loopful of the fluid from the preceding was transferred on December 18th to a tube of freshly prepared rabbit blood acid agar. This was set aside as before at 35°. On the fifth day several small rosettes were found. On the twelfth day the culture was very rich in rosettes and single organisms. On the seventeenth day many granular round bodies and only a few motile trypanosomes were observed. 570 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. Generation III. — On December 23d a freshly prepared medium, same as that used before, was inoculated with a loopful of the liquid from the preceding culture. The first examination, made on the eighth day, showed many active rosettes, as well as single forms. On the thirteenth day it was exceedingly rich in trypanosomes. Generation IV. — This was inoculated on December 31st on the same medium as before, prepared the day before. Examined on the fifth day it was found to be extremely rich in rosettes and single forms. The same result was obtained on the eighth day. Generation V. — This was made on January 5th. A loopful of the previous culture was transferred to the freshly prepared rabbit blood acid agar. On the eighth it was found to be very rich in trypanosomes. On January 12th (the seventh day after inoculation) round granular bodies were noted. The culture when last examined, on the sixteenth, still showed active organisms. Generation VI. — The same medium (four days old) was used as for the preceding. It was inoculated on the eighth of January, and like the others, was kept at 35°. On the second day numerous small beginning rosettes were seen. On the fourth day the rosettes were more numerous and larger. On the eighth day the culture was very rich. Early on January 17th the tempera- ture of the room and hence of the incubator rose to about 40°. The cultures were exposed for several hours to this temperature. As soon as the condition was recognized the tubes were removed and examined. This particular culture showed large masses of granules, evidently the products of a rapid disintegration of the trypanosomes. A few motile organisms, however, were present. Eleven sub-cultures were made from this tube on January 10th and 12th. Some of these were examined on the sixteenth and showed good development. The rise of temperature on the seventeenth caused the death of all but two of these sub-cultures. These two showed very few motile trypanosomes. In order to save the life of the organism a large number of sub-cultures were made from the three tubes which stiU showed a few motile trypano^mes. Most of these sub-cultures were placed in the incubator at 35°, while several were set aside at room-temperature. Of those placed in the incubator, not one developed. On the other hand, the sub-cultures which were set aside at room-temperature developed and thus gave rise to the next generation. Generation VII. — ^This, as just indicated, was developed in the room from the over-heated, original sixth generation. The inoculation was made on January 18th. On January 27th many large motile rosettes were found. On February 6th the culture was very rich in rosettes. On April 6th niunerous round bodies were present but no motile trypanosomes. Generation Vila. — As mentioned above, a sub-culture from the sixth genera- tion was made on January 10th. On the sixteenth many rosettes were present. WARD J. McNEAL AND FREDERICK G. NOVY. 571 The high temperature of the seventeenth destroyed all but a few motile forms. On the eighteenth, transplants were made from this. Those placed in the incubator failed to show any growth, while one of the transplants kept at room-temperature gave generation Villa. Generation T7//.— This A\-as inoculated on January 27th and set aside at 37°. An acid blood agar, ten ilays old, was used. On February 2d, very few beginning rosettes were found. On February 6th, the tenth day, no living forms were recognized. On January 30th, the culture was transplanted to two rabbit blood agar tubes. One of these was set aside in the incubator, where it failed to develop; the other was kept at room-temperature and gave generation IX, which see. Generation Villa. — This was made on January 18th from Generation ^TIa, and was kept at room-temperature, about 25°. On the ninth day no gro\Hh could be detected. On the nineteenth day only one four-cell rosette was foimd. On April 6th the tube was found to be very rich in round bodies and in actively motile rosettes. The accidental rise in temperature is interesting in several ways. In the first place, it caused the death of most of the trypanosomes present. Those which did survive the exposure were clearly very feeble, or as we might gay, attenuated. This is seen in the failure of the incubator sub-cultures to develop. It is also seen in the slow development of the room cultures, especially Genera- tion Villa. Again, it will be noted that Generation VIII died out more rapidly than is ordinarily the case, showing that the organism had not acquired its fuU strength after the revivification at room-temperature as Generation "\TI. Furthermore, it is very important to note that this accident furnishes a proof, if indeed a proof is necessary, that we are dealing with genuine niulti- pUcation of trypanosomes, and not with mere transference of the orgar^ipms from one tube to another. The latter supposition is obviously inadmissible in view of the large number of generations which we have carried out. !fJow- ever, it is certain that the exposure to heat destroyed all but very few qf the organisms, and that the seventh and subsequent generations start from a very small number of surviving trypanosomes. That the exposure to 40° actually destroyed the trypanosomes whicl^ were present was proven by subsequent experiment. A large number of cultures of trypanosomes were present in the incubator at the time. These were known to have had more or less well-developed growths. When, as a result pf the exposure, the cultures were found to be presumably dead, eight of these tubes were taken and inoculated with material from the fourth generation gfown at room-temperature. The tubes thus reinoculated, as well as eight! other tubes, which were exposed to the heat, were set aside in the incubator. ' None of the latter, or control tubes, showed the shghtest sign of developijaent, whereas the reinoculated ones gave fair growths. While, therefore, the tem- 572 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. perature was sufficient to destroy the organisms it did not appreciably alter the nutrient qualities of the medium. Generation IX. — As indicated above, the over-heated sixth generation, when transplanted to fresh media, failed to give growths in the incubator, but did so in tubes kept in the room at about 25° (Generation VII). The latter, when transplanted and developed at 37° gave a slight growth which rapidly died out, showing that the trypanosomes were still weak from the effects of the over-heating. This fact is again demonstrated by the failure of the sub- cultures from this eighth generation to take hold in the incubator, whereas a sub-culture kept in the room gave a slow but fair growth. This culture, therefore, constitutes the ninth generation in this series. The medium used for this generation was the same as for the preceding; that is, it was a freshly prepared, slightly acid agar to which defibrinated rabbit blood was added at 50°, after which the tube was solidified in an inchned position. The inoculation was made on January 30th, and the sealed tube was kept in the dark at about 25°. On February 6th, a few actively motile rosettes and some division forms were present. On February 14th about the same condition was found. On April 6th (sixty-sixth day) large masses of rosettes were present, showing a very rich growth. Generation X. — This was made from the preceding on February 15th. The medium used for this and for the eleventh generation was a one per cent nutrient agar containing 20 c.c. of normal sodium hydrate per liter, and hence the reac- tion was distinctly alkaline. Rabbit blood was added to the agar as in previous cultures. The tube was kept at 34°. When examined on February 20th, several small rosettes were present, which were motile, but had a granular appearance. On the twelfth day the culture showed several rosettes and many actively motile single forms of varying size. On February 28th the entire fluid contents of the tube was injected into a wild rat, which, however, did not become infected. This negative result may be due to the use of the alkaline medium, since, as pointed out, the trypanosomes showed granular alterations. This condition was not found in a companion culture (Generation Xa), which was made on the usual acid agar blood medium. Moreover, the latter, as will be seen, did cause an infection in a rat. Generation Xa. — This was started on the same day and from the same source as the preceding, and was developed at the same temperature. The agar in the medium employed was prepared without the addition of alkali. When examined on the fifth day the culture showed many actively motile, small rosettes and single dividing forms. The trypanosomes on this acid medium, unlike those on the alkaline blood agar, were perfectly clear. On the twelfth day only a few very active clumps could be found. On the following day, February 28th, the culture fluid was injected into a wild rat which, examined ten days later, showed agglutination rosettes in its blood. On March 12th, WARD J. McNEAL AND FREDERICK G. NOW. 573 trypanosomes were present in the blood as single individuals, not in rosettes. When examined again, on April 7th, no parasites could be detected in the rat's blood. While, therefore, the alkaline culture failed to infect a rat, the acid one did. It should be noted, however, that the infection as judged by the number of parasites in the blood and by the duration was not as severe as when fresh trypanosome blood is used. Generation XI. — The same medium was used for this as for Generation X. The inoculation was made on February 27th, and the tube was kept at 34°. Ten days later the culture was rich in large rosettes. On the 19th day a few motile single forms and a few small rosettes were present. Generation XII. — This was made from the preceding on March 9th. The medium employed was the same as that used for Generation Xa; the addition of rabbit blood, however, was not made until two days before the medium was used. The first examination, made on March 13th, showed several small motile rosettes and some small, clear, single trypanosomes. On March 17th small rosettes, as well as single forms, were numerous. Most of the organisms at this date were granular and non-motile. None were present on March 30th. Generation XIII. — This was made on March 17th on the same acid agar as before, to which rabbit blood was added three days previous to use. The culture was developed at 34° till March 28th, after which it was kept at 37°. Examined on the sixth day, the culture was exceedingly rich in very active rosettes. Some round bodies were also present. On the thirteenth day active "single forms and granular round bodies were found. Generation XIV. — ^This was inoculated on March 23d. The one per cent agar used contained 5 c.c. of normal alkali per liter. On the fourth day, at 34°, the culture was very rich in rosettes. A few round bodies which appear to indicate maximum development were also present. The tube was now placed at 37°. On the tenth day very few small rosettes in process of degen- eration were found. On the fifteenth day only a few bunches of round bodies and no motile forms could be detected. Generation XV. — For this the same medium was used as for the preceding. The inoculation was made on March 27th and the culture developed at 37°. On the thirtieth several fair-sized rosettes were seen. On the following day the culture was quite rich in rosettes, which were composed of very actively motile individuals. On April 7th a few bunches of round bodies with granular motile trypanosomes were noted. Generation XVI. — This was inoculated on March 31st, on the same medium as used before. On the 3d day, at 37°, a few actively motile rosettes and some motile single forms were present. On the sixth day the rosettes were of moderate size. Very few motile forms and some round bodies were found on the ninth day. Generation XVII. — The agar used contained 20 c.c. of normal alkali per 574 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. liter. The culture was started on April 6th. Several small, actively motile rosettes were found on the third day. On the eleventh day a moderate number of motile single forms and a few bunches of round bodies were seen. Generation XVIII. — This was made on April 10th, and the medium used was the same as that employed for the preceding culture. When examined on the seventh day it was very rich in actively motile rosettes and in single forms. Round bodies were also present. Generation XIX. — ^An agar containing 15 c.c. of normal alkali was used. The inoculation was made on April 18th, and the culture was developed at 37°. On the third day a few actively motile rosettes were found. The organ- isms were clear in appearance. On the seventh day very numerous rosettes, consisting of actively motile and granular cells were observed. No round bodies. Generation XX. — This was inoculated on April 22d to the same medium as the preceding (37°). Three days later many active rosettes and long, clear trypanosomes were found. On the 5th day the culture was very rich in rosettes. The individuals, as a rule, were perfectly clear; a few only showing granules within the cell. After transplanting this culture to several tubes, all of the remaining fluid (0.25 c.c.) was injected intraperitoneally into a white rat (A) on April 27th. Rat A, when examined on May 1st, the fourth day after inoculation, showed no parasites in its bloofl. On May 3d, only one trypanosome was found in the preparation. Examinations made on May 4th, 6th, and 8th were negative. Clearly no infection took place. This result is markedly at variance with those obtained in the case of Rats B and C. Generation XXa. — The sixteenth generation was transplanted on April 6th to blood agar to which no alkali was added. In three days this seventeenth generation showed many small rosettes. On April 17th a few bunches of round bodies and some motile forms were present. From this culture on April 10th a transplant was made to the same medium freshly prepared. This eighteenth generation showed on April 17th many very active rosettes and a few round bodies. A nineteenth generation was planted on the eighteenth to the same medium, freshly prepared. In three days a few large rosettes developed. On the seventh day single forms were present, but not numerous. Generation XXa was made from this tube on April 22d, and when examined three days later, no trypanosomes could be found. On the fifth day no rosettes, but few actively motile single parasites, were observed. On April 27th all of the fluid (0.6 c.c.) was injected into the peritoneal cavity of a white rat (B). Rat B was examined on May 1st with negative result. On the sixth day after inoculation, about two trypanosomes were found in each field of the microscope. Developmental forms were present. On the eighth day the blood WARD J. McNEAL AND FREDERICK G. NOVY. 575 showed about thirty tiypanosomcs in each field besides some agglutination forms. On May 8th, the eleventh day, and on May 17th, the twentieth day, the blood contained about fifteen parasites in each field. The rat was found dead on the morning of May 18th (twenty-first day) . This rat, it will be noted, received an acid agar culture of the same age as rat A. While the former failed to infect, this one clearly caused a slight infection which, however, resulted fatally. Cultures made from the heart-blood failed to show the presence of bacteria. Generation XXh. — This sub-culture was derived from the original eighteenth generation, which was transplanted on April 18th to blood agar, the latter containing 20 c.c. alkah per Hter. On the third day a few rosettes and some single forms, but no round bodies, were found in Generation XlXa. On the seventh day many large rosettes which were very active were present. The cells were shghtly granular. Generation XXb was made from this on April 22d to blood agar of hke alkahnity. On the twenty-fifth a small number of single forms was found. On the twenty-seventh very many rosettes were present. The trypanosomes were mostly clear, though some showed granules. On the same day the entire fluid contents (0.7 c.c.) was injected intraperitoneally into a white rat (C). Rat C was examined on May 1st with negative result. On May 3d the blood showed about eight organisms per field. Developmental forms were present. On May 4th there were about twenty trypanosomes in each field and a few rosettes (developmental?) were also present. On the next day the number of trypanosomes was enormously increased, about two hundred being estimated in each field. On May 8th, the eleventh day, the number of parasites was apparently increased, and developmental forms were very common. The rat on this day was quite sick; it offered no resistance when handled and its respiration was about 150 per minute. The same condition persisted on the 9th, the respiration being increased to 180 per minute. The rat probably died shortly after 1 p. m., on May 10th, thirteenth day, that being the last time it was seen until it was found dead at 8 a. m. the following morning. Owing to the prevaiUng high temperature, decomposition was already advanced, and hence, of course, bacteria were found in the blood and organs. Nevertheless, it seems quite certain that the enormous development of trypanosomes was responsible for the death of the animal. Autopsy showed an enlarged spleen, several large cheesy cysts in the lungs, and a large worm in the intestines. It may be that the fatal outcome in this case was due to a lowered resistance of the animal. Generation XXI. — This was started on April 27th. The usual rabbit blood agar was used, except that the agar was made without the addition of alkah. This, and the following cultures, were developed at 36°. When examined on May 1st the culture showed about eight rosettes on the cover-glass. On 576 ON THE CULTIVATION OF TRYPANOSOMA LEWISI. May 7th, the tenth day, many active rosettes were present. On May 11th the rosettes were less numerous and were very granular. Generation XXIa. — ^The medium used for this culture was the same as that employed for the preceding, except that the agar contained three per cent of pepton, while in all cultures previous to this it contained only one per cent pepton. The two media were inoculated at the same time, from the same source, and were kept side by side at the same temperature. On May 1st several small rosettes, were found. On May 3d the rosettes were fairly numer- ous and actively motile. On May 7th very few single forms were present, and on May 11th no trypanosomes could be found. Generation XXII. — A blood agar, the latter prepared without the addition of alkali, was inoculated from No. XXI on May 3d. On May 7th several small motile rosettes were found. On May 11th the tube showed many granular roiuid bodies besides several small rosettes. On May 16th no motile forms and only bunches of granules could be detected. Generation XXIIa. — This was planted at the same time as the preceding, on a freshly prepared blood agar, the latter containing three per cent of pepton, as in the case of culture XXIa, from which, moreover, the inoculation was made. On May 7th, the fourth day, many small, extremely active rosettes were found. On May 11th, the eighth day, the tube showed many actively motile single forms, also rosettes and granular round bodies. On May 16th, the thirteenth day, onty two motile single forms, besides masses of granules, could be found. Generation XXIII.