(JJocnell Httiueraitg Hihrarg Dtljara. Ncm $nrk THE CHARLES EDWARD VANCLEEF MEMORIAL LIBRARY BOUGHT WITH THE INCOME OF A FUND GIVEN FOR THE USE OF THE ITHACA DIVISION OF THE CORNELL UNIVERSITY MEDICAL COLLEGE MYNDERSE VAN CLEEF CLASS OF 1874 1921 Cornell University Library QL 835.K93 The anatomy and physiology of capillarie 3 1924 001 323 660 B Cornell University 9 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/cu31924001323660 YALE UNIVERSITY MRS. HEPSA ELY SILLIMAN MEMORIAL LECTURES THE ANATOMY AND PHYSIOLOGY OE CAPILLARIES SILLIMAN MEMORIAL LECTURES PUBLISHED BY YALE UNIVERSITY PRESS ELECTBICITY AND MATTER. By Joseph John Thomson, D.Sc, LL.D., Ph.D., F.R.S., Felloiv of Trinity College and Cavendish Profes- sor of Experimental Physics, Cambridge University. (Fourth printing.) THE INTEGRATIVE ACTION OF THE NERVOUS SYSTEM. By Charles S. Sherrington, D.Sc, M.D., Hon. LL.D. Tor., F.R.S., Bolt Professor of Physiology, University of Liverpool. (Sixth printing.) RADIOACTIVE TRANSFORMATIONS. By Ernest Rutherford, D.Sc, LL.D., F.R.S., Macdonald Professor of Physics, McGill University. (Second printing.) EXPERIMENTAL AND THEORETICAL APPLICATIONS OF THERMODYNAMICS TO CHEMISTRY. By Dr. Walter Nernst, Professor and Director of the Institute of Physical Chemistry, Uni- versity of Berlin. PROBLEMS OF GENETICS. By William Bateson, M.A., F.R.S., Director of the John Innes Horticultural Institution, Merton Park, Surrey, England. (Second printing.) STELLAR MOTIONS. With Special Reference to Motions Determined by Means of the Spectrograph. By William Wallace Campbell, ScD., LL.D., Director of the Lick Observatory, University of California. (Second printing.) THEORIES OF SOLUTIONS. By Svante Arrhenitjs, Ph.D., ScD., M.D., Director of the Physico-Chemical Department of the Nobel Insti- tute, Stockholm, Siceden. (Third printing.) IRRITABILITY. A Physiological Analysis of the General Effect of Stimuli in Living Substances. By Mas Verworn, M.D., Ph.D., Pro- fessor at Bonn Physiological Institute. (Second printing.) PROBLEMS OF AMERICAN GEOLOGY. By William North Rice, Frank D. Adams, Arthur P. Coleman, Charles D. Walcott, Walde- mar Lindgren, Frederick Leslie Ransome and William D. Matthew. (Second printing.) THE PROBLEM OF VOLCANISM. By Joseph Paxson Iddings, Ph.B., ScD. (Second printing.) ORGANISM AND ENVIRONMENT AS ILLUSTRATED BY THE PHYSIOLOGY OF BREATHING. By John Scott Haldane, M.D., LL.D., P.R.S., Felloiv of New College, Oxford University. (Second printing. ) A CENTURY OF SCIENCE IN AMERICA. With Special Reference to the American Journal of Science 1818-1918. By Edward Salisbury Dana, Charles Schuchert, Herbert E. Gregory, Joseph Barrell, George Otis Smith, Richard Swann Lull, Louis V. Pirsson, William E. Ford, R. B. Sosman, Horace L. Wells, Harry W. Foote, Leigh Page, Wesley R. Coe and George L. Goodale. THE EVOLUTION OF MODERN MEDICINE. By Sir William Osler, Bart., M.D., F.R.S. (Second printing.) RESPIRATION. By J. S. Haldane, M.D., LL.D., F.R.S., Fellow of New College, Oxford, Honorary Professor, Birmingham University. AFTER LIFE IN ROMAN PAGANISM. By Franz Cumont. THE ANATOMY AND PHYSIOLOGY OF CAPILLARIES. By August Krogh, Ph.D., LL.D., Professor of Zoo-physiology, Copenhagen Uni- versity. THE ANATOMY AND PHYSIOLOGY OF CAPILLARIES BY AUGUST KROGH, Ph.D., LL.D. PROFESSOR OF ZOO-PHYSIOLOGY COPENHAGEN UNIVERSITY NEW HAVEN YALE UNIVERSITY PRESS LONDON ■ HUMPHREY MILFOKD ■ OXFORD UNIVERSITY PRESS MDCCCCXXII {AeA'.\ £_ — 1_ *-. p >\ '4-Jl 1 f "■& ' L ' "^ MB 5sJ~^ ~~^~'^~~ x p^ W^ v *^ J 1 ^ — X^^Sv ~~ ; \ \P ^, Vi =7 1 ^^S ""V ^s- : r \\ k^sT/S3 |crj - _- " 'n - •— ■} ^_ > _^ Pig. 1 . Small arteries (black), capillaries and veins from striated muscle. After Spalteholz. about the fibres. The capillaries unite into venules in- tercalated regularly between the arterioles, and the whole system of veins reproduces and follows almost exactly that of the arteries. All the veins down to the smallest branches are provided with valves allowing the blood to flow in the direction of the heart only. Short pieces of secondary arteries and veins, with DISTRIBUTION 7 arterioles, venules and capillaries, are shown in Fig. 1, reproduced from Spalteholz. When the muscle contracts its form is greatly al- tered, the fibres becoming much shorter and propor- tionally thicker. The whole of the vascular system is beautifully adapted to these changes : the arterial and 7 rfErS^/f gg* ^Hffi&" mQ®& y|jp§ ; 1- •■% m^mwm §fi^5HBr A '?W'H IKIigSfl i •■JflHgpJP^f 4rejPPW( Fig. 2. Transverse section from the injected m. gastrocnemius of a horse. X156. Fig. 3. Transverse section of injected muscle from the tongue of a cat. X268. venous networks insure the supply and drainage of almost every point, even if a number of anastomoses are temporarily blocked. The capillaries, which in the resting muscle are practically straight, become much twisted. The blood is driven out by compression from a number of the venous branches, and, when the muscle relaxes again, these can be filled from their peripheral ends only. Since muscular contractions are usually 8 CAPILLARIES more or less regularly alternated with relaxations, the system of valves makes of the veins of every muscle a very effective pump, capable of maintaining a low pressure in the muscle capillaries. The significance of this arrangement will come up for study later, but at present our attention must be focused on the capil- laries. It is apparent from Fig. 1 that sections, cut at right angles to the muscle fibres, must represent the capillaries as dots which can be counted and the dis- tribution of which can be studied. Such transverse sections are given in Figs. 2 and 3 ; and an inspection of them shows that the capillaries are present in very large numbers and are distributed among the muscle fibres with conspicuous regularity. A quantitative test of the regularity of the arrangement can be obtained by counting the number of capillaries in a large num- ber of small equal areas chosen at random from sec- tions of the same muscle. I give as an example a series of countings, each made on an area of 0.0300 square mm. (mm. 2 ), from five different transverse sections of the m.gastrocnemius of a horse. I S o 4 5 45 34 38 38 31 40 34 42 43 33 42 40 43 47 43 41 46 41 49 39 44 44 46 33 36 36 41 49 38 Average, 42 39 42 42 36 An inspection of this table shows the remarkable regularity of distribution, and, when it is treated mathematically, the average number of capillaries in the area measured works out as 40.5 ± 5, a dispersion of not more than 12 per cent. Dividing by 0.03 we get the number of capillaries per square mm. trans- DISTRIBUTION 9 verse section as not less than 1350 with a mean error of ± 31. The transverse section of an ordinary pin or hairpin is about 0.5 mm. 2 It requires some mental effort to conceive how there can be room within such a pin for about seven hundred parallel tubes carrying- blood in addition to about two hundred muscle fibres. In other animals the number of capillaries per mm. 2 may be even larger. It is well known that mammals have a higher metabolism than cold-blooded verte- brates, and small mammals a higher metabolism than large ones, and there appears to be some relation . between the metabolic activity and the number of capillaries per mm. 2 of muscle. In a dog's m.semimem- branosus the number worked out from 30 countings as 2630 ± 51, while the dispersion or ' ' standard devia- tion" of the single counting was not more than 10.6 per cent. Even larger figures were found for guinea pigs' muscles, and I have no doubt that in the smallest mam- mals the number of capillaries per sq. mm. is well above 4000. In cold-blooded animals, such as the cod and the frog, much smaller figures are found, averag- ing only about 400. To obtain some insight into the meaning of figures such as these let us consider very briefly the problem concerning the supply of oxygen to the muscular tis- sue. The oxygen molecules have to travel outwards from the capillaries, and the longest distance a mole- cule has to go must be half the distance between neigh- bouring capillaries, denoted R. This works out in the case of the frog's muscle (400 capillaries per mm. 2 ) as R = 28/u (calculated from the centre of each capillary) and in the case of the dog's muscle (2600) as R = 11^. I shall show later in some detail how these figures can be utilized for a calculation of the oxygen pressure head necessary for supplying the muscles. If we con- 10 CAPILLARIES sider the exchange of dissolved substances between the blood and the muscle lymph, this depends evidently on the available capillary surface, and, assuming the av- erage diameter of capillaries 2r as equal to the diam- eter of a red blood corpuscle, we obtain the following figures for the total surface of capillaries in 1- cubic centimeter of muscle : Approxi- Muscle mate capil- Surface Surface of weight laries per R ir cm. 2 Volume 1 cc. blood Muscle kilos. sq. mm. ^ >>■ per cm. 3 per cent. cm. 2 Frog, 0.05 400 28 15 190 7.1 2700 Horse, 500 1400 15 5.5 240 3.3 7300 Dog, 5 2600 11 7.2 590 10.6 5600 On the same assumptions the volume of blood in the muscle capillaries works out as between 3.3 per cent (horse) and 10.6 per cent (dog) of the muscle volume, and the surface of 1 cc. blood contained in capillaries as 2700 cm. 2 (frog) to 7300 cm. 2 (horse).' It is evident that very large exchanges of substances can take place in a short time through such enormous surfaces. Sup- posing a man's muscles to weigh 50 kg. and his capil- laries to number 2000 per sq. mm., the total length of all these tubes put together must be something like 100,000 kilometers or 2y 2 times round the globe and their total surface 6300 sq. meters.- It is evident that much more work could be done, and ought to be done, on what I should like to term the quantitative anatomy of muscle capillaries. A number of different animals ought to be examined and a num- ber of different muscles from each animal. The regu- larity or otherwise of the capillary arrangement ought to be made out and definite relationships established between the capillary supply and the amount of work required of muscles. I would suggest, for instance, com- parisons between the capillaries in the muscles of the & DISTRIBUTION 11 hind legs and the hearts in domesticated rabbits and hares. The supply of blood to the human skin. The vascular system of the skin has also been very carefully studied by Spalteholz (1893) and illustrated by numerous figures, the magnifications of which are correctly given, as Professor Spalteholz himself has very kindly informed me. The skin is supplied from the underlying tissue through a large number of small arteries. In all places where the skin is movable these arteries are very sinuous and will be able to supply blood even when greatly stretched. In the deepest layer of the cutis the arteries form a richly anastomosing irregular plexus, from which small arteries rise perpendicularly through the skin, to form, a little below the papillae, the sub- papillary arterial plexus. This plexus is on the whole regular, with oblong meshes, more or less parallel to the papillary ridges. The meshes are of somewhat dif- ferent size in different regions, varying from 0.2 to 2 mm. 2 They are smallest in the hand and foot, where the skin is regularly exposed to pressure. From the subpapillary plexus still smaller arteries spring to supply the papillary capillaries. These do not anastomose, and each supplies a small but variable number of papilla?. Spalteholz has found the area sup- plied to vary in the sole of the foot between 0.04 and 0.27 mm. 2 Each papilla is normally provided with a central capillary loop, the arterial branch of which is gen- erally narrow, while the tip of the loop and the venous branch are often 0.02 mm. or more in diameter. The length of the loops varies generally from 0.2 to 0.4 mm. The number of capillaries is very small compared with the muscles, but regular countings have, I believe, 12 CAPILLARIES never been made, and the only figure at my disposal is 20 capillary loops counted in my laboratory by Miss Carrier (1922) in an area of about 0.5 mm. 2 on the back of the human hand. The papillary capillaries supply the germinative layer of the epidermis with all the substances necessary for its continuous growth. '^PSfc^ -22 ?=*&. *H&f§iO i^i\ i i. j ^T s I ^"^ / ^21l *>-> ^^^-.«'- : :;. ^■Jl' i ■'' a '' v vt-5 (JSMjr t ,'■ 1 .s^ V - ^ \ Pig. 4. First subpapillary venous plexus with a few nar- row arterial branches and capillary loops. After Spalteholz. X41. According to the countings made the average distance from the capillaries to this layer should be from 50 to 100m. The venous branches of the papillary capillaries combine to form venules which return to a first sub- papillary plexus of small veins, situated just below the papillae. This plexus, with some of the arterial branches at the same level and a few of the capillaries, is shown DISTRIBUTION 13 in Fig. 4. All the venules making up this plexus are of about the same size and only a few hundredths of a mm. wide. The first subpapillary plexus of veins is connected by very numerous short anastomoses with a second close-meshed network, lying at about the level of the arterial subpapillary plexus and made up, like the first, of very narrow venules. Passing down through the cutis Spalteholz describes two more plexuses with larger meshes and made up on the whole of larger veins. In the last of these, lying in the boundary zone between the cutis and the sub- cutis, valves begin to appear, while in all the veins throughout the skin itself valves are absent. It is a very characteristic feature of the vascular supply to the skin that there are practically no capil- ( laries, except those in the papillae. Throughout the; rest of the skin the exchange of substances must take place through the walls of the veins, which are ex- tremely thin. In the deeper layers of the skin the larger veins are accompanied by numerous smaller vessels, branching off from them and returning to them at a lower level. The vascular surfaces available for exchange of sub- stances in the human skin have not been made out. The total surface of the capillaries proper is extremely small, being something between 1 and 2 cm. 2 per cm. 2 of the skin surface. Even when the whole surface of the veins is assumed to be available for exchange the total surface will fall far short of the capillary surface available in muscles, and the average dis- tances from the tissue elements to the vessels will be much larger. This is no doubt the anatomical expres- sion of the fact that the metabolic level of the skin is low and probably not very variable. This point will be referred to again in Lecture XI. 14 CAPILLARIES The capillary system in the intestinal villi. Through the columnar epithelium covering the sur- face of the villi in the small intestine practically the whole of our daily food is transported. Once inside the epithelium of a villus there are two ways open to the absorbed solution of food-stuffs. The dissolved sub- stances can pass either into the network of capillaries below the surface or into the lymph system, repre- sented in each villus by the central lacteal. The distri- bution of the substances between these two channels must depend to a large extent on the surface develop- ment of the capillary system and its relation to the epithelial surface of the villus. No real insight into the physiological processes can be obtained unless these surfaces are, at least approximately, made out. An attempt in this direction was made long ago by Mall (1887), but it cannot be denied that his calcula- tions are rather summary and his results approximate only. Mall found in the dog's small intestine 16 villi per mm. 2 These are approximately cylindrical, with a height of 0.5 to 0.6 mm. and a diameter of 0.2 to 0.25 mm. I calculate the surface of each villus to be about 0.43 mm. 2 Mall describes a small artery entering each villus and running right up to the top, where it splits up sud- denly into 15 to 20 capillaries which travel down along the internal sin'face of the epithelium, forming a close- meshed network (Fig. 5). The average diameter of the single capillaries is given as 8/*. At a height varying between one-third and one-half of the height of the villus, part of the capillary blood is taken up by veins, but the capillary network continues downward, with the difference that there are fewer capillaries and their average diameter is only 5/*. DISTRIBUTION 15 According to Mall the capillary system in the upper two-thirds of the villus could be represented by 30 parallel tubes 0.4 mm. long and 8/* in diameter, and the capillaries in the lower third by 15 tubes of 0.2 mm. length and 5/* diameter. From these figures the total Fig. 5. Villi from the small intestine of dog. Height about 0.5 mm. After Mall. capillary surface in each vilhis works out as 0.35 mm. 2 or 82 per cent of the surface of the epithelium. On each mm. 2 of the internal surface of the intestine we find an epithelial surface of the villi amounting to 7 mm. 2 and a capillary surface of 5.6 mm. 2 In my laboratory Dr. Vimtrup has begun a study of the quantitative anatomy of the absorbing system in 16 CAPILLARIES the rabbit's intestine. The results will be published in detail at a later date,' and I shall give only the measure- ments undertaken on a single villus in the duodenum, Fig. 6. Villus from small intestine of rabbit, showing capillary network of one side. X68. shown in Fig. 6. As seen in the figure the villi in this part of the rabbit's intestine are not cylindrical, but have an elliptical transverse section. The capillary sys- tem is very irregular compared with the one described DISTRIBUTION 17 by Mall. It is supplied by two small arteries and drained mainly by four veins. Tbe larger vessels run at some distance from tbe epitbelial surface, but it is of considerable importance from a physiological point of view that almost all the capillaries are practically cemented on to the bases of the epithelial cells. On a large scale drawing, on which only the capil- laries of one side of a villus have been represented by single lines, the total length of capillaries can easily be determined by running a small measuring wheel along them. On a surface of 0.84 mm. 2 the total length of capillaries amounted to 2.47 mm. The average diame- ter of each capillary was found in the injected speci- men to be 12|U, giving a total surface of 0.93 mm.- or 109 per cent of the epithelial surface. Multiplying the total length of the capillaries by their average diame- ter, instead of by their circumference, we obtain the projection of capillaries on the epithelial sur- face as 0.295 mm. 2 or 34 per cent, which means that about one-third of the epithelial cells will give off the substances passing through them directly into the capillaries, while two-thirds will deliver them into the intercapillary spaces. The total surface of one villus works out as 2.2 mm. 2 ,, with a capillary surface of 2.4 mm. 2 and a capillary projection of 0.75 mm. 2 With from 6 to 10 (average about 8) villi per mm. 2 internal surface of the duo- denum we find an epithelial surface per mm. 2 of 17.6 mm. 2 and a capillary surface of 19.2 mm. 2 , both of which figures are more than double those found by Mall for the dog. Some use will be made of these results in the final lecture, where the mechanism of the distribution of absorbed substances between the blood and the lymph will come up for discussion. 18 CAPILLARIES The rete mirabile annexed to the oxygen gland in the eel. I shall finally describe briefly a very peculiar ar- rangement of capillaries, a true rete mirabile, found in the swim bladder of fishes and presenting the most extreme development of capillary surface known to me. In the wall of the swim bladder of most fishes there is an oxygen gland, the function of which is to take oxygen from the blood and secrete it into the swim bladder, where the oxygen pressure can become very high. Between the vessels of the oxygen gland and the Fig. 7. Diagram of rete mirabile and capillary loops from swim bladder of eel. general circulation of the fish the organ of which I am now speaking is interposed. Its form varies greatly, but its essential structure is the same everywhere, accord- ing to Woodland (1911). The description I am about to give refers especially to the organ of the eel and is rep- resented in the diagram, Fig. 7. The artery supplying the oxygen gland splits up at a definite point into nu- merous branches, which divide further into an enor- mous number of parallel capillaries. These capillaries run as straight tubes for a definite distance and then suddenly unite again and form an artery which goes to the gland, where it supplies an ordinary network of arterioles, capillaries, venules and veins, uniting finally into a single vein, which comes back to the distal end of the arterial rete, splits up just like the artery into DISTRIBUTION 19 straight parallel capillaries, which are intercalated with the most astonishing regularity between the arte- rial capillaries and which, finally, unite at the proxi- mal end of the rete to form a single vein. Fig. 8. Transverse section of rete mirabile from swim bladder of eel. Arterial capillaries gray. Venous capillaries black. When the vessels are cut open between the gland and the rete and cannulas introduced into the proxi- mal artery and vein it is easy to inject the rete with two differently coloured gelatines and to obtain prepa- rations which, when cut in suitable sections, will give pictures like the one reproduced in Fig. 8. The arterial capillaries are gray, the venous black, and you will notice how each venous capillary is regularly sur- 20 CAPILLARIES rounded by a number of arterial capillaries which are somewhat narrower. Countings and measurements on such sections have given the following results in the case of a medium- sized eel. There were two parallel retes, each of which had a cross-sectional area of 8.0 sq. mm., while the capillaries of which it was composed had a length of 4 mm., a very great length when it is remembered that the capillaries in muscles, which are otherwise among the longest, are seldom more than y 2 mm. long. The total volume of the two capillary systems is therefore only 64 cubic millimeters. What surfaces can it be possible to get into a vol- ume of 64 mm. 3 , the size of a drop of water? As the capillary system is of the most extraordinary regularity throughout, I have only counted the capil- laries in two contiguous narrow rectangles 0.33 mm. long and 0.033 mm. broad. In the one I found 60 venous and 77 arterial, in the other 60 venous and 81 arterial capillaries, giving a total of 120 venous and 158 arterial in the area of 0.0218 mm. 2 , or per mm. 2 of the cross- section 5500 venous and 7250 arterial capillaries. For the whole of both organs this will give 88,000 venous and 116,000 arterial capillaries with aggregate lengths of 352 and 464 meters. In another area the capillaries were drawn by means of a prism at a high magnification to make out the relative cross-sectional areas of venous and arterial capillaries and of interstitial tissue, including the capillary walls, respectively. The area measured in this way was only 0.00386 mm. 2 It contained 21 venous capillaries occupying an area of 0.00149 mm. 2 , or 38.6 per cent of the whole, and 34 arterial, with an area of 0.00108 mm. 2 , or 28.1 per cent. The interstitial tissue works out to be just 33.3 per cent of the whole. On the basis of the countings given above the area should con- DISTRIBUTION 21 tain 21 venous and 28 arterial capillaries, a very satis- factory agreement with the numbers 21 and 34 ac- tually found. Dividing by the average number we find 71 and 39 square microns (p 2 ) as the average cross-sectional area of a venous and an arterial capillary, respectively. Taking these cross-sections to be circles, which is, as you will see on Fig. 8, not quite correct, we find the diameters to be respectively 9.5 and 7.1/u and the cir- cumferences (corrected for their not being circles), 30 and 22.5/*, respectively. When these figures are multi- plied by the aggregate lengths we obtain the interest- ing result that the venous and arterial capillary sur- faces in these organs are equal, being respectively 106 and 105 cm. 2 , while the total volume of the venous capillaries is about 25 and that of the arterial about 18 cubic millimeters. A plea for the study of quantitative anatomy. The measurements and calculations here given have been undertaken because they were directly required for physiological researches, carried out or planned in my laboratory. They are, all of them, crude and not very accurate, because the work was quite outside our regular routine, and we did not want to go any farther than just necessary for our immediate purposes. I believe, however, that, when taken up in earnest by competent anatomists, the field of quantitative anatomy will prove to be a rich and fruitful one. Many determinations of vascular and glandular sur- faces are urgently needed as a basis for quantitative physiological work. I shall mention as instances only the glomerular and tubular surfaces of the kidney, the active surfaces of various glands and the dimensions and numbers of sarcomeres in muscle-fibres. Quite apart, however, from the needs of plrysiology, I can- 22 CAPILLARIES not but think that quantitative anatomy will prove a very attractive subject for its own sake, especially when conducted as a comparative science. In a cat's kidney, weighing 13 g\, Miller and Carlton (1895) have counted 16,000 glomeruli. In a dog's kidney, weighing 34.5 g., Brodie (1914) found 142,000 glomeruli. I have been unable to find any error in either of these calcu- lations, and I for one feel very curious about the possi- bility and significance of such an enormous difference. LECTURE II THE INDEPENDENT CONTRACTILITY OF CAPILLARIES THAVE described in the first lecture the wonderful arrangement of capillaries and tried to make it clear that enormous surfaces are by them made available for the exchange of substances between the blood and the tissues. I alluded also very briefly to the problem with which we are now confronted : Supposing these facilities for exchange to be necessary and just sufficient to provide for the needs of an organ when that organ is doing work at its maximum rate, how can they avoid being wastefully in excess of the require- ments when the organ is absolutely or comparatively at rest ? The current view of the capillary circulation, at least until a few years ago, was that the capillaries are passive, that blood is flowing continuously through all of them at rates which are determined by the state of contraction or dilatation of the corresponding arte- rioles, and that the dilatation of an arteriole will cause a rise of pressure in the corresponding capillaries, which will become passively expanded, to contract again by their own elasticity when the pressure is re- duced. By the varying resistance in the arteries and arterioles the supply of blood to an organ can un- doubtedly be regulated in accordance with its require- ments, but an increase in current must always be accompanied by a corresponding increase in capillary pressure, and when the requirements are small the 24 CAPILLARIES quantity of blood constantly present in a large number of the capillaries would serve no useful purpose. A much more effective distribution would obviously be obtained if the capillaries themselves were contractile, if in a resting organ only a limited number of capil- laries, distributed at suitable regular intervals, were kept open so as to admit blood and provide the neces- sary surface area for the exchange of substances. This hypothetical conception was to me personally the start- ing point and guide in the experimental study of capil- lary contractility. Since, however, I was by no means the first to discover or even to demonstrate that capil- laries show independent contractility, the most suitable course will be to present the evidence in the order in which it was published. The older experiments on capillary contractility. While working on the problem of inflammation, brought about experimentally in the web of the frog, Lister (1858) observed that capillaries could become enormously dilated, but he states definitely that in his opinion the dilatation is brought about by the in- creased pressure brought to bear on their walls by the dilatation of arteries, and the first to notice an independent contractility of capillaries was, so far as I have been able to make out, Strieker (1865), working on the excised nictitating membrane of the frog, in which he observed irregular spontaneous contractions and relaxations of single capillaries. He tried to evoke