— This was started on May 12th from No. XXII. The agar used had 10 c. c. of normal alkali per liter. On May 16th the tube was examined with negative results, whereas the companion culture No. XXIIIa, on the same day, showed a good growth. On May 18th a few bunches of round bodies and very few trypanosomes were found. Generation XXIIIa. — ^The medium used was the same as that employed for No. XXIII, except that three per cent pepton was added to the agar. The inoculation was made on May 12th from No. XXIIa. On May 16th many small rosettes, consisting of twenty to one hundred actively motile cells, were present. On May 18th this culture showed several small actively motile rosettes. On May 23d only a few motile forms were present. Generation XXIV. — This was made on May 18th. On May 23d several small rosettes and motile single forms were present. It would seem as if the addition of three per cent of pepton to the agar, in the case of Nos. XXIa, XXIIa, and XXIIIa, favored the growth of the trypanosomes as compared with the corresponding three cultures in which the ordinary one per cent agar was used. The increased and rapid growth may, however, be only apparent, owing to the fact that in making the trans- plantations, more trypanosomes were carried over in the former than in the WARD J. McNEAL AND FREDERICK G. NOVY. 577 latter series. Further trials will be made to ascertain definitely whether or not pepton is favorable to the growth of the trypanosomes. It is evident from the data given that rat trypanosomes can be cultivated at incubator temperature. In the interval between December 4, 1902, and May 23, 1903 (170 days), we have been able to develop twenty-four generations. CONCLUSIONS. It will be seen from the foregoing that the artificial cultivation of rat trypanosomes is an accomplished fact. For the first time it has been possible to obtain strictly pure cultures of a pathogenic animal parasite. The demon- stration of the causal relationship of this parasite to disease has been effected. The cultures at room-temperature, while slower, are surer than those which develop in the incubator. Moreover, it is possible to secure growths at room- temperature when inoculation of the same material fails to infect animals. The cultures which are .developed in the incubator tend to die out in from two to three weeks, whereas those grown in the room may retain their vitality for months. In one case the trypanosomes were alive after 306 days. The virulence of the trypanosomes, when cultivated at room-temperature, and under the conditions employed by us, is not decreased, as is seen from the following tests: Thus, an initial culture kept at room-temperature for 113 days was found to be virulent. Again, an eight-day-old fourth generation, of a series which we have not described, was found to infect a rat. This inoculation was made on the 144th day after the series was started. Further- more, the tenth generation, of a series which was under cultivation for an entire year, was hkewise virulent. The few experiments which have been made with cultures developed in the incubator indicate the possibility of altering the virulence of these organ- isms. The reaction of the meditmi, as well as the temperature employed, are important factors in this regard. This question of the alteration of the viru- lence of the trypanosomes will be the subject of future study. It is probable that the methods which we have employed are appUcable to other trypanosomes and that it may be possible in this way to demonstrate their relation to disease, and to secure attenuated varieties of these organisms which may be of use in protective inoculations. STUDIES ON NEUEOGLIA TISSUE. G. CARL HUBER, M.D., Junior Professor of Anatomy and Director of the Histological Laboratory, University of Michigan. The Neuroglia Cells and Neuroglia Fibers of Vertebrates. It is my purpose to present here the results of observations which have extended over a period of several years, and which deal with the structure of the neuroglia tissue of types from several classes of vertebrates. At the time when this investigation was undertaken, it seemed desiiable to extend the more recent studies of neuroglia tissue, made by the aid of special differential staining methods and confined in the main to an investigation of the neuroglia tissue of man, so as to include observations on the neuroglia tissue of types from several classes of vertebrates. A confirmation, or otherwise, of the results obtained by observers who have investigated the neuroglia tissue of man by means of methods of differential staining, by observations made on the neuroglia tissue of vertebrates other than man, studied by similar meth- ods, seemed of sufficient importance to warrant such an undertaking, and it was hoped might tend to harmonize the various views concerning the structure of this tissue now current. In a former brief communication, the writer drew attention to the fact that the majority of the differential or selective staining methods of neuroglia tissue, to be more specific, neuroglia fibers, were selective and applicable to human neuroglia tissue only, and then only when the tissues employed were fresh. The difficulty of obtaining fresh human tissue, more especially of the central nervous system, must be apparent. It seemed, therefore, of importance to establish the fact that certain of these selective staining methods gave specific reactions with neuroglia tissue taken from vertebrates other than man. Animal tissue may, of course, be readily ob- tained; opportunity is thus given for a wider study of this tissue. It may be parenthetically stated that the ultimate aim of these investigations was to gain data, which must of necessity be at hand, in order that a further re- search might be undertaken, in which it was the purpose to deal with the be- havior of the neuroglia tissue in areas of degeneration of the central nervous system, resulting from mechanical interference — mainly by section. This work is now in progress and will form the subject of a further communication. The literature dealing with the neuroglia tissue is extensive. Anything like a complete review of it would extend this article beyond the limits set for it. The necessity for even a comparatively complete review of it is obvi- 578 G. CARL HUBBR. 579 ated by the fact that a number of writers, who have in recent years recorded their observations on neurogUa tissue, have given comprehensive reviews of the articles treating of the neuroglia tissue which antedate their own. We may here especially mention Weigert, who, in his classic monograph on neu- roglia, discusses, in a special chapter, fully and critically, the earlier literature bearing on the subject. It may be of interest, however, and add to a clearer understanding of the matter to be presented, if we sketch hastily certain epochs or periods in the development of our knowledge of the neuroglia tissue. Rudolph Virchow was the first to recognize the existence of a substance or tissue in the central nervous system, which differed from ordinary connective tissue, and which he designated as "Nervenkitt" and later as "Neuroglia." Mrchow's description of this tissue, found in a number of brief notices appear- ing at intervals of several years, is somewhat difficult to interpret at the pres- ent time. He undoubtedly recognized neuroglia fibers and nuclei, but the description of these elements is subordinated in his accounts, and emphasis is given to a soft, amorphous, or at times granular ground substance, the " Nervenkitt. " That ^^irchow was the first to call attention to the " Neuroglia " as a distinctive tissue there seems no question. I cannot, however, agree mth Babes, who, in discussing Virchow's contribution to our knowledge of the structure of nem-oglia tissue, states " dass die neuere Forschung nichts von den grundlegenden Angaben Virchow's erschiittern konnte. " V. KoUiker, and after him more especially Deiters, called attention to the cellular elements of the neviroglia; the latter especially, by means of isolation methods, was able to observe branched cells in the supporting tissue of the central nervous system. In recognition of which facts these cellular elements of the neiiroglia were by earlier writers now and then designated as " Deiters' cells." Golgi described the neuroglia cells much more fully and brought out more clearly than his predecessors the fact that the supporting tissue of the central nervous system was composed of cells and cell-processes. His observations were made on preparations, stained or unstained, gained by isolation methods, and on sections. Both v. Kolliker and v. Lenhossek call attention to the importance of this contribution; the latter expresses himself as follows regarding it: "The portrayal of these elements, as also their pic- torial reproduction, is so excellent .... that they should be designated as 'Golgi cells.' " Jastrowitz, who also recognized the cellular character of the neurogUa, suggests the name of spider-cells (Spinnenzellen). Boll speaks of them as brush-cells (Pinselzellen). In this connection we may mention also Ranvier's contributions to our knowledge of the neuroglia, which, as appears to the writer, have received due consideration only by Weigert, and, as will be seen later, may be regarded as more in accord with my own views of the structure of the neurogUa than that of observers who preceded Ranvier and the majority of those who have written on this subject since the publication of his obser- 580 STUDIES ON NEUROGLIA TISSUE. vations. Ranvier's observations — ^I shall follow here what may be regarded as his latest contribution on this subject — were made on tissues taken from the spinal cord of the ox and the dog and treated by Miiller's fluid; sections were then stained in picrocarmin and mounted in glycerin under a thick cov- er-glass, which was at first pressed down, then repeatedly elevated and low- ered onto the section. In this way a certain amount of " teasing " and an isolation of the elements was obtained. In specimens thus prepared were made the following observations : " The elements of the neuroglia appear as cells with centrally placed nucleus, granular protoplasm, and numerous processes; the processes do not present ' natural ends, ' but are either cut or torn. Care- ful observations showed that the fibers (processes of the cells) do not arise from the cells, but merely pass through them; they pass by the side of the nuclei and are sunken into the surrounding protoplasm. When the fibers leave the cells they are accompanied for a distance by protoplasmic branches, which often unite two fibers." Ranvier further gives several methods by means of which sections of the spinal cord may be so treated as to bring out neuroglia fibers, the cells or the cell nuclei. The citations here given belong to what v. Lenhoss^k has termed the pre- Golgi period (vorgolgische Periode) in the development of our knowledge of the structure of the neuroglia tissue. New incentive to the study of the neuroglia tissue in the adult, as well as to the study of its development, was given by the discovery of the fact that, by means of the chrome-silver method (Golgi method), this tissue could be stained with as much or even greater readiness than the nervous elements, and numerous investigators in a com- paratively short space of time presented observations made on the central nervous system of embryos and ad\ilt animals treated by the chrome-silver method; in many of these investigations, the neuroglia tissue was the subject of special consideration, while in others, it was incidentally studied with the other cellular elements. If, from this portion of the neuroglia literature, I have selected for special consideration the contributions made by v. Len- hossek, I did so because he expresses very well the general consensus of opinion of the investigators who have studied neuroglia tissue by means of the chrome- silver method; primarily, however, because he has introduced a nomenclature now widely used. For all cells of the central nervous system whose function is supportive he suggests the general term spongiocytes (Spongiocyten), and for the form of these cells as found in the higher vertebrates, the term astro- cytes (Astrocyten) ; the term "Neuroglia" is retained as a "collective term" for all these elements. V. Lenhossek describes the astrocytes as relatively small elements — especially in the spinal cord of man, somewhat larger in other mammals and vertebrates — ^with nucleus and narrow border or zone of protoplasm, from which arise numerous processes which pass out in all direc- tions, the cell body often appearing as a nodal point from which the processes G. CARL HUBBR. 581 radiate. Two general types of astroc>'tcs are recognized — long-rayed and short- rayed astrocytes; the former are found in both the white and the gray matter of the central nervous system, the latter iu the gray matter only. Astrocytes with characteristic grouping of the processes were recognized; such in which the processes arise from one or from two opposite poles were especially considered. The chrome-silver method was for many years the only method at our disposal, by means of which the cellular elements of the neuroglia were brought to hght ^^ith clearness, and to this method we are indebted for the results obtained by Golgi, v. Lenhossfek, v. Kolhker, Ramon yCajal, Retzius, Salay Pons, Van Gehuchten, Nansen, Andriezcn, Eurich, and others. Collectively their investigations embrace a study of the neurogUa tissue, or the supporting tissue of the central nervous system, its development, its differentiation, and its structure and arrangement in full development as found in the different classes of vertebrates. I shall, in discussing my own results, find occasion to refer more particularly to the results of observations made on the neuroglia by means of the chrome-silver method, and shall discuss their bearing on the va- rious ^'iews of the structure of this tissue and shall content myself at this time with giAdng a brief account of a summary of this literature as found in one of the contributions of v. Lenhoss^k, which, although presented as a quotation, is rather a statement of the thought expressed than a literal quotation. After considering the structure of the supporting tissue of the spinal cord of verte- brates other than mammals, he expresses himself as follows: "The .sup- porting tissue of the spinal cord of all vertebrates consists of cells — of ependy- mal cells and more or less branched glia cells — all of which are of ectodermal origin and develop in the spinal cord itself. Certain inherent differences are recognized in these cells in different animals, also in their arrangement, so long as in comparing them they are considered only in the condition as found in the fully developed cord; these differences are, however, readily ex- plained as soon as they are considered in their development, for then it becomes apparent that the simpler types of the supporting tissue of the spinal cord, as found in the lower vertebrates, correspond to the transitory developmental stages of this supporting tissue as found in the higher and highest vertebrates. " A study of the " Golgi Uterature, " so far as this pertains to neuroglia tissue, warrants the statement that observers who have used the chrome-silver meth- od in the investigation of neuroglia tissue have reached the conclusion that this tissue is composed of cellular elements, possessing processes varying in nrmiber, size, and arrangement — neuroglia cells and neuroglia fibers, the latter arising from and going over and forming a part of the protoplasm of the neu- roglia cells. As apparent exceptions to this general statement, we may men- tion the results obtained by Andriezen and Reinke. Andriezen recognizes in the brain cortex two varieties of neuroglia cells. One of these varieties he designates as the "neurogha fiber cell. " He was able to make out in some of 582 STUDIES ON NEUROGLIA TISSUE. his preparations the fact that certain of the neuroglia fibers are not processes of cells, but " pass right through the cell body. " These are the astrocytes with long processes of other writers. The other variety of neuroglia cell, mentioned by him in his descriptions under the name of protoplasmic glia cell, is said to be of mesoblastic origin. These cells possess short, stout, and very shaggy branches, or processes, which vary greatly in size. Reinke modified the chrome- silver method as follows: The tissue was treated with chrome-silver in the usual way and was then dehydrated and embedded in paraffin and sectioned. The sections were fixed with albumin fixative, stained with Heidenhain's iron-lack hematoxylin and counterstained with eosin. In preparations of the white matter of the spinal cord thus prepared he made observations sum- marized by him as follows: "The neuroglia tissue consists of cells and fibrils. The cells possess numerous processes, some of which are branched and which run in part transversely and in part obliquely; the majority of them, how- ever, run vertically, that is, parallel to the nerve fibers. These processes are well stained by the chrome-silver method. The fihrils differ morphologically, physically, and chemically from the cell processes. They are, however, developed from the protoplasm of the cells and lie partly in and partly on the protoplasm and have a direction which in the main is opposite to that of the cell processes. For the most part these fibrils, the length of which is unknown, are emancipated from the cell-bodies. The fibrils differ in thickness and prob- ably do not anastomose. These are the fibrils which are so clearly brought out by the Weigert method. " Reinke interprets his resvilts as harmonizing the contradictory views expressed concerning the structure of the neuroglia tissue. I shall consider his results more fully later. The chrome-silver method is in its essentials a method which colors by precipitation; cell protoplasm and processes are in no sense differentiated, but give exactly the same color reaction; furthermore, the nuclei of the cells are not brought to view and "a method which colors by precipitation is a ■priori incapable of giving us the information which we require; it must of necessity be confusing in its pictures of structural detail, " as has been cor- rectly stated by Taylor. This being the case, we may readily appreciate the desire of certain observers to discover a method by means of which neuroglia tissue might be stained differentially. Selective staining methods, in the sense that certain tissue groups or tissue elements are given a characteristic color, while other tissue groups or elements remain unstained or are given another color, are being more and more regarded as necessary procedures in the investigation of normal and pathologic tissues. They are, to quote Wei- gert, "the keys which unlock the doors closing the passages leading to scien- tific treasures. " In the introduction to his large monograph on the structure of normal neurogha tissue of man, from which the above quotation was taken, Weigert reviews the difficulties experienced by him in developing a selective G. CARL HUBER. 583 staining method for neuroglia tissue and thc>n publishes in full the method evolved, also his observations on the structure of the normal neuroglia and its topographic distribution in the central nervous system of man. In prepara- tions of the central nervous system stained after Weigert's selective neuroglia stain are seen numerous blue fibers. Besides these the nuclei of all cells are colored blue, and now and then the rod blood colls in the blood-vessels. Among the nuclei stained there are such which from their position— white matter of the spinal cord— can be regarded only as the nuclei of neuroglia cells; of these Weigert recognizes two A-arieties — large vesicular nuclei with granular chro- matin and smaller ones in which the chromatin presents itself as a deeply staining, homogeneous mass; transition forms between the two varieties were also recognized by him. The blue-colored fibers are the neurogha fibers. They present a more or less regular course, straight or nearly straight or with bold curves, and present smooth borders without varicosities and do not end in conical enlargements. They vary in thickness and probably do not branch and anastomose. Weigert's statements as to the relations of the nuclei and fibers are as follows: The fibers course in the immediate vicinity of the large vesicular nuclei, passing them a little to one side, or after approach- ing them bending more or less sharply to pass in some other direction; others pass over or under the nuclei and still others stop abruptly in the neighbor- hood of these nuclei. The neuroglia fibers never touch the nuclei, but are always separated from them by a narrow space ("kleinen Zwischenraum"), which one may regard as filled by the protoplasm of the neuroglia cells, not stained by this method. (In a footnote, Weigert states that by means of other methods, for instance neutral carmin, it is possible to bring to view the protoplasm of neuroglia cells in preparations stained by his method. In his own work, however, as also in the writings of others who have used this method, the protoplasm of the neuroglia cells is almost entirely disregarded, a point which should here be taken into consideration, and to which I shall later draw special attention.) Very characteristic are the figures seen when numerous fibers are associated with one of the vesicular nuclei, figures which, with a little imagination, enable one to recognize the astrocytes seen in chrome- silver preparations. The smaller nuclei with deeply staining chromatin very seldom show any special relation to the neuroglia fibers, perhaps never. Wei- gert smnmarizes his results, so far as they pertain to the relation of the neurog- lia fibers to the neuroglia nuclei (neuroglia cells) as follows : " The neuroglia fibers, which have been hitherto regarded as the processes of Deiters' cells, differ chemically from the protoplasm of these cells. This difference in the chemical constitution of the "cell-processes" is apparent in the immediate vicinity of the cell-nucleus, as well as at some distance from it. The majority of the so-called cell-processes are not cell-processes, since two such apparent processes form a continuous fiber, which is in no way interrupted by the cell 584 STUDIES ON NEUROGLIA TISSUE. body, as would be the case were they true processes arising individually from the cell body. In a word, there is no question here of cell processes or cell extensions, but of fibers which are fully differentiated from the protoplasm. " Weigert's publication of his selective staining method for neuroglia tissue and of the observations made by him with this method, giving results so at variance with the results obtained by former investigators, gave a new impetus to the study of neuroglia tissue. Reference should, however, here be made to the fact that Mallory published a method for staining neuroglia tissue differentially about the time that Weigert gave his method to the public. Mallory's method is essentially a modification of Weigert's method for stain- ing fibrin, applied to tissues properly hardened and mordanted. Certain of the figures illustrating Mallory's brief account (Figs. 1 to 5, Plate III) portray a relation of neuroglia fibers and nemroglia cell nuclei identical to that shown in Weigert's figures; however, in Figs. 6 and 7 of the same plate, the cell pro- toplasm is also stained and in the text occurs the statement : " In the majori- ty of the ' glia cells, ' the fibers appear merely to touch and not to end in the protoplasm of the cells, for they are to be traced in two directions. " The results obtained by Weigert were confirmed by Pollack, who, in a short note, called attention to the importance of the method, and were further corroborated by Aguerre, who added details concerning the shape, size, and structure of the neuroglia cell nuclei ; and again by Krause and Aguerre, who described at some length the distribution of the neuroglia in the human spinal cord; and further by Krause, who, in a monograph in which he discussed the neurogha tissue in the spinal cord of apes, again subscribed to the general conclusion reached by Weigert. A further discussion of the results obtained by Krause may be dispensed with at present; I wish, however, to call attention to the following observation made by him : In the spinal cord of one of the species of apes (Ateles) studied by him, he found that it was very easy to stain, besides the neuroglia fibers and nuclei, also the protoplasm of the neuroglia cells, especially of those lying in the neighborhood of the blood-vessels. In such preparations the nuclei of the cells are stained deep blue, the protoplasm light blue, and the neuroglia fibers dark blue and "it may be seen that the neuroglia fibers run in the peripheral layer of the protoplasm." Such neuroglia cells are often star-shaped, the protoplasm accompanying the respective fibers for a distance and being closely applied to the fibers and gradually disappear- ing. In over-stained preparations, the fibers are not differentiated from the protoplasm and appear as processes. In Fig. 8, Plate I, of his article are shown two neuroglia cells in which nucleus, fibers, and protoplasm are stained. Erik Miiller studied the supporting tissue of the central nervous system of amphioxus, myxine, several selachians, teleosts, reptiles, amphibia, and mam- mals, using a method similar to that used by him in his studies of secretory capillaries. The tissues were fixed as for the chrome-silver method and were G. CARL HUBER. 585 then stained in iron-lack hematoxylin. This method gives satisfactory results only with tissue derived from amphioxus, myxine, sharks, and teleosts; in a somewhat modified form only an incomplete staining of the neuroglia tissue of other vertebrates. As a result of his investigations, Miiller reaches the conclusion that the neuroglia tissue of the lower vertebrates "is composed of cells and fibers, which, however, stand in such relation to each other, that all the fibers are to be regarded as processes of cells." The fibers differ mor- phologically and physico-chemically from the cell protoplasm, as is shown by the staining reaction ; they are however cell-processes, and have no inde- pendent existence. He feels warranted, from observations made, in concluding that in the higher vertebrates the neuroglia tissue presents the same structure. Essentiallj' the same method was used by Joseph in his investigations of the supporting tissue (nem'oglia) of the nervous system of annelids. He records in detail the appearances presented by the elements of the neuroglia in the several species of invertebrates studied, and discusses the bearings of the observations made by him on the question of the structure of the neuroglia in general. It will suffice here to give in substance some of his general con- clusions. Joseph finds in annelids and in other invertebrates that the neu- roglia is composed of neuroglia fibers, which show striking similarity to the neuroglia fibers of vertebrates and which are in permanent relation to the neurogha cells from which they are developed. This relation of the neuroglia fibers to the cells is such that they occupy a position which is on or in the peripheral layers of the protoplasm of the neuroglia cells; they do not pass over into the protoplasm of the neuroglia cells, but retain their characteristic form and structure and pass from one process of the cell to another, without losing their identity. These facts are clearly brought out by Joseph in his detailed description and also in the numerous figures which accompany his article. From a study of the neurogha of types from the several vertebrate classes, which he does not, however, regard as complete, he concludes that a similar relation of neuroglia fibers to cells exists in the vertebrates. Yamagiwa modified Stroebe's differential axis-cylinder stain by preceding the staining with anilin-blue by one with saturated alcohohc solution of eosin and finds that thus used this method may be regarded as a differential neu- rogha stain. Results obtained with this method, especially in the study of a ghoma, lead him to say that. he regards the neuroglia fibers as differentiated intercellular structures, which are, however, not in all instances completely separated from the nem-ogha cells. Yamagiwa 's figures, though rather crude, show this relation of neuroglia fibers and protoplasm of neuroglia cells. As it is my purpose to discuss only the structure of normal neurogha tissue, I shall not here consider the extensive literature dealing with the structui-e of neurogha tissue in the pathologic states of the central nervous system and in gliomata. Mention may parenthetically be made of the observations recorded 586 STUDIES ON NEUROGLIA TISSUE. by Taylor, who worked with the Mallory method, of Storch and Bonome, who used the Weigert method; these, as far as pertains to the structure of neuroglia tissue, confirm in the main the views expressed by Weigert and Mallory, and finally of Stroebe, who investigated the origin and structure of ghomata of the brain, using Mallory 's phosphotungstic-hematoxylin method; the latter describes and figures neuroglia cells and processes. Even a cursory study of the literature here presented enables the reader to form the opinion that the various views expressed concerning the structure of neuroglia tissue are dependent on the method or methods used by investi- gators in their study of this tissue. The whole question is really a matter of technique*. The following brief summary of the views expressed concerning the structure of neuroglia tissue may serve to emphasize this fact : 1. The earlier observers, who recognized nexiroglia cells and fibers, describe the fibers as processes of the cells; we may especially mention the eariier work of Golgi. On the other hand, Ranvier described neioroglia cells and neuroglia fibers not processes of cells, though in contact with them. 2. Investigators who used the chrome-silver method in their study of the neuroglia, in its development and fully formed state, find in vertebrates neu- roglia cells (and ependymal cells) with the neuroglia fibers as processes of these cells. (Andriezen finds certain neuroglia fibers which pass through the cells, ' The more important of the methods which may be used in the study of neuroglia tissue are the foUomng: The Chrome-Silver Method. — This method, in its various modifications, is so well known that it need not receive further consideration. Reinke's modification of it is referred to in the text. Weigert' s Selective Neuroglia Fiber Staining Method. — Fix tissues in ten per cent for- malin four days; place small pieces in chrome-alum solution (chrome-alum, 5 g., distilled water 200 c.c, acetic acid [C.P.] 10 c.c, neutral acetate of copper, 10 g.), six to eight days in warm oven at 38° C. ; wash in distilled water, dehydrate in graded alcohol, and embed in celloidin; bring sections into distilled water and transfer to 0.33 per cent aqueous solu- tion of permanganate of potassium, ten minutes, rinse in distilled water and place sec- tions in chromogen solution (chromogen 5 g., distilled water 100 c.c, concentrated formic acid 5 c.c. ; when dissolved filter and add to 90 c.c. of filtrate 10 c.c. of a ten per cent aqueous solution of sodium sulphite) twelve hours, and in a second chromogen solution (chromogen 5 g., distilled water 100 c.c.) two to six days ; rinse sections in distilled water and stain for a few minutes in methyl-violet-oxalic acid solution (methyl-violet 2 g., seventy per cent alcohol 100 c.c, heat and decant after cooling and add 5 c.c. of a five per cent aqueous solution of oxalic acid to each 100 c.c. of the solution) ; rinse in normal salt solution; trans- fer sections to an iodo-iodide of potassium solution (iodin 4 g., distilled water 100 c.c, iodide of potassium 6 g.) for five to ten seconds; wash thoroughly in distilled water; dry sections by placing them on a slide and pressing filter-paper down on them; differentiate in aniliu-xylol mixture (equal parts of anilin oil and xylol) ; rinse in xylol and mount in xylol-balsam. All nuclei and neuroglia fibers stain blue. Mallory's Selective Neuroglia Fiber Staining Methods. — Fix tissues in ten per cent formalin four days; place in saturated aqueous solution of picric acid four days; place in G. CARL HUBER. 587 and Reinke finds neuroglia fibers and fibers which may be regarded as pro- cesses of cells.) 3. Observers who employed the selective neuroglia fiber staining methods suggested by Weigert and Mallory, describe neuroglia fibers which differ mor- phologically and in their chemical constitution from the protoplasm of the neuroglia cells, fibers which are in no sense processes of the cells and are not interrupted by the protoplasm of the cells; they are intercellular structures. The majority of investigators who have used these methods disregard the protoplasm of the neuroglia cells which is not stained; Mallory and Krause con- sider it only incidentally. 4. Erik Miiller, who stained the neuroglia tissue with iron-lack-hematox- ylin, describes the neuroglia fibers as processes of the neuroglia cells, which differ however morphologically and in their physico-chemical properties from the protoplasm of the neuroglia cells. 5. Joseph, who also used iron-lack-hematoxylin as a stain in his studies of the neurogUa of invertebrates, describes neuroglia cells and neuroglia fibers, the latter differing morphologically and chemically from the protoplasm of the neurogha cells, which are branched structures. The fibers are, however, in permanent relation with the protoplasm of these cells, being found on or in the peripheral layer of the protoplasm. I have grouped in Fig. 1 a number of typic figures, selected from various five per cent aqueous solution of ammonium bichromate four to six days in warm oven at 38° C; dehydrate and embed in celloidin; sections may be stained in Weigert's fibrin stain and differentiated with equal parts of anilin oil and xylol or they maybe treated as follows: Place sections in 0.5 per cent aqueous solution of permanganate of potassium twenty minutes ; wash in distilled water one to three minutes ; place in one per cent aqueous solution of oxalic acid thirty minute.?; wash in distilled water; stain in phosphotungstic- acid-hematoxylin solution (hematoxylin 1 g., distilled water 80 c.c, ten per cent aqueous solution of phosphotungstic acid [Merk], 20 c.c, peroxide of hydrogen [U.S. P.], 2 c.c.) for twelve to twenty-four hours; rinse in distilled water and place for five to twenty minutes in an alcoholic solution of ferric chloride (ferric chloride 30 g., thirty per cent alcohol 100 c.c.) ; rinse in distilled water and dehydrate quickly, clear in oil of bergamot, and mount in xylol-balsam. Erik Mailer's Method. — As has been stated, this observer hardened the tissues as for the chrome-silver method and stained the sections in iron-lack-hematoxylin (Heidenhain) . Joseph used essentially the same method. Yamagiwa's Method. — The tissues are hardened in Miiller' s fluid and then in absolute alcohol and embedded in celloidin ; stain sections for twelve hours in an alcoholic solution of eosin; stain for six hours in a saturated aqueous solution of water-soluble anilin-blue; wash in one per cent alcoholic solution of potassium hydrate until the color of the sections is a reddish brown; wash in water until sections are pale blue; wash in weak alcoholic solution until the sections have a reddish tinge; clear in origanum oil and mount in balsam. Benda's Selective Neuroglia Staining Method. — Benda has for some years concerned himself with perfecting selective staining methods for differentiating certain constituents 588 STUDIES ON NEUROGLIA TISSUE. contributions on neuroglia tissue, which present in a pictorial way the facts above given. Weigert and Mallory specifically state that the neuroglia staining methods suggested by them are applicable only for the staining of human neuroglia tissue and only in case the tissues are relatively fresh. The reason for this is difficult to give. It may be due tO' some slight difference in the physical properties or chemical composition of the neuroglia fibers of man and other vertebrates. I do not doubt that the neuroglia fibers of vertebrates other than man take the stain when treated after Weigert's or Mallory's methods; thus far, however, no differentiating agent has been discovered which does not remove the stain from the neuroglia fibers quite as readily as from the other tissue elements. At the beginning of this investigation, much time was spent in an endeavor to find a suitable differentiating agent, or otherwise modify these methods so that they might be applicable for the staining of the neuroglia tissue of vertebrates, but in general without success. It was then found that with Benda's sodimn sulphalizarate-toluidin blue method, the neuroglia tissue of vertebrates generally could be stained differentially. I am not aware that Benda is familiar with this fact; he has not, so far as I know, made mention of it. In preparations of the central nervous system of verte- brates treated after this method, the following observations may be made. Well-differentiated preparations present to the naked eye a bluish red or of the protoplasm of cells, and has recently published a number of staining methods, by aU of which neuroglia fibers may be more or less successfully differentiated. According to him, certain hematoxylin solutions, used after proper fixation and mordanting of the tissues, may be used for neuroglia stains ; also hematoxylin staining, followed by staining . with an acid-anilin water crystaUviolett solution. These will not be considered here. I wish, however, to call especial attention to the following method for staining neuroglia tissue, suggested by Benda, since it has certain advantages not possessed by other selective neuroglia stains. Fix small pieces of tissue in ten per cent formalin; place in Weigert's chrome-alum solution (formula given above), four days in warm oven at 38° C. ; wash in water twenty-four hours; dehydrate in graded alcohols; embed in paraffin; cut thin sec- tions and fix these to slides with the albumin-glycerin fixative ; remove paraffin and place sections in mordant consisting of a four per cent aqueous solution of ferric alum; rinse thoroughly in two tap waters and one distilled water; place in a sodium sulphalizarate solution (add to distilled water a sufficient quantity of a saturated solution of sodium sulphalizarate in seventy per cent alcohol to give it a sulphur-yeUow color) twenty-four hours; rinse in distilled water; stain for fifteen minutes in a 0.1 per cent aqueous solution toluidin blue, which should be heated after the sections are in the stain until the solution steams ; aUow the stain to cool ; rinse in distilled water ; wash in a one per cent aqueous solu- tion of glacial acetic acid for a few seconds or in acid alcohol (six drops of hydrochloric acid, seventy per cent alcohol 100 c.c.) for a few seconds; dry sections with filter-paper; dip sections a few times in absolute alcohol ; differentiate in creosote, ten minutes to an hour — control now and then under the microscope; wash in several xylols and mount in xylol- balsam. Neurogha fibers blue, chromatin of neuroglia cell nuclei a purplish blue, pro- toplasm of neuroglia cells brownish red to bluish red. G. CARL HUBER. 589 Fig. 1. Six figures selected from as many contributions on the structure of neuroglia tissue, illustrating the different views which are held concerning the structure of this tissue. A. Neuroglia ceU with processes. (Fig. 5, b, Plate I. — Golgi. Untersuchungen iiber den feineren Bau des centralen und peripherischen Nervensystems. Atlas.) B. Neuroglia cell and processes as seen in chrome-silver preparations (Astrocyte with 590 STUDIES ON NEUROGLIA TISSUE. long rays). (Fig. 20. — V. Lenhossek. Der feinere Bau des Nervensystems im Lichte neuester Forschungen. 