contractions by suitable stimuli, but it was only very occasionally that he was successful in this. Since the membrane was excised the blood pressure could have nothing to do with the movements observed, but on the other hand the apparent fitfulness of these and the fact that they were observed in conditions which could not be regarded as physiological, did not inspire con- CONTRACTILITY OF CAPILLARIES 25 fidence. His statements were challenged by several authors (Cohnheim, 1867), and, though they were con- firmed and somewhat extended by others, no very defi- nite evidence was forthcoming until the publication of the beautiful researches of Roy and Graham Brown (1879). These authors constructed the ingenious ap- ■ — r s^----' n ^— <^L- K \ ' <*? \ L r 1 ^ T7 b ■■i J 1 Fig. 9. Roy and Brown's apparatus for measuring capillary pressure. paratus shown in Fig. 9 for applying pressure to a transparent tissue, such as the web of the frog, and thereby measuring the blood pressure in the vessels of that tissue. The apparatus consists of a chamber (a) closed below by a glass plate (b) and above by a deli- cate and very flexible, but inelastic, membrane (d). The air pressure inside the chamber can be raised to any desired height through the tube (c) and meas- ured by a suitable manometer. The transparent tissue to be studied is arranged on top of the membrane and pressed up against the adjustable cover-glass (h). In their experiments with this apparatus Eoy and Brown found, as one would expect, that the pressure which was just sufficient to cause the collapse of one 26 CAPILLARIES capillary might be insufficient for its immediate neigh- bours, but they observed further that the pressure relations were constantly changing. "If, for example, we take a curarized frog, and having arranged the compressing apparatus in the manner above described, and having sketched roughly the position and relations of the various capillaries which can be seen in the field of the microscope, we mark on our drawing the order in which the capillaries cease to admit the passage of blood-corpuscles on gradually increasing the extra- capillary pressure; if having done this, we lower the pressure to Avhich the portion of tissue is subjected to 0, and, after leaving everything untouched for, say, half an hour, and again investigate the order in which the capillary vessels of the same part become imper- vious on raising slowly the applied pressure, we usually find that there is a more or less marked differ- ence in this respect between the two observations. Occasionally it is found that those capillary vessels which closed in the one observation with a relatively low pressure on their exterior are, in the other ob- servation, those which remain longest pervious to the blood-flow; and this, although every possible precau- tion has been taken to insure that the conditions should remain unchanged." These changes they rightly ex- plain as due to spontaneous changes in the calibre of the single capillaries, and in some cases they have been able to measure these changes directly. They find, further, that an extracapillary pressure which is just insufficient to bring about the total collapse of a capil- lary has no appreciable influence upon its diameter; that is, it may cause a shrinkage amounting to at most 15 per cent, which shows that elastic expansion by inside pressure does not play more than a compara- tively insignificant part in determining the diameter of capillaries, and this is even more forcibly brought CONTRACTILITY OF CAPILLARIES 27 out by the fact that a sudden diminution of the internal pressure to about does not cause any appreciable contraction of normal or even of exceptionally dilated capillaries. During the later years of the nineteenth century, the doctrine of independent capillary contractility was, as far as I have been able to make out, tacitly or openly accepted by many pathologists and clinicians, who are certainly very often brought face to face with cases of hyperemia which are, to say the least, difficult to understand on the basis of any other theory; but the general attitude of physiologists working on circula- tion problems by means of blood pressure, plethysmo- graphy and other methods was, during the same period, very sceptical and remained so in spite of the fresh, and to my mind conclusive, evidence on the subject brought to light by Steinach and Kahn (1903). Steinach and Kahn again examined excised tissue, the nictitating and other membranes from the frog and the omentum from young cats. By suitable elec- trical stimulation they were able to induce contrac- tions of true capillaries in all these tissues. The con- tractions were sometimes sharply localized, sometimes extending over long distances. By varying the inten- sity of the stimuli they could determine the degree of contraction from a just perceptible narrowing to a complete closing of the capillary. They found that the stimuli acted after a latent period of some seconds, that the contracted capillaries dilated slowly after ces- sation of the stimulation, and that the same capillary could be made to contract repeatedly up to 10 or even 20 times. When the nervous connection of the nictitating membrane with the body was kept intact, though the natural circulation through it was abolished, they were able to induce contractions by stimulating the dorsal 28 CAPILLARIES sympathetic. In this case the time of latency was pro- longed and the capillaries responded only after the arteries had been brought to contraction by the same stimulus. It seems very strange that the experiments of Steinach and Kahn did not arouse any active interest in the physiology of capillaries and were on the whole disregarded by physiologists, though their results were incorporated in one physiological text-book of repute, vis., Tigerstedt's Lehrbuch der Physiologie. In the following years, until 1917, changes in the diameter of capillaries, which appeared to be inde- pendent of the arterial blood pressure, were observed and described by some authors, but the real significance of the facts observed was not as a rule grasped, 1 and generally the sources of error inherent in the micro- scopic observation of capillaries were not recognized or properly guarded against. It will be useful here to anticipate to a certain extent the results of later work and discuss briefly the prin- cipal errors which may vitiate the results of direct microscopic observations of capillaries. When low-power objectives are used it is generally difficult and often impossible to observe directly the walls of capillaries. What one sees is the stream of blood corpuscles, and it is difficult to realize that these do not always occupy the total bore of the channel, but may run in an axial current, as they do in most cases in arterioles and venules. When the velocity of flow is altered, which may happen independently of the state of contraction of the capillary in question, the form of flow may change in consequence and mimic in a very plausible way a change in diameter. When the capillaries in the microscope field become i The experiments of a single author from this period, Heubner (1907), will be referred to in a following lecture. CONTRACTILITY OF CAPILLARIES 29 emptied of corpuscles the impression is produced, and is difficult to resist even when the capillary walls can be distinctly seen, that a contraction has taken place. A reduction in the number of corpuscles, which may even amount to their complete washing away from a certain capillary field, is often brought about by the process which I have termed plasma skimming: When part of a small artery branching from a larger vessel is made to contract without the contraction being com- plete, as indicated in the diagram, Fig. 10, the current Fig. 10. Plasma skimming by contraction of a branch of an artery. of blood through it may seem to cease altogether, and at the same time the corpuscles are washed out from the corresponding capillaries. This is due to the distri- bution of the elements in the larger artery, which shows the well-known axial current of corpuscles with a marginal zone of clear plasma. When the current in the branch artery is sufficiently reduced by the contrac- tion, the plasma from the marginal zone is simply skimmed off by it. In favourable circumstances it can be observed how at each pulse the column of corpuscles bulges into the mouth of the branch artery and retreats again immediately afterwards, leaving only a few cor- 30 CAPILLARIES puscles, which become detached and are passed swiftly along through the contracted portion of the artery. Plasma skimming may, of course, take place to any intermediate extent between giving clear plasma and blood with a normal number of corpuscles, and in the artery skimmed the corpuscle count must rise to a corresponding extent. When pronounced, it may con- stitute, as indicated, a very serious source of error in fe observations on capillary contractility. In a minor degree it is probably of very frequent occurrence in all such organs where the arterioles show frequent "spontaneous" variations in diameter and may be responsible for many irregularities in the numbers of red corpuscles observed in counts from "capillary" blood. The modem study of capillary contractility. In 1917 Ebbecke published his paper on the local vasomotor reactions of the skin and internal organs, the outcome of several years' careful observation and experimentation and deep thinking, and this publica- tion marks the beginning of a new epoch in the study of the capillaries, because he was the first to recognize clearly the full significance of the facts brought to light. Ebbecke 's paper contains a wealth of informa- tion, and I shall have to refer to it very often in the course of these lectures, but at this stage I am con- cerned only with the evidence of independence between the reactions of capillaries and arteries. Ebbecke describes the following experiment on frogs which were curarized and kept moist, while the web of one foot was pinned out and allowed to dry up slowly at rates which could be conveniently varied. He finds that at first the circulation is very slow, the arteries narrow and many capillaries completely closed, while others allow the passage from time to CONTRACTILITY OF CAPILLARIES 31 time of a single red corpuscle. During the first half hour the arteries dilate, a number of new capillaries appear, and the current of blood becomes very rapid through all the vessels. So far the observations could be taken to favour the view that the capillaries become passively distended by the blood pressure, but at this stage the arteries begin to contract again and become gradually very narrow, while the capillaries become more and more dilated, and the number of visible capillaries is further increased until, finally, it is three to four times the initial. This opening up and dilatation must necessarily be independent of the internal capil- lary pressure, which at this time is very low. Ebbecke points out the fact that redness or paleness of the human skin depends upon the quantity of blood present in the cutaneous capillaries and venules, that is, upon their state of dilatation or contraction, while the temperature of the skin is chiefly dependent on the rate of blood flow through the skin, and he goes on to show that the skin of the hand, for instance, may be very warm, without being red, while the action of cold may produce a state of pronounced hypersemia characterized by a bluish colour, which indicates that the flow of blood is so slow that its oxygen is used up to an unusual extent. Ebbecke concludes that in the warm, pale hand we have dilated arteries and arte- rioles without dilated capillaries and in the cold, blue hand the arteries are strongly contracted and the capil- laries (and venules) dilated. In 1917 a paper was also published by Cotton, Slade and Lewis, in which important evidence is brought for- ward to show that the capillaries of the human skin are able to contract and dilate independently of the arterioles. These authors have studied the dermo- graphic reactions of the human skin, of which I shall have more to say later in the course of these lectures. 32 CAPILLARIES A slight stroke along the skin with a blunt point produces in most individuals a white line, while a heavy stroke produces a red line which may be bor- dered by white. These reactions take place after a time of latency of several seconds, they reach a maxi- mum in half a minute or more and fade away after several minutes ; their borders are very sharp and cor- respond closely to the area directly stimulated. Cotton, Slade and Lewis are of the opinion that the sharp defi- nition of the white and red bands suggests their origin as capillary reactions, since reactions of arterioles must result in an irregular border Line, but the descrip- tion given in my first lecture of the vascular system of the skin shows that the meshes of arterioles and venules are small enough to render such a conclusion invaLLd. They have obtained very de fini te evidence, how- ever, by studying the reactions after suddenly cutting off the blood supply to an arm by means of a sphygmo- manometer armlet in which the pressure was raised well above the arterial. When the blood had become completely stagnant they still succeeded in getting quite definite reactions after the usual period of latency, and they maintain rightly that in this case the very vessels which by their content of blood are responsible for the colour of the skin, must have contracted or dilated. There can be no doubt that these vessels are the capillaries and venules. They compare the white tache after gentle stroking to the whitening produced mechanically by gentle pres- sure. The latter begins to suffuse immediately when the pressure is released and disappears completely in a few seconds, because blood runs into the open capil- laries from all sides. The former develops after the stroking and is maintained for a comparatively long time. It must be impossible, therefore, for the blood to CONTRACTILITY OF CAPILLARIES 33 get into the vessels responsible for the colour : they must be actively closed. The paper of Dale and Richards (1918) contains a detailed and very closely reasoned comparison between the actions of three "depressor" drugs, leading up to the conclusion that one of them must produce a relaxa- tion in the tone of arterial smooth muscle, while the other two must produce relaxation of the capillary wall. I propose to state the reasoning of Dale and Eichards in some detail, not only because it is an excep- Fig. 31. Effects of intra -aortic, intra-caval and intra-portal injections of 0.01 mgm. histamine. After Dale and Eichards. tionally beautiful example of physiological analysis, but also because it serves to expose the inherent fal- lacy of the classification of substances acting on the circulation as either "pressor" or "depressor," a point on which I shall have more to say hereafter. The analysis takes its starting point from the obser- vation that in carnivorous mammals infinitesimal doses of histamine, adrenaline or acetyl-choline in- jected into the blood stream produce an evanescent fall of arterial blood pressure. In larger doses, on the other hand, acetyl-choline has a pure "depressor" action, which can be ascribed to dilatation of arteries, adrenaline shows the well- 34 CAPILLARIES known "pressor" effect, clue to contraction of arteries, and histamine induces in the animals which are sus- ceptible to its action the characteristic symptoms of shock. Dale and Richards show first, by measuring the time of latency between the moment of injection of a small dose of histamine and the onset of the fall of blood pressure, that this drug, like the others, acts chiefly on the peripheral vessels of the systemic circu- lation, the latent period being much shorter with intra- aortic than with intra-caval or intra-portal injections. They go on to examine the effect of small doses of the substances studied by simultaneous arterial blood pressure records and plethysmograph records of the volume changes of selected parts, usually one of the hind legs. On animals with intact nerves they find that acetyl-choline will always produce dilatation, while the other two drugs may sometimes produce dilatation, sometimes contraction. When the nerves to the limb are cut, an increase in volume results, due, chiefly, at least, to a dilatation of the arteries, and when the three substances are tested under these conditions the dilator effect of acetyl-choline is diminished, while there is constantly a considerable dilator reaction after histamine or adrenaline, as exemplified in the figure. When arterial tone is re-established after complete degeneration of the nerves to a leg the three drugs all give definite vasodilatation. When a leg, and prefer- ably a clenervated leg, which can be relied on to give constant dilatation on injection of any of the sub- stances, is made anaemic by occlusion of the vessels, a great increase in volume takes place when the blood is again admitted. In this stage the fall in general blood pressure caused by histamine or acetyl-choline is accompanied by passive decrease in the volume of the limb. The normal dilator effects return when the swollen limb shrinks, but at different rates for the CONTRACTILITY OF CAPILLARIES 35 different substances, and a stage may be found in which acetyl-choline causes dilatation, while histamine causes passive contraction, corresponding to the fall in blood pressure. All these facts point to the conclu- Fig. 12. Effect of 0.01 mgm. histamine. Volumes of leg Tvith nerves freshly cut and of normal leg; blood pressure. Time markings, 10 seconds. After Dale and Eiehards. 36 CAPILLARIES sion that, while acetyl-choline acts (as was known beforehand) through relaxation of arterial tone, the dilator effect of histamine and adrenaline must be localized in another part of the circulatory system and probably in the capillaries. This conclusion is supported by a series of perfu- sion experiments. It was found that while the dilator action of acetyl-choline can easily be demonstrated plethysmographically on a limb perfused with oxy- genated gum Einger, so long as the arteries retain their tone, and will be accompanied by a large increase in outflow from the perfused limb, the dilator action of histamine would only take place when an adequate supply of oxygen was secured by the addition of red corpuscles to the perfusion fluid, and when this con- tained a sufficient amount of adrenaline (1 part in a million to 1 in 10 millions). In such conditions a large increase in volume of the perfused limb will be accom- panied by a small increase in flow. Special perfusion experiments on a preparation of the superior mesenteric artery and all its ramifications up to the line where they enter the intestines showed conclusively that the effect of histamine on arteries is under all conditions a contraction, diminishing the rate of outflow. The conclusion drawn from these perfusion experi- ments is again that the dilator action of histamine must be exercised beyond the arteries on the capil- laries and can be exercised only when the tone of these vessels is kept up by a sufficient supply of oxygen and by the presence of a tonic substance, such as adrenaline in suitable concentration. Dale and Richards further strengthen their conclu- sion by some interesting and important evidence of a more direct nature. On cats with unpigmented feet they have made the observation that the pads of the foot CONTRACTILITY OF CAPILLARIES 37 in a denervated leg become distinctly warmer, but at the same time paler, than the pads of the normal foot. By dipping each foot in 10 cc. of cold water and noting the increase in temperature of the water they have shown conclusively that there is through the dener- vated foot a more rapid flow of blood, and that the pale colour must therefore mean that the capillaries are contracted in spite of the increased pressure. In such a cat, injection of 0.01 nig. histamine causes a definite flush (dilatation of capillaries) in the denervated foot, while acetyl-choline has no distinct effect on the colour. When a pledget of cotton wool soaked in 0.1 per cent histamine is applied locally to the surface of the cat's pancreas, in which there are no vessels of such a size as to be visible to the naked eye, a distinct red flush is produced in 10 to 20 seconds, showing that the minute vessels are dilated. Dale and Richards are careful to point out that the evidence which they have brought forward does not warrant any sharp distinction between the capillaries proper and the smallest arteries, and it might, I think, be argued that the reactions which they have observed could be reactions of arterioles, which would imply, however, that these shoidd behave quite differently from those larger arterial branches which are visible to the naked eye. My own first contribution to the problem of capillary contractility was published in Danish in 1918, about a month after Dale and Eichards' paper, and some- what later appeared in the British Journal of Physi- ology (1919). It was undertaken to test the hypothe- sis of a regulation of the supply of blood to muscles through the opening and closing of individual capil- laries. I found it possible to observe at least the super- ficial capillaries of muscles both in the frog and in mammals through a binocular microscope, using strong 38 CAPILLARIES reflected light as a source of illumination. Besting muscles observed in this way are usually quite pale, and the microscope reveals only a few capillaries at fairly regular intervals. These capillaries are so nar- row that red corpuscles can pass through only at a slow rate and with a change of form from the ordinary flat discs to elongated sausages. When the muscle un- der observation is stimulated to contractions a large number of capillaries become visible and dilated, and the rate of circulation through them is greatly in- creased. When the stimulus has lasted only a few seconds the circulation returns in some minutes to the resting state ; the capillaries become narrower and most of them are emptied completely, while a small number remain open. Since capillaries, even in a group fed by the same arteriole, do not all behave in the same way, the changes obviously cannot be due to arterial pressure changes. In resting muscles of the frog the average distance between open capillaries, observed simultaneously through the microscope, was estimated at 200 to 800/*, but after contractions this could be reduced to 70 or 60/t. In the guinea pig average distances of about 200/* could be observed during rest. The exposure to the air and the strong light always increased the cir- culation, and it was often possible to see the circula- tion begin in one capillary after another. It might be argued that the observations here re- corded could be explained as the results of dilatation of arterioles alone, on the assumption that the capil- lary paths offer various degrees of resistance, that a few are opened by a low pressure, while the majority of capillaries belonging to an arteriole require a higher pressure and are opened only when that arteriole is dilated. Such an assumption would involve as a con- sequence that a reduction of the pressure to by cut- CONTRACTILITY OF CAPILLARIES 39 ting the artery must produce an elastic contraction and emptying of the postulated high-pressure capil- laries. Numerous observations have shown that all the capillaries may remain open when a piece of muscle is cut out after stimulation. The measurement of distances between open capil- laries made upon living specimens could not, of course, be very accurate, and the degree of regularity of their distribution could not be satisfactorily made out by simple inspection. I had, therefore, to try and devise a method by which the state of the vessels at any given moment could be studied after fixation. This I succeeded in doing by injecting an India ink solution, dialysed against a Ringer solution to make it isotonic with the blood and deprive it of the toxic substances, added to the commercial preparation. When a suitable quantity of India ink is introduced into the circula- tion of a living anaesthetized animal it is evenly mixed with the blood, and if the animal is suddenly killed by stopping the circulation a few minutes later, and prepa- rations are made from the muscles and other organs, these show the capillaries which were open at the time. On frogs I found by this method that there were large differences between different organs in the num- ber of open capillaries. The skin, liver and brain were always well injected, with all, or nearly all, capil- laries open. The tongue was generally white and nearly bloodless, when not stimulated before being re- moved. The empty stomach and intestines had only a small number of open capillaries. The injection of muscles was variable, but in most of the resting- muscles few capillaries only were open, while muscles which had been tetanized before stopping the circula- tion were almost black from the large number of in- jected capillaries. Countings of the open capillaries on transverse sections of muscles injected vitally in this 40 CAPILLARIES way demonstrated the fairly regular distribution of open capillaries. In stimulated muscles I counted in one case of an extensor tarsi muscle 195 capillaries per mm. 2 , while the corresponding unstimulated muscle from the other leg had so few that an accurate count could not be obtained. There were certainly not more than 5 in a mm. 2 Other resting muscles showed higher numbers, however, especially the rectus abdominis muscle, which had also been observed in the living state to be always well supplied with blood, and three countings from which gave, respectively, 115, 155 and 180 open capillaries per mm. 2 , which is from 30 to 40 per cent of the total number. On guinea pigs consid- erable differences were observed between the degrees of vascularization of different resting muscles, prob- ably connected with the length of time since they had been in activity. Some flat muscles were examined pro- visionally in the fresh state by counting under a low power the capillaries visible in a single field without using the vertical adjustment. "When the positions of capillaries along the eyepiece micrometer are noted, a fairly good idea of the regularity of their distribution can be obtained. I give some instances of such countings. Capillaries in scale divisions Muscle from abdominal 1 1Q u 1Q gl M 2g 31 ^ ^ 3g ^ 5Q wall upper layer: V ^ fll 6g ^ ^ ?5 ?g 1 division = 21.8/*. ' Largest distance 6 divisions, smallest 2, average 3.8 = 83,u. Diaphragm: 20 22 23 25 27 30 32 35 37 39 41 42 44 46 48 50 1 division = 8.8m- 51 53 55 57 59 62 64 66 68 70 71 73 75 77 79 81 Largest distance 3, smallest 1, average 1.96 = 17/x. The first muscle had been at rest, but the diaphragm, being the chief respiration muscle, had been working vigorously up to the death of the animal. Several CONTRACTILITY OF CAPILLARIES 41 'Countings of optical transverse sections of the same muscles, which is, of course, a more accurate method, gave for the muscle from the abdominal wall the fig- ures 86, 70, 92 capillaries per mm. 2 , corresponding to a distance of 125/* between them, and for the diaphragm 2700, 2550, 2450 capillaries per mm. 2 , corresponding to .a distance of 18/*. 200 700 <2$00 ZL 100 too jtO Fig. 13. Preparations from vitally injected muscles from guinea pig. Optical transverse sections. Fig. 13 shows equal optical sections from three differ- ent muscles with the number of capillaries per sq. mm. as noted. In this figure the approximate diameters of the open capillaries are also indicated, and it should be noted that in the working muscle they vary consider- ably, while in the resting muscle, with 200 open capil- 42 CAPILLARIES laries, these are extremely narrow. The table gives a number of measurements in micromillimeters. When it is remembered that the red corpuscles of the frog are on an average 22/* long, 15/x broad and 4/* thick in the middle, while those of the guinea pig are 7.2^ in diameter and about 2^ thick, it seems almost incredible that they can pass through capillaries of the smallest dimensions given and even that they can pass through the average open capillaries of resting muscles. That they do so can be easily observed in the living muscles, especially after injection of India ink, and they are seen to become sometimes extremely elongated. Frog, m sartorius. Guinea Pig- Besting Stim ulated Abdominal wall Diaphragm 7.3 5.2 7.2 7.4 2.2 4.0 4.4 4.2 3.5 2.4 7.6 7.3 4.1 1.8 4.S 4.0 2.5 4.1 6.0 S.O 3.0 3.8 7.4 2.8 10.6 3.6 7.0 4.3 4.2 3.0 S.2 5.S 2.9 4.3 7.6 4.2 4.5 3.5 3.3 7.4 4.6 2.1 4.1 6.7 2.7 6.5 6.7 10.4 5.6 2.7 9.6 6.7 3.7 3.1 2 2 3.5 5.5 3.3 5.5 10.7 2.5 2.9 4.2 4.6 4.0 2.9 7.4 5.6 2.9 5.0 4.7 5.5 4.4 4.0 4.9 7.0 2.8 3.3 3.1 5.4 Average, 4. 3/i 6 8 M 3 5m 5 0m I have said enough, I believe, to make it abundantly clear that capillaries are not merely passively dis- tended by arterial blood pressure, but possess a tone of their own and may show contractions and relaxa- tions independently of the corresponding reactions in the arteries. I think it right, however, to record one more experiment which demonstrates in a crucial man- ner that the whole length of a capillary from an arte- riole to a venule can be contractile, that it cannof, when contracted, be forced open by the available arterial CONTRACTILITY OF CAPILLARIES 43 pressure, but can be made, by suitable stimulation, to relax and open up to a pressure which is much lower. The lower surface of the frog's tongue is covered by a smooth mucous membrane and, when suitably stretched, shows a very wide-meshed network of capil- laries, small arteries and veins. While in the mouth Fig. 14. The frog's tongue pinned out for microscopic examination. Natural size. the tongue is usually pale and almost bloodless. When the tongue of a narcotized frog is pinned out, as shown in Fig. 14, it becomes stimulated by the process and a large number of ves ; appear. When the tongue is left to itself in a mois atmosphere the vessels contract again, the tongue becomes pale and many of the capil- laries are completely closed and cannot be seen at all with the magnification practicable with the binocular 44 CAPILLARIES erecting microscope. If the surface of the tongue is now very lightly scratched with a hair or a fine glass- needle along a small vein (Fig. 15, 1) a reaction can be obtained like that shown in Fig 15, 2. A small branch opens up and is filled with blood, which becomes stag- nant. By continuing the stimulus in front of the column of stagnant blood the relaxation is carried further (Fig. 15, 3), the blood flows slowly on in the direction from the vein and at last connection is established with an arteriole, resulting, of course, in a sudden onset of current in the opposite direction, towards the vein. Fig. 15. Opening up of a capillary by repeated weak stimulation. The effect of internal pressure on the calibre of capil- laries. While the calibre of capillaries is, as we have seen, mainly determined by their own tonus, it is, of course, also affected to a certain degree by the blood pressure. "When the. main artery supplying one side of a frog's tongue is completely blocked before an area is stimu- lated no visible dilatation takes place until the blood is admitted. When the arterial pressure is greatly dimin- ished by compression of the artery the dilatation after CONTRACTILITY OF CAPILLARIES 45 mechanical stimulation will take place slowly, and after the opening of the artery a further slight dilata- tion will take place and a few capillaries may be opened up which had up to that remained closed. In the frog's web the arteries can be brought to dila- tation by the application of a drop of acetyl-choline and to a partial contraction by adrenaline, and when the capillaries are watched during these changes the effect upon their calibre is seen to be very slight and sometimes scarcely noticeable. When muscles, especially those of mammals and fishes, are artificially injected in the fresh state it is found to be very difficult to obtain a complete injection, and microscopic examination of some of the injected specimens has revealed the fact that of the number of capillaries supplied by the same arteriole a minority only have become injected, in spite of the high pres- sure employed. The power to resist an internal pressure is devel- oped to a very different degree in the capillaries of different tissues. It appears to be very high in muscles and comparatively low in the frog's skin and web. The cutaneous capillaries and venules in the arm of man become somewhat dilated when the venous pres- sure is raised by means of a Eecklinghausen cuff to about 30 cm. water pressure. They appear to resist the pressure for some minutes before they give way. Accurate measurements of the relations between the calibre of capillaries and the pressure to which they are exposed are very desirable, but have not been made so far. LECTURE III THE STRUCTURE OF THE CAPILLARY WALL HAVING- established the fact of independent capillary contractility we are again face to face with a problem, this time mainly histo- logical : By what means can the contractions be car- ried out? If we turn to the histological text-books for information we get a rather disappointing reply. The walls of the arterioles and small veins consist gen- erally of three histologically different layers, vis., an inner tube made up of flat polygonal endothelial cells, an outer thin coat of connective tissue fibres, contain- ing cells, and a middle portion consisting of one or more layers of smooth, ring-shaped muscle cells. When we approach the capillaries the outer coat first dis- appears, the muscle fibres become fewer in number and do not form a continuous layer, and finally we have left only the endothelial tube. This is built up of very thin cells of a polygonal or usually of an elongated rhomboidal shape, compared by Stohr to steel pens pointed at both ends. The cells are cemented together at their edges to form a completely closed tube. The delimitation of the cells can be made very distinctly visible microscopically by treatment with nitrate of silver and subsequent reduction. (Fig. 20.) The cement will then show up as black lines. By suitable staining each endothelial cell can be shown to possess a nucleus, generally of oval form and projecting somewhat above the internal as well as on the external surface of the cell. STRUCTURE OF THE CAPILLARY WALL 4-7 A structure like this makes it easier to understand why the majority of physiologists have declined to accept any evidence for capillary contractility, and leads, when the evidence becomes overwhelming, to the assumption of a mechanism of contractility entirely different from that possessed by the larger vessels. The mechanism assumed, I believe, by most of the physiologists who have observed the contractility for themselves (Hooker, 1920), was suggested by Strieker (1876) on the basis of his observation that the out- side diameter of capillaries did not become appre- ciably altered, even when the lumen was greatly diminished. This observation leads almost unavoidably to the conclusion that the decrease in internal diame- ter must be brought about by a swelling of the proto- plasm ; that, in other words, osmotic or imbibition proc- esses are responsible for the variations in the internal diameter of capillaries. The existence of contractile cells in the capillary wall. There exists, however, another and very different conception regarding the contractile mechanism of capillaries. A few years after Strieker's first communication, Rouget (1873), who had independently observed the contractility of capillaries in young tadpoles, studied histologically the capillaries in the hyaloid membrane of the frog's eye and found on the outside of the endo- thelial tubes certain oblong nuclei, arranged in the direction of the tube and surrounded by a zone of protoplasm with ramified elongations which embrace the capillary in a number of places like so many hoops. In a second communication (1874) he states having observed these cells also on the living capillaries of young newt larvae and having seen them contract. 48 CAPILLARIES He takes them to be related to the smooth muscle cells of larger vessels. Bouget's papers were practically completely disre- garded and soon, as it seems, completely forgotten. Nearly thirty years later Sigmund Mayer (1902) re- discovered the branched cells on the capillaries of the hyaloid membrane and also on those of the intestine of amphibia and stated that a continuous series of cells of intermediate form could be demonstrated between these and the normal spindle-shaped cells of the arterial muscular coat. It did not inspire confi- dence, however, that these interesting structures could be made visible only by vital staining with methylene blue, and even then only occasionally, and the further statement, that a system of branched cells, similar to that on capillaries, or even isolated cells of the same kind, could occasionally be found where no capillaries could be detected, could not fail to arouse the suspi- cion that they had no real physiological connection with capillaries. The suspicious attitude of histolo- gists was probably further strengthened by the fact that, although Mayer concluded his preliminary paper by the announcement that a detailed publication, ac- companied by figures, was to appear shortly, this promise was never fulfilled. Mayer's paper gave, however, the immediate im- pulse for the physiological investigations under- taken by Steinach and Kahn, and although these authors have not made any positive contribution to the histological problem, they have shown conclusively by measurements of dilated and contracted capillaries that the rival theory of capillary "contractility" by imbibition of the endothelial cells cannot be correct, since they find that the outside diameter of contracting capillaries, far from remaining constant or even in- creasing, as that theory demands, is very considerably STRUCTURE OF THE CAPILLARY WALL 49 diminished. They give a number of examples, of which I reproduce a few. Outside diameter of capillaries in frogs ' nictitating membrane Lumen Dilated Contracted when contracted M M 26 12 still open 24 7 closed 22 6 very narrow 19 3 closed These observations have been repeatedly verified in my laboratory, and the position a couple of years ago was the following: Of the two theories brought for- ward to explain the capillary contractility, one — the imbibition theory — was untenable for physical rea- sons, while the existence of the histological elements demanded by the other was, to say the least, extremely doubtful. Evidently a solution of this difficulty was essential for a successful attack on the many physio- logical problems connected with capillary contractility. If the contractility were due to muscles, the innerva- tion, for instance, or the reactions to stimuli must be supposed to be very different from what one would expect if it were due to imbibition processes in the endothelial cells or to a mechanism as yet undiscov- ered. I therefore asked a young histologist, Dr. Vim- trup, to take up the problem, and he has just published a paper (1922) in which the main problem, at least, is settled. Utilizing the fact that it is possible to bring about experimentally in certain tissues, and notably in the muscles and mucous membrane of the tongue of the frog, any desired degree of capillary contraction and dilatation, Vimtrup has examined sections from such tongues suitably fixed and stained. On the outside of the capillaries he finds certain nuclei slightly, but dis- 50 CAPILLARIES tinctly, different from the ordinary endothelial nuclei. The form of these nuclei varies with the state of con- traction of the capillary. On a dilated capillary they are broad and very thin; by contraction they become narrower and thicker, their cross-section approaching the form of a circle. The protoplasm belonging to these nuclei can be made visible by suitable staining, but even then it requires high-power immersion lenses and — especially on dilated capillaries — a good light, pref- erably excentric, to see it in its entirety. On a dilated capillary the protoplasm surrounds the nucleus as a continuous layer on the capillary wall, but it diminishes in thickness towards the periphery, which is very irregular and sends out a number of very fine branches along and especially around the capil- lary wall. The branches show at their base a definitely triangular cross-section, but soon become flat. Some- times they become broader and divide, but the ends are always very thin and pointed. Most of the branches lie athwart the capillary and are of such length that they reach those from the other side. Some of the branches, however, run along the capillary, and both the protoplasm and the nucleus are, as a rule, stretched in this direction. From the ends of the continuous pro- toplasm a broad branch usually runs along the capil- lary, giving off very fine secondary branches encircling the vessel. The distance from the last of these branches to the centre of the nucleus can, in large cells, amount to about 100/*. On a capillary of normal diameter, which will allow the red corpuscles to pass without much deformation, the appearance of the protoplasm is rather different. The continuous mass around the nucleus is distinctly thicker, with a more conspicuous structure; the con- tours of the branches are more distinct ; the triangular form of their cross-section is very pronounced, the STRUCTURE OF THE CAPILLARY WALL 51 sides being somewhat concave. On very narrow capil- laries these changes are further accentuated. The pro- toplasm is packed about the nucleus and the branches are short and stout, though still encircling the capillary and ending in sharp points. On some capillaries the branches have a different appearance, in that they show numerous anastomoses and form a sort of network. When the branched cells are followed along a capil- lary towards an arteriole or a venule their shape gradually becomes different. They become shorter ; the nuclei do not lie exactly in the direction of the capil- lary, but more or less slantingly around it; the branches are reduced in number and length along the capillary. On the arterioles themselves every stage of transition can be found between normal spindle-shaped muscle cells with an elongated nucleus, encircling the vessel, and others with an oblique nucleus and the pro- toplasm split up into a few ramifications. These latter become more numerous on approaching the capillaries, and there is no sharp distinction between them and the richly branched cells characteristic of the capillary. With suitable stains a fibrillar structure of the proto- plasm of all these cells can be made out. Vimtrup has further elaborated the method of supra- vital staining with methylene blue employed by Mayer, and succeeded in making it perfectly reliable. Its appli- cation is limited, however, to thin membranes, which can be examined in toto without being cut into sections, such as the web, bladder and nictitating membrane of frogs. I reproduce three of Vimtrup 's figures showing smooth muscle cells from an arteriole and the complete series of transitional cell forms connecting these with the typical branched cells on a capillary. Fig. 16 shows the point of branching of an arteriole with circular, smooth muscle cells around the endo- 52 CAPILLARIES thelial tube. A few of these cells are simple spindles, while the protoplasm of others is broken up into two or more parallel threads, and others again show a quite irregular ramification. The nuclei (a) of the more or less regular forms are arranged as arcs of a circle at right angles to the direction of the vessel, while the nuclei of the others occupy a slanting posi- tion. The nuclei of the endothelial cells (b) are visible in the figure, but are not very distinct. "if, Fig. 16. Transition between arteriole and capillary. Fig. 17 shows a large capillary with typical branched cells in considerable number, though much fewer than on the arteriole. The threads, which run from the nuclei (a) round the circumference of the capillary, probably represent the muscle fibrils and not the entire protoplasm of these cells. The granular appearance of the threads has probably nothing to do with the struc- ture of the fibrils, but is a simple deposition of methy- lene blue. In Fig. 18, which shows a small and slightly contracted capillary of 8/* diameter, the cells are fewer bo £ .9 So, a 2 >o X >> bD c6 fl 3 3 B a 5 3 ° 1 S ^ '3 ^ c3 o a: fq a a I 3« a - X CO 6C 54 CAPILLARIES in number. Their nuclei are arranged lengthwise on the vessel, and in one place the elongation of the cell along the capillary wall can be distinctly seen. Red blood corpuscles, r, are also shown. Pig. 19. Two Rouget cells (« and 6) as seen on capillaries in living newt larvae, b is contracting, c is a red corpuscle. X500. After Vimtrup. As there can be no doubt that the richly ramified muscle cells on the capillary wall are the same as those originally found by Eouget in the hyaloid membrane, Vimtrup has named them after the first discoverer, and we shall speak of them henceforward as Rouget cells. After having studied these cells in stained prepara- tions so as to be completely familiar with their distri- bution and appearance, Vimtrup has further succeeded in observing them on living capillaries and has been able to follow in a single cell the changes of form STRUCTURE OF THE CAPILLARY WALL 55 taking place by contraction. The best object for these observations he has found to be the tail of young newt larvae (Triton punctatus), which can be arranged for observation even with oil immersion lenses. By a sim- ple and ingenious arrangement Vimtrup has been able to immobilize these larvae and keep them in excellent condition for hours without having recourse to nar- cosis. In these animals spontaneous contractions and dila- tations of single capillaries or parts of capillaries are of frequent occurrence, and it is a significant fact that a contraction always begins just where the nucleus of one of the Rouget cells is located. During the process of contraction the cell shows the different forms de- scribed as characteristic in the case of stained prepa- rations, but the movements themselves are generally too slow to be followed by the eye. This can be done, however, when contractions are induced by stimulation of nerves in the web of a small frog. After a latent period of about fifteen seconds the Rouget cell under observation will show an increase in the refraction of light, and a few seconds later the contraction proper will begin. The nucleus of the cell is observed to sink a little into the capillar)-, and on the opposite wall several small indentations make their appearance. Some of the ramifications, as a rule, be- come distinctly visible, when the capillary is already somewhat contracted, and it can be observed that their positions correspond to the indentations seen in the endothelial wall. After two to three minutes' stimula- tion a maximum, though usually incomplete, contrac- tion is generally obtained, but about this time a very curious change takes place in the tissues, which lose their normal transparence and become so opaque that no structural details can be observed. "When the stimu- lation has ceased the tissues regain their normal trans- 56 CAPILLARIES parence and the contracted Eouget cells relax in the course of a few minutes. The changes in the endothelium by contraction and dilatation. On special preparations Vimtrup has studied the ap- pearance of the endothelial cells and their nuclei, cor- responding to the different states of contraction and Fig. 20. Capillaries from frog 's web. Border lines of endothelial cells silvered. Black pigment cells. X300. dilatation of capillaries. On dilated capillaries the edges of the rhomboidal endothelial cells are more or less straight, the cells themselves are large and their nuclei are thin, oval discs, which project only very slightly, if at all, on the inside of the capillary tube. On somewhat contracted capillaries the endothelial cells are smaller, their edges are sinuous, as shown in Fig. 20, their nuclei are more or less ovoid and project STRUCTURE OF THE CAPILLARY WALL 57 into the lumen of the capillary tube. On capillaries which are strongly contracted the endothelial wall be- comes folded. That folding takes place was strongly maintained by Steinach and Kahn, and Vimtrup, too, has observed living capillaries, the appearance of which strongly suggests a folding of the wall, but he found it impossible to obtain decisive evidence on the point. Examining the same object as used by Steinach and Kahn, the nictitating membrane of young frogs, and stimulating the vessels to contraction by faradic cur- rents, Mr. Rehberg has now succeeded in my labora- tory in demonstrating with absolute certainty the folding of the endothelium in strongly contracted capil- laries. When a capillary is watched under a high power during the contraction the formation of the folds can be very distinctly seen. Occasionally the folds protrude into red corpuscles which can be squeezed out into very strange forms. In the arterioles with their larger in- ternal surface the folding of the endothelium by con- traction is even more pronounced and very easy to observe. An attempt to combine all the observations into a coherent conception of the normal anatomy of a capil- lary will give about the following result : Like the walls of the smallest arteries and veins, the capillary wall consists of two distinct elements — the endothelial tube and the outside muscular coat. The important differ- ence between capillaries and larger vessels lies in the arrangement of the muscles, which in the arteries and veins form a more or less continuous layer, greatly in- creasing the thickness of the wall and offering a con- siderable resistance against the exchange of sub- stances between the blood and the surrounding lymph spaces or tissue cells, while in the capillaries the mus- cular coat is arranged more or less in the form of a 58 CAPILLARIES wide-meshed network, leaving the larger part of the endothelial surface uncovered and adapted for the passage of substances with a minimum of resistance. The muscular coat of the capillaries possesses, like smooth muscles generally, a definite tonus, or "pos- ture," in the terminology of Sherrington (1920), sub- ject in the organism to nervous, hormonal and other influences, and the diameter of a capillary is, on the whole, determined by the state of contraction of its muscular coat. The changes observed in the configura- tion of the endothelial cells and in the shape of their nuclei make it clear, however, that the endothelial tube itself must possess some normal form and calibre which it will assume when the outside and inside pres- sure is absolutely the same. "When the outside pressure is the higher, the tube collapses. When the Rouget cells contract, it becomes folded, and when they relax and the inside pressure is ever so slightly higher than the outside, it is dilated, and the endothelial cells are passively stretched, their surface is enlarged, their thickness is diminished and their nuclei are flattened out. We have here a close analogy to the behaviour of the alveolar epithelium in the lungs of man, according to Marie Krogh's (1915) determinations of the pulmo- nary diffusion constant. These determinations of the quantity of carbon monoxide which will diffuse into the blood from the alveoli, show that when the lungs are expanded beyond a certain point the epithelial sur- face becomes larger and thinner by stretching, but when they are allowed to collapse below that point the surface area and thickness remain constant — the sur- face is folded. The elastic properties of cells as illustrated by the red corpuscles. The force necessary to stretch the endothelium — STRUCTURE OF THE CAPILLARY WALL 59 apart from the muscular coat — must be extremely slight, since the capillaries can, as I have shown, be opened up by a minimum of pressure, when their muscles are relaxed. 30 -60 v 70/i J Fig. 21. Muscle capillaries from guinea-pig, vitally in- jected with India ink. Walls of capillaries not shown. It would be difficult to believe that such extreme softness could be combined with such perfect elasticity as that which the endothelial cells and their nuclei must have, if the red corpuscles did not furnish an example in which this combination of qualities can easily be shown to exist. I mentioned in the preceding lecture the fact that red corpuscles can become greatly deformed during their passage through narrow capillaries. Fig. 21 60 CAPILLARIES shows red corpuscles of the guinea pig in muscle capil- laries of different calibre, vitally injected with India ink. These corpuscles are normally flat discs 7p in diameter and between 1 and 2/* thick. In capillaries of 4 to 5/* diameter they pass through easily by having their edges rolled (Fig. 21, 5). In still narrower capil- laries they become sausage-shaped, and their length can be increased to 14//. or more. When they escape from such narrow vessels their shape must become nor- mal again, since free deformed corpuscles are never observed. The pressure necessary to bring about the deforma- tion in narrow capillaries must be comparatively low, since the flow does not stop in a single narrow capil- lary, even when the same arteriole supplies several others through which the corpuscles can pass freely, but a definite estimate cannot be attained. It is well known that red corpuscles will pass through filter paper, the pores of which will hold back quantitatively precipitates consisting of particles which are much smaller than the corpuscles. There can be no doubt, though the passage has never been directly observed, that the corpuscles are greatly deformed during the passage. The pressure available for such a passage cannot exceed the height of fluid in the funnel. The most direct evidence of the wonderful plasticity and elasticity of red corpuscles is obtained when they are watched in a current, where they can be caught against a projecting edge and bent by the pressure of the current flowing past them. On a cinema film taken in my laboratory of the flow of blood through the alveolar capillaries in the frog's lung, we find a corpuscle caught on a very sharply projecting edge, as shown in Fig. 22, copied from the film. It remained hanging on the edge during about four seconds (seventy pictures). At first it is riding nearly along STRUCTURE OF THE CAPILLARY WALL 61 the short axis, and both ends are bent down in the direction of the current, but the pressure is not suffi- cient to press the content of the corpuscle out towards both ends. Somewhat later it slides along towards one of the ends and becomes rather sharply bent into the shape of a large and a small sac. Finally, when the current becomes slack in the interval between two beats of the heart, the corpuscle slides off the edge. The four lower figmres show four consecutive stages of this Fig. 22. Eed corpuscle riding cm projecting edge in frog's pulmonary capil- laries and finally sliding off in four lower pictures. From moving picture film. process with an interval of 0.06 seconds. In the last figure, about 0.3 seconds after the release, the shape has practically returned to normal, in consequence of the elastic properties of the corpuscle. When a riding corpuscle is watched on the cinema film or directly through a high-power microscope, the current appears to be so rapid that it is not very sur- prising to see such effects, but it must be remembered that the rate of flow is magnified in the same propor- tion as the objects. On the film the actual rate can be approximately measured by noting the shifting posi- 62 CAPILLARIES tion of free swimming corpuscles from one picture to the next. In the case now described, the rate, during the period the corpuscle remained hanging, did not exceed 0.12 mm. per second, corresponding to about 400 mm. per hour, a rate which is too slow to be fol- lowed by the naked eye. I am not familiar enough with hydraulic problems to venture upon a calculation of the actual pressure to which a current of this rate has exposed the corpuscle, but it is evident that it must be of a very low order. The differences in structure of different capillaries. The descriptions given above of the endothelium and Rouget cells apply to the capillaries of amphibia gen- erally, where the contractile elements have been ob- served on so many different capillaries that they can safely be taken to exist in almost all tissues. They have been seen by Rouget on the capillaries in the hyaloid membrane of the frog's eye and on those in the tails of tadpoles and newt larv?e ; by Mayer on smooth muscle capillaries in the wall of the intestine and the urinary bladder of frogs and newts, and Vimtrup has found them further in cutaneous capillaries, in those of the mucous membrane and striped muscle of the tongue and the nictitating membrane of frogs. As a rule, espe- cially large and powerful Rouget cells are found at or near the points of branching of capillaries. The Rouget cells appear to differ considerably in shape and structure, according to the organ where they are found. In the tail of amphibian larvas and in the web of grown frogs they are of a rather primitive ainceboid character, with