1895.) C. Neuroglia cell ^^'ith neuroglia fibers as processes. (Fig. 20, Plate V. — Erik Miiller. Studien iiber Neuroglia. Arch. f. mik. Anat., Bd. LV, 1900.) D. Neuroglia cell with completely differentiated neuroglia fibers, in contact with or completely embedded in the peripheral layer of the protoplasm of the cell. (Fig. 365. — Ranvier's technisches Lehrbuch der Histologie, Uebersetzt von Nicati und Wyss, 1888.) E. Neuroglia cell nucleus and neuroglia fibers as seen in preparations of neuroglia tissue when stained after Weigert's stain. (Fig. 1, D, Plate I. — "Weigert. Beitrage zur Kenntnis der normalen menschUchen Neuroglia, Frankfurt, A. M., 1895.) F. Neuroglia cell with differentially stained neuroglia fibers, in contact with or embedded in the protoplasm of the cell (Invertebrate tissue). (Fig. 34, Plate IV. — Joseph. Untersuchungen iiber die Sttitzsubstanzen des Nervensystems. Arbeiten a. d. zoologisch. Instit., Wien., Bd. XIII, 3 Heft, 1902.) brownish red color, large masses of neuroglia showing as blue areas. Under proper magnification, the neuroglia fibers stand out very clearly and are stained a deep blue color; in over-differentiated preparations, they may appear of a light blue, reddish purple, or reddish brown color, depending on the extent of bleaching, but may generally be clearly seen. The chromatin of the nuclei of the neuroglia cells generally presents a purple-red color ; they may, however, appear brownish red. In well-differentiated preparations, the protoplasm of the neuroglia cells has a reddish brown color; since, however, it is at times necessary to interrupt the differentiation before the toluidin blue has been removed from the protoplasm of the neuroglia cells, this may have a bluish tinge; this does not materially detract from the clearness of the preparations. Myelin, axis cylinders, and nerve cells have a brownish red color, and the nucleoli of nerve cells are of a deep purplish blue. CoUaginous connective tissue stains a reddish pink, its nuclei purplish blue. The red blood cells are often a dark green color. A word may here be said concerning the manner in which sodium sulphalizarate and toluidin blue act in this staining reaction, since it will in part explain the color effects obtained. Benda believes that the sulphah- zarate acts as a stain and also as a mordant for certain basic anilin stains. If, for instance, a sulphalizarate stained preparation is further stained with a basic anilin stain (methylen blue, toluidin blue), no contrast staining is obtained, but all the tissue elements previously stained in the sulphahzarate are now so deeply stained with the basic anilin stain that the histologic ele- ments are not discernible, and the stain is so tenacious that it is not removed by the ordinary differentiating procedures. It is possible in this way to stain tissue elements which ordinarily show no affinity for the basic anilin stains. If, however, the proper differentiating procedures are instituted, certain tissue elements give up the basic anilin stain before others, and where this is removed the color given the tissue elements by the sulphahzarate again appears. In preparations stained after the sulphalizarate-toluidin-blue method, there is G. CAUL HUBER. 591 obtained a blue coloration which covers up the red color, but the latter is not displaced by the former. If none or very little of the toluidin blue is removed from the tissue elements, they possess a blue color; if all has been removed, they present a brownish-red color (sulphalizarate), and the exttmt of the removal of the toluidin blue determines the nuance of the elements not stained blue or reddish brown. The relative ease with which the neuroglia fibers are stained in the different types of \'ertebrates studied is determined by the tenacity with which the various tissue elements, more particularly the neuroglia fibers, retain the toluidin blue. My own investigations of the structure of the neuroglia tissue embrace a study of this tissue in the following vertebrates : Mammals — man, dog, cat, and rabbit; birds — dove; reptilia — tortoise {Emys meleagaris) ; amphibia — frog {Rana Catesbiniana and Rana halecina). In the special description which follows, I purpose to discuss more particularly the structure of the neuroglia tissue of the white matter of the spinal cord of the vertebrates studied. The neurogUa cells and neuroglia cell nuclei are here less likely to be confused with other cells — smaller nerve cells of the gray matter for instance — generally also, the neuroglia tissue is here more readily stained than in other parts of the central nervous system. My observations on the human neuroglia are limited when compared with those made on the neuroglia of other vertebrates, o\A-ing to the fact that it has proven difficult to obtain tissue sufficiently fresh to admit of differential nexu-oglia staining. The observations made on the human nein-oglia will, therefore, be considered after discussing the neuroglia tissue of the other vertebrates studied. The topographic distribution of the neuroglia tissue, in the vertebrates studied, was not the subject of this investi- gation and is not here considered. Dog. — ^The neuroglia tissue of the dog consists of neuroglia cells and neu- roglia fibers. When stained after Benda's sulphalizarate-toluidin-blue method, the neuroglia fibers present a deep blue color, the nuclei of the neuroglia cells a purpUsh blue and the protoplasm a brownish red color. On Plate I, in Figs. 1 to 12, are portrayed certain t3rpic varieties of neuroglia cells found in the white matter of the spinal cord. The size and shape of the neuroglia cells vary greatly as also the amount of protoplasm. A layer of pro- toplasm — the cell body of the neurogha cells — surrounding the cell nucleus can be made out in the great majority of the cells. This may, however, be present in such small quantity as to form only a narrow zone, and free- nuclei without apparent protoplasmic covering are also met with. Since, however, the great majority of my observations were made on sections which do not exceed 6 /* in thickness, it is very probable that certain of these free nuclei are in reahty eccentrically placed nuclei, the greater portion of the respective cell being found in other sections. Cells in which the protoplasm is present to an appreciable extent constitute the greater 592 STUDIES ON NEUROGLIA TISSUE. portion of the neuroglia cells seen in the white matter of the spinal cord. The shape of such cells, as seen in sections, varies with their location. When situated in the main mass of the white matter of the spinal cord, and seen in cross sec- tions of the cord, they present an irregular triangular and quadrangular form, with more or less pronounced concavity of the sides, against which abut the contiguous nerve fibers and with protoplasmic processes which arise from the angles and which extend for a variable distance between the nerve fibers. (See Figs. 1, 2, 3, 4, and 6 of Plate I.) Now and then anastomosis between the protoplasmic branches of two neuroglia cells is observed (Fig. 6). Neuroglia cells situated by the side of the small bundles of nerve fibers which leave the gray matter to form the nerve roots, or by the side of pial septa or blood-vessels are often of spindle shape. (Figs. 5 and 7.) In longitudinal sections of the white matter of the cord many of the neuroglia cells present an irregular rectangular or oval form with short protoplasmic processes. Such cells are often grouped in rows, varying in length and comprising six to twelve cells and situated between the nerve fibers as shown in Fig. 11. The shape of the neuroglia cells of the gray matter of the cord is somewhat more difficult to determine, since the color taken by the gray matter is in general similar to that presented by the protoplasm of the neuroglia cells. The pr-evailing shapes observed are irregularly round or oval, with short protoplasmic processes, although many cells without apparent processes are seen ; on the whole neuroglia cells of the gray matter have less protoplasm than those of the white matter. The nuclei of the neuroglia cells vary greatly in shape, size, and struc- tiu-e, the size varying in proportion to the size of the cells except in cells which have very little protoplasm. In cross and longitudinal sections of both the white and gray matter, the majority of the nuclei seen are of nearly roxmd or oval shape, some are horseshoe-shaped, and others polymorphous. The major- ity of the nuclei, both in the white and gray matter, may be described as vesic- ular with the chromatin arranged in larger or smaller granules. Generally such nuclei show one or several large chromatin masses, which appear like nucleoli, but are probably not true nucleoli, as they stain like the remaining chromatin. It' should be stated, however, that the method used is not espe- cially adapted for bringing out the details of nuclear structure. Another variety of neurogha cell nuclei observed may be characterized by the fact that they are of -relatively smaller size, and stain more deeply, and it is often not possible to make out nuclear structure. They are of round or oval shape and belong to what] I, have designated as free nuclei, though frequently "cells" of this type have vesicular nuclei (Figs. 8 and 9). The vesicular nuclei and the small, deeply staining nuclei may be regarded as the two extremes of a series with intervening transitional forms. The nucleus as a whole stains more deeply than the protoplasm of the cell and is readily distinguished from it. The neurogha_fibers stain a deep blue color. They present smooth bor- G. CARL HUBER. 593 ders, show no varicosities, and any one fiber presents throughout about the same diameter. They vary somewhat in size, but are in the dog relatively small. That they are sohd and not tubular structures may readily be observed when fibers are seen in cross section, when they appear as deeply stained blue dots. So far as I have been able to determine, the neuroglia fibers do not branch and anastomose. Whether the points of termination of fibers, as seen in section, are in reahty the ends of the respective fibers, is quite impossible to say. In fibers which appeared to terminate in the sub-pial layer of neuroglia tissue or on the adventitial sheaths of blood-vessels, I have not observed what might be regarded as a terminal conical enlargement, nodule, or foot plate, as is now and then seen in chrome-silver preparations. As the question of relation of the nem'ogha fibers to the neurogha cells or nuclei is one which is as yet not answered in the same way by all investigators who have considered it, it was the subject of special inqmry in this research and the majority of the figures presented were selected with special reference to it. In considering these figures, it should be borne in mind that they reproduce the appearance presented by sections which are on an average 5 /i in thickness. When compared with the figures accompanying the articles of Weigert, Krause, Aguerre, and others, my figm-es show fewer fibers in relation to any one neuroglia cell or nucleus than are shown by their figures, readily explained by the fact that these show the appearance presented by neuroglia tissue as seen in sections varying in thick- ness from 10 i" to 20 /*, and when compared with the neuroglia cells as seen in chrome-silver preparations, the difference is still more striking; all familiar with such preparations are aware that they are cut to best advantage at a thickness varying from 40 ;i to 80 /i. The appearance of the majority of the neurogUa fibers, in the figures here given, in the form of short segments is also accounted for in the same way. I have deemed it necessary to offer this explanation, which applies to all the figures accompanying this article, to avoid the liability of misinterpretation. To explain why sections about 5 /i in thickness were used throughout this work, it may be stated that it was early found that it was easier to obtain good differentiation of the neuroglia fibers in thin sections; and fiuther, a determination of the relation of the neuroglia fibers to the protoplasm of the neuroglia cells is more readily made in relatively thin sec- tions, for it is necessary to use relatively high power in the study of neuroglia tissue. Generally the relation of the neurogha fibers to the protoplasm of the neurogha cells is readily ascertained; this relation may here be described with reference to certain specific cells. In Figs. 3, 4, and 6 are represented what may be regarded as typical neurogha cells from the white matter of the spinal cord of the dog, with vesicular nuclei, protoplasm, and protoplasmic processes. As may be observed, certain of the neuroglia fibers lie on or in the peripheral layer of the protoplasm, and where a sufficient length of any one fiber is shown, such a fiber may be traced to the cell, often along one of its processes, along the 594 STUDIES ON NEUROGLIA TISSUE. side of, or over, or under the cell, and again away from the cell, along another of its processes. Other neuroglia fibers are traced to the cell, where they end abruptly, in or on a protoplasmic process, or on the protoplasm in the immediate vicinity of the nucleus. Such fibers always, however, end abruptly; there is no gradual transition of fiber to the protoplasm of the cells, as one may assume would be the case if the fibers were actual processes of the cells, but the ends of the fibers in relation with the protoplasm of the cells always have, in well differ- entiated preparations, the appearance of cut ends, and are always traced to their termination Tsdth the utmost clearness, and are readily distinguished from the protoplasm. In case the differentiation is less complete, in which event the protoplasm of the neuroglia cells has a bluish or purplish blue tinge, the shade and depth of stain depending on the extent of differentiation, cut fibers ending in or on the protoplasm now and then have the appearance of actual processes of cells, the blue-colored fiber going over into the blue-colored protoplasm. That this is, however, not a normal relation is readily seen in properly differen- tiated preparations, as above stated. Neuroglia fibers, which pass a cell in its immediate vicinity without touching its protoplasm or its protoplasmic pro- cesses are met with, and such as pass over or under a cell and now and then its nucleus are often seen. It is now and then difficult and sometimes even in thin sections quite impossible to reach a definite conclusion as to whether a certain neuroglia fiber is situated in the peripheral layer of the protoplasm of a neuroglia cell or lies upon it ; especially is this the case when fibers are seen from the side, in longitudinal view. When the fibers are seen in cross section, that is, when they appear as deeply stained, circumscribed dots, their relation to the proto- plasm may often be ascertained with certainty ; and it may be seen that certain of the fibers are undoubtedly situated in the peripheral layer of the protoplasm of the neuroglia cells, while others abut against the protoplasm. In Fig. 10 is reproduced a cell, which is interpreted as representing a neuroglia cell so cut as to exclude the nucleus. Numerous neuroglia fibers are here seen in cross sec- tion — black dots — some of which lie in the peripheral layer of the protoplasm, others of which are placed against it very closely. This figure was drawn from a section not more than 4 ^t in thickness and a play of the micrometer screw of the microscope scarcely disturbed the relation of the drawing reflected with the camera lucida and the image, so that it was possible by careful focus- ing to control the accuracy of the drawing made. See, also. Figs. 3, 4, and 6; also, other figures on the remaining plates. In the larger neuroglia cells, with relatively large amoimts of protoplasm, the relation of neuroglia fibers and protoplasm of neuroglia cells, as here described, is often clearly made out. In Figs. 1, 2, and 5 are represented such cells. The cell reproduced in Fig. 5 is situated by the side of one of the small bundles of nerve fibers leaving the anterior horn of the gray matter to form the anterior root. The cell is, no doubt, cut parallel to its long axis. In the section, there is a slight separation G. CARL HUBER. 595 of the tissue elements in the immediate \'icnuity of the cell, so that it is practi- cally isolated from the structures surrounding it. That it is a neuroglia cell and not a nerve cell is assured by its position — peripheral portion of the white matter. As is shown in the figure, short segments of a number of neuroglia fibers are seen in close relation A\'itli the two prominent protoplasmic processes possessed by the cell ; others pass along its long borders and still others cross the cell and nucleus. In Figs. S antl 9 aw shown two neuroglia cells with only narrow borders of protoplasm, which in the former figure is nearly covered up by the neuroglia fibers. All neuroglia fibers which were in the immediate vicinity of the " free nuclei " are represented in the figures. In thicker sections and with suitable staining, the two cells would present an appearance closely resembling that shown by Weigert in his figures of neuroglia tissue, especially the latter figure, in which is shown a looping of the neuroglia fibers in the neighborhood of the nucleus as described by this observer and seen by others, ^lany of the nuclei of neuroglia cells, which appear as free nuclei of vesicular structure, are in reality eccentrically placed nuclei, only a portion of the respec- tive ceU showing in the section, the protoplasmic portion of the same cell appearing in other sections and presenting an appearance shown in Fig. 10. Small, deeply staining nuclei, without apparent protoplasmic covering, and showing no definite relation to neuroglia fibers, are also met with, but are not numerous in my preparations. It has been stated above that neuroglia fibers which have no apparent connection with the protoplasm of neuroglia cells, but pass by the side of, or over, or under the cells, at some slight distance from the protoplasmic border of the cells, are frequently met with. Attention may be drawn to Fig. 7, with reference to this fact. It is now and then possi- ble, even in thin sections, to trace individual neuroglia fibers uninterruptedly through more than a field of the microscope. The observation may now and then be made that such fibers are in close relation with the protoplasm of one neurogha cell, either in or on the peripheral layer of its protoplasm, a point which is often diSicuIt to determine in longitudinal view of the fibers, and that if such fibers are traced for a distance they will be observed as passing over or by the side of some other neiu"Oglia cell in close proximity to it. Two such fibers are clearly shown in this figure, and two others for which such relation may be projected. A careful study of Figs. 11 and 12 will also show this. The question natiu-ally arises whether or not all neuroglia fibers are in per- manent relation with neuroglia cells. This question will be considered in the general discussion of the structure of neuroglia tissue which follows, since reference will again be made to it in the further presentation of my own obser- vations. In Fig. 12 is shown a small portion of a neuroglia septum, found in the lateral column of the white matter of the cord and seen in longitudinal section. Fibers and cells were traced with the aid of the camera lucida; rather too few than too many fibers are represented. The relation of the 596 STUDIES ON NEUROGLIA TISSUE. neuroglia fibers to the neuroglia cells, to which attention has been called, may be seen, though, owing to the complexity of the network of fibers, not so clearly as when single cells are figured and studied. Attention may be called to the numerous fibers which show no apparent relation to the neuroglia cells; many appear as short segments, others may be traced for longer distances. The figure presents the appearance presented by a section about 5 /« in thickness. Cat and Rabbit. — ^More space has been given to the consideration of the appearance presented by the neuroglia tissue of the dog than is deemed necessary for the presentation of observations made on the neuroglia tissue of the other vertebrates studied, since in the former I have discussed once for all certain general questions which will need no fm"ther consideration. Figs. 13 to 19, Plate II, pertain to the neuroglia tissue of the cat and Figs. 20 to 24 to the neuroglia tissue of the rabbit. In both the cat and the rabbit the neuroglia fibers are quite readily stained differentially. The protoplasm of the neuroglia cells retains a purplish red tinge. The general shape and character of these cells are shown in the figures presented, and since these do not differ materially from those described for the neuroglia cells of the dog, they need no especial description here. The nuclei also present similar charac- teristics — vesicular nuclei with the chromatin in granules and smaller more deeply staining nuclei form the two extremes of the varieties seen. In Figs. 18 and 19 (cat) and Fig. 24 (rabbit) are represented nuclei of the latter variety, one in each figure. Their general characteristics, as here represented, are like those of similar nuclei found in the neuroglia tissue of the dog, for which no special figures were given. Figs. 13, 14, and 15 portray neuroglia cells which may be regarded as typical for the white matter of the spinal cord of the cat. Cells with distinct protoplasmic branches were selected for the sketches. In Fig. 17 is shown a neuroglia cell with two nuclei; such polynuclear cells are not numerous, though met with now and then. In Fig. 18, which is from a longitudinal section of the spinal cord of the cat, is shown the arrangement of neuroglia cells in short columns, placed between the nerve fibers of the white matter, as mentioned for the dog. Fig. 19, also from a longitudinal section, represents a section through a neuroglia septum, parallel to its sur- face. The neuroglia fibers of the cat bear the same relation to the protoplasm of the neiu-oglia cells as described for the dog; they are situated on or in the peripheral layer of the protoplasm of the neuroglia cells. In Fig. 15 are shown two nem-oglia cells, with nevirogha fibers which pass from one cell to the other. Especial attention needs to be drawn to the cell represented in Fig. 16. V. KoUi- ker, in a recent edition of his " Handbuch der Gewebelehre des Menschen, " describes the neuroglia cells, which he terms Golgi cells, as consisting of two portions — a cell-body containing the nucleus, which is intimately associated with a cell plate from which arise the cell processes. He suggests the hypoth- esis that " a Golgi cell with a portion of its cell protoplasm develops a cell G. CARL HUBER. 