much undifferentiated proto- plasm and few muscle fibrils. Their total length varies in the larval capillaries from about 60 to 200/t, and in the frog's web from 40 to 80/*. The number of cells STRUCTURE OF THE CAPILLARY WALL 63 per mm. of the capillary in the web has been found in a single count to be about 70. | In the tongue and nictitating membrane the cells are I more robust and distinctly muscular, with a large num- ber of fibrils. Their total length varies between 40 and 80/*, and from 20 to 50 have been counted per mm. In the hyaloid membrane the cells are extrernefy elongated along the capillary, giving off in a very regu- lar manner hooplike branches encircling the capillary tube. The normal thickness of the endothelium in a capil- lary which is neither dilated nor contracted, appears to be somewhat less than 1/* (about 0.8^ probably). On a dilated capillary the thickness is reduced by stretching in proportion to the increase in diameter. In mammals Rouget himself describes the contrac- tile elements as occurring in the retina and fatty tissue in rabbits and ruminants and in the brain of ruminants, while Vimtrup has found them in the intestine of mice, in the interstitial connective tissue of man and quite recently (August, 1922) in the cutaneous vessels of man. Finally, Prof. E. Muller, of Stockholm, showed me in 1920 the nuclei of these cells on capillaries in the intestine of the cat. In none of the mammals has the investigation hitherto gone beyond the demonstra- tion of the existence of Rouget cells. Nothing is known about their shape, distribution and number, and we have here a rich field for further study. A special search is to be made for the Rouget cells on the capillaries in the glomeruli of the kidney, in the liver and in the lungs both of amphibia and mammals. Personally, I expect the glomerular capillaries to be of normal structure, and their contractility has quite recently been demonstrated by Richards (1922), but the liver capillaries are so aberrant that I shall not venture even a guess regarding their contractile ele- 64 CAPILLARIES ments, though they have been shown by Ebbecke to possess contractility. (See Lecture VIII.) The endothelium of the hepatic capillaries appears to be a syncytium with numerous nuclei, but without cell borders as in embryonic capillaries. At short, rather regular, intervals, the star-shaped cells of v. Kupffer (1876, 1899), the function of which appears to be phagocytic, form an integral part of the capillary wall. An even more remarkable feature in the struc- ture of the hepatic capdlaries are the ' ' canaliculi ' ' in- side the liver cells, which communicate directly, ac- cording to the researches of several authors (Brovicz, 1899 ; Schafer, 1902 ; Herring and Simpson, 1906), with the lumen of the capillaries. According to v. Kupffer the liver capillaries are surrounded by an adventitial layer — perhaps constituting a pericapillary lymph space — in which we will have to make a search for the contractile cells. The liver has the power to absorb microscopic par- ticles from the blood at a very rapid rate, and the capil- lary walls are much more permeable to dissolved sub- stances than in other organs, but it is not known exactly how these peculiarities are correlated with the above-named structural features. The relation of capillaries to arterioles and venules. In all the capillary fields which have been closely ex- amined there is a very distinct difference between arterioles and capillaries. The arterioles have a well- developed muscular coat of simple circular fibres, cov- ering the entire surface of the endothelial tube. Pro- ceeding from an arteriole to a capillary there is a short zone of transition, as shown in Fig. 16, in which the number of muscle cells is reduced and the endothelium uncovered, but, on the whole, there is a sharp demarca- tion between the muscular arteriole and the capillary, STRUCTURE OF THE CAPILLARY WALL 65 the distinctive feature of which is the large extent of free endothelial surface in proportion to the area cov- ered by the contractile elements. Between the capillaries and the venules the differ- ence is much less pronounced. The muscular coat of many venules is so thin and incomplete that the ex- change of substances through their walls must be con- siderable, and in some cases the vessels, which are anatomically classed as veins, are from a physiological point of view to be regarded as capillaries, since, in spite of their comparatively large size, they have a coating of Eouget cells instead of simple muscle fibres, while the greater part of their surface is uncovered. The subpapillary veins of the human skin are in reality such "giant-capillaries." Spalteholz (1893) de- scribes them as having no rnuscular coat, and Vimtrup has now succeeded in observing on stained prepara- tions the nuclei of their Eouget cells. In following lec- tures it will be shown that in their general behaviour and reactions to various stimuli these vessels have a very close similarity to the capillary loops connecting them with the arterial system. The possible existence of direct communications be- tween arteries and reins. In the anatomical literature of fifty years ago there are a number of references to the existence of "deri- vating channels," represented by small arteries open- ing directly into somewhat larger veins. Such con- nections have generally been observed on injected specimens and without verification by a study of the structure of the vessels in question. As single capil- laries can easily become greatly dilated by injection under pressure and give the appearance of wide channels, evidence of this kind is obviously untrust- worthy. 66 CAPILLARIES Hoyer (1877) describes the existence of direct anas- tomoses between arteries, generally of about 0.02 mm. diameter, and slightly larger veins as occurring only in certain places in the body of mammals. He has found them near the edge of the ears in rabbits, dogs and cats, in the tip of the snout in several animals, as well as in the tip of the tail, and finally in the tips of fingers and toes in man and other mammals. In many of his cases he has studied the structure of the vessel wall in the places of anastomosis after suitable stain- ing, and he describes and pictures vessels of undoubted arterial structure opening directly into others which must be accepted as veins. He has made a search for these "derivating channels" in many other places, but apart from the well-known case of the arteries opening directly into the corpora cavernosa, he has found them only in such projecting parts as might require a considerable supply of blood to keep them warm, when exposed to low temperatures. Though the evidence for the existence of these anas- tomoses presented by Hoyer cannot be accepted as con- clusive, they may very well exist and fulfil a function of some physiological importance just in those places where he has seen them. I believe, therefore, that it would be worth the trouble to search for them again. Unfortunately it is, I am afraid, impossible to see them in living animals. According to Hoyer 's descrip- tion, they are too deeply situated for that, even in the ears of rabbits, where the superficial circulation in the skin can be very clearly seen under the microscope. In the numerous observations made in my laboratory upon this object, we have only once seen a combina- tion of vessels in which there might be a direct com- munication between a small artery and a vein, though the two observers could not come to a final conclusion on the point. STRUCTURE OF THE CAPILLARY WALL 67 The non-existence of "vasa serosa." Some authors, who have observed that capillaries are sometimes bloodless, while the blood is flowing freely through a number of others around them, or that blood may be flowing through single capillaries, while the majority are closed, assume that certain capillaries will not normally admit red corpuscles, but only plasma, and that they are opened up when the pressure becomes high. Such capillaries have been de- scribed by Oohnstein and Zuntz (1888), who made their observations on the web of the frog, as "vasa serosa." It should be pointed out, however, that capillaries can only under exceptional circumstances and for short periods admit a slow current of plasma without admit- ting corpuscles. This is an unavoidable consequence of the extreme softness of the red corpuscles. When a capillary containing one or more red corpuscles con- tracts to such a degree that the corpuscles cannot pass along, they will be squeezed into a form which fills up entirely the lumen of the vessel and prevents any pas- sage of plasma along it. It is only when a capillary happens to contain no corpuscles at the moment of becoming so narrow that these elements can no longer be squeezed through, that it may for a time admit a current of plasma which must necessarily be very slow. Even in that case it will generally not last long before a corpuscle is carried into the mouth of the vessel and blocks the passage. In my observations of the circula- tion in specimens vitally injected -with India ink, I have found that the submicroscopical particles of this substance can pass, as a rule, through only those capil- laries which are also open to the passage of corpuscles. Smith, Arnold and Whipple (1921) conclude from the extremely valuable series of comparative blood- volume determinations undertaken at the Hooper Foundation that there must be a systematic difference 68 CAPILLARIES between the composition of the blood in the larger ves- sels and in the capillaries. To obtain agreement be- tween their determinations of corpuscle volume and plasma volume on the one hand and the hematocrit determinations of the ratio of corpuscle volume to plasma volume on the other, they assume that the blood in the capillaries is more dilute than the blood in larger vessels, and they ascribe this supposed difference partly to the existence of "vasa serosa" in the sense of Cohnstein and Zuntz, but mainly to the flow of the blood in an axial stream containing the corpuscles and a "still space" of pure plasma. While the blood cer- tainly flows in this way in small arteries and veins, as exemplified above, and may also do so in wide capil- laries, my impression, after having observed the blood flow in a very large number of capillaries, is that the ' ' still space " is in average capillaries practically non- existent : the capillary walls are regularly brushed by the corpuscles. As far as I am able to judge, the "still spaces," which actually do exist, are insufficient quan- titatively to account for the discrepancies noted by the authors. LECTURE IV THE INNERVATION OF CAPILLARIES WHEN, as we have seen, the capillaries have, like the arteries, a definite muscular coat, we must expect these muscles to be innervated. Anatomically it has been shown long ago (Beale, 1860) and repeatedly confirmed on numerous tissues in different animals (Grlaser, 1920), that the capillaries are regularly accompanied by nerves. Generally there are two fine, non-medullated fibres, one on each side of the capillary and connected by a number of anasto- moses, crossing the capillary at angles of about 45° (Krimke, 1884). The nerves show small swellings at irregular intervals, but ganglion cells have not been found, and the actual connections with the capillary wall are uncertain, as the descriptions of the authors differ considerably. Krimke asserts definitely that each capillary has its own nerves, which do not anas- tomose with those of other capillaries, but this seems very doubtful. In many preparations single nerve fibres are found to run in a close spiral along some portion of a capillary in addition to (or instead of) the two comparatively straight fibres accompanying the vessel. As far as I have been able to find, the nature of these different fibres has never been ascertained, but Glaser seems to take them to be sympathetic. In larger vessels and especially in the arteries there is a rich adventitial plexus of fibres with numerous cells which seem to be connected with the fibres, but o ^\ **S v> -^v— *%,_: • «, -- V iCit-'*''^! " '"•""t^". Si &k'4 f-*y&k. jijr ■•■'-• *^b ' -* J ' ' ' ' * imp fifO>--.^ J '*- '^A *jf$ JA i • V ■ ' ^^Q C- ■' ./. M x*r''f^ ■ US^ 1 IT**"* _- ■_ ftf. ' ft ■■•* * 1 Iff ■- »- 1 ■Si^. _> ■ ] m. 's/'x^- ' : j '4S' J& ^^L ' ' w/^M r-J~ZKem WJ ^^k '/ ■ ■ ■ £ ~J 'vC^^Cj 3r M \^3^% i-i ->; .^^B ^L *i ¥&J&3& ■ ■*>?■■*•:. *^H ^l ^~m ^^M ■^^v^. be THE INNERVATION OF CAPILLARIES 71 which cannot be ganglion cells, since they do not pre- vent the degeneration of the whole plexus, when this is cut off from the dorsal sympathetic ganglia, by the post-gangiionic fibres of which it is made up, accord- ing to the degeneration experiments of Eugling (1908). The fibres of the main plexus probably do not anasto- mose, but they give off numerous branches forming a secondary plexus inside the vessel wall, and it is quite possible, though not, I believe, definitely established anatomically, that a real nerve net is formed by anas- tomoses between the fibrils. From this plexus (or nerve net) the muscle fibres of the vessels are inner- vated. The arrangement appears to be very similar to that described by Hoffmann (1907) for the musculature of the intestine and bladder. Hoffmann maintains that the secondary plexus in these organs is a true network of fibrils, but he has been unable to decide whether each fibre forms a separate fibrillar end plexus or there is a general anastomosing between all fibrils. E. Miiller (1920) has demonstrated the existence of true fibrillar nets in the vagus system in the stomach of sharks. The sympathetic innervation of capillaries. By analogy with the larger vessels we would expect the capillaries to be likewise innervated through the dorsal sympathetic system and to contract when their sympathetic fibres are stimulated. You will remember that this is just what was found by Steinach and Kahn on the frog's nictitating membrane, and their experi- ments were, in fact, guided by these considerations. In recent years the capillaries of several other tissues have been shown to be innervated through sympa- thetic fibres. Hooker (1920) found that electrical stimulation of the cervical sympathetic brought about a very pronounced constriction of capillaries and 72 CAPILLARIES venules, as well as of arteries, in the skin of the cat's ear, and in my laboratory we have made the same ob- servation on the ear of the albino rabbit, where very accurate observations can be made by transmitted light. Leriche and Policard (1920) report a contraction of the capillaries at the base of the nail in man, when the sympathetic fibres running in the adventitial coat of the humeral artery are mechanically stimulated. The reaction is said to be practically instantaneous. On the hind legs of frogs we have made a more de- tailed study of the sympathetic innervation (Krogh, Harrop and Eehberg, 1922). We have stimulated the lower ganglia (8 to 10) of the sympathetic chain, which brings about a constriction first of the arteries and a few seconds later also of the capillaries of the web. Generally we have not succeeded in making the capil- laries contract until obliteration. As mentioned in the preceding lecture, the contractions can be observed with sufficiently high magnifications to start from those points where the nuclei of the Rouget cells are situated. The capillaries in the muscles of the leg are also influenced by stimulation of the sympathetic ganglia and sometimes contract completely. There is reason to believe that every single Rouget cell is sup- plied from a sympathetic fibre and can be made to contract through it. Sympathetic tonus of capillaries. When the sympathetic ganglia are removed in the frog or the sciatic cut below them, the capillaries of the web dilate more or less permanently, from which fact it follows that they are, like the arteries, tonically innervated through the sympathetic system. Some- times the dilatation takes place at once, as is the rule for the arteries, but more often it takes half an hour or more to develop. This might be supposed to be due THE INNERVATION OF CAPILLARIES 73 to the stimulus of the nerve section, but mechanical stimulation has usually only a slight effect on sympa- thetic fibres and would wear off at any rate in a short time. The probable explanation is to be sought rather in the influence of another tonic mechanism of which I shall have something to say later on. While the tonus of the arteries is usually re-estab- lished a few days after the abolition of sympathetic influences, the capillaries are, as a rule, very slow to regain their normal state, and in one case even re- mained dilated for a period of 100 days. In the cat's ear Hooker (1920) failed to observe any influence upon the calibre of the capillaries and venules of section of the cervical sympathetic, but in the rabbit's ear Rehberg and I have seen them dilate definitely after the corresponding operation. In this case we have no direct evidence that the arterial dila- tation resulting from the section cannot be responsible, through the increase in capillary pressure, for the observed reaction, but we have other evidence, to be given later, that the capillaries of the ear are under the permanent influence of a strong sympathetic tonus. Like the arteries the capillaries of the rabbit's ear regain their tonus a day or two after section of the cervical sympathetic. When the tonus is re-established after section of the nei'ves, we have observed in the frog that the regula- tion of the capillary tonus in the web is, as a rule, very imperfect. States of strong contraction alternate with more or less complete relaxation. Breslauer (1919) has made corresponding observations on patients with nerve lesions and mentions the "spontaneous" changes from ischaemia to hyperemia and vice, versa. The tonic sympathetic innervation of capillaries is probably of very widespread occurrence in the verte- brate organism. There are, however, a few organs in 74 CAPILLARIES which it appears to be absent. One of these is the tongue of the frog, though the evidence is not quite conclusive. Electrical stimulation of the lingual nerves may cause contraction of some arteries but has no con- strictor influence on capillaries, and, when the propa- gation of impulses along the nerves is blocked by cool- ing to the freezing point or actual freezing of the nerves, usually some dilatation of arteries occurs, but the response of capillaries is so slight that it is prob- ably to be explained by the increased blood pressure to which they become subjected by the arterial relaxation. It is possible, however, that dilatation would have come had the blocking been maintained for a longer period. Dilator innervation of capillaries. "When the lingual nerves of a frog (especially the giossopharyngeus) are stimulated mechanically by pinching, a distinct hyperemia is produced in the tongue, due to dilatation of capillaries as well as of arteries. 1 In a few cases the dilatation of the capil- laries has preponderated to such an extent that the current of blood through them became visibly slower. This dilator effect develops after a latent period of sev- eral seconds or even one-fourth to one-half a minute. It subsides very gradually during fifteen minutes or more. These vasodilator reactions call to mind the "antidromic" vasodilatation produced in the legs of mammals by mechanical stimulation of the sciatic and shown (Bayliss, 1901 and 1902) to be localized in pos- terior root fibres. Bayliss' experiments have given no definite proof that the capillaries are involved in his 1 1 have not succeeded in producing this capillary hyperemia by elec- tric stimulation, but I find that Bruck (1909) has obtained it in one half of the tongue by faradie stimulation of the corresponding glossopharyn- geal nerve. THE INNERVATION OF CAPILLARIES 75 antidromic dilatations, and in the case of the frog's tongue there is no definite proof that the stimuli are propagated along posterior root fibres. In the case of the frog's hind legs, however, it can easily be demon- strated that stimulation of posterior roots will cause dilatation of capillaries as well as of arteries in the web and skin. This was first shown by Doi (1920) and confirmed by Krogh, Harrop and Rehberg (1922), who found, however, that it is probably only a limited, though rather large, number of the capillaries in the web which will respond to stimulation of posterior roots. That the dilatation of arteries cannot be respon- sible for the capillary reactions observed was shown by Doi by the application of a dose of acetyl-choline. This drug will cause a maximal dilatation of the arte- ries and thereby prevent any response on their part to the nerve stimulation. In these circumstances the observed dilatation of capillaries can only be due to relaxation of their own tonus. In the capillaries of frogs' muscles the antidromic dilator innervation appears to be slight and sometimes absent. Bayliss found on mammals that the dilator effect of stimulating posterior roots, as measured by a plethysmograph, was practically abolished when the leg was skinned. It should be remembered that prob- ably very few posterior root fibres go to the muscles compared with the large number supplying the skin. In experiments made by Rehberg and myself we failed to elicit any dilator response by electrical or mechanical stimulation of the peripheral ends of the sensory nerves to the rabbit's ear. There is, however, even in this case some evidence, to be mentioned later, pointing to an innervation of capillaries by sensory fibres and, however this may be, we have in the clinical symptoms of the skin disease known as Herpes zoster very definite evidence that a large number of capil- 76 CAPILLARIES laries of the limbs, trunk and at least the larger part of the head are in direct connection with posterior root fibres and become dilated when these are stimulated,, probably mechanically, by pathological processes. Herpes zoster is characterized by such processes,, usually of an inflammatory character, taking place in one or more of the spinal ganglia or in the Gasserian ganglion. The initial symptoms are dilatations of small vessels, including capillaries and venules in the zone directly innervated by the fibres from the ganglion af- fected, and the correspondence is so close that a reflex dilatation, produced by the pain usually associated with the initial stages of Herpes zoster, can be ex- cluded. The only possibility seems to be a conduction of impulses along posterior root fibres directly to the skin vessels, and the evidence, such as it is, points to the conclusion that it is a general feature in verte- brates that vessels of the skin and notably a number of capillaries are directly connected with posterior root fibres, and that impulses travelling along these will cause dilatation of the vessels concerned. Bayliss has shown by absolutely conclusive evidence in the case of the mammals examined by him that the dilator fibres in question could not be distinguished from normal bipolar sensory fibres with their cells in the spinal ganglion connected both with the periphery and with the spinal cord. In the case of frogs we have found that no degeneration takes place when the pos- terior roots are cut between the spinal ganglia and the cord. It is scarcely necessary to emphasize the fact that our knowledge of capillary innervation is extremely imperfect. Only a very few organs have been studied in the frog and one or two mammals. Nothing is known about the innervation of capillaries in such organs as the intestines, the glands or the central nervous sys- THE INNERVATION OF CAPILLARIES 77 tern. They may or may not be subjected to tonic im- pulses from the dorsal sympathetic; they may or may not be provided with special dilator fibres belonging to the cranio-sacral or even perhaps in some cases to the sensory system. There is a wide and probably very fruitful field for future research. The information at hand will enable us, however, to get some insight into the mechanism of certain vascular reflexes which I shall now proceed to describe and discuss. Fig. 25. Keflex erythema following a prick with a needle. After L. E. Matter. "Reflex erythema" and some allied vascular reactions. When a needle is drawn across the skin in man so as to provoke a more or less painful sensation, or when a sensation of pain is brought about in any other way, for instance, by the action of high or very low tem- perature or of many different chemicals, the stimulus will result after a short period of latency in a red area 78 CAPILLARIES of irregular form and variable size surrounding the point directly stimulated. This point may itself show various reactions, according to the nature and strength of the stimulus. With these we are not now con- cerned, but we will confine our attention to the sur- rounding erythema. A microscopical examination of the erythematous skin and a comparison with normal skin reveals the fact that the redness is due to a large number of skin capillaries and small veins which are normally closed, or at least very narrow, but are now open and comparatively wide. In these dilated vessels there is a rapid current of blood, showing that the arteries and arterioles of the affected area have also become dilated. An erythema of similar appearance and also due to dilatation of both capillaries and arteries can be pro- duced in the skin of many animals and can be studied in all its details; for instance, in the depilated ears of albino rabbits and in the tongue and skin of the frog. The mechanism of this erythema has been studied in man by Miiller (1913) and Ebbecke (1917) ; in the rabbit's ear by Eehberg and myself, and in the frog by Harrop, Rehberg and myself (1922). In the mammals we have to do with a true reflex. The erythema does not develop if the skin is made insensitive by cocainization or if the sensitive nerve fibres to the point stimulated have been cut. It is pre- vented further if the spinal cord at the level where it is entered by these fibres has been destroyed and finally if the sympathetic fibres supplying the vessels concerned are cut. The reflex arc of the reaction runs, therefore, from the organs of pain in the skin, along the posterior root fibres to the spinal cord and back to the vessels along sympathetic fibres. The path of the reflex within the cord is a short one, and the sensation of pain, by which THE INNERVATION OF CAPILLARIES 79 the reflex is normally accompanied, has nothing di- rectly to do with it. Miiller has described cases in which a short portion of the spinal cord has been crushed. The reflex erythema was abolished in the sur- face zone corresponding to the crushed portion, but could be provoked both above and below, as shown in Fig. 26. The parts of the body below the injury were, of course, insensitive. Fig. 26. Absence of reflex erythema in zone corresponding to crushed portion of spinal cord. After Miiller. In the Loven reflex (1866) we have evidently the same reflex arc as in the erythema now dealt with. Loven found that vasodilatation in the skin could be produced by stimulation of the central end of a sensory nerve through reflex relaxation of sympathetic tonus. Mr. Eehberg and I have observed on the rabbit's ear that the capillaries are directly involved in this reflex as well as the arteries. 80 CAPILLARIES Ebbecke (1917) states (p. 31) that reflex erythema cannot be provoked on the surface of internal organs like the liver or kidney. When the tongue of a deeply narcotized frog (E.esculenta) is pinned out and a needle drawn across its surface, a hyperaemic zone 2 to 4 mm. broad will begin to develop after a latent period of a few sec- onds. In an anaemic tongue a number of closed capil- laries are opened up and greatly dilated, while the arteries supplying the hyperseniic zone become dilated way back in the tongue. This reaction is abolished, as in mammals, when the surface of the tongue is anaes- thetized by cocainization, and one would expect, of course, that it should be of the same reflex character as in mammals. This is not the case, however, since sec- tion of all the lingual nerves does not influence it in the least. The possibility of a true reflex is therefore excluded. That the spreading of the reaction from the line directly stimulated must take place along nerve fibres can be concluded from the time of latency and the rapid rate of spreading after the latent period, and is proved further by the fact that degeneration of the nerves, after section, will prevent the development of the erythema and confine the reaction of the vessels to the line directly stimulated. It follows from these facts that we have to do with a local nervous mecha- nism in which sensitive nerve endings are probably in- volved (since cocainization will abolish the reaction) and also, of course, those fibres by the mechanical stimulation of which dilatation could be produced and which are, according to the evidence presented abovej. assumed to be sensory. Two different local mechanisms can be conceived and are illustrated by the diagram (Fig. 27). In I the single nerve fibres are supposed to split up into fibrils of which some are in connection with the contractile THE INNERVATION OF CAPILLARIES 81 elements of the vessels. If one or more of the fibrils or their (hypothetical) end organs are stimulated, the stimulus will spread to all the branches of that nerve fibre and cause relaxation of the capillary and arterial muscle cells with which it is connected. This is the original conception of an antidromic local reflex adopted by Bruce (1910) to account for his observa- tions on the conjunctiva of mammals. The diagram II Fig. 27. Alternative diagrams of axon reflex. corresponds to a conception to which Bardy (1915) was led in studying the details of local vascular reac- tions to stimuli in the conjunctiva. It assumes a stimu- lation of sensitive nerve endings which is carried through special branches of the fibres concerned to local vasomotor ganglion cells and from these on to the vessels. The experimental results obtained on the frog's tongue can be best explained on the simpler assump- tion of sensitive nerve fibres being directly connected with the contractile cells. On the hypothesis that the stimulus is carried to an autonomic ganglion cell one must expect that all the elements innervated from this cell would react simultaneously and to the same ex- tent, and there would be no reason to expect that the point stimulated would always show the most intense reaction. It is possible, however, by very cautious mechanical stimulation of a single point, to restrict the response to a length of capillary corresponding 82 CAPILLARIES to 2 to 4 Rouget cells and to graduate it from this mini- mum by very small steps to a maximum involving a capillary area of 5 mm. 2 , or more, and 5 to 10 mm. of artery supplying this area. We must assume, there- fore, that the nerve fibres which accompany the smaller arteries and arterioles give off fibrillar branches to their musculature, before they are finally split up in a number of branches supplying the mucous membrane and its capillary vessels. A stimulus applied to one of the branches must travel in a central direc- tion along this, but whenever it comes to a point of bifurcation it will be distributed between the stem and the other branch. Judging by analogy from what is known to happen in the nerve nets of lower animals, we may suppose it to be weakened by the distribution, and this assumption will explain in a simple manner the fact that the area to which the reaction spreads depends (at least roughly) upon the strength of the initial stimulus. In some later experiments Mr. Rehberg has found that chemical stimulation of a point in the frog's tongue by means of silver nitrate gives rise to a dis- tinct hyperemia which involves a large part of the sur- face. This reaction, which is also a local reflex, per- sisting after the lingual nerves have been cut, has its exact counterpart in the skin, and its significance will be discussed presently. In the web and skin of the frog (R.temporaria) the reactions to local stimuli are somewhat more compli- cated. A sharply localized, very weak, mechanical stimulation may cause dilatation of a capillary over a short distance, but a stronger stimulus will often cause contraction. Both the dilatation and the contraction take place after a latent period of some seconds, and the contraction may occur practically simultaneously in several neighbouring capillaries at a distance of at THE INNERVATION OF CAPILLARIES 83 least several tenths of a millimeter beyond the point directly stimulated under the microscope, and may involve also a part of the artery supplying the con- tracting capillaries. The spreading of the reactions over longer distances can be observed only by stimulation of the arteries. A very weak mechanical stimulation, such as that pro- duced by gentle rubbing with a soft hair across an artery, may bring about a dilatation of that artery over a length of several millimeters, but a stronger stimulus, like the prick of a needle, may cause the same length of artery to contract even to complete obliteration. If the artery is hit and pierced the biological significance of this latter reaction reveals itself beautifully. After a few seconds' latency the artery becomes closed and remains so for many minutes, and when it opens again the drop of blood shed through the wound has clotted and further bleeding is prevented. While a reaction like the reflex erythema in mam- mals cannot be provoked in the frog's web by mechani- cal stimulation, it can easily be brought about by cor- rosion, e.g., with a small crystal of silver nitrate. After a latency of some ten seconds all the arteries between the two toes, where the crystal is placed, and often a large number of those between neighbouring toes, will dilate. A few seconds later a large number of capil- laries will also show some dilatation, which is in the main independent of the increase in pressure and blood flow resulting from the widening of the arteries. We have made a large number of experiments to study the mechanism of these various vascular reac- tions in the frog's skin. The first point to be tested is, of course, whether the reactions are due to some local mechanism or are of a reflex character. Since all of them, including the erythema, which spreads over the whole web after 84 CAPILLARIES chemical stimulation, are not in the least influenced by section of the sciatic nerve, it follows that they must be accounted for by local mechanisms. According to the numerous histological investiga- tions made, it is extremely improbable that local gan- glion cells exist in the extremities, and we have at our disposal to account for the reactions, in so far as they must be taken to be of a nervous nature, only the two sets of nerves: the sympathetic, artificial stimulation of which causes contraction of arteries and capillaries, and the posterior root system, stimulation of which causes dilatation. We have no definite evidence that the local dilatation of capillaries after mechanical stimulation is brought about through nerves, since no definite spreading beyond the point directly stimulated has been ob- served. The contraction of capillaries on strong me- chanical stimulation is evidently due to the same mechanism as the corresponding arterial contraction, but is much more difficult to deal with experimentally. These two reactions are, therefore, left out of account in the following discussion, and we confine our atten- tion to the remaining three : the contraction of arteries over long distances after strong mechanical stimula- tion ; the dilatation of arteries over long distances after weak mechanical stimulation; and the general "ery- thema ' ' produced by chemical stimulation. In all these cases spreading takes place, after a short time of latency, at such a rate and to such distances that con- duction of the impulses along nerve fibres appears to be the only possibility. The hypothesis which suggests itself in this case is the existence of a distinct axon reflex mechanism in the sympathetic as well as in the sensory system of fibres. "We assume that when a mechanical stimulus of sufficient force acts upon sympathetic fibres or fibrils THE INNERVATION OF CAPILLARIES 85 it will set up a wave of excitation which will travel along in both directions from the point stimulated, spread to all the fibrils belonging to the same fibre and cause contraction of all the muscle cells supplied by these fibrils. The reactions observed do not compel us to assume the existence of a true fibrillar network of sympathetic fibrils, but can be explained by the plexus of fibres found in the arterial wall, on the as- sumption that each of these fibres supplies a large number of muscle cells along the artery and some of them, further, a certain number of Rouget cells on capillaries. In order to explain the dilatory reactions we must postulate posterior root fibrils or end organs sensitive to mechanical and to chemical stimulation. A mechani- cal stimulation of suitable strength will cause dilata- tion over some distance, but the reaction must be capricious, because the same stimulus will affect also the antagonistic sympathetic system. A chemical stimulus, on the other hand, will affect the dilator sys- tem alone and, when the chemical action continues, there will be summation of stimuli resulting in a con- siderable effect. To understand the spreading of the effect over practically the whole web of one foot it is necessary to assume either that each neurone branches to practically the whole surface of the foot, or that the neurones are connected up peripherally by their fibrils forming a true nerve net, in which an excitation wave set up at one point can spread in all directions, suf- fering only the decrement due to its distribution over an increasing number of paths. The first of these assumptions seems very irrational, while the second, though it has at present very little trustworthy histo- logical evidence to support it, is by no means inher- ently improbable. To test these hypotheses concerning the mechanisms 86 CAPILLARIES of the reactions we have made a considerable number of degeneration experiments. If our assumptions are correct it should be possible by operative removal of the sympathetic ganglia of one of the hind legs of a frog to abolish the local contraction reactions and by removal of the spinal ganglia to abolish the dilator reactions. Our first series of degeneration experiments must be classed as incomplete nerve sections. "We did not at the time suspect the existence of nerve nets or the sup- ply of sympathetic fibres along the arteries of the leg independently of the large nerves. We expected that section of the sciatic would abol- ish, after a suitable period for degeneration, all the local reflexes. We found, however, that they became weakened only after a period of about forty days, dur- ing which they were first normal, then very distinctly increased and then abnormal. In a few frogs, which lived 100 to 150 days after the operation, the local reflexes finally became nearly normal again. We expected that simple section of the anterior and posterior roots of the ninth and tenth nerves supplying the leg would have no effect on the local reflexes, since the trophic centres of the nerve fibres involved would remain intact ; and we found this expectation fulfilled. We expected that operative removal of the ninth and tenth spinal ganglia would abolish the dilatation reac- tion after weak mechanical stimulation, but we found that it was only weakened for a variable period from the fiftieth day to 50 or 100 days later, while in simi- lar experiments, in which the eighth to tenth sympa- thetic ganglia were removed, the result was a similar weakening, for a variable period between the fiftieth and one hundred and twentieth day, of the contraction response. When the anatomical descriptions (Eugling, 1908) THE INNERVATION OF CAPILLARIES 87 are studied closely it seems probable that in all these experiments a certain number of nerve cells, of the kind we wished to remove, have remained in undis- turbed connection with the end organs in the web, and this can explain the results, when it is assumed, fur- ther, that these cells will prevent the complete degen- eration of the whole net of fibrils with which they are in connection, and gradually cause its recovery. In some later experiments we have tried to obtain a complete denervation of one leg of a frog. In some of these we tore out the sciatic from the knee upwards to the spinal canal and removed also the eighth and tenth sympathetic ganglia. In others we cut also the femoral artery to get rid of the periarterial plexus through which fibres might come down to the web. 1 Unfortunately these latter animals died before the degeneration could become complete, but on the former we found the local dilator reflexes completely abolished after eighty days ; no dilatation of arteries could be ob- tained by mechanical stimulation, and the action of a silver nitrate crystal was absolutely limited to the area into which the silver diffused and was precipitated as chloride of silver. The contraction of arteries on strong mechanical stimulation was present, however, and quite normal, and the absence of degeneration of the sympathetic system points strongly to its innerva- tion along the arteries themselves, and to the exist- ence of a sympathetic network of fibrils which require i We had failed formerly (Krogh, Harrop and BehbeTg, 1922) to find any evidence of a supply of nerve fibres along the femoral artery of the frog. By a strictly local mechanical stimulus applied to the artery under the microscope we are able, however, to obtain contraction over a distance of 2 to 3 mm. in both directions from the point stimulated. At the stimulated point itself a very sharply localized dilatation may often occur. The contraction persists for several minutes, just as in the small arteries of the periphery. Quite small frogs are the best for this experi- ment. 88 CAPILLARIES only the maintenance of connection with a few fibres to prevent degeneration. This conception of peripheral nerve networks, each made up of the richly anastomosing fibrils of the cor- responding nerve fibres, is, of course, so novel that it will have to be tested in several ways, both histologi- cally and physiologically, and found confirmed, before it can be accepted as anything but a provisional work- ing hypothesis. Personally I believe, however, that it will prove to be a fruitful hypothesis, and I would like to suggest that it might be worth while to study the cutaneous senses in man on the basis of the exist- ence of such networks connecting up all the sense points of each special sense and determining the "local sign" of any sensation by the combination of fibres, the "conductive pattern" (Sherrington, 1920, p. 179) through which the impulse is carried up to the brain. We have learned now that the nervous mechanism by Avhich a local erythema is produced after applica- tion of a painful stimulus is very different in mam- mals from the one found in the frog and probably in other cold-blooded vertebrates. Neither in the one nor the other has the sensation of pain anything to do with the reaction, which, in the lower animals, depends upon a strictly local axon reflex in sensory fibres, while in the higher it is brought about by a true spinal reflex with a short reflex arc. This is in keeping with the gen- eral trend of evolution in the vertebrate nervous sys- tem, where a "tendency" towards increasing centrali- zation is unmistakable. One would expect, however, that the older, "lower" type of reaction would remain, at least as a "rudiment," in the higher animals. And so it is, in fact. Its existence can be demonstrated, even in man himself, though the combination with the true reflex erythema has made its detection rather difficult. On the rabbit's ear Mr. Rehberg and I have found THE INNERVATION OF CAPILLARIES 89 that mechanical stimulation under the microscope with a very fine needle or a hair will cause immediate dila- tation of the capillary loop stimulated and, after a latency of a few seconds to half a minute, depending largely upon the temperature of the ear, dilatation of surrounding capillaries to a distance of about 1 mm., and finally of the artery supplying these capillaries. The capillaries are seen to dilate, and afterwards the current through them becomes rapid. The strength of stimulus necessary to bring about this reaction is so small that it can scarcely be felt at all on the human skin, and no reflex erythema develops. That we have to do with a local mechanism can be shown by section of the auricular anterior and posterior nerves, which does not immediately affect the reaction in the least, though the ear becomes completely anaesthetic, and the true reflex erythema is, of course, abolished. With the prog- ress of degeneration in the divided (chiefly sensory) nerves the reaction becomes definitely weakened after twelve days and appears to be completely absent after nineteen days. I do not think that this reaction, which is so slight that it can be observed at all only by means of the microscope, has any functional significance, but would rather consider it as a true physiological rudi- ment. In other cases, however, the axon reflex erythema seems to be functionally effective. Bruce (1910) has studied the inflammation produced in the conjunctiva of rabbits by a drop of mustard oil. He finds that the symptoms, which include maximal capillary dilatation, can be greatly diminished by local anaesthetization with cocaine; that they are not affected by simple cutting of the sensory nerves to the conjunctival surface, but fail to appear when the cut nerves are allowed to de- generate. He rightly concludes that the symptoms must be due to a local nervous mechanism and puts 90 CAPILLARIES forward the hypothesis of an axon reflex in sensory fibres. Bruce 's results have been, on the whole, con- firmed and amplified by Bardy (1915), who has thought it necessary to assume the more complicated nervous mechanism referred to above. I shall not enter upon a discussion of these theories, the more so as I think the problem will require a renewed experimental inves- tigation with the help of the microscope and with the application of less drastic methods of stimulation. In a recent very interesting and valuable paper Breslauer (1919) has studied on the skin of man the reactions to mustard oil. On normal skin mustard pro- duces, as is well known, a painful sensation and a con- siderable local hyperemia which one would expect to be of a true reflex nature. Breslauer finds that a true reflex does indeed play a considerable part in the reac- tion and is responsible for its spreading to an area considerably in excess of that to which the mustard is applied, but the reaction on the place of applica- tion he finds to be unaffected in his patients in cases where the corresponding part of the spinal cord or the posterior nerve roots have been destroyed and also in fresh cases of section of the corresponding spinal nerve, while it fails to appear after degeneration of this nerve. In special experiments Breslauer showed that a degeneration of the contractile elements of the vessels themselves was out of the question: they re- acted promptly to the application of local cold by an initial contraction, followed by relaxation if the cold was sufficiently intense. We are, therefore, also in these cases, driven to the assumption of a local nervous mechanism. That this mechanism involves sensory fibres has been shown by Breslauer in experiments made on his own skin, in which he produced a local anaesthesia by novocaine and found that the reaction to mustard failed to appear in the anaesthetized area. THE INNERVATION OF CAPILLARIES 91 The local anaesthesia lasted only about twenty minutes, and it is very significant that the reaction began to appear when the subjective sensation of pain showed that the nerve endings had recovered from the novo- caine paralysis. The temperature regulating mechanism. Another vascular reaction, which is very conspicu- ous in the ear of the rabbit and plays a very impor- tant role also in the skin of man, is worth studying from our point of view, because it brings out very clearly the essential difference between arteriomotor and capillariomotor control of the circulation. "When a rabbit is narcotized and tied up on its back it is unable to maintain its body temperature and must be kept warm artificially. If the rectal temperature is increased above normal the ears become very hot, that is, their temperature is only 1° to 2° Centigrade below the rectal temperature. The larger arteries of the ears are strongly dilated, and the pulse in them can be both seen and felt. When the animal is cooled and the body temperature has fallen to a certain point, somewhat below the normal, depending to some extent on the depth of the narcosis and somewhat different in differ- ent animals, the arteries of the ears suddenly contract and their temperature drops to a few degrees above the room temperature. These reactions are largely independent of the temperature to which the ears are exposed and are determined by the body temperature, which they serve to regulate, since the heat loss is obviously diminished when the temperature of the ears is nearly the same as that of the air and increased when it approaches that of the internal organs. The enormous changes in blood flow through the ears involved in these reactions are accompanied by comparatively slight changes in the colour of the ears, 92 CAPILLARIES and microscopic observation reveals the fact that the capillaries are involved to a slight extent only and sometimes not at all. The large increase in blood flow, taking place when the body temperature rises above normal, is brought about by a dilatation of arteries and arterioles. If the number of open capillaries was very small beforehand some more will be opened up, but not to anything like the extent to Avhich they may become opened and dilated by local stimulation. The venules and veins will, of course, become somewhat dilated by the increased flow through them, and this brings about some reddening of the ears to the naked eye inspection, but not enough to obscure the fact, which comes out very clearly by a comparison between this reaction and the local erythema, that the colour of the skin depends mainly upon the state of dilatation of the cutaneous capillaries and not upon the rate of blood flow through them, while the temperature of the skin is determined primarily by the rate of blood flow, which, in its turn, depends upon the state of contrac- tion of the arteries and arterioles. These rules are of wide applicability and can be used as a guide to obtain an approximate estimate of the vasomotor reactions in many organs by simple ocular inspection and temperature estimates or measure- ments. A definite change in the blood colour of an organ always implies a change in the state of contrac- tion of its capillaries, but says nothing about the state of the arteries or the rate of blood flow. The temperature changes can be utilized only in or- gans the metabolism of which is insufficient to maintain them at body temperature and which are consequently heated to a considerable extent by the flow of blood through them. In such organs a large temperature deficit means a low rate of blood flow, while a small deficit indicates a rapid flow. Stewart (1911) has elabo- THE INNERVATION OF CAPILLARIES 93 rated an ingenious method by which the heat elimina- tion of a hand or foot measured by a simple calorimet- ric device is utilized for quantitative determinations of the rate of blood flow through these limbs. The mechanism of the vascular reactions in the skin serving for regulating the body temperature is toler- ably well known. Our experiments show that the effer- ent nerve paths belong to the dorsal sympathetic system, since in the rabbit section of the cervical sym- pathetic on one side abolishes the temperature regulat- ing changes in the corresponding ear. The afferent path is in the case of man very generally assumed to be along fibres from the temperature sense organs in the skin, but, though a reflex of this kind undoubtedly ex- ists, its role as a part of the mechanism for tempera- ture regulation has been somewhat overestimated. It can be shown experimentally on rabbits that a heat- ing of the skin to a temperature above that of the body does not produce any considerable increase in the blood flow, unless the body temperature is at least normal, while, on the other hand, an increase of the body temperature will have this effect even when the skin is surrounded by very cold air. The stimuli sent out along the sympathetic system to regulate the body temperature must take their origin within the body, and the experiments of Barbour (1921) and others make it almost certain that a nerve center, which is directly sensitive to temperature changes and governs the heat regulation, exists in the corpus striatum of the brain. The psychic vascular reactions. When a rabbit is tied up, without being narcotized, and the ears arranged for microscopic observation of the circulation, the animal will usually, when kept warm, remain perfectly quiet over long periods if 94 CAPILLARIES not disturbed. Any sudden noise, a slight shaking of the table or even a sudden, strong light may, however, frighten the animal and cause some muscular move- ments. At the same time, a very distinct vascular reac- tion often takes place in the ears : The capillaries and arterioles contract and the ears may become visibly paler, even to the naked eye. This contraction lasts only for a few (2 to 5) seconds and is followed by a pronounced capillary hyperaemia, which subsides gradually in the course of a minute or more. In some cases the vascular reaction is the only visible conse- quence of the disturbance. The afferent path for this reflex is evidently one of the specific sensory nerves (auditory or optical), the efferent path both for the contraction and for the subsequent dilatation of the vessels is through the dorsal sympathetic, since sec- tion of the sympathetic fibres abolishes the reaction in the ear concerned, while section of the sensitive nerves of the Qar has no influence. A light narcosis completely prevents any reaction of this type, and we have good reason, therefore, to believe, though a defi- nite proof will require some further experimentation, that we have to do in this case with a reflex arc com- prising typical brain centres and of essentially the same type as those which are responsible for the emo- tional vascular reactions in man. Everybody knows that these reactions manifest themselves through distinct and sometimes quite sud- den changes in the blood colour of the skin. This means that they are brought about chiefly by tonus variations in the capillaries and venules. Though the point has not been specially investigated, I believe that the arte- rioles are also usually and perhaps always involved, since the blush appears to be accompanied by a rise in skin temperature, while the skin which has turned pale from fear or sorrow is proverbially cold. THE INNERVATION OF CAPILLARIES 95 In analogy with the results obtained on the rabbit's ear I assume that emotional paleness is brought about, at least mainly, by an increase in the sympathetic tone of the small skin vessels, while blushing is due to a reflex relaxation of this tone. Blushing is sometimes ascribed to an antidromic in- nervation through posterior root fibres. Such an inner- vation seems improbable, in view of the fact that oper- ative removal of the sensory trigeminus nerve in man does not abolish or even affect the emotional vascular reactions in the corresponding part of the face. LECTURE V THE REACTIONS OF CAPILLARIES TO STIMULI TT is well known that the state of contraction or relaxation of smooth muscles can be influenced by various stimuli, and the Eouget cells forming the contractile coat of capillaries are no exception to this general rule. They can be strongly acted upon, as we have seen, by mechanical stimuli, they are influ- enced in intricate ways by the temperature of the surrounding medium and by temperature changes, they may relax more or less completely under the in- fluence of a strong light. The osmotic concentration of the surrounding fluid, its hydrogen and hydroxyl ions, numerous inorganic or organic substances of a more or less well-defined constitution, which may be added to it, have a more or less definite action, and finally we have to deal with the action of substances produced within the organism itself and sent with the blood as "chemical messengers," or hormones, to regu- late the state of contraction of the capillaries in various organs. To attempt a rational classification of the factors influencing the contractile elements of the capillary wall is — at present at least — a hopeless task. One can- not help thinking with a little sigh of the conventional and delightfully simple classification of substances act- ing upon the general blood pressure into a "pressor" and a "depressor" group and being tempted to classify the factors acting upon the Eouget cells as "constrictor" and "dilator," respectively, but such a REACTIONS TO STIMULI 97 classification would be nearly as wrong in principle and might prove quite as harmful in its application