597 plate from which the cell processes arise; this plate originally, and as long as the cell processes possess the power of growth, is intimately connected with the nucleated poi'tion of the Golgi cell; in many instances, however, the cell plate attains a different consistency, and perhaps also a different constitu- tion, and under certain conditions may sever its connection with the nucleated portion of the Golgi cell." The cell shown in Fig. 16 may, it appears to me, be regarded as substantiating the above hypothesis, suggested by v. KoUiker, so far as concerns this cell, which is found just external to the neuroglia layer surrounding the central canal of the spinal cord. Cells of this type are, how- ever, not numerous, and were only occasionally seen by me, although cells of a similar form were now and then met with. They cannot, therefore, be regarded as typical. They are of interest when considered in connection with the so-called free nuclei of the vesicular type, to which attention has been drawn, two of which are shown in Figs. 8 and 9 of Plate I. If the cell shown in Fig. 16 had been cut in a plane parallel to the long processes possessed by the cell, and at right angles to the one shown in the figure, at least one of the sections would have shown a "free nucleus," surrounded by a very narrow layer of protoplasm, one or several other sections only non-nucleated proto- plasmic masses. In Figs. 20, 21, and 22 are shown cells which may be regarded as typical for the neuroglia tissue of the rabbit, giving the general character of the neu- roglia cells, nuclei, and neurogUa fibers, as seen in cross sections of the white matter of the spinal cord. Fig. 24 is from a longitudinal section of the spinal cord, again showing the arrangement of the neurogha cells in short irregular coltinms, situated between the nerve fibers. In the rabbit, as in the dog and the cat, the neurogUa fibers are seen in close relation with the protoplasm of the neurogha cells, either in the peripheral layer of the protoplasm of these cells or abutting against it. Neurogha fibers which are in close proximity to the protoplasm of neuroglia cells, but without touching it, are also frequently seen. In Fig. 20 are seen three neurogha cells as found in a cross section of the white matter of the spinal cord and shown in the relative positions which they occupied in the section. This figure shows very clearly the relations of neu- rogha fibers which are in contact with the protoplasm of one nemoglia cell and which in their course pass in close proximity to other neuroglia cells. Had this section shown any one of these cells and not the others, it would be but natural to regard neiirogha fibers which pass in close proximity to that cell as associated with it and it woxild be more natural to do this were the preparations stained after Weigert's method, in which event the proto- plasm of the neuroglia cells would not be brought to view. In Fig. 23 is represented a type of neuroglia cells which is now and then met with — very large cells, with large vesicular nuclei and with the chromatin climiped in one or several large granules or globules. Such cells I have seen 598 STUDIES ON NEUROGLIA TISSUE. not only in the spinal cord of the rabbit, but also in the spinal cord of the cat and the dog. The cell reproduced in the figure is situated in the peripheral portion of the white matter, just internal to the sub-pial neuroglia layer, and in the neighborhood of the place of entrance of the posterior root fibers. From its position, it can be regarded only as a neuroglia cell, unless the views con- cerning the location of nerve cells are materially altered. The notches seen in the sides of the cell are occupied by nerve fibers varying in size. Only a few neuroglia fibers are foimd in relation with this cell ; some of these are seen in cross section, others in oblique section, and still others have a longitudinal course bordering the peripheral layer of the 'protoplasm. Neuroglia cells of this size are not numerous, and I have observed them only in the white matter of the spinal cord, never in the gray. Dove. — I have had more difficulty in staining differentially after the sodium sulphalizarate-toluidin blue method the neuroglia tissue of birds than was experienced with the neuroglia tissue of the other vertebrates studied. Such tissues stain less uniformly; in cross sections of the cord, the peripheral por- tions of the sections will often show quite satisfactory staining of the neuroglia fibers, while toward the center of the sections these fibers do not show differential staining. It is necessary to control carefully the removal by means of creosote of the excess of the toluidin blue stain, as the neuroglia fibers do not retain the toluidin blue stain much more tenaciously than does the protoplasm of the neuroglia cells or the coUaginous connective tissue. The best results are obtained when the extraction of the stain, as controlled under the microscope, is continued until the neuroglia fibers are differentiated to the extent that they are readily made out — dark blue color — ^the protoplasm of the neuroglia cells still retaining a purplish blue tinge, the medullary sheaths showing abrown- ish red, and axis cyhnders a brownish red to a dark blue color. Blue-stained axis-cylinders are readily distinguished from neuroglia fibers of the same or similar color by their size and by their position with reference to the medullary sheaths. In Figs. 32 and 33 are represented the appearances presented by the neuroglia of birds, as seen in cross and longitudinal sections of the white matter of the spinal cord of the dove. The sections from which the figures were drawn were differentiated with reference to the neuroglia fibers and the protoplasm of the neuroglia cells. The great majority of the neuroglia cells of the dove, as seen in cross sections of the cord, show a relatively small amount of protoplasm, with protoplasmic processes which extend for a short distance between the surrounding nerve fibers. The nuclei of such cells are of vesicular structure with one or several larger chromatin granules and a varying number of smaller ones. This structure of the nuclei is more clearly seen in prepara- tions more fully differentiated than those from which the figures were taken. Free vesicular nuclei are also met with, as also relatively small, deeply staining ones, without apparent protoplasmic covering. Here and there in the white G. CARL HUBER. 599 matter of the cord, more often in the gray matter, are seen neurogUa cells of irregularly round or oval form without distinct protoplasmic processes. The neurogUa fibers of the dove are relatively fine and appeair in properly differ- entiated preparations as distinct fibers and not as processes of neuroglia cells. However, in the study of many sections, numerous examples of neuroglia cells are met with which would lead one to suppose that the neuroglia fibers are true processes of the cells, going over into the protoplasm of the cells; a familiarity mth the staining method used and with the results obtained when the requisite differentiation is obtained enables one to say that such appear- ances do not represent the actual conditions. The relations of the neuroglia fibers to the protoplasm of the neuroglia cells are the same in the dove as those described for the mammalia studied — in or on the peripheral layer of the proto- plasm. This is shown in the figures presented. In Fig. 32, which is from a cross section of the spinal cord of the dove, are shown five typical neuroglia cells -n-ith protoplasmic processes; the cells shown occupy in the section the same relative position given them in the figure ; cells and fibers were sketched without moving the field and all the neuroglia fibers seen in contact with the cells and in their immediate vicinity are reproduced. Fig. 33 is from a longi- tudinal section, showing a portion of a group of neuroglia cells with the neuro- glia fibers associated with them. Tortoise. — The neuroglia tissue of the spinal cord of the tortoise is quite readily stained differentially with the sulphalizarate-toluidin blue method, although care needs to be exercised in the extraction by means of creosote of the toluidin blue ; when this is done, it is often possible to remove entirely the blue stain from the protoplasm of the neuroglia cells, leaving this a brownish red color, and still retain a blue staining of the neuroglia fibers. As in prepara- tions of the neuroglia stained after this method, the color shown by the proto- plasm of the neuroglia cells merely indicates the extent to which the differ- entiation may be carried without removing the blue stain from the neuroglia fibers, and does not, as previously explained, indicate any pecuUarity of its structure, and is, therefore, of minor importance. I have selected for the figures illustrating the neuroglia cells of the tortoise two (Figs. 30 and 31) that were drawn from preparations in which the neurogha fibers were well differentiated and in which the protoplasm of the majority of the neurogha cells presented a bluish tinge. The neurogha cells of the tortoise are relatively large ; the amount of protoplasm varies somewhat and the protoplasmic processes when present are short and slender. The majority of the nuclei are of round or oval form, of vesicular structure, with the chromatin arranged in one or several larger gran- ules and a varying number of smaller granules. Vesicular nuclei without apparent protoplasmic covering and relatively small ones which stain deeply- free nuclei— are also met with, though they are not numerous. The neuroglia fibers of the tortoise, which vary in thickness, stain a deep blue and are readily 600 STUDIES ON NEUROGLIA TISSUE. differentiated from the protoplasm of the neuroglia cells. Certain of the neu- roglia fibers are situated in the peripheral layer of the protoplasm of the neu- roglia cells, others are in contact with the protoplasm, most clearly seen in cross sections of the fibers. (See the cell at the right in Fig. 30.) When the neuroglia fibers are seen in longitudinal view, they are traced along the bor- ders of the neuroglia cells, over or \mder them without being interrupted by the protoplasm. In Fig. 30, which is from a cross section of the spinal cord of a tortoise, are shown three typical neuroglia cells, showing unessential differences in the shape of the cells and in the relationship of the neurogUa fibers to the cells. In the reproduction of the cell shown at the left of the figure may be seen short segments of a number of neuroglia fibers, which are in close relation to the protoplasmic processes of the cell, and which appear in the figure to end in the protoplasm of the cell and to constitute true processes. That this is, however, not the case is shown by their clear differentiation. In case the protoplasm of the cell were stained more deeply blue and the seg- ments of the neuroglia fibers shown in the figure not so clearly differentiated, a neuroglia cell with neuroglia fibers as true processes would be closely simu- lated. In Fig. 31 is shown a group of neuroglia cells from a longitudinal sec- tion of the spinal cord of the tortoise, showing the arrangement of the cells in short columns, situated between the nerve fibers, an arrangement to which reference has previously been made. A closer study of the figure will show many neuroglia fibers, which appear in relation with more than one neuroglia cell, a phenomenon to which attention has been called in the description of the neuroglia tissue of the other vertebrates considered. Frog. — ^The neuroglia tissue of the frog consists of relatively large neuroglia cells and relatively thick neuroglia fibers. It is quite readily stained after the sulphalizarate-toluidin blue method; the neuroglia fibers take a deep blue color and the protoplasm of the neurogHa cells a light blue or purplish blue tinge. The neuroglia cells have a relatively large amount of protoplasm, and as a rule distinct protoplasmic processes. The nuclei of the neuroglia cells are large, of round, oval, or horseshoe shape, and of vesicular structure, with the chromatin clumped in one or several larger granules and a var3dng number of smaller ones. Free vesicular nuclei or such surrounded only by a narrow zone of protoplasm are here and there seen, and relatively small, deeply stain- ing nuclei are met with, but are not numerous, and are perhaps more readily seen in longitudinal sections than in cross sections of the cord. The neurogha fibers of the frog are relatively thick and in well-stained and well-differentiated preparations stand out very clearly. A large per cent of them have a trans- verse direction in the white matter of the spinal cord and in cross sections of the cord are often traced iminterruptedly for long distances. There are on the whole relatively fewer neuroglia fibers in the neuroglia tissue of the frog than in that of the other vertebrates studied; they bear, as will be shown. G. CARL HUBER. 601 the same relation to the protoplasm of the neuroglia cell as described for the other vertebrates studied. Figs. 25 to 29 are from the neuroglia tissue of the white matter of the spinal cord of the frog. In Figs. 25, 26, and 27 are represented three typical nevn-oglia cells, with large vesicular nuclei, proto- plasm and protoplasmic processes. In each of the three figures arc shown cross-cut neuroglia fibers, certain of which are in the peripheral layer of the protoplasm, others are in contact with it, and still others (to the right in Fig. 27) are separated from it by a narrow space. Certain of the neuroglia fibers seen in longitudinal view run for longer or shorter distances along the sides of the cells, in contact with or in the peripheral layer of the protoplasm; others cross the cells and still others appear to end in or on the cell-protoplasm as in the region of the large protoplasmic process shown in Figs. 26 and 27; that the ends of these fibers in contact with the protoplasm of the cells are cut ends and not natm-al ends a study of the preparations clearly shows. It is often difficult to determine, even by careful focusing, whether neuroglia fibers which appear as crossing the neuroglia cell and nucleus (Figs. 26 and 27) are in actual contact ^ith the cell protoplasm or separated from it by a narrow space; to determine this, it would be necessary to see cell and fibers in cross section. In Fig. 28 is shown what appears to be a vesicular nucleus not sur- roim.ded by protoplasm, with the neuroglia fibers seen in its immediate vicinit}'. Of the neuroglia fibers shown, probably only three (the one under and the one over the nucleus and the one just below it) are in relation with the nucleus, while the other fibers pass in its neighborhood. Such nuclei are here and there met with in the white matter of the spinal cord of the frog, more often in the gray matter, and may be regarded as eccentrically placed nuclei, the protoplasm not showing in the respective section. In Fig. 29, which is from a longitudinal section, are shown four neuroglia cells with vesicular nuclei and one smaller deeply staining nucleus (upper portion of figure) arranged in a short column. Relatively few neuroglia fibers having a longi- tudinal course appear in the figure, which bears out the statement above made that the greater per cent of the neuroglia fibers of the white matter of the spinal cord of the frog have in general a transverse direction. Man. — ^My observations on the neuroglia tissue of man are confined to sections stained by the sulphaHzarate-toluidin blue method and made from small pieces of tissue derived from the brain cortex, cerebellum, medulla, and upper end of the spinal cord of a middle-aged man and fixed in ten per cent formahn a few hours after death. These tissues were regarded as normal. They were not removed and fixed with reference to the staining of the neuroglia tissue and had been in the ten per cent formalin solution for several months before they were prepared for neuroglia staining. In tissues obtained from other individuals at post-mortem, I have not been able to obtain any uniform staining of the neurogha tissue, as they were not sufficiently fresh. Weigert 602 STUDIES ON NEUROGLIA TISSUE. calls attention in the first communication concerning his differential neuroglia fiber stain and also in his large monograph to the fact that it is necessary to employ fresh tissues in order to obtain satisfactory staining of the neuroglia fibers — "dass man nur ganz frisches Material von guter Consistenz benutzen darf, " and again, "that the more central portions of a piece of tissue are not suitable for neuroglia staining unless the hardening fluid completely permeates and fixes the tissues within twenty-four hours" are phrases used by him. Mallory expresses himself likewise, and Benda states at the beginning of his contribution on the staining of neuroglia tissue that one of the most important observations made by him during his investigation of neuroglia tissues was "that when the gha fibers were well fixed, a satisfactory staining of them may be obtained by several methods, unsatisfactory staining of them being generally the result of faulty conservation of the tissues. " In the prepara- tions made from the tissue pieces to which I have above referred, and especially in cross sections from the upper end of the spinal cord, I was able to obtain very satisfactory differentiation of the neuroglia tissue, although I was gener- ally not able to extract the toluidin blue by means of creosote to the extent possible for the neuroglia tissue of the other mammals studied and still retain sharp differentiation (dark blue color) of the neuroglia fibers; in case all the blue coloration was removed from the protoplasm of the neuroglia cells, the neuroglia fibers presented either a light blue or reddish brown color. T^ter some experimentation, it was found that the most instructive preparations were those in which the extraction of the toluidin blue was carried on to the extent of obtaining a clear differentiation of the neuroglia fibers, these retaining a dark blue color while the protoplasm of the neuroglia cells retained a light blue or purplish blue color. The reason why a further differentiation was not possible in these preparations is to be found, I believe, in the fact that the tissues from which they were made were exposed to the action of the formalin solution for several months and were probably not perfectly fresh when placed in this fixing solution. I have found that in staining sections made from pieces of the central nervous system of the other mammalia studied, which had remained in the formahn solution for from four to eight weeks, instead of that number of days, before they were treated preparatory for the staining, it was often difficult to obtain differential staining of the neuroglia fibers, while in sections made from the same tissues, hardened and fixed in formalin for only a few days and then carried through the various steps of the method as prescribed by Benda, excellent results were obtained. The blue staining of the protoplasm above noted has the advantage of bringing this to view more clearly, and unless altogether too deep, does not interfere materially with the determination of the relation of the neuroglia fibers to the protoplasm of the neurogha cells. The figures presented as illustrations of the neuroglia tissue and cells of man (Figs. 34 to 42) are drawn from cross sections of the upper portion G. CARL HUBER. 603 of the spinal cord, which have an average thickness of 5 ij-. This selection was made to have them conform with the other illustrations presented. The neuroglia tissue of man consists of neuroglia cells and neuroglia fibers. The cells are variously shaped, generally irregularly stellate with a variable amount of protoplasm and with protoijlasmic processes. Irregularly round or oval cells, without distinct protoplasmic processes, are met with and also cells with scarcely any protoplasm. The two types of neuroglia cell nuclei, which Weigert recognized and described as "vesicular nuclei with granular chromatin and smaller ones in which the chromatin forms a homogeneous dark mass," I am also able to recognize; also the transitional forms mentioned by him. The great majority of the neuroglia cell nuclei observed by me in cross sections of the human spinal cord are of round or oval form. Aguerre considers very fully in his contribution to the neuroglia literature the size and shape of the nuclei of neuroglia cells, and calls special attention to irregular polymorphous nuclei of vesicular structure which he observed. He classifies the neuroglia cell nuclei according to their size as follows: a, small nuclei, 3 /i to 4 II, to which variety belong the small, deeply staining nuclei ; h, medium sized nuclei, 6 /t to 8 /*, smaller nuclei of the vesicular type; c, large nuclei of the vesicular type, often measuring 14 ij-. This. difference in the size of the neurogha cell nuclei I have also observed, as pertains to the neuroglia tissue not only of man, but also of the other vertebrates studied; there appears to me, however, to be no need of making such classification as above given, except perhaps for purposes of description, since, as a rule, a difference in the size of the nuclei is not accompanied by a difference of structure. Polymorphous nuclei of the vesicular type were not as frequently met with by me as might have been expected from the account given by Aguerre. As, however, the sections studied by him have a thickness of about 20 ^t, while my own sections average 5 /* in thickness, and would thus show only a segment of one of the larger neuroglia cell nuclei (measuring 10 ij- to 14 ij-), it is probable that this difference of observations made may in part be explained by the difference in thickness of the sections studied, my sections showing segments of nuclei, while in his sections the nuclei were shown in their entirety. As above stated, the majority of the neuroglia cell nuclei observed by me in human neuroglia tissue are of round or oval form and of vesicular structure and are surrounded by a zone of protoplasm which is generally readily made out. Nuclei pre- senting this structure surrounded by only a narrow zone of protoplasm are also met with (see nucleus in upper portion of Fig. 39). I have not especially figured the small, deeply staining nuclei found in human neuroglia tissue; they resemble closely the ones found in the neuroglia of the other vertebrates studied (see Figs. 18 and 24) and discussed in the preceding pages, and need, therefore, no further considerations My own observations on human neurogha tissue, in so far as these pertain to the relation of the neuroglia fibers to the 604 STUDIES ON NEUROGLIA TISSUE. protoplasm of the neuroglia cells, enable me to say that this is in general the same as in the neiiroglia tissue of other vertebrates; namely, that the neuroglia fibers are situated in or in close contact with the peripheral layer of the protoplasm of the neuroglia cells. This may be substantiated by brief references to certain specific cells shown in the figures presented. In Figs. 34 and 35 are shown two typical neui'oglia cells, with vesicular nuclei, surround- ed by a distinct zone of protoplasm and with protoplasmic processes. Certain of the neuroglia fibers there seen are over the neuroglia cells, others may be traced along protoplasmic processes and in continuity with their proto- plasm, still others appear to end in the protoplasmic branches or on the proto- plasm of the cells ; of the cross or obliquely cut fibers seen, certain ones are in the peripheral layer of the protoplasm, others in close contact with it, and still others at some little distance from it. In Figs. 36 and 38 are portrayed rela- tively large neuroglia cells with large vesicular nuclei; in the former figure especially, in which numerous cross-cut neuroglia fibers appear, this relation of fibers to cell protoplasm is clearly brought out. The cells shown in these four figures may serve to illustrate the general appearances presented by neu- roglia cells and fibers as seen in cross sections of the white matter of the human spinal cord. Fig. 37 represents a cell situated in the immediate vicinity of a coimective tissue septum; all the fibers surrounding the cell were reproduced in the figure, although it is very probable that the majority of them are not directly associated with the cell. This figure gives the appearances presented by neuroglia tissue when present in larger amounts; in such locations it is quite impossible, especially so in thin sections, where many short segments of fibers are seen, to gain a clear idea of the relations of the neuroglia fibers to the neuroglia cells and to ascertain what fibers are associated with any particular cell. Neuroglia cells possessing more than one nucleus were not frequently seen by me in sections from human tissue. In Fig. 41 is given an example of one such cell. Aguerre describes and figures neuroglia cells from the human spinal cord with two and with three nuclei, and believes that such cells indicate a proliferation of the nein-oglia tissue. My own preparations do not, how- ever, support this view, as I could not observe nuclei which I could regard as fixed in process of division, even amitotic division. I see no reason for attach- ing special significance to the occasional presence of neuroglia cells with more than one nucleus. In Fig. 40 is reproduced a neuroglia cell of such peculiar form as to make it desirable that special consideration be given it. The cell is from a cross section of the white matter of the human spinal cord and from a relatively thin section (not more than 5 ij-) ; its boundaries were clearly made'out. That it is a neuroglia cell is assured by its position. Of special interest to me, aside from the large protoplasmic process shown, are the nxmierous neurogUa fibers seen in cross section, as also their relation to the protoplasm and the protoplasmic processes. Attention may, however, be called to the cross-cut G. CARL HUBER. 605 neuroglia fibers which are situated in or in contact with the terminal portion of the protoplasmic processes and at some distance from the nucleus. In sections preceding and following the one from which the figure was drawn, such neuroglia fibers would very probably appear between nerve fibers and show no relation to neuroglia cells. In every section there are found numerous neuroglia fibers which show no relation to neuroglia cells, more numerous the thinner the section studied. Joseph has also called attention to the apparent dispro- portion of neuroglia fibers to neuroglia cells. Such cells as the one here repro- duced, the figure giving only a fragment of the entire cell, may serve to give some idea concerning the number of neuroglia fibers which may be in contact with any one neuroglia cell. In Fig. 39 is reproduced a small group of neuroglia cells, found between two nerve fibers of a small bundle of nerve fibers cut in the longitudinal direction. In the account of my own investigations on the structure of the neuroglia tissue, I have, as previously stated, given more particular attention to the neuroglia tissue of the white matter of the spinal cord, and only incidentally has reference been made to the neuroglia of other regions, and no consideration has been given to the topographic distribution of the neuroglia, as the research has concerned itself mainly with the structure and relation of its elements. However, the impression should not be gained that the neurogUa elements of other regions of the central nervous system are not stainable by the method employed, for such is not the case. The neuroglia tissue of the white matter of the cord has proven on the whole easier to stain (probably because it is more readily fixed; I base this statement on the observation made by Benda, as above quoted) and the results obtained are in general more readily interpreted. The observations made by me in numerous preparations of the cerebellum and cerebrum of vertebrates studied corroborates in general the statements I have made concerning the structure of the neuroglia tissue. In the gray matter of the preparations studied, including the cortical gray matter of the cerebellum and cerebrum, as also the gray matter of the spinal cord, neuroglia cells with nuclei of vesicular structure, surrounded by a zone of protoplasm are frequently met with, also nuclei of such structure with very little proto- plasm, as well as the relatively small, deeply staining nuclei of neuroglia tissue. The protoplasm of the neuroglia cells of the gray matter is on the whole less abundant than that of the nem'oglia cells of the white matter and the proto- plasmic processes are, as a rule, smaller and less conspicuous ; the great majority of the neiu-oglia cells of the gray matter present, therefore, an irregularly roimd or oval form, while the larger per cent of the neuroglia cells of the white matter when seen in cross section present an irregularly triangular and stellate form. Such a general statement is, however, as may be readily understood, not to be subjected to closer analysis, as the exceptions are numer- ous. The question of the difference of shape of the neurogha cells as found in 606 STUDIES ON XEUROGLIA TISSUE. the various regions of the central nervous system may, however, be regarded as one of secondary importance, as these cells as a rule assume the shapes of the spaces which they occupy and these are governed by factors not depend- ent on the inherent structure of the histologic elements, which enter into the formation of the central nervous system. A comparison may here be drawn between the neuroglia cells and the fixed connective tissue cells of white fibrous connective tissue — fixed cells in areolar connective tissue, corneal and tendon corpuscles — ^the forms of these cells being also dependent on the shapes of the spaces which they occupy. The neuroglia fibers of the gray matter need no special consideration; they are in every respect like the neuroglia fibers of the white matter — in their structure and in their relation to the protoplasm of the neuroglia cells. Numerous attempts were made by me to make use of the Benda sulpha- lizarate-toluidin blue method in the study of the development of the neuroglia tissue, but thus far this method has not given me satisfactory results when applied to embryonic tissue. Here and there in such preparations, a few differentially stained neuroglia fibers may be found, but as a rule, only the sulphalizarate coloration (reddish brown) is observed in sections which have been differentiated in the creosote. My attempts to modify the method have not led to more satisfactory results, and need not, therefore, be given consid- eration here. In the preceding account, no mention has been made of the histologic ele- ments of the ependyma. The sulphalizarate-toluidin blue method may be used to advantage in the study of this tissue and presents certain distinct advantages over other neuroglia methods which might be selected for the study of this tissue, since it stains the cell bodies (protoplasm) of the ependy- mal cells as well as their nuclei and the neuroglia fibers. It is, however, not my purpose to give special consideration in this communication to the cells of the ependyma, nor to consider the literature bearing thereon. The few observations concerning them here presented were gathered from an incidental study of these cells made on sections of the spinal cord, the structure of the neuroglia of which has been discussed in the preceding pages. These prepara- tions show that the ependymal cells have an irregular columnar form and have distinct protoplasmic processes which arise, as a rule, from the basal portions of the cells, rarely from their sides, and extend for a variable distance into the underlying neurogha tissue. These cells possess relatively large nuclei of a vesicular structm-e, with the chromatin in larger and smaller granules. The neuroglia fibers which are in relation with the ependymal cells are not con- tinuations of their protoplasmic processes. Such fibers are, by reason of their reaction to the stain, readily differentiated from the protoplasm of the cells (the fibers staining dark blue, the protoplasm reddish brown by the method used) and may often be traced along the borders of the protoplasmic processes G. CARL HUBEll. 607 and along the sides of the ependymal cells to near their free borders. The relation of the neuroglia fibers to the protophisiu of the- c-peudymal cells is, therefore, the same as that described for neuroglia fibers and the protoplasm of nem-ogUa cells, namely, the fibers are in contact with or situated in the periph- eral layers of their protoplasm. The so-called ependymal fibers, which I have designated as neuroglia fibers, are in their fully developed state to be regarded as neuroglia fibers, which they resemble in every respect. This, if I interpret Weigert correctly, corroborates his obser\'ations on the ependymal fibers. Weigert mentions groups of small granules, which are colored blue in his neuroglia preparations, and which are situated on the inner borders of the ependymal cells; these he regards as cuticular formations. These granules are clearly seen in preparations of ependymal cells, stained after the sulphalizarate-tolui- din blue method, in which they possess a dark blue or purplish blue color. That they are not deformed cilia, an idea which Weigert also confutes, is clearly shown in many of my preparations, where the presence of well-formed eiUa in addition to the granules may often be made out. In my own prepara- tions of the ependymal cells, when these were well fixed and well stained, such granules appear, not on the inner border of the ependymal cells, but in the protoplasm of the cells near the inner border, which leads me to question whether they should- be regarded as cuticular formations. They present the appearance of the mitochondria described by Benda and are interpreted as such bjr me. After ha^•ing thus presented, seriatim, the results of observations made by me on the neuroglia tissue of the vertebrates included in this investigation, we may now briefly discuss these results in a more general way, and consider their bearing on the various views concerning the structure of the neuroglia as found in the literature and as set forth in the summary of the literature previously given. In the foregoing pages I have designated the supporting tissue of the central nervous system as neuroglia tissue and have followed in this respect Erik Miiller, who discussed, at some length, the advisability of regarding the supporting tissue of the central nervous system as a special tissue (eine beson- dere Gewebeart),in place of regarding it as a collection of cells, and of retaining the name neuroglia as a "collective name" for a special variety of cells — astrocytes, spongiocytes, and ependymal cells — found in the central nervous system, as suggested by v. LenhossSk. As Erik Miiller correctly states, the neuroglia possesses the characteristics of a tissue, and should be regarded as such. (I may here, parenthetically, call attention to the fact that in describ- ing the method of Weigert and in making reference to it, I have spoken of it as a differential neuroglia fiber stain; this to emphasize the fact that by means of this method only certain constituents of neuroglia tissue are brought out.) In presenting my own results, I have spoken of the neuroglia tissue as 608 STUDIES ON NEUROGLIA TISSUE. composed of neirroglia cells and neuroglia fibers, the method used bringing out cleariy in the great majority of the cells a zone of protoplasm which sur- rounds the nucleus. This zone of protoplasm constitutes, as I trust has been imderstood, the cell body of the neurogha cells. This phraseology was used to call especial attention to the protoplasm of the neuroglia cells, which is not stained by means of the Weigert method and is in his discussion almost wholly disregarded. Attention has also been drawn to the relation of the neuroglia fibers to the protoplasm of the neuroglia cells. In the preparations of the neuro- glia tissue of all the vertebrates studied by me, whenever it was possible to gain good differentiation of the sections, I was able to observe that certain of the neuroglia fibers which appeared to be associated with any one neuroglia cell were embedded in the peripheral layer of the protoplasm of that cell, others were in close contact with it, and still others were seen at a slight distance from the cell and not in contact with its protoplasm. This, I believe, the figures presented will substantiate. Attention may, however, again be drawn to the fact that these figures were drawn from sections averaging 5 m in thickness. The neuroglia fibers are in no sense processes of the cells, but are fully differ- entiated from the protoplasm of the cells. To use Weigert's words, "The majority of the so-called cell processes are not cell processes, since two such apparent processes form a continuous fiber, which is in no way interrupted by the cell body, as would be the case were they true cell processes arising from the cell body. In a word, there is no question of cell processes or cell extensions, but of fibers which are fully differentiated from the protoplasm. " In the case of nem-oglia fibers seen in longitudinal view and which are to be traced by the side of, or over, or under neuroglia cells, this may be clearly made out in the neuroglia tissue of all vertebrates studied by me ; in the case of neuroglia fibers seen in longitudinal view, but which appear to end in the protoplasmic processes of the cells or on the protoplasm of the cells, it is now and then difficult to determine whether such fibers are to be regarded as true processes continuous with the protoplasm or whether the ends in contact with the protoplasm are to be regarded as the cut ends of neuroglia fibers. In well- differentiated preparations of the spinal cord of all the vertebrates studi,ed, I have not, however, found sufficient reason for hesitating in interpreting such neuroglia fibers as fully differentiated from the protoplasm, the apparent ends of the fibers in contact with the protoplasm representing cut ends and indicating that at such points the fibers left the plane of the section. Such fibers, when not fully differentiated, often appear as true processes of cells. In preparations of neuroglia tissue, differentially stained, there are always found a large number of neurogha fibers, which show no apparent relation to the neurogha cells seen in the section; certain of these are seen in longitudinal view; the majority, however, appear as obhquely cut or cross-cut fibers. It is, as may be readily understood, quite impossible to make definite statements G. CARL HUBER. 609 as to whether such fibcn-s aie in reality without relation to the protoplasm of neurogha cells or whether they only appear so in the sections; or, to express this in different form— are all neuroglia fibers in p(>rmanent relation with the protoplasm of the neuroglia cells, or only a certain per cent of them? Joseph, who has considered this question, reaches the conclusion that in the neuroglia tissue of invertebrates more carefully studied by him, all neuroglia fibers are in permanent relation with neuroglia cells, although it is not possible for him to prove this with mathematical exactness. The disproportion of neuroglia fibers to neuroglia cells in general, more particularly in regions where this tissue is present in larger amounts — neuroglia septa, surrounding the central canal in the cord and the sub-pial layer — has been noted by observers who have worked with the differential neuroglia stains and is used as an argument in favor of the existence of neuroglia fibers completely separated from neuroglia cells. ■ In comparing the results obtained with the special neurogha tissue staining methods with those obtained with the chrome-silver method, the much greater number of nem-oglia fibers (processes of neuroglia cells) seen in relation with any one neuroglia cell, as observed in preparations made by the latter method, is worthy of consideration. It should be remembered, however, that the chrome-silver method stains by precipitation and that neuroglia fibers, cell processes, and neuroglia cells are in no sense differentiated, and in all proba- bility that neuroglia fibers passing in the immediate vicinity of a neuroglia cell would be included by the precipitate formed around the cell body of such cell and made to appear as cell processes; and furthermore, that sections of tissues treated by the chrome-silver method are generally cut very much thicker than is permissible when the special neuroglia staining methods are used. Data gained from chrome-silver preparations of neuroglia tissue may at best be used only as suggestions in interpreting the appearances presented by neuroglia tissue stained after the selective neuroglia tissue staining methods, and are of little value in determining the relation of neuroglia fibers to neuroglia cells. Attention may also be drawn to the fact that we possess at present no definite data concerning the length, or average length, of neuroglia fibers. In differentially stained sections, only segments of fibers are seen, and the chrome-silver method is so precarious that results gained with it are not to be trusted, so far as concerns the length of fibers. One needs only to be reminded of the maimer in which axis cylinder processes of neurones are stained by means of this method — sometimes for long distances, again for short distances, or not at all — in regions where the length of the axis cylinder may be conjectured with reasonable exactness. The question under consid- eration is therefore one which cannot be conclusively answered, and any statement made concerning it must be of the nature of a hypothesis. In a former communication I made the statement that there are certain neu- 610 STUDIES ON NEUROGLIA TISSUE. roglia cells, possessing protoplasmic blanches, the neuroglia fibers of which are not completely separated from the protoplasm, but are in continuity with it, or even pass through it. Further observation leads me to amplify this statement, and to formulate the hj^pothesis — ^that all neuroglia fibers are per- manently in relation with the protoplasm of neuroglia cells, either in contact with, or passing through, the peripheral layer of the protoplasm of the cells. I find it, however, difficult to give specific reasons for reaching such a conclu- sion; in general, the following had weight: 1. The relatively large number of neuroglia fibers which are now and then seen in contact with, or in the periph- eral layer of a neuroglia cell. Such observations have a wider significance when considered in connection with the fact that even in thin sections one may now and then trace neuroglia fibers through more than one field of the microscope (under -^ oil immersion), thus showing that they possess con- siderable length, and would appear in numerous sections were the plane of the sections at an angle to that occupied by the fiber; and further, that while in general the course of many neuroglia fibers is relatively straight, still dis- cursions to the extent of 10 /^ to 20 /i are met with in nearly all such fibers; other fibers have a wavy course, some appear in the form of loops, still others are more irregular in their course; all such fibers, even though their general direction is parallel to the plane of the section, would of necessity appear in more than one section, if these averaged a thickness of 5 ,", all of which goes to show that the apparent disproportion of neuroglia fibers to neuroglia cells is more apparent than real. 2. It is now and then possible to trace neuroglia fibers which are in close relation to the protoplasm of one cell to the immediate vicinity of another neuroglia cell, and some distance beyond it, without com- ing in contact with the protoplasm of the second cell. Such appearances serve to throw light on the numerous instances met with in which neiu'oglia fibers are seen in the immediate vicinity of a neuroglia cell, without being in actual contact with its protoplasm, and which are not observed as in rela- tion with any other neuroglia cell. Such observations have assisted me in reaching the conclusion that very probably all neuroglia fibers are in perma- nent relation \\dth the protoplasm of neuroglia cells. Brief consideration may now be given to neuroglia cells with vesicular nuclei surrounded by only a narrow zone of protoplasm, or such in which no protoplasm can be distinctly made out. Many of these are, no doubt, eccen- trically placed nuclei, the section including only the nucleus, while others are, I believe, cells possessing verj^ little protoplasm, and showing no apparent relation to neurogUa fibers; a consideration of these will be deferred until mention has been made of the small, deeply staining nuclei, found in neurogha tissue, and described by Weigert, Aguerre, and Krause, and mentioned by me in the preceding pages. Krause states that probably these small nuclei possess the same amount of chromatin as do the vesicular nuclei with the chromatin G. CARL HUBER. 611 ill granules, but in more compact arrangement. The general statement con- cerning such nuclei is that they show no definite relation to the neuroglia fibers, and do not form centers around which the fibers are grouped. The following hypothesis has suggested itself concerning the nature of these nuclei : they may be regarded as undeveloped neuroglia cells, neuroglia cells which have not, as yet, taken part in the formation of neuroglia fibers, and that may in post-fetal life gradually develop into neuroglia cells with vesicular nuclei, the development of the nucleus being accompanied by an increase in protoplasm and by the formation of neuroglia fibers. Weigert states that, while the neuroglia cell nuclei belonging to either of the two chief varieties (vesicular nuclei and small, deeply staining nuclei) are so characteristic as to be easily distinguished, a certain number of neuroglia cell nuclei are found regarding which it is difficult to state to which variety they belong, "and which, if it gives pleasure, njay be regarded as transitional forms." I have previously stated that the chief varieties of neuroglia cell nuclei here mentioned may be regarded as the two extremes of a series of nuclear forms met mth in neuroglia tissue. Considered in this light, the neuroglia cells with nuclei of the vesicular type with only a narrow zone of protoplasm and without special relation to neuroglia fibers, may then be looked upon as developing neuroglia cells, representing a more mature stage than that represented by the small cells with deeply staining nuclei, from which they are probably developed. Of especial interest in this connection are observations made by Schaper, pertaining to the early development and differentiation of the structural elements of the nervous system. He recognizes, in the mantle layer of the developing spinal cord, and probably in other regions of the nervous system also, two varieties of cells, distinguished by their size and by the structure of their nuclei — neuroblasts with large vesicular nuclei, each of which shows a distinct nucleolus, and cells which he designates as spongioblasts (using the term in a different sense from that suggested by His), with relatively smaller nuclei, showing a dense, coarsely granulated nuclear structure; the latter he interprets as the mother cells of neuroglia tissue, while the former are the cells which develop into the ganglion cells. The small, deeply staining nuclei of the neuroglia tissue of the adult present a structure which is similar to that of the nuclei of the spongioblasts as described by Schaper. In making this comparison, however, it should be borne in mind that his observations were made on tissues fixed with reference to the finer details of nuclear structure, while formalin solutions (used for fixing the tissues studied by me) are not regarded as especially adapted for this purpose. In the mantle layer there are fiorther found, even at a time when distinct differentiation of certain of its cells may be made out, certain undifferentiated cells, with nuclei of vesicular structure, with fine chromatin granules which show no definite arrangemeiit 612 STUDIES ON NEUROGLIA TISSUE. and without nucleoli. Schaper looks upon such cells as undifferentiated germinal ceUs of the mantle layer, certain of which retain the undifferentiated character until late in embryonic life, or even later; such cells possess the power of proliferation, and furnish new structural elements to the nervous system. He considers it very probable that a certain per cent of these cells are present during the entire life of the organism, and may furnish the elements for the regeneration processes which follow loss of substance in the central nervous system as a result of operation or of pathologic changes. In regard- ing the small, deeply staining nuclei of neuroglia tissue as undeveloped struc- tural elements of the neuroglia, I am corroborating, as appears to me, Scha- per's observations and speculations concerning the presence in the central nervous system of undifferentiated germinal elements; on the other hand, his observations substantiate the interpretation placed by me on these small, deeply staining nuclei. In comparing the results of my own observations with those obtained by other investigators who have studied neuroglia tissue, it may be stated in a general way that the observations here recorded agree in the main with those made by Ranvier and by Joseph (the latter on the neuroglia of invertebrates), and corroborate, so far as concerns the neuroglia fibers, the general conclu- sions reached by Weigert and Mallory and others who have used their differ- ential staining methods, and that they are not in accord with the recorded observations of investigators who regard the neuroglia fibers as processes of neuroglia cells. In comparing my own results with those obtained by observers who have used the chrome-silver method in the study of neuroglia tissue, it is only neces- sary to draw attention again to the fact that this method colors by precipita- tion, and can in no sense be regarded as a differential staining method so far as concerns nuclear and protoplasmic structure, and the results obtained with it cannot be compared with those obtained by methods for differential staining of neurogha fibers. In chrome-silver preparations, the relation of neuroglia fibers (cell processes) to the protoplasm of the cells cannot be made out except in rare instances, as is shown by the account given by Andriezen. Reinke's observations need further consideration. As previously stated, this observer, by modifying the chrome-silver method, was able to make obser- vations which led him to conclude that the neuroglia tissue was composed of neuroglia cells possessing numerous and in part branched processes, which stained by the Golgi method, and of neuroglia fibers, the direction of which was different from that of the processes, and which were not stained by the Golgi method. In the figures presented by me, neuroglia cells will be observed stained after the sulphalizarate-toluidin blue method, in which the protoplasm and protoplasmic processes, as well as the neuroglia fibers were stained; the latter accompanied the protoplasmic processes, the protoplasm extending along the G. CARL HUBER. 613 neuroglia fibers for a variable distance", gradually becoming attenuated, and ultimately disappearing. This Ranvier observed and described; so, also, Joseph. I cannot, therefore, agree with Reinke when he states that the pro- toplasmic processes of the neuroglia cells and the neuroglia fibers have in the main different directions. Furthermore, the great number of protoplasmic processes shown by Reinke leads mc to question his interpretation of the appearances observed by him, as usually only a few processes are seen in con- nection with any one cell. If I may bo allowed to interpret the results obtained by Reinke from the figures which he presents, the following conclu- sions seem permissible: It appears to nic that this observer obtained only partial staining of the neuroglia cells and fibers (processes) by means of the chrome-silver method. I base this statement on the coloration presented by Figs. 1 and 2, Plate I, of his article. With these figures he confirms the results of other observers who have used the chrome-silver method in their study of neuroglia tissue, who also describe neuroglia cells with the neuroglia fibers as processes of these cells, the protoplasmic processes described by him being none other than the processes (neuroglia fibers) seen by other observers using this method. By means of the process of dehydration, embedding and fur- ther staining (see his modification of the chrome-silver method) of the tissues already treated by the chrome-silver method, Reinke removes, as it appears to me, in part, or perhaps wholly, the chrome-silver precipitate, and restains the protoplasm of the neuroglia cells and certain of the neuroglia fibers. The second part of the method used by him is not unlike the method used by Erik Miiller in his investigation of neuroglia tissue (Heidenhain's iron-lack- hematoxylin staining on tissues fixed as for the Golgi method), also similar to the method used by Joseph and certain of the methods suggested by Benda. Figs. 4 and 5 of Plate I of Reinke's work appear to me to show this conclu- sively; indeed, these figures may be regarded as confirming the observations made by Ranvier, Joseph, and myself, in that they show distinct neuroglia fibers which are in close relation to neuroglia cells, embedded either in the peripheral layer of the protoplasm, or in contact with it. Reinke's method does not bring out processes of neuroglia cells and neuroglia fibers, two dis- tinct structural elements, but the same structural elements presented in a different manner. His results are of interest to me in showing the contra- diction in results obtained in preparations treated after the chrome-silver method and those obtained by the differential neuroglia fiber staining methods, and this in the same tissue, since, as appears to me, he first stains the tissues after the chrome-silver method and observes zreuroglia cells and processes; in his further treatment of the tissues he removes the chrome-silver precipi- tate and stains differentially the neuroglia fibers in relation with the proto- plasm of the neurogUa cells, which no longer appear as cell processes. We may now turn to the results obtained by Erik Miiller in his investiga- 614 STUDIES ON NEUROGLIA TISSUE. tion of the neuroglia tissue, using a method devised by himself (fixation of tis- sues as for the Golgi method and staining with Heidenhain's iron-lack hema- toxylin). This observer, as previously stated describes the neuroglia tissue as consisting of neuroglia cells with processes, which he designates as neuroglia fibers, which, while differing in their physico-chemical properties from the protoplasm of the neuroglia cells, are yet continuous with it and end in it. As Erik Mijiller's observations pertain mainly to the neuroglia tissue of the lowest vertebrates and as my own do not include observations on the neuroglia of amphioxus, myxine, selachians, and teleosts, the forms particularly studied by him, an accurate comparison of results cannot be made; since, however, he feels comdnced that the structure of the neuroglia tissue of the higher vertebrates is the same as that of the lower vertebrates, I shall venture to consider Erik Muller's observations in the light of my own. As has been repeatedly stated in the preceding pages — in preparations stained aftei the sulphahzarate-toluidin blue method and not fully differentiated by means of creosote — I have often observed neurogUa fibers, which might be regarded as forming true processes of neuroglia cells, such in which neuroglia fibers having a deep blue color appeared to end in the protoplasm of neuroglia cells having a bluish color, the intensit)- of the stain depending on the extent of the differentiation. Such cells present an appearance which is not unlike those figured by Erik Miiller. In sufficiently differentiated preparations, however, I have always been able to make out distinct neuroglia fibers, which were either not interrupted by the protoplasm of the neuroglia cells with which they were in relation, or, if cut in relation with the protoplasm of the neurogha cells, presented cut ends which were readily differentiated from the protoplasm, with which they were not continuous. Erik Miiller states that especially with reference to the neuroglia fibers of myxine, he long adhered to Weigert's view of independent neuroglia fibers; a closer study, however, enabled him to observe many instances of characteristic union between neu- roglia fibers and the protoplasm of neuroglia cells. A careful study of the excellent figures accompanying his article enables me to see here and there neuroglia cells in which the neuroglia fibers are sketched, not as processes of the cells, but as distinct fibers (Figs. 4 and 25, for instance). In considering Erik Muller's results in the light of my own observations, I have gained the convic- tion that he has interrupted the differentiation of his preparations at a time when the neurogha fibers appeared to best advantage and in greatest number, and that in such preparations the stain has not been sufficiently extracted from the protoplasm of the neuroglia cells to admit of sharp differentiation between neuroglia fibers and protoplasm of neurogha cells. It should be remembered that the neuroglia fibers and the protoplasm of the neuroglia cells are both stained by the hematoxylin in the method used by Erik Miiller, though to different degrees of intensity. It is here 'again a question of the G. CARL HUBER. 615 relative tenacity witli wliicli tlie different structural elements retain the stain. In the method used b>' this observer, tlie neuroglia fibers are not caused to hold the stain much more tenaciously than is the protoplasm of the neuroglia cells, and differentiation can, therefore, not be carried to the extent of removing the stain completely or nearly so from the protoplasm of the neuroglia cells without also removing it from the neuroglia fibers, as can be done in the Weigert and Mallory methods, nor to the extent of giving the neuroglia fibers one color and the protoplasm of the neuroglia cells another, as in preparations stained after Benda's sulphalizarate-toluidin blue method. This phase of the neu- roglia staining technique, I have considered more fully in discussing Benda's method. Preparations in which Erik Miiller recognized distinct neuroglia fibers I would regard as better differentiated than those in which the fibers appear as processes of the neuroglia cells. The conclusions reached by him are based, it would appear to me, on the appearances presented by neuroglia tissue treated by a method, which, to a certain extent, stains the tissue differ- entially, but by means of which it is not possible to obtain complete differen- tiation of the majority of the neuroglia fibers, which, therefore, appear as processes of the neuroglia cells, continuous with the protoplasm of these cells. The results obtained by the Benda sulphalizarate-toluidin blue method sup- plement and extend those of Erik Miiller and give them their correct interpre- tation; his results cannot, therefore, be considered as contradictory to those obtained with the above method, nor, as concerns the neuroglia fibers, to the results obtained by the methods of Weigert and Mallory. My own results, as presented in this communication, confirm in the main the results obtained by Weigert and Mallory, gained by special differential staining methods; also those of Pollack, Krause, and Aguerre. Weigert, in his comprehensive monograph on the structure of the neuroglia tissue of man, discusses in fuU and very critically his results and their bearing and relation to those of other investigators who have preceded him in the study of neuroglia tissue. I need not, therefore, consider this phase of the question. As is well known, Weigert and others who follow him regard the neuroglia tissue as composed of distinct and independent fibers — intercellular structural ele- ments — and neuroglia cells, although only the nuclei of these are considered. The appearances presented by neuroglia tissue when stained after the Weigert method leave no room for doubt concerning the nature of the neuroglia fibers; they are distinct fibers and not processes of cells, and are in no way interrupted by the protoplasm of the neuroglia cells, as would be the case were they true processes. With the exception of Krause, who studied the neuroglia tissue of apes and half-apes, the investigations of the above-named observers pertain to the neuroglia tissue of man. My own observations warrant me in stating that very probably in all vertebrates (I cannot speak from personal observa- tion of the neuroglia tissue of the lowest vertebrates), the fibrillar structures 616 STUDIES OX XEUROGLIA TISSUE. of neuroglia tissue are to be regarded as distinct fibers, completely differen- tiated from the protoplasm of the neuroglia cells, and not as processes of the neuroglia cells. These results, if I may be allowed this interpretation of them, appear to me to warrant the further statement that they supplement and extend those of Weigert and others who work with his method in that consideration is given to the relation which the neuroglia fibers bear to the protoplasm of the neuroglia cells. Weigert and others who have accepted his views concerning the structure of neuroglia tissue do not especially consider the relation of the neuroglia fibers to the protoplasm of the neuroglia cells, since in neuroglia-tissue preparations stained after his method, the protoplasm of the neurogha cells is not brought to view. In a certain sense this must be regarded as to the distinct advantage of this method, since it enables a sharper differentiation of the neuroglia fibers and a more certain determina- tion of the fact that we are here dealing with fibers not interrupted by the proto- plasm of the cells and thus not with cell processes. Having, however, reached definite conclusions concerning this fact, it is necessary to ascertain the rela- tion of these neuroglia fibers to the other structural elements of neuroglia tissue before our knowledge of this tissue can be regarded as complete. The figures accompanying this article may, on hasty examination, appear to differ from those presented by Weigert, Mallory, Aguerre, and Krause in connection with their contributions to the structure of neuroglia tissue ; on closer examina- tion, however, it will be observed that, as concerns the neuroglia fibers and the nuclei of neuroglia cells, they portray essentially the same structure and rela- tions, if consideration is given to the fact that my own figures are drawn from much thinner sections; it is evident, therefore, that in all these figures the same structural elements are presented, my own differing from those of the observers above mentioned in that the protoplasm of the neuroglia cells is also shown, enabling a presentation of the relation of the neuroglia fibers to the protoplasm of the neuroglia cells. As previously stated, I regard it as very probable that all neuroglia fibers are in permanent relation with the protoplasm of neuroglia cells, either embedded in the peripheral layer of their protoplasm or in close contact with it. Taylor, in his discussion of the structure of human neuroglia tissue, considers particularly the results obtained with the Weigert method and asks the question, whether this observer "does not take a step beyond the knowledge furnished by his method, " when he considers the neu- roglia fibers as intercellular substance, the fibers being completely separated from the neuroglia cells. He takes the view that this method is incapable of giving information concerning the relation of neuroglia fibers to neuroglia cells. The criticism of Robertson on results obtained mth the Weigert method may also be considered here. In sections stained by this method, nem-oglia fibers which pass uninterruptedly across neuroglia cells (nuclei) this observer regards as processes of other neuroglia cells; a neurogha fiber which loops in G. CARL HUBER. 617 the neighborhood of a neuroglia cell, he regards as composed of two processes joined by a fold of the cell membrane, which stains like the processes, and thus a continuous fiber is simulated. In neurogha fibers in which the proto- plasm of the neuroglia cells is stained one color (brownish r(>d) and the neu- roglia fibers another (dark blue) we have evidence which goes to show that the explanation given by Robertson concerning neuroglia fibers which are said to form a loop in the vicinity of the nuclei of neuroglia cells is not tenable. His explanation of the appearances presented in neuroglia preparations stained after Weigert's method is refuted also by the fact that when neuroglia fibers which are in contact with or are embedded in the peripheral layer of the pro- toplasm of neuroglia cells are seen in cross section, they appear as sharply circumscribed dots, and do not give the appearance of folds in cross section. The view that the neuroglia fibers are completely dissociated from the protoplasm of the neuroglia cells, as expressed by Weigert and others who have accepted this view concerning the structure of neuroglia tissue, appears to me, when viewed in the light of my own observations, as untenable, since at least a large proportion of the neuroglia fibers are in permanent i elation to the protoplasm of neuroglia cells; whether this is true for all the fibers is a question which, at present, can be answered only by way of hypothesis. A parallel may here be drawn between the relation of the neuroglia fibers to the protoplasm of the neuroglia cells and other fibrillar structures to the proto- plasm of cells of ectodermal origin, in the great majority of which, as is known, the fibrillar structures are embedded in the protoplasm, a fact which I need not further discuss, since Joseph has given it consideration in his contribution to the structure of the neuroglia tissue of invertebrates. I am also tempted to consider here the mode of development of coUaginous connective tissue and of other connective tissue fibers and of their relation to the cellular ele- ments of fully developed connective tissue ; however, since the known methods for differential staining of neuroglia fibers are not applicable to embryonic tissue, and since we possess no accurate data concerning the mode of develop- ment of the nem:oglia fibers (in chrome-silver preparations, the relation of neuroglia fibers to neuroglia cells cannot be ascertained), and since a com- parison of the mode of development of the fibrillar structures of these tissues and their relation to the cellular elements would result in speculation, this may be dispensed with. I must, also, leave unanswered for the present the question whether the neuroglia fibers of fully formed neuroglia tissue of the adult possess the property of proliferation (by longitudinal splitting) and retain the power of growth, and if it should be concluded that the neuroglia fibers are in permanent relation with the protoplasm of neuroglia cells, whether knowing this fact would aid in finding the correct answer. That neuroglia fibers, even under normal conditions, differ markedly in thickness was observed by Weigert, who also states that in certain pathologic conditions, and espe- 618 STUDIES ON NEUROGLIA TISSUE. cially in the cerebral cortex of cases of progressive paralysis, there are found especially thick neiiroglia fibers, indicating that these fibers retain the power of growing in thickness. Appearances indicating a division of neuroglia fibers I have not observed in the neuroglia tissue of the vertebrates studied by me. Brief consideration may finally be given to the views expressed by Ranvier and Joseph, concerning the structure of neuroglia tissue and the relation of neurogha fibers to neuroglia cells. The former, many years ago, working with methods much simpler than the ones now at our disposal, reached the conclusion that the neuroglia tissue of the spinal cord consisted of neuroglia cells and neurogha fibers, and that the processes of neuroglia cells described by observers who preceded him in the study of this tissue were fibers differ- entiated from the protoplasm of the cells. One must express admiration at the accuracy of these observations of Ranvier, when we consider the simple tech- nical methods at his disposal, even though it be conceded that he was not so fortunate in his interpretation of the structure of the neuroglia tissue of the brain, which he described as composed of cells with the neuroglia fibers as cell processes. Joseph's investigations on the structure of the neuroglia tissue pertain more particularly to a study of this tissue as found in the invertebrates. He found the neiiroglia tissue of invertebrates to consist of nem'oglia cells and neurogha fibers, and makes the statement concerning the relation of the latter to the former, which I shall give as found in the original : " Die Lags der Gliafasern ist eine solche, dass sie imi der Aussenschichte des Zellplasmas an — oder eingelagert sind; sie gehen nicht, in der Nahe des ZelUeibes angelangt, in das Protoplasma iiber, wie dies Erik Miiller annimmt, sondern behalten ihren stofflichen Character bei, indem sie weiterlaufend von einem Fortsatz auf den andern iibergehen (Weigert)." Joseph expresses here very clearly the posi- tion which the writer has taken concerning the relation of the neuroglia fibers to the neurogha cells. His conclusions are of especial interest to me, since they are based, in the main, on observations made on the neuroglia tissue of invertebrates, while the observations recorded in this communication pertain to the neurogha tissue of vertebrates. Conclusions. I may, in conclusion, present the following very brief summary of the results obtained by me in the investigation of the structure of the neuroglia tissue of vertebrates : 1. The neurogha tissue of vertebrates consists of neuroglia cells and neu- roglia fibers. 2. The shape of the neurogha cells varies. In cells that are found singly or in small groups, it depends to a certain extent on the arrangement of the structural elements surrounding the cells. The protoplasm of neurogha cells G. CARL HUBER. 619 may be brought to view with suitable staining methods, as also the proto- plasmic processes; the latter vary in shape, size, and number. The majority of the neuroglia cell nuclei are of vesicular structure, with chromatin in granules. They represent the nuclei of neuroglia cells which show a definite and characteristic relation to neuroglia fibers. A varying number of rela- tively small nuclei, which stain more deeply, are also found. They generally appear as free nuclei, without apparent protoplasmic covering. They are regarded as belonging to undeveloped neuroglia cells and show no special relation to neuroglia fibers. 3. The neuroglia fibers differ in their physico-chemical properties from the protoplasm of neuroglia cells, from which they are completely differentiated, lis is shown by their differential staining. The neuroglia fibers are, however, in permanent relation with the protoplasm of neuroglia cells, either in close contact with it or embedded in the peripheral layer of the protoplasm of these cells. They are not processes of the neuroglia cells, but form continuous fibers, which are not interrupted by the protoplasm of the cells. 4. The ependj'mal fibers show a similar relation to the protoplasm of the ependymal cells. LITERATURE. Aguerre. — Untersuchungen iiber die menschliohe Neuroglia. Archiv f. mik. Anat. u. Entwickelungsg., Bd. LVI, 1900. Andriezen. — ^The Neuroglia Elements in the Human Brain. Brit. Med. Journ., Vol. II, 1893. Babes. — Ueber Neurogliawucherung. Deutsche Med. Wochenschr. Jahrg. XXVII, 1901. Benda. — Erfahrungen iiber Neurogliafarbungen und eine neue Farbungsmethode. Neurologisches Centralbl., Bd. XIX, 1900. Benda. — Die Mitoohondriafarbung und andere Methoden zur Untersuohung der Zellsubstanzen. Anat. Anzeiger, Brganzungsheft, Bd. XIX, 1901. BoU. — Die Histologie und Histiogenese der nervosen Centralorgane. Archiv f. Psychiatr. u. Nervenkr., Bd. IV, 1874. Bonome. — Bau und Histogenese des Pathologischen Neuroglia-Gewebes. Virchow's Archiv., Archiv. f. Path. Anat., Bd. CLXIII, 1901. Dejters. — Untersuchungen iiber Gehim und Riickenmark des Mensohen und der Saugetiere. Braunschweig, 1865. Eurich. — Studies on the Neuroglia. Brain, Vol. XX, 1897. Eurich. — Contribution to the Comparative Anatomy of the Neuroglia. Journ. Anat. Phys., Vol. XXXII, 1898. Golgi. — Beitrage zur feineren Anatomic der Centralorgane des Nervensystems. Unter- suchungen iiber den feineren Bau des centralen und peripherisohen Nervensystems. Fischer. Jena, 1894. Jastrowitz. — Ueber Encephalitis und Myelitis im ersten Kindesalters. Archiv f . Psych, und Nervenkr., Bd. Ill, 1871. Joseph. — Untersuchungen ueber die Stiitzsubstanzen des Nervensystems. Arbeiten aus d. Zoolog. Instit. d. Universitat, Wien., Bd. XIII, 3 Heft, 1902. 620 STUDIES ON NEUROGLIA TISSUE. V. KoUiker. — Handbuch der Gewebelehre des Menschen. 6 Aufl., Bd. II, p. 136, 1893. R. Krause. — Untersuchungen iiber die Neuroglia des Affen. Anhang z. d. Abh. d. Konigl. Preuss. Akad. der Wissensch. zu Berlin, 1899. R. Krause and Aguerre. — Untersuchungen iiber den Bau des Menschliohen Riicken- markes mit Besonderer Beriicksichtigung der Neuroglia. Aaat. Anzeiger, Bd. XVIII, 1900. V. Lenhoss^k. — Die StiitzzeUen des Riiokenmarkes, Chapter VI, pp. 176 to 247 inclu- sive. Der feinere Bau des Nervensystems im Lichte neuester Forschungen, 1895. (v. Lenhossfek gives here a very complete review of the Golgi literature pertaining to neuroglia tissue. References to many of tlie papers, to which I have incidentally referred, will be found there.) MaUory. — Ueber Gewisse Eigenthiimliche Farbereactionen der Neuroglia. Centralbl. AUg. Path, und Path. Anat., Bd. VI, 1895. MaUory. — ^A Method of Fixation for Neuroglia Fibers. Journ. of Experiment. Med., Vol. II, 1897. Erik MiiUer. — Studien iiber Neuroglia. Arch. f. mik. Anat. u. Entmckelungsg., Bd. LV, 1899. Pollack. — Einige Bemerkungen iiber die Neuroglia und Neurogliafarbung. Arch, f. Anat. u. Entwickelungsg., Bd. XL VIII, 1897. L. Ranvier. — Technisches Lehrbuch der Histologic, iibersetzt von Nicati and Wyss., 1888. Reinke. — Ueber die Neuroglia in der weissen Substanz des Riickenmarks vom er- wachseneu Menschen. Arch. f. mik. Anat. u. Entwickelungsg., -Bd. L, 1897. Robertson. — Note on Weigert's Theory Regarding the Structure of the Neuroglia. Journ. of Ment. Sc, Vol. XLIII, 1897. Stroebe. — Ueber Entstehung und Bau der Gehirngliome. Beitrage z. Path. Anat. u. z. allgem. Path., Bd. XVIII, 1895. Storch. — Ueber die pathologisch-anatomisohen Vorgange am Stiitzgeriist des Cen- tralnervensystems. Archiv. f. Path. Anat., Bd. CLVII, 1899. Taylor. — A Contribution to the Study of Human Neuroglia. Journ. Experiment. Med., Vol. II, 1897. Rudolph Virchow. — Ueber das Granulierte Arisehen der Wandungen der Gehirnven- trikel. Zeitschrift fur Psychiatrie., 1846. (Not seen in the original. Quoted from Weigert's large monograph, 1895.) (Virchow also considers this tissue in connection with other intercellular and fibrillar structures in Arohiv. f. Path. Anat., Bdr. Ill, V, VI, and VIII.) Weigert. — Beitrage zur Kenntnis der normalen menschlichen Neuroglia. Fest- schrift, Frankfurt a. M., 1895. (In this monograph, Weigert considers in full and critically the earlier literature pertaining to neu- roglia tissue. I am greatly indebted to him for many references which otherwise might have been over- looked.) Yamagiwa.— Eineneue Farbung der Neurogha. Archiv. f. Path. Anat., Bd. CLX, 1900. Plate Fig. 7. Fig. 10. Fip-. 9, Plate Huber. Fig. 14. Fig. 20. IV Fig. 25, 'iff. 2S, fis. 26. /. ig, 27. Fiff. 28, Fig. 35. Fig. 40. Fig. 37. Fiff. 4 2, Fig. 33 Plate II. — (Cat and Rabbit.) Figs. 13 and 14. — ^Two typical neuroglia cells with protoplasm and protoplasmic pro- cesses clearly shown and with nuclei of vesicular structure ; numerous neuroglia fibers in relation with the cells. Prom cross section of the white matter of the spinal cord of the cat. Fig. 15. — ^Two typical neuroglia cells, with certain neuroglia fibers, each of which is in relation with both cells; from cross section of the white matter of the spinal cord of the cat. Fig. 16. — Neuroglia cell with eccentrically placed nucleus; from cross section of spinal cord of cat. Fig. 17. — NeurogUa cell with two nuclei ; from white matter of spinal cord of cat. Fig. 18. — A group of neuroglia cells, arranged in a short column, from the white matter of longitudinal section of spinal cord of cat. One relatively small, deeply staining neuroglia cell nucleus is seen, about the middle of the figure. Fig. 19. — Longitudinal section of neuroglia septum, from white matter of spinal cord of cat. Figs. 20, 21, and 22. — Typical neuroglia cells with protoplasm and protoplasmic pro- cesses and with nuclei of vesicular structure, from cross section of the spinal cord of the rabbit. Fig. 23. — Relatively large neuroglia cell, with large nucleus, of vesicular structure, from the peripheral portion of the white matter of the spinal cord of a rabbit as seen in cross section. Fig. 24. — ^A group of neurogha cells, arranged in the form of a short column, from longitudinal section of the white matter of the spinal cord of a rabbit; one small, deeply staining nucleus is seen to the right of the figure. Plate III. — (,Frog, Tortoise, Dove, Man.) Figs. 25, 26, and 27. — Tjrpical neuroglia cells from cross section of the white matter of the spinal cord of a frog. Fig. 28. — Neuroglia cell with vesicular nucleus, without apparent protoplasmic cov- ering; from cross section of the white matter of the spinal cord of the frog. Fig. 29. — ^A group of neuroglia cells as seen in longitudinal section of the white matter of the spinal cord of a frog; at the top of the figure, one relatively small, deeply staining nucleus. Fig. 30. — Three typical neuroglia cells from cross section of the white matter of the spinal cord of the tortoise. Fig. 31. — ^A group of neurogUa cells arranged in the form of a short column, from a longitudinal section of the white matter of the spinal cord of the tortoise. Fig. 32. — Five neuroglia cells, shown in their relative position, from the white matter of the spinal cord of a dove. Fig. 33. — Neuroglia cells from longitudinal section of the white matter of the spinal cord of a dove. Figs. 34 and 36. — ^Typical neuroglia cells, with protoplasm and protoplasmic processes and with vesicular nuclei, from cross sections of the white matter of the human spinal cord. Figs. 36, 37, and 38. — Relatively large neuroglia cells with large vesicular nuclei, from cross section of white matter of human spinal cord. Fig. 39. — A group of neuroglia cells, situated between the nerve fibers of a bimdle of nerve fibers seen in longitudinal section, from cross section of human spinal cord. Fig. 40. — ^Large, irregular neuroglia cell with long protoplasmic branch, -mth which numerous neuroglia fibers are in relation. From white matter of cross section of human spinal cord. Fig. 41. — ^Neuroglia cell with two nuclei, from white matter of human spinal cord, in cross section. Fig. 42. — Neuroglia cell with large vesicular nucleus, with relatively small amount of protoplasm and short protoplasmic processes. Numerous neuroglia fibers, observed in cross and obUque sections are seen in relation with the protoplasm and protoplasmic processes.