as the pressor-depressor classification itself, because it might put together in one class factors which differed in kind of action and separate others which dif- fered only in degree. We are confronted further with the difficulty which also sometimes may upset the most beautiful pressor-depressor classification, that one and the same substance or physical factor may act, accord- ing to concentration and circumstances, sometimes as a constrictor and sometimes as a dilator. As is generally the case in physiological research, we have a double purpose in studying the reactions of capillaries to physical and chemical stimuli : we want to find out the mechanism (in the broadest sense of that term) of every single reaction studied, and we want to find out the meaning, the part played by the reactions in the delicate regulations by which the organism and the organs are adapted to the ever changing environmental conditions. Neither of these purposes must be lost sight of, and in many cases they are, of course, so closely connected that they can- not be separated ; but generally, I think, it is conducive to good economy to keep them as separate as possible, both in research work and in the presentation of its results. I shall try, therefore, to deal with the reactions first by attempting to analyze their mechanisms. The mate- rials available for such an analysis are very scanty, however, and in many cases we can only reach the negative and veiy unsatisfactory conclusion that the mechanism is at present entirely unknown. Direct and indirect capillary reactions. When a substance acts upon a capillary field and brings about, say, dilatation, we can, according to 98 CAPILLARIES present physiological notions, distinguish between the following mechanisms, any of which may possibly be involved in the reaction, either alone or in combination with one or more of the others. We may have either a direct action or an indirect action through nerves or other propagating tissue. The first of these mechanisms involves several quite distinct possibilities. The substance may act upon the contractile cells themselves, it may act upon their myoneural junctions, the somewhat hypothetic ele- ments interposed between the nerve fibres and the muscle cells and generally acted upon through the nerve fibres, or it may, finally, act upon the tissues generally, altering in a way which at present eludes definition the "environment" of the contractile cells. We may hope some day to be able to utilize these pos- sibilities to obtain some understanding of discrepan- cies and differences in reactions which are at present inexplicable, but at present we cannot, as a rule, get beyond the stage of classing as "direct" all those reac- tions which are confined to the vascular area where the stimulus is actually applied and spread only in so far as the stimulating substance itself diffuses out beyond the place of application. As "indirect" reactions we class those which take place, usually after a short period of latency, outside the area directly stimulated. In indirect reactions the stimulus must be propagated in some way. It is pos- sible that the Eouget cells on some capillaries may form a syncytium in which a contraction may be propa- gated from one cell to another, but the observations so far made have failed to reveal any direct connec- tions between the cells. A propagation of this kind would reveal itself as a contraction progressing slowly along a capillary from a single point. In the frog we have never seen this form of contraction, but there are REACTIONS TO STIMULI 99 a few observations on the skin of man (Ebbecke, 1917) which may possibly be explained in this way. Gen- erally, however, there can be no doubt that in indirect reactions we have to do with a propagation along nerve fibres. Here again we have to distinguish, whenever possible, between the two different mechanisms already described : we may have to do either with a local axon reflex or with a true reflex. In the first case we shall, according to our present notions, expect a spreading of the effect over a very limited area, closely surrounding the place directly stimulated, while in the case of true reflexes the reaction may take place over a much larger area with irregular and indefinite limits or even in areas quite distinct from the one directly stimulated. Whenever the reaction to a stimulus is essentially the same outside as inside the area stimulated, it will be natural to conclude that the stimulus acts primarily on nerves, and in all cases where no spreading what- ever occurs we must assume a direct action on excit- able tissue without the intervention of nerves, but, even Avith the greatest care in making experiments and the greatest caution in their interpretation, it is often extremely difficult to distinguish between a direct ac- tion and an indirect one through axon reflexes, and the possibility must always be borne in mind that both forms may be involved in the same apparently simple reaction. The reactions of capillaries to heat and cold. The influence of temperature and temperature changes upon small blood vessels, including capillaries, in several organs of rabbits has been studied by Bicker of Magdeburg and his school. In 1910 M. Natus of that school described a method 100 CAPILLARIES (based upon the researches of Kiihne and Lea) for studying the circulation in the rabbit 's pancreas under conditions which were at least approximately normal. The field of observation was constantly irrigated with 0.9 per cent saline, the temperature of which could be varied at will. The normal resting pancreas is described as pale, with narrow arteries. The lobules are provided with a network of capillaries which, in the periphery of the lobule, are so narrow that the red corpuscles can only pass through one by one or in single file, while the central capillaries are wider and usually allow two coi'puscles to pass side by side. When the pancreas is irrigated with saline of 22° Centigrade the only visible change reported is a slow- ing of the blood stream, due probably to the increase in viscosity of the plasma as a direct consequence of the lowered temperature, but at still lower temperatures (7° to 5°) there is definite evidence of a considerable contraction of arteries, Avhile the emptying of capil- laries, observed and reported as contraction, may pos- sibly be due mainly to plasma skimming. High tempera- tures (42° to 56°) always bring about dilatation of the capillaries. At the higher temperatures (48° to 56°) this dilatation soon becomes maximal and stasis devel- ops. The effects of high temperatures on the arteries are more complicated. Increases in temperature to about 45° will usually cause a dilatation of arteries, though contractions may take place occasionally. At still higher temperatures contraction becomes pre- dominant, though the initial reaction may be dilatation. At 56° the artery under observation contracted com- pletely in half a minute, then opened a little to contract again, but after three minutes a relaxation began which soon became maximal. The experiments of Natus have been repeated and REACTIONS TO STIMULI 101 confirmed, so far as high temperatures are concerned, by Eicker and Eegendanz (1921), who have further studied the effects of high temperatures on the ear and the conjunctiva of rabbits. After removal of the hairs they have dipped the upper third of one ear of an albino rabbit into hot water for a few minutes and studied both the immedi- ate and after effects of this treatment. They find that temperatures from 52° to 70° produce a very pro- nounced hyperemia, with dilatation of both arteries and capillaries. At the lowest temperature tested (52°) some of the capillaries return to normal shortly after the exposure, while at all higher temperatures they remain dilated. The arteries, on the other hand, con- tract and usually become for a while completely closed after the exposure. This contraction takes place imme- diately after the exposure at temperatures below 60°. After exposure to 60° it takes place a minute later; after 63° to 65° the contraction comes after two and one-half minutes and after 70° the arteries are unable to contract. On the conjunctiva they have tested also the effects of less drastic temperatures, from 40° upwards. In this tissue nothing but dilatation is observed as a result of the application of heat. Irrigation with saline of 40° for three minutes has no microscopically visible effect. By the application of 42° to 44° the arteries become somewhat dilated, some new capillaries are opened up and all visible capillaries become somewhat dilated, but the velocity of the blood flow decreases, which must mean that the larger arteries supplying the area contract. At still higher temperatures all arteries and capillaries become strongly dilated, and stasis develops in a large number of the capillaries. In no case has contraction of arteries been observed as an after effect. 102 CAPILLARIES The experiments of Natus and of Bicker and Regen- danz give no information regarding the mechanism or mechanisms of the temperature reactions. It is well known that local exposure of the skin of man to low temperature causes a definite paleness of the exposed skin, and it can easily be verified micro- scopically (Bruns and Konig, 1920; Carrier, 1922) that a contraction alike of capillaries, venules and arte- rioles is involved in this reaction, which can be ob- served in the hand when put in a water bath at 20°. By a prolonged exposure to somewhat lower tempera- tures, and the lower the temperature the sooner, the capillaries and venules relax, sometimes after repeated contractions of short duration. At first the supply of arterial blood may be sufficient to maintain a red hyperemia, but generally the arterioles contract fur- ther, and the exposed skin becomes cyanotic. This relaxation on exposure to cold is probably a direct action of the temperature on the contractile cells, which become partially paralysed. The initial contraction by cold is of a complex nature, partly reflex, partly due to local mechanisms. As men- tioned in the last lecture, Breslauer has found that the local paleness can be produced by cold in completely anaesthetic areas, the spinal nerve supply of which has degenerated, which shows the existence of a local mechanism. On the other hand, reactions have been observed by Wernoe (1920) and also by Zak (1920) which must, according to present notions regarding innervation, be accepted as reflex, though their mechanism is by no means clear. Wernoe found in cases of intestinal disease that the well-known hypersesthetic skin areas, the locations of which correspond to the seat of the disease, can be objectively demonstrated by exposure of the abdomen of the patient to room tem- perature. They react in this case by a blanching which REACTIONS TO STIMULI 103 is distinctly more pronounced than that of the rest of the surface. In some experiments on the rabbit's ear undertaken by Eehberg and myself we have found that the dila- tation brought about by the local application of heat is also partly a reflex, partly due to local mechanisms. The reflex hyperemia can be abolished by section either of sensory nerves or of the sympathetic. The local effect persists after section and degeneration of all the nerves to the ear, including those running in the wall of the main artery, and we have reason to believe, therefore, that we have to do in this case with a direct action on the muscle cells. The local dilatation produced by moderate heat (about 50°) on a dener- vated ear will persist for hours after the stimulation has ceased. In a few, not very conclusive, experiments on the frog's tongue I have found (1920) that the local appli- cation of 35° causes a pronounced dilatation of capil- laries and arteries, which effect is abolished by cocain- ization. The reaction is, therefore, probably due to the local axon reflex mechanism referred to in the previous lecture. It is, perhaps, worthy of note that in the case of the frog local cooling caused also dilatation, instead of the contraction I had expected. Tlie reactions of capillaries to light. The influence of light upon capillaries was studied about a quarter of a century ago (1900) by my great countryman, Xiels Finsen. His experiments are few in number and of a simple description, but they have been devised, performed and interpreted with that unerring instinct which is rightly termed genius. In one of his experiments, which I shall reproduce in some detail, Finsen exposed his own forearm, which was white and unpigmented, to the light of a 40,000- 104 CAPILLARIES candle-power arc lamp at a distance of 50 cm. for ten minutes and, since the heat at this distance was rather disagreeable, for ten minutes more at a distance of 75 cm. On the arm, of which Fig. 28 shows a photo- graph, a circular plate of rock crystal and a series of glass plates — red, yellow, green, blue and clear — were mounted with glue, while two letters N F and other figures were painted on the arm with India ink. The Fig. 28. The arm of Finsen arranged for the experiment on the influence of light upon human skin. After Pinsen. patches of glue are visible on the photograph under the rock crystal and the clear glass. After the twenty minutes' exposure the plates, etc., were removed, and the arm was now uniformly red without any visible difference between the covered and uncovered parts. This erythema, which was due to the heat radiation, was considerably diminished after two hours, when the colour of the arm was still quite uniform, but after three hours the characteristic light-erythema began to appear in the uncovered parts, reaching a maximum after twelve hours and decreasing again after two days. The India ink and all the different glass plates, including the clear glass, had been able to protect the REACTIONS TO STIMULI 105 skin completely against the effect of the light. The erythema below the rock crystal was just as pro- nounced as on the uncovered surface, except in the place where the thin layer of transparent glue had afforded some protection. The edges of the figures were sharp and well defined and corresponded closely to the paintings. The rock crystal lets through the whole spectrum of ultraviolet rays, while all the glass plates keep back the ultraviolet rays and absorb more or less of the visible rays. According to determinations made, the clear glass employed let through about 90 per cent of all visible rays. The India ink, of course, kept back all light, but increased to some extent the heating effect of the rays. The experiment shows, therefore, that the light-erythema was produced by ultraviolet rays. In other experiments Finsen showed that the visible rays from the violet and blue end of the spectrum also act upon the human skin and may produce light- erythema by sufficient concentration and length of exposure. The after history of the experiment described above is very characteristic and important from the point of view of capillary physiology. The erythema disap- peared gradually in the course of about ten days and was followed by desquamation of epidermis and brown pigmentation, leaving, of course, the covered areas in their original white colour. The pigmentation faded away very slowly. The letters were distinctly visible after two and a half months, and even after five months the areas corresponding to the glass plates could still be discerned. After about six months, how- ever, the arm was uniformly white, but even then a definite after effect of the light upon the skin capil- laries could be made out: when the arm was rubbed 106 CAPILLARIES the skin became red, but the redness was less pro- nounced in the areas which had been covered during the experiment. The light had produced an increase in excitability of the capillary wall towards the mechani- cal stimulus of rubbing. Two of Finsen's pupils, Dreyer and Jansen (1905), studied the light-erythema on the frog's tongue, ex- cluding the temperature effect by filtering the light and irrigating the tongue with cold saline. They screened off the light by tinfoil, exposing only an area of 2.5 x 5 mm. for periods varying from about five to thirty minutes. In an exposed area all vessels became dilated in a few minutes and in the capillaries stasis developed rapidly. The limit between the exposed area and the unexposed surrounding was very sharp, any vessel crossing the frontier showing a sudden decrease in diameter to the normal width. In a similar experiment Jansen (1906) pressed the frog's tongue between quartz plates, so that it remained quite bloodless during the exposure. The normal reac- tion — dilatation and stasis — developed in the exposed area when the blood was again admitted afterwards. In a further series of experiments Dreyer and Jan- sen cut through the cervical sympathetic on one side on albino rabbits and then exposed two small spots on both ears to concentrated chemically active light, one for ten, the other for thirty minutes. The spots were made anaemic during the exposure and carefully cooled. As in man the visible reaction began after several hours' latency, but always first on the side where the sympathetic had been cut. With the ten minutes' ex- posure the reaction on the operated side became very pronounced, while it was very slight on the control side. With the long exposure the reaction, though begin- ning much later on the control side, became quite as strong after about one to two weeks, and the authors REACTIONS TO STIMULI 107 note that the final restitution was always reached about one week earlier on the operated side. Finsen's experiments on man as well as the experi- ments on the frog's tongue show conclusively that we have to do with a direct action of the chemically active light rays, since there is no trace of a spreading of the reaction through nerve fibres and no indication of any reflex action. The action must be of a very pecidiar na- ture, however, since there is, especially in mammals, such a long delay before any change becomes visible. It would appear that the power of resistance of the small vessels to the internal pressure from the blood becomes very gradually lowered by some unknown process which is induced by the light, but is able to go on afterwards without further stimulus. I have been told by medical men employing local treatment with concentrated light, that it is possible to predict within one-half hour the period which will elapse between the exposure and the appearance of the erythema. The results of the nerve section experiments are probably to be explained by the vasodilatation result- ing from the cutting off of tonic impulses. The pres- sure in the arterioles and capillaries will be abnor- mally high, and the5 r will give way earlier when weakened by the light. When the skin of man is exposed to strong light at any time during a certain period after a light -inflam- mation, the reaction is greatly weakened. This is usually ascribed to the resulting pigmentation, which keeps out the light from the sensitive tissue elements. This is not the whole truth, however. C. With (1920) bas made some interesting experiments on the reac- tion to repeated treatment with the Finsen light of vitiligo areas in which no pigmentation develops. Even here a definite immunity to the light followed the inflammatorv reaction due to the first treatment. 108 CAPILLARIES £he reactions of capillaries to chemical stimuli. Numerous substances have been described which act upon the capillary wall and induce a change in its state of contraction. Almost all these substances have a more or less powerful dilator action and produce, when locally applied, a state of inflammation. Heubher (1907) has made a sharp distinction be- tween inflammatory poisons and what he terms capil- lary poisons. According to him, inflammatory poisons have an irritative action upon all tissue elements and give rise to more or less distinct reactions even in tis- sues where vessels are absent. One of the reactions, the importance of which is specially emphasized by Heub- ner, is the pain produced by the action of inflamma- tory poisons on sensitive nerve endings. As true capil- lary poisons Heubner classes those substances which have a selective action on the capillary wall, the tonus of which they diminish or abolish, while they do not stimulate pain receptors and are indifferent to non- vascular tissue. By their action on capillaries they may give rise when applied locally to secondary symptoms which resemble those of inflammation. I believe that the principle underlying Heubner 's classification is sound, though there are many sub- stances which cannot at present be referred with any certainty to one group or the other. I have little doubt, however, that the group of inflammatory poisons is heterogeneous and comprises substances which differ fundamentally as to their mode of action. The reactions to inflammatory poisons. I shall give some examples to illustrate the circula- tory changes brought about by inflammatory poisons. Bicker and Begendanz (1921) have studied the local action of a number of substances on the pancreas and conjunctiva of the rabbit. Among these is iodine, dis- REACTIONS TO STIMULI 109 solved in saline by the addition of two parts NaJ to every part of iodine, and applied as an irrigating current of indifferent temperature over the area studied microscopically. They find that irrigation of the pancreas with 0.002 per cent iodine for three minutes will bring about contraction of both arteries and capillaries, while by the application of stronger solutions, 0.02 to 0.3 per cent, the initial contraction will give place, about fifteen minutes after the applica- tion of the weakest of these solutions and about three minutes after the application of the strongest, to gen- eral dilatation of the capillaries. After application of a 2 per cent solution no contraction of capillaries can be made out, though the arterioles still contract for a short period. All the stronger solutions show after effects per- sisting for days and characterized by leucocytosis, stasis in the capillaries and small bleedings. With the weakest solution (0.002 per cent) the inflammatory symptoms after fifty-five hours are slight only, but even in this case there is a persisting dilatation of capillaries with a considerable increase in the number of white corpuscles present. In the inflammatory stage the application of adrenaline (0.1 per cent) has not the usual effect of producing arterial contraction, but will bring about a general stasis in the area affected. In the conjunctiva iodine acts differently in so far as the initial constrictor effect is extremely slight and of very short duration when it is present at all. It has, in fact, been observed only once after 0.005 per cent iodine. The dilatation is pronounced even after 0.0005 per cent iodine, and even this weak solution shows very pronounced inflammatory after effects lasting eight days. During the inflammatory stage adrenaline does not produce contraction but dilatation of arteries and capillaries and general stasis. 110 CAPILLARIES To account for the reactions after iodine and a large number of similar reactions after the application of other substances Richer assumes a definite arrange- ment of vasomotor nerves, consisting in a set of con- strictor fibres supplying both arteries and capillaries and a corresponding antagonistic set of dilator fibres for the same vessels. He assumes further that very weak stimuli will be ineffective on the constrictor fibres, but will cause dilatation by stimulating their antagonists. Stronger stimuli may produce contrac- tion through their action on constrictor fibres, but these are assumed to be easily paralysed by strong stimulation, especially in the capillaries and arterioles. After the paralysis of the constrictors of the smaller vessels the response of the dilators, which are taken to be much more resistant, will, in combination with an assumed constriction of larger arteries, supplying the field of microscopic observation, account for the phe- nomena observed at this stage. In proof of this system of suppositions only considerations of a very general kind are offered. No attempt has been made to see if the responses were able to spread, or if they could be abolished or modified by degeneration or paralysis of the nerves taken to be involved. In these circumstances it is not worth the trouble to examine in detail whether the reactions observed correspond to the explanation or not. The real and great significance of the experiments undertaken by Ricker and his collaborators lies in their objective study of the prolonged after effects pro- duced by inflammatory stimuli. These effects cannot be deduced from any special reaction of the nerves or the vessels taking place during the application of the "stimulating" substance, but we must assume that a change is initiated in the general activities of the tis- sues, involving probably all the elements present. The REACTIONS TO STIMULI 111 proliferation of connective tissue cells, the destruction or alteration of epithelial cells, the changes in reactiv- ity of nerves and vessels, as evidenced by the abnormal responses to adrenaline and other stimuli, all bear wit- ness to processes going on of such a complicated nature that attempts at a causal analysis must, in the present state of our knowledge, appear hopeless. It is a very characteristic phenomenon in the inflam- matory reaction that the circulatory and other dis- turbances observed several hours or days after the application of the stimulus are much more pronounced than the immediate effects. One is strongly reminded of the capillary reactions to chemically active light, which produces in the skin of mammals practically no immediate effect, while the later reaction is essentially of an inflammatory character. A comparison between two of the experiments of Ricker and Regendanz, in which practically the same initial symptoms were observed, while the after his- tory was very different, will furnish a useful illus- tration. 1. In the course of ten minutes a small quantity of pulverized pumice stone is thrice introduced into the conjunctiva of a rabbit, which is thereupon carefully washed with saline. The immediate effect is a strong hyperemia of the surface with some cedema and, after a few minutes, stasis over the whole surface. The deeper vessels are not affected. The next day there is still a general hypersemia, no oedema. The deeper layers of capillaries are also some- what dilated. The current of blood is rapid. On the fourth and fifth day there is no macroscopic hyperse- mia. Microscopically the superficial vessels are slightly dilated, but the current of blood is now slow. The number of open capillaries is slightly increased. The deeper vessels are on the fourth day slightly di- 112 CAPILLARIES lated. The reaction to adrenaline is at this stage the normal constriction. 2. Abrine, 0.01 mg. in 0.1 cc. saline, is instilled into the conjunctiva. The eye is kept open for five minutes, during which time the fluid disappears. There is some hyperemia and slight oedema of the surface. All the vessels which are microscopically visible are dilated. The current is rapid. This initial effect is even slighter than that seen after the mechanical insult. There is no stasis, but the substance has evidently penetrated into the tissue and affected also the deeper vessels. Eight hours later there is niacroscopically a pro- nounced hyperemia and slight oedema of the con- junctiva. Microscopically all the visible vessels are strongly dilated. In some parts of the surface there is capillary stasis, in all others a slow current. The next day the conjunctiva contains pus and is very hyperse- mic. Stasis has developed in almost all superficial capillaries and a number of capillary bleedings have appeared. During the next two days the symptoms of inflam- mation, especially the bleedings, increase, but there- upon recovery processes set in, and after fifteen days the appearance of the conjunctiva is practically nor- mal. A second instillation of 0.01 mg. abrine under- taken on the fifteenth day gives rise to even more pro- nounced inflammatory symptoms, which even thirty- two days later have not returned completely to normal, in that the capillaries, especially in the deeper layers, are still dilated. The application of adrenaline at this stage produces an immediate hyperseniia, but after two or three minutes the constrictor effect on the arteries of the deeper layer asserts itself. That the tissue cells themselves are strongly af- fected by the poison is evident from the fact that the REACTIONS TO STIMULI 113 cornea becomes opaque and new capillaries develop and grow into the periphery of the cornea. Though the mechanical insult in the first of these two experiments has given rise to a slight inflamma- tion, the difference between them is sufficiently striking to make any further comment seem unnecessary. It can be demonstrated for certain substances and is probably the case for a number of others that nervous reactions play some part in the initial inflammatory symptoms. In my own experiments (1920) on the frog's tongue, iodine (1 per cent in potassium iodide), applied in very small drops under the microscope, pro- duced a violent contraction of the underlying muscles. The capillaries of the mucous membrane became strongly dilated, and the dilatation spread to a con- siderable distance. The arteries were affected also, but their dilatation was not so pronounced as that of the capillaries. After the application of cocaine to the mucous mem- brane, so as to paralyse the sensory nerves and nerve endings, a small drop of iodine had no effect, and in another experiment, in which iodine was applied at 1.5 mm. distance from a cocainized area, the reaction be- came much weakened over the cocainized area, though it did spread into it. In a further experiment iodine was applied to the mucous membrane after section and degeneration of the lingual nerves. No contraction of muscles took place ; the capillaries directly affected became some- what dilated, but no spreading could be observed. It is clear from these experiments that nerves are involved in the reaction to iodine. The nerves respon- sible in this case are probably the sensory fibres which produce dilatation through local axon reflexes. In the mustard oil experiments by Bruce, Bardy and Breslauer, referred to at some length in the preceding 114 CAPILLARIES lecture, it is likewise obvious that a considerable part of the reaction is due to the effect of the poison upon sensitive nerve endings, but experiments made by Bicker and Regendanz on normal, anaesthetized and denervated eyes of rabbits show indisputably that mus- tard acts as a powerful poison also to the vessels them- selves and in all probability to the tissues generally. 1 i In my paper on the vascular reactions in the frog's tongue (1920) it is erroneously stated that mustard oil has no influence upon the capil- laries of that organ. The action is very pronounced. LECTURE VI THE REACTIONS OF CAPILLARIES TO STIMULI (CONTINUED) Capillary poisons. AS a type of substances which show a selective /\ action on the contractile elements of the capil- jL_ _^_lary wall Heubner has studied the gold salt, AuCl 4 Xa + 2ELO, and describes the effects of an intravenous injection of a lethal dose of this substance in mammals (rabbit, cat and dog). The lethal dose for a rabbit is about 15 mg. per kg. body weight. For the carnivora it is about three times as high. The arterial pressure begins to fall during the injection and con- tinues to fall, reaching zero in a few minutes (not above ten), when the animal dies, though the heart- beat may continue for a short time afterwards. At autopsy the parenchymatous organs, as also the lungs and muscles, appeared to contain an abnormal quantity of blood, and bleedings from minute vessels were fre- quent. The intestines, especially the empty stomach, duodenum and jejunum, of the carnivora were abnor- mally injected with numerous patches of a deep red in the mucosa, in spite of the fact that the animals had been without food for a day before being used for the experiments. In several cases considerable quantities of blood were found in the abdominal cavity, though no macroscopic wounds could be detected. The microscopic examination of the organs showed dilated and very numerous capillaries and dilated 116 CAPILLARIES venules and very numerous microscopic bleedings from capillaries and venules, especially in the liver, lungs and kidneys. The small arteries were everywhere con- tracted, in most cases even to obliteration of the lumen. These post-mortem observations, together with the sudden fall of arterial pressure preceding death, show clearly that Ave have to do with a relaxation of the capillaries and venules to such an extent that only a fraction of the blood poured into them could return to the heart to keep up the circulation. A picture of the reaction which is even clearer has been obtained by Heubner through microscopical observation of the mesentery of curarized frogs before, during and after injection of gold salt into a femoral vein. By careful preparation Heubner secures a nor- mal circulation in the mesentery. The injection of 0.5 cc. 0.1 per cent gold salt makes no immediate dif- ference, but after one-half to one minute, when the poison reaches the mesentery, a very large number of new capillaries are quite suddenly opened up. The network of available capillaries is increased about three to four fold, and macroscopically the colour of the exposed piece of intestine changes from pale to a distinct red. Very soon afterwards the circulation be- comes visibly slower and ceases altogether after a few minutes. The animal is "bled to death into its own capillaries," as Heubner himself aptly puts it. Heubner mentions a number of other substances which act as capillary poisons. These include a number of double salts of heavy metals, belonging to the gold and iron groups, arsenic and the organic bases emetine and sepsine (Faust, 1904). Several of these show, of course, other toxic properties beyond that exercised upon the capillary wall, which latter becomes domi- nant only when they are introduced in suitable concen- tration directlv into the blood. For some of them, REACTIONS TO STIMULI 117 especially when given in closes which are not imme- diately lethal, the capillaries of the intestine appear to constitute the place of predilection: they produce an enormous hyperemia of the mucous membrane with numerous capillary ecchymoses. In the class of capillary poisons must be reckoned also histamine, by means of which Dale and Richards (1918) made the beautiful study of capillary reactions referred to at length in my second lecture. It will be remembered, as the result of this study, that small doses of histamine injected intravenously in cats (or other carnivora) produce an evanescent fall of blood pressure, which analysis showed to be due to capillary dilatation. Dale and Laidlaw (1919) have investigated the even more striking effects of the introduction into the cir- culation of larger quantities of histamine. The blood pressure changes resulting from the injection in cats of a single large dose (1 to 2 mg. histamine per kg.) are complicated by the effects of the poison on other organs, but the general effect is a very pronounced fall in arterial blood pressure, which reaches in a few minutes a level of 50 to 30 mm. By a slow infusion of a dilute solution of histamine the initial irregularities of the action can be avoided, and the resulting blood- pressure curve shows an initial rapid fall in pressure, corresponding to that observed after a single small dose, and a subsequent slower decrease, during which the pulse waves become gradually smaller, until finally they disappear almost altogether. When the heart is inspected at this stage it may still be beating vigor- ously, but it drives very little blood into the systemic arteries, because the supply of blood through the sys- temic veins, which are found to be flaccid and empty, has failed. At this stage an intravenous injection of a sum- 118 CAPILLARIES ciently large quantity of saline will bring about a rapid recovery of the output of the heart and a con- sequent increase in the blood pressure, showing that neither the force of the heart nor the peripheral arte- rial resistance has been impaired. Pig. 29. Effect of 10 mg. histamine on lilood pressure in dog. Pig. 30. Continuation of Fig. 29. Eleven minutes after the injection. After Dale and Laidlaw. The blood, not being returned to the heart, must evi- dently be stored somewhere, and Dale and Laidlaw made a systematic search for it. As described above, it was absent from the great veins, and the arterial system was found to contain very little blood. ' ' There remained the capillaries and venules, and here, at last, signs of accumulation of blood could be detected. The signs were clearest in the case of the abdominal vis- cera, possibly on account of the greater transparency of the tissues. If the abdomen was opened in the earlier stages of the main fall of the blood pressure, the intes- REACTIONS TO STIMULI 119 tines looked diffusely and somewhat duskily red, and the network of venules stood out distinctly. The pan- creas had a purplish, congested appearance." In the soeletal muscles no definite accumulation of blood could be made out in the ordinary experiments, but, when a state of plethora was brought about by transfusion of blood from another cat, the general peripheral engorgement produced by histamine was very striking. The sceletal muscles were deep red and obviously contained much blood. There can be no doubt, therefore, that the muscle capillaries become relaxed by histamine, though they are probably able to withstand the toxic influence a little longer than the intestinal. A very instructive comparison has been made be- tween the effect of histamine and another "depressor" drug, acetyl-choline. By a carefully adjusted, very slow infusion of a dilute solution of acetyl-choline into the veins of a cat under ether it is possible to produce a persistent low arterial pressure of 40 to 50 mm. by vasodilatation alone without any complicating reac- tions. In such an experiment the output of the heart remained thoroughly efficient. The intestines showed a pink flush and pulsated obviously, and the great veins were well filled, but not overdistended. Perhaps the most striking contrast with the effect on the circulation produced by histamine, when similarly administered, was presented by the events following the stoppage of the infusion. After acetyl-choline the arterial pressure began immediately to rise from the low level, at which it had been maintained, and with an extraordinary rapidity regained or surpassed the level at which it had stood before the infusion began. After similar treat- ment with histamine the typical picture is the circula- tory "shock" described. Nothing can be more striking than this contrast 120 CAPILLARIES between the result of arteriole dilatation and capillary dilatation, respectively. 1 Studying the depilated ears of cats under the micro- scope by reflected light Hooker (1920) has been able to follow the capillary reactions after intravenous injec- tion of a suitable dose of histamine. "The results were uniformly clean-cut and decisive. Within a few minutes after the injection the capillaries and venules were filled with stagnant blood and definitely dilated. The dilatation was distinctly more conspicuous in the venules. These changes developed in conjunction with the fall in arterial blood pressure." For the same purpose Rich (1921) has employed a very different method. Flooding the peritoneal cavity of cats with Zenker's fluid, he obtained a prompt fixa- tion of the omentum in which the capillaries could afterwards be studied microscopically. If the tissue was thus fixed immediately after the intravenous in- fusion of histamine, marked capillary and venous dila- tation and engorgement could be demonstrated. This vascular change was entirely absent in control experi- ments in which salt solution was infused instead of histamine. Miss Carrier (1922) in my laboratory has tested the local effect of histamine on the skin of man. At the i For many years the circulation was regarded as if the heart action and the vaso-motor (arterio-motor) mechanism were the only variable factors, and as if any failure or depression not due to the former must be due to the latter. Observations by Yandell Henderson (1908) upon an experimental form of shock first demonstrated clearly the occurrence of what he termed failure of the ' ' veno-pressor mechanism, ' ' characterized by decreased and finally inadequate venous return to the right heart, re- sulting in decrease in cardiac filling and discharge, while the arteries at the same time were found to be, not relaxed, but constricted. Other papers by Henderson and his collaborators (1909, 1910) brought forward evi- dence in favour of the view that it is this mode of circulatory failure rather than ' ' vaso-motor failure ' ' which is characteristic of shock and, although there are a number of matters still needing further elucidation, this conception has now come to be generally accepted. REACTIONS TO STIMULI 121 base of the nail histamine (ergamine phosphate 1: 1000), introduced into the skin by means of a glass capillary a few hundredths of a mm. in outside diame- ter, produced some widening of the nearest capillaries with hastening of the stream. The increase in the cur- rent might be evidence for a dilatation of arterioles were it not that the arterial loop of the nail capillaries (see Fig. 35) is often so narrow that a definite increase in its diameter may very well diminish the resistance, especially when other capillaries, supplied from the same arteriole, are not simultaneously dilated. When applied in the same way on the back of the hand, the introduction of this minimal quantity of histamine gave rise, after a latent period of a few seconds, to a sharp, painful itching, lasting one to five minutes and producing in turn a definite reflex erythema over an irregular area of 3 to 12 cm. 2 The erythema, of course, prevented the estimation of any direct dilator effect produced by the histamine, but the pain aroused by the histamine is of interest as an illustration of the difficulty of any hard and fast classification of sub- stances acting on capillaries. Histamine appears, according to Dale and Laidlaw (1910, 1911), to be a poison for the capillaries of cer- tain animals only — dog, cat, monkey, fowl and man — - but not for herbivorous mammals (rabbit, guinea pig); Doi (1920) has found it active by intravenous injec- tion for the capillaries of the frog's web, but I (1921) have been unable to observe any activity by local appli- cation in any concentration to web capillaries, and a single perfusion experiment made by Rehberg and myself, in which histamine was added in the concentra- tion of 1 : 1000000 to a perfusion fluid made up of gum Ringer and corpuscles, failed to give any evidence of dilator activity. Even in an animal like the cat, where histamine is 122 CAPILLARIES undoubtedly active, its dilator action on the capillaries depends upon certain conditions which are, I believe, far from being completely understood. It will be remembered, from the second lecture, that Dale and Richards found that the dilator action of his- tamine could be demonstrated in perfusion experi- ments only when the perfusion fluid contained enough red corpuscles to provide an adequate supply of oxy- gen and, further, a small amount of adrenaline, but failed to appear when these conditions were not ful- filled. Dale and Richards ascribed this failure to loss of tone of the capillaries, but the explanation appears to me very doubtful, since they were able by means of their beautiful technical arrangement to change with- out interruption from the normal circulation of the animal to the artificial perfusion, and the "loss of tone" or spontaneous dilatation must take some time to develop. In a recent paper Burn (1922) has shown that after section and complete degeneration of the nerve fibres to one leg of a cat the limb does not show the usual plethysmography response — increase in vol- ume — to a small dose of histamine intravenously ad- ministered, but shows passive contraction, due to the fall in arterial blood pressure, instead. By depriving a leg of the sympathetic innervation alone Burn was able to show that lack of sympathetic tonus did not interfere with the normal response, and he concludes, therefore, that capillary tone has been seriously impaired by the interruption of the connection with posterior root fibres. Though the circulation in the leg deprived of its posterior root fibres may very well have suffered, any considerable dilatation of capil- laries cannot, I think, have taken place, because that would involve a change in appearance — especially of the pads of the foot, which were regularly inspected — which could not have escaped notice; they would, at REACTIONS TO STIMULI 123 the very least, have become of a deeper red colour. There appears, therefore, to be something beyond loss of tone which is able to prevent the response of capil- laries to histamine. Is it allowable to pnt forth the con- jecture that this unknown something might also be responsible for the absence of histamine dilator reac- tion in most animals 1 Histamine seems to be normally produced in the mucous membrane of the small intestine (Barger and Dale, 1911). Other substances with a histamine-like action on capillaries are produced in diverse tissues under pathological conditions. These will be dealt with later (Lecture XI). The reactions of capillaries to narcotics. In connection with the capillary poisons, the action of narcotic substances on capillaries may appropri- ately be referred to. In experiments on the tongue of the frog I found that urethane shows a very powerful dilator action on capillaries, while it is indifferent towards arteries and arterioles. When a microscopically small drop (that is, a small fraction of 1 cubic millimeter) of 25 per cent urethane is applied to a capillary in the spread ventral surface of the frog's tongue a complete relaxation occurs. The capillary may become filled quite gradu- ally from an arteriole which remains so narrow that the corpuscles are coming through one by one. The relaxed capillary may reach a diameter finally of 50^ and show a peculiar varicose appearance. In such a capillary complete stasis develops, and even after sev- eral days the vessel may remain filled with a mass of corpuscles. Application of 25 per cent urethane in larger quantity may produce a maximum dilatation also of the muscle capillaries below the mucous mem- brane, but this is accompanied by much twitching of 124 CAPILLARIES the muscles, which may of itself cause dilatation of the capillaries. Application of 5 per cent urethane pro- duces the same reaction in the superficial capillaries, but it develops more slowly, and even a 1 per cent solution may produce a moderate dilatation when re- maining on for several minutes. This latter concentra- tion is only about twice to thrice as high as the ure- thane concentration in the blood of a completely narcotized froe;. Fig. 31. Diagram of the effect of urethane on capillary in frog 's tongue. Experiments with urethane on the cocainized sur- face of the frog's tongue show that the effect of strong solutions, at least, is partly indirect. On the anaesthetic surface the reaction develops much more slowly and, after 5 per cent urethane, the dilatation never becomes maximal and no stasis occurs. In the skin and web of the frog the action of urethane upon the diameter of capillaries is less striking, but a definite dilator action of a 5 per cent solution can be made out. In the bladder we have failed to observe any effect of urethane solutions up to a strength of 25 per cent. The effect of the volatile narcotics, chloroform and others, on the tongue and web of the frog are similar to those of urethane, but chloroform shows in addition a very peculiar effect upon the red corpuscles, which become strongly adhesive to one another and to the REACTIONS TO STIMULI 125 capillary walls. Dale and Laidlaw (1919) found in their experiments on histamine shock that while un- anaesthetized cats will stand comparatively large doses of histamine without developing shock, that is, without any extreme and irreparable dilatation of the capil- lary system taking place, anaesthetized animals are vastly more susceptible. This point has been worked out by the British Committee on Surgical Shock (Medi- cal Research Committee, 1919), who find both in experiments on animals and in observations on pa- tients suffering from traumatic toxaemia that the onset of shock symptoms are greatly accelerated and may be directly induced by anaesthetization. They find, further, that the depth of the anaesthesia is very important : the deeper the anaesthesia the greater liability to shock. The kind of anaesthetic used, chloroform, ether, ure- thane, seems to be irrelevant, but it is definitely stated that anaesthetization with nitrous oxide gas can be used with impunity. These observations are to be under- stood in the light of the above experiments. The ordi- nary anaesthetics have themselves a dilator effect upon capillaries which is just imperceptible by itself at the concentration inducing complete anaesthesia, but is, nevertheless, sufficient to cause more or less complete loss of tone in the capillaries when its effect is added to that of some other capillary poison. It is quite con- ceivable and even likely that a synergistic action may take place, in the sense that the combined effect of two such substances is greater than could be inferred from simple addition of the effects of each substance. So far I have dealt with the effects on the contractile mechanism of capillaries of substances which are, on the whole, foreign to the organism. Some of them are of interest only from a toxicological point of view, while others may be of considerable importance from pharmacological and even pathological standpoints. 126 CAPILLARIES We come to deal now with substances which are always present in the organism and may be supposed to play some part in its normal life as regulators of capil- lary circulation. Unfortunately, our present knowledge regarding the action of such substances is of a very fragmentary character, and what I can say about them will serve more as a reconnaissance into a very prom- ising field of research than as a definite map of the essential features of that field. Hydrogen ions. A large number of investigations, among which spe- cial reference should be made to the famous experi- ments of Chauveau and Kaufmann (1887) on the leva- tor labii muscle of the horse and to the beautiful researches of Barcroft (1907, 1915) on the submaxil- lary gland, have shown that the flow of blood to active organs is increased, and tbat this functional hypere- mia is brought about by some reaction on the part of the active tissue itself. It is generally believed that this reaction is simply the increased formation of acid metabolites, especially carbonic acid, which is insepa- rably bound up with increased activity, but I am afraid it must be admitted that the foundations of this latter belief are not very secure. It has been shown repeat- edly that the addition of very dilute acids to perfusion fluids will increase the volume perfused at a given pressure, or, in other words, produce a decrease in the resistance, which must be due to arterial dilatation, but in a recent paper Fleisch (1921) has shown that in almost all the experiments hitherto made the hydrogen ion concentrations employed have been many times higher than they can ever occur in living tissue. Let me recall to you that in normal arterial blood there is 1 gram equivalent of hydrogen ions in 22 mil- lion liters of blood, that is, the hydrogen ion concentra- REACTIONS TO STIMULI 127 tion is 1 : 10 7,35 normal, or, as it is generally expressed, the p H of the blood is 7.35. The normal p H of mixed venous blood is 7.3. The acidity of the tissues them- selves is perhaps a little higher and with maximum ac- tivity is may conceivably rise to p H = 7. The p H of the solutions generally used to demonstrate the dilator ef- fect of acids in perfusion experiments has been in the neighbourhood of five or even four ; that is, they have been from a hundred to several thousand times more acid than the blood can possibly become. Fleisch him- self has made a very beautiful series of experiments with perfusion fluids, which were buffered by the addi- tion of phosphates so as to maintain a definite p H , which could be adjusted to any desired figure, and he has shown that an increase in hydrogen ion concentra- tion even from p H = 7.6 to p H = 7.5 will produce a dis- tinct increase in the volume perfused. The increases obtained by Fleisch with hydrogen ion concentrations within physiological limits are certainly too small, however, to account for the increase in blood flow through active organs, and it must be argued further that his perfusion fluids have contained too little oxy- gen to supply the needs of the tissues. Lack of oxygen is, as we shall see, a very serious complicating factor, and I find it impossible to accept as binding the evi- dence at present available for the view that the in- creased supply of blood to active organs is brought about exclusively or even mainly by the vasodilator action of acid metabolites. From the point of view of these lectures we are interested not so much in the absolute increase in blood flow through active organs as in the simultaneous open- ing up and dilatation of capillaries, which probably take place in all organs during activity and which have been demonstrated most clearly in the case of muscles. We shall examine, therefore, the available evidence 128 CAPILLARIES regarding acids, as substances which bring about capil- lary dilatation. We shall have to consider first a few experiments in which acid solutions have been brought in contact with the tissues from outside. These will require some introductory words of explanation. Any acid or any substance having an acid reaction will, when brought in contact with living tissue, bring about a wandering, in all probability by simple diffu- sion, of hydrogen ions into the tissue. These hydrogen ions will increase in turn the hydrogen ion concentra- tion of the tissue, but to what extent that will take place cannot be predicted with any certainty, because the living tissues and the blood contain the so-called ''buffer" substances (notably bicarbonates) and resist by chemical combination any increase in their hydro- gen ion concentration. All that can be said, therefore, is that within the tissue, so long as it is living, the increase will probably be very small compared with the hydrogen ion concentration of the acid applied. It is important to bear this fact in mind, when the fol- lowing experiments are considered. When a small drop of 1 per cent acetic acid is ap- plied to the ventral surface of the spread tongue of a frog, dilatation of both arteries and capillaries takes place, and the reaction spreads at once to a distance of a couple of millimeters beyond the area directly af- fected. When the surface of the tongue is cocainized, or when the lingual nerves have been cut, the reaction is still present, but it is much diminished in intensity, especially with regard to the capillaries, and it does not spread. It follows from these observations that a certain, so far undefined, but probably considerable, increase in the hydrogen ion concentration mil affect the sensitive nerve endings of the tongue and also directly the smooth muscle elements of the arterioles and to a less extent of the capillaries also. REACTIONS TO STIMULI 129 In another experiment — performed for me by Dr. Harrop — buffer mixtures of definite hydrogen ion con- centrations were made up and applied to the ventral surface of the tongue in reaction basins, which are paraffined brass rings of 3 to 4 mm. diameter and 2 mm. broad, put on to the tissue under investigation and filled with the fluid, the effect of which is to be studied. The buffers used were Sorensen's mixtures of sodium N citrate with — hydrochloric acid. To our surprise we had to use very acid mixtures to obtain any dilator effects on the capillaries. A mixture of 5 volumes cit- rate with 5 volumes HC1 having a p H = 3.65 had no effect, and even for the next mixture, 4 citrate + 6 HC1 with a p H of 2.96, the result was very doubtful, while 3 citrate + 7 HC1, p H = 1.94, gave a definite, but slight, N dilatation after a short latency. Pure — HC1, p H = 1, gave a well-marked dilatation. A definite change in the hydrogen ion concentration of the tissue can be brought about by changes in its carbon dioxide tension. We have at present no reason to believe that C0 2 acts in any other way than by form- ing with water a very weak acid, and, in any case, stronger acids, which are formed in or brought into the tissue, will react in the first instance with the bicar- bonates present and thereby raise the C0 2 tension. The spread tongue of a frog was exposed to C0 2 by means of the apparatus shown in Fig. 32 — a slight modifica- tion of Roy and Brown's apparatus for measuring capillary blood pressure, mentioned in my second lec- ture. The gas or gas mixture is led into the lower chamber, the top of which is covered with a thin peri- toneal membrane, thence through the tubing to the upper chamber and from this to a tube dipping in water. The spread tongue is inserted between the two 130 CAPILLARIES chambers and the slight gas pressure maintained is sufficient to make the system gas tight. When a current of pure C0 2 gas is led through this apparatus it will produce, after about one minute, a considerable hy- peremia, with dilatation of both arteries and capil- laries and a rapid current of blood. A mixture of 10 per cent C0 2 in air produced in one experiment after a couple of minutes a distinct increase in the circula- Fig. 32. Apparatus for application of gases to capillaries under the microscope. tion through a few capillaries selected for observa- tion, but the general increase over the whole field was too slight to be definitely ascertained, and no widen- ing of capillaries could be made out. In these experi- ments the C0 2 tension of the tissue has corresponded, at least very nearly, to the C0 2 percentage of the gas mixture. The C0 2 diffuses very rapidly through the tissues, as will be shown in some detail in a following lecture (IX). It will enter into chemical combination to a certain extent at first, but saturation will be reached very soon, and, though some C0 2 is undoubt- edly carried off by the blood, this is compensated by the constant formation of C0 2 in the tissue. The actual hydrogen ion concentration reached in such an experi- REACTIONS TO STIMULI 131 ment can be made out with considerable accuracy by saturating frog's blood with C0 2 at the same pressure and determining its p H by a suitable method. 1 For practical purposes it is sufficient, however, to consider the C0 2 tensions. Normal frog's blood pos- sesses a C0 2 tension between 1 and 2 per cent ; it will be in diffusion equilibrium as regards C0 2 with an atmosphere containing between 1 and 2 per cent C0 2 . A C0 2 tension of 10 per cent, therefore, means a de- gree of acidity which probably never occurs in the nor- mal frog, and it follows that, though the vessels, and especially the arteries, of the frog's tongue are dilated under the influence of increased hydrogen ion concen- tration, this reaction cannot play any considerable role in the normal regulation of the blood supply and ( probably plays no part whatever as a regulator of the capillary circulation. We have now to consider some experiments and ob- servations on the rabbit's ear made by Mr. Rehberg and myself. They were undertaken for another pri- mary purpose, to which I shall have to return later, but incidentally they can be used to give some informa- tion on the problem under consideration. In the first of these experiments the rabbit was breathing through a tracheal cannula, while the de- pilated transparent ears were watched both macro- scopically and microscopically by transmitted light. When the "dead space" of the lungs was increased by adding a tube of 15 cc. to the tracheal cannula, a very pronounced hypera?mia of the ears, comprising capil- laries as well as arteries, would set in after about two minutes. At the same time the blood became visibly cyanotic. When the extra dead space was removed the 1 In such a determination, made later, it was found by Mr. Easmussen that saturation with 10 per cent CO, would lower the p H of a frog's blood from 7.5 to 7.1. 132 CAPILLARIES hyperaemia would disappear in the course of one-half to one minute. By increasing the dead space to 35 cc, the effects could be intensified and brought about more quickly. The cyanosis in such an experiment became very pronounced. These experiments show that an in- creased venosity of the blood brings about both arte- rial and capillary dilatation, but they leave it unde- cided whether the dilatation is central or peripheral in origin, and whether it is due to an increase in CO; tension or to lack of oxygen, or both. In a further experiment, a rabbit was made to breathe air with 10 per cent CO, from a spirometer. The CO- percentage of the expired air rose to 11.47 per cent, but, in spite of this considerable increase in the hydrogen ion concentration, no dilatation of capil- laries or arteries could be detected in the ears. "When the rabbit was given about 14 per cent CO^ to breathe, a definite hyperaemia could be made out, however. This corresponds to a decrease in the p H of about 0.4, ac- cording to a determination made by Mr. Easmussen. We next tested the influence of oxygen lack by let- ting the animal breathe from a small, closed-circuit respiration apparatus, in which the carbon dioxide was absorbed while the oxygen percentage was allowed to fall. As soon as cyanosis began to appear hypera?mia of the ears developed and became very much pro- nounced as the oxygen content of the blood was dimin- ished. In a case like this, organic acids are undoubtedly produced and will be present in the blood, perhaps in considerable quantities. They will not, however, be able to raise the hydrogen ion tension of the blood to any appreciable extent, since they will liberate CO^, which will be eliminated through the lungs. The experi- ments of the British Committee on Shock (Medical Research Committee, Xo. 25, p. 272), among others, clearlv show that verv large amounts of acid can be REACTIONS TO STIMULI 133 introduced into the blood of a living animal without producing any perceptible change in hydrogen ion con- centration, so long as the respiratory center maintains its activitj-. In our case, the C0 2 percentage of the expired air remained low (3.00 per cent when the 2 had fallen to 7.17 per cent and the hyperaemia was marked), so that it can be definitely asserted that the hydrogen ion concentration of the blood cannot have been responsible for the hyperaernia, which must be associated in some other way with the lack of oxygen. We have made a few experiments to see whether the hyperaemia due to lack of oxygen is central or periph- eral in origin. In some of these we had cut the syinpa- , thetic on one side in the neck and cut further the ante-i rior and posterior nerves to the ear itself. No perceptible difference could be observed between th(j hyperaemia in the two ears as a result of lack of oxy- gen. We have some reason to believe, however, from a later experiment on the same ear, that the sympathetic innervation had not been abolished. 1 In another ex- periment of the same kind the arterial hyperaemia on the operated side was definitely less pronounced. Both arteries and capillaries remained comparatively nar- row, but a large number of hitherto closed capillaries were opened up, so that we must conclude that an essential part, at least, of the capillary dilatation was due to a peripheral effect of the oxygen lack. It is worth referring to in this connection that Eoth- lin (1920) has found that the tonus of isolated rings from larger mammalian arteries is a simple function of the oxygen tension of the Ringer solution in which they are suspended. As their oxygen consumption is i Fletcher (1898) has shown that the sympathetic fibres to the rab- bit, 's ear do not all pass through the cervical sympathetic. Several of them pass from the stellate ganglion along the ramus vertebralis on to the third cervical nerve and thence to the ear. 134 CAPILLARIES probably slight and their power of altering the p H of the solution by their respiratory exchange certainly negligible, Eothlin's results point, as he himself has emphasized, to the oxygen pressure as having a tonic influence upon the arterial wall. It is quite possible, nay probable, that metabolic products formed during activity have a dilator effect upon capillaries and arteries, but it seems to me very unlikely that such an effect is due to their supposed acid properties. I am painfully aware of the unsatisfactory character of what I have been able to say about the effect of acids on the capillaries and about the extremely important problem of how the regulation of the supply of blood and its distribution through the capillary network is brought about during activity, but I hope that I have at least made out a case for a renewed investigation, in which the point of view of oxygen lack must not be lost sight of, though there are good reasons, to be given in some detail in a later lecture (IX), for the belief that there will not generally be any oxygen lack during activity. Adrenaline. The action of adrenaline upon capillaries is, per- haps, even more puzzling than the action of that sub- stance upon arteries. Until quite recently it came pretty near to being an axiom in physiology that ad- renaline was strictly sympathicomimetic and could be relied on to have the same effect on any tissue as that produced by a stimulation of the corresponding sympathetic fibres. Since a number of capillaries have been shown to be innervated through the sympathetic, the stimulation of which causes prompt contraction, this proposition would involve that the capillaries REACTIONS TO STIMULI 135 should respond by contraction when a suitably dilute solution of adrenaline was applied to their walls. Cer- tain capillaries do so respond, but others undoubtedly do not. Since the reactions of capillaries to adrenaline are bound up more or less closely with those of arte- ries, it will be necessary in the following discussion to deal also with certain aspects of the response of arte- ries to adrenaline. In the tongue of the green frog (R.esculenta) where, as I must admit, the existence of a sympathetic innervation to the vessels has not been demonstrated, the capillaries invariably respond to the application of 0.1 per cent adrenaline, in small or large drops or in reagent basins, by distinct and, as a rule, very pro- nounced dilatation. Most of the smaller arteries and arterioles also dilate, while the larger arteries remain unaffected. A few small arteries have been observed to contract for a brief period and in one of these cases a simultaneous dilatation of the corresponding capil- laries was distinctly seen. In the tongue of the brown frog (R.platyrrhina) numerous arteries are susceptible to the action of adrenaline and contract strongly, but several, and especially the largest, arteries are not affected and the capillaries dilate as in R.esculenta. In the internal organs of the frog (R.esculenta: stomach, intestine, bladder) numerous arteries and arterioles are refractory towards adrenaline, and capillaries have not been observed to respond at all. In the muscles, on the other hand, all the arteries and arterioles so far tested have responded very promptly to the application of even the smallest drops of adrenaline, while careful observation shows the capillaries to remain unaffected in spite of the fact that they can be brought to contraction by stimulation of the sympathetic. It is necessary to emphasize the 136 CAPILLARIES careful observation, because in this test plasma skim- ming is particularly apt to occur and will give the ap- pearance of capillary contraction, which is the more difficult to guard against, because the walls of the capillaries are generally very difficult to see between the muscle fibres. In the skin and web of the frog the reactions to adrenaline have been observed more closely and a very definite discrepancy between them and the responses to sympathetic stimulation has been made out. The capillaries are all refractory towards adrenaline, though prompt contractions can be evoked by stimu- lation of the sympathetic. Most of the arterioles and small arteries are also refractory, though the larger arterial branches (above about 0.1 mm. external diame- ter in a dilated condition) contract readily. Usually the limit between the refractory and responsive portions of arteries is quite sharp, as can be demonstrated by the following experiment. Small drops (0.001 mm. 3 ) of 0.1 per cent adrenaline are placed on the skin or web just over the superficial branches of arteries. When such a drop is placed over an arteriole it shows no constrictor effect. By follow- ing up the artery, putting drop after drop over its course, a point is finally reached where the constrictor reaction begins. When the drops are placed a short dis- tance apart and half a minute or more allowed to elapse between the single tests, it is often observed that the artery begins to contract at a certain distance proximally from the last drop, and the contraction spreads slowly in a proximal direction, while the artery just below the last drop remains open. By repe- titions of the experiment the limit of the adrenaline constriction remains the same on the same artery, and we have convinced ourselves in special tests that the arterial branches which are refractorv to adrenaline REACTIONS TO STIMULI 137 respond just as promptly to sympathetic stimulation as do the others. W. Jacobj (1920) has found that dilute adrena- line solutions (from 0.03 per cent downwards) are inactive when applied to the web of the frog, but that a previous treatment with 5 per cent veronal sodium or 1 to 8 per cent sodium carbonate for some minutes will render even extremely dilute adrenaline solutions (down to one in a million) effective. Jacobj ascribes this effect of alkaline solutions to an increased permea- bility of the frog's epidermis and shows that the absorption of other substances (strychnine) is greatly facilitated by the treatment with alkali. I have con- firmed the observation of W. Jacobj, but I find that those arteries which were refractory before also re- main so after the treatment. They are not, therefore, as might be supposed, less sensitive to adrenaline, but insensitive. Abderhalden and Gellhorn (1921) have studied the combined action of adrenaline and certain substances, "optones," obtained by hydrolysing the tissue of endo- crine glands, on the web of the frog. They find that when they put onto the web of a frog, after suitable treatment with alkali, a mixture of dilute adrenaline (1 : 30000 to 1 : 100000) with a 1 to 3 per cent solution of some of their optones, the arterial contrac- tion is abolished or greatly diminished. These "op- tones" from the thyroid, ovary, pituitary and other organs appear, therefore, to be antagonistic towards adrenaline, while they have generally no distinct effect upon the circulation when applied without adrenaline. This result may, perhaps, have some bearing upon certain observations on the effect of adrenaline on tissues in a state of inflammation, recorded by Bicker and Regendanz. In their experiments on the effects of strong chemi- 138 CAPILLARIES cal stimuli, of which some examples were mentioned in the preceding lecture, these authors have found, as a rule, that the arteries and capillaries in the tissues investigated (pancreas and conjunctiva) react abnor- mally on the application of adrenaline (0.1 per cent) during the period when the inflammation is in prog- ress, and may even continue to do so at a period when the symptoms of inflammation have subsided and the circulation appears to have returned to normal. This has been tested especially on the conjunctiva. The abnormality may consist only in an appreciably slower contraction of the arteries than normally, but gener- ally it is much more pronounced and consists in a dilatation of arteries and capillaries in many cases leading to stasis. A constrictor effect of adrenaline upon capillaries has been demonstrated in the case of man and the cat. Several statements, notably by Ricker and Regendanz, which seem to imply a constrictor effect of adrenaline upon the capillaries of the rabbit and other animals, are not sufficiently clear to be accepted as evidence, the more so as the sources of error have not been recog- nized. Cotton, Slade and Lewis (1917) have found that the introduction of a few drops of adrenaline beneath the skin of man will produce, after a latency of fifteen seconds to one minute, an intense blanching. They argue that this might be due to constriction of arte- rioles and washing out of the capillaries with plasma, but, when they find that the same blanching can be pro- duced by adrenaline five or more minutes after the circulation has been brought to a standstill by means of an occluding armlet, they rightly conclude that the capillaries themselves must contract to be emptied under these circumstances. The constrictor effect of adrenaline upon human REACTIONS TO STIMULI 139 skin capillaries has been verified microscopically by Miss Carrier (1922). By introducing a minute drop of adrenaline (1: 1000 or 1: 10000) close to one of the capillary loops at the base of the finger nail she could induce a complete local contraction of both the venous and arterial limbs of such a loop, while the tip re- mained open and contained some stagnant corpuscles. The circulation starts again in about five minutes, but the contracted part of the capillary is not back again to its original diameter for forty-five minutes. Higher dilutions, up to 1 : 5 millions, were tested to see if it were possible to obtain a dilator effect, but the results were negative. Hooker (1920) reports observations of contractions of capillaries and venules in the cat's ear after intra- venous injection of 0.06 mg. adrenaline. In the rabbit's ear the effects of local application have been tested by Rehberg and myself with entirely negative results, as far as capillaries and venules were concerned. It may be accidental, but the possibility cannot be excluded that it has some significance that the capillaries in man and in the cat respond both to adrenaline and to histamine, while those in the rabbit and the frog do not. Reference must be made finally to the puzzling ex- periments of Dale and Richards, mentioned in some detail in a former lecture, in which it was demon- strated that a minute dose (0.004 mg.) of adrenaline, injected intravenously on a cat, produces an evanes- cent vasodilator reaction, which must probably be lo- cated in the skin capillaries, while a larger dose gives the usual vasoconstrictor response, which, as we now may infer from Hooker's experiments, involves the capillaries as well as the arteries. Basing themselves on the further inference from their perfusion experiments with histamine, that the 140 CAPILLARIES presence of adrenaline in the perfusion fluid in con- centrations between 1 : 1 million and 1 : 10 millions is necessary to maintain the tonus of the perfused capil- laries, Dale and Richards put forth the suggestion that adrenaline, in the concentration in which it is normally present in the circulating blood, may be responsible for the maintenance of capillary tonus. While I admit that the arguments brought forward by Dale and Richards do show that the idea ' ' cannot be set aside as inherently incredible or contrary to all experience," and even that it is not unlikely that adrenaline may have something to do with the maintenance of capil- lary tonus in those animals the capillaries of which are definitely susceptible to the constrictor action of adrenaline, I think it very improbable that it should be the general hormone for the maintenance of capillary tonus, the more so as another substance having a defi- nite tonic action on capillaries can be shown to be present in the circulating blood of mammals generally. The experimental evidence on which this assertion is based will be given in some detail in the next lecture. LECTURE VII THE HORMONAL CONTROL OF THE CAPILLARY CIRCULATION WE come now to the substance which I regard as the principal hormone for maintaining the tonus of capillaries, and I shall present the evidence for the existence and origin of such a sub- stance in the order in which it has been acquired. The effects of cutting off the supply of blood. When the supply of blood to one of the limbs of man or to the ear of a rabbit is temporarily cut off, the skin, which has been deprived of blood, becomes hypersemic with dilated capillaries as soon as the blood is again admitted. This reaction will be described in some detail and analysed in the following lecture, and reasons given why, in these cases, the hyperemia must be taken to be due mainly to the lack of oxygen suf- fered by the tissues while deprived of blood. It was with a very intense sense of curiosity, therefore, that I made the corresponding experiment of cutting off the supply of blood to the spread tongue of a frog. In this case there would be no oxygen lack, because an amount of oxygen quite sufficient for the needs of the tissues would get in by diffusion from the atmosphere through the large surface exposed. I had observed repeatedly that blood having a venous colour when flowing into the tongue through the lingual arteries, leaves the tongue through the veins in an arterialized condition, and in the experiments to be described the 142 CAPILLARIES stagnant blood in the vessels remained arterial throughout. The result of the first experiment of this kind was that the renewed admittance of blood, after clamping the lingual arteries for about five hours, resulted in a very distinct hyperemia with some dilatation of arteries and pronounced dilatation of the capillaries. This experiment was repeated and varied in several ways, and in all cases where the supply of blood had been completely cut off for at least three hours a subsequent dilatation of arteries, and espe- cially of capillaries, resulted. The fresh supply of blood generally restored the original tonus of the ves- sels in twenty to thirty minutes, the arteries contract- ing first and the capillaries some minutes later. In one experiment, however, in which the artery had been in- completely blocked for seven hours and thereafter completely blocked for fourteen hours, the capillaries had been so far damaged that the rapid circulation observed just after opening the artery soon slowed down, and complete stasis developed in the strongly dilated capillaries. In another fourteen-hour experi- ment the blocking of the artery had been incomplete, and a very slow current could be observed in the larger arterial branches. It is very significant that in this case the hyperemia on admitting the blood was con- fined to the capillaries and was very slight even here. To make quite sure that oxygen lack had nothing to do with the reaction, a few experiments were made in which the rate of 2 diffusion was increased fourfold by keeping the frogs in an atmosphere of pure oxygen. The results were quite the same as with air. It appears to follow from these results that there is something in the blood which is necessary for the maintenance of the contractility of the Rouget cells and which disappears slowly when the supply of fresh blood is cut off. I have described how the tongue of a HORMONAL CONTROL 143 frog, which is normally very anaemic, becomes hyper- semic by the mechanical stimulus of spreading. The subsidence of this hyperemia, taking place when the tongue is left to itself, can be conceived as due to the abundant supply of the unknown factor in the blood, and we reach the conclusion that in the anaemic tongue any single capillary cannot remain closed for an indefi- nite period. When it gets no blood its tonus will dimin- ish, and it must finally become relaxed and admit a current of blood which will allow it to regain its tonus. Every capillary must, therefore, alternately open and close, and the positions of open capillaries must be continuously changing, with the result that the whole tissue is uniformly irrigated when considered over a period of sufficient length. An extremely favourable object for the closer study of this reaction of the capillaries to blood was found in the frog's web. When the femoral artery is clamped, the current of blood in the web is usually diminished so as to be barely visible in the larger arteries, though it rarely stops completely, as it generally does in the tongue. When the clamping is maintained for ten to fifteen minutes both arteries and capillaries become gradually dilated and filled up with the blood flowing slowly in. The diameter of individual capillaries can be observed to increase from about 5,u to about 20/*. When the clamp is removed a rapid current of blood passes through the dilated vessels, which, after five minutes, are again distinctly contracted and may return to the normal state in ten minutes or less. When the blood supply is cut off for a period of one to two hours the capillaries relax completely, and stasis may develop when the blood is again admitted. As in the case of the tongue, lack of oxygen has nothing to do with the reaction. The web is amply sup- plied with oxygen by diffusion from the atmosphere, 144 CAPILLARIES and the stagnant blood retains its bright arterial colour and only turns blue when the atmospheric oxygen is shut oft' by means of a cover slip. The hypothesis of a hormone x. These observations on the frog's web formed the basis for a hypothesis which could be brought to the test on the same tissue, showing such definite and prompt reactions. Supposing the existence in the frog's blood of a substance which we will provisionally designate by the symbol x, acting on the Rouget cells on the outside of the capillary wall, this substance x must be able to diffuse through the capillary wall and, since stagnant blood rapidly loses its power of maintaining the tonus of capillaries, Ave must assume that the x-substance is rapidly destroyed — probably oxidized — by the sur- rounding tissue or by the Rouget cells themselves. Both these postulated properties point to the conclu- sion that x must be a comparatively simple substance with a molecule of quite moderate size, similar, in fact, to such hormone molecules as adrenaline or thy- roxine, the structure of which has been made out by the splendid efforts of the organic chemists. Now, these well-known hormones, as well as others which are still chemically undefined, like insuline, for instance, are from the zoological point of view unspe- cific; their existence has been demonstrated in a com- paratively large number of different animals, and we must assume that the identical substances are present in the blood of every single vertebrate. If this reason- ing holds also for x it should be possible to maintain the capillaries' of the frog's web in a state of contrac- tion by perfusion with blood from a mammal, and it might be possible by suitable treatment to divide up HORMONAL CONTROL 145 mammalian blood, which is obtainable in large quan- tities, into fractions, to get a fraction, containing x, separated from a large number of the other constitu- ents of the blood and, perhaps, to concentrate the x fraction so as to allow a closer study of its properties. Preliminary experiments with mammalian blood. The first point was to see if there would be any difference between the effect of an artificial perfusion fluid, like frog's Ringer, and mammalian serum or plasma upon the capillaries of the frog's web. In order to make the capillaries distinctly visible it was neces- sary to add some colouring matter to the perfusion fluid, and the simplest plan Dr. Harrop and I found to be the addition of blood corpuscles. We made up, therefore, our comparison perfusion fluids by adding carefully washed red corpuscles from ox blood to Ringer's fluid or to 3 per cent gum Ringer, while the fluids to be tested were dilutions of defibrinated ox blood with Ringer. The very first experiment showed clearly the difference between the artificial fluids and the blood. On perfusion with Ringer or gum Ringer the capillaries in the web of the perfused leg would begin dilating almost at once, and, generally, after fifteen minutes the dilatation became maximal and stasis developed. Ox blood, on the other hand, could maintain the capillaries in a normal state of contrac- tion for a variable time between forty minutes and two hours. In some cases there would be some initial dila- tation, because the blood flow had to be stopped while the perfusion cannula was being inserted into the femoral artery of the frog. "With the artificial perfu- sion fluids this dilatation always went rapidly from bad to worse. With the ox blood (usually diluted to 2/3) recovery frequently took place. 146 CAPILLARIES Preliminary dialysis experiments. After these preliminary perfusion experiments \re felt pretty certain that we were on the right track, and the next thing to do was, obviously, to see whether we could get our x-substance out of the blood by dialy- sis, since the hypothesis demanded that it should be a diffusible substance. We put about 100 cc. of Einger's solution into a dialysing collodion bag, lowered the bag into a vessel with one liter of defibrinated ox blood and kept both fluids well stirred for about twenty- four hours. Thereupon we took out the dialysate. added the necessary quantity of washed corpuscles, and com- pared its action in perfusion experiments with that of ordinary Einger — corpuscles. The results were quite convincing : the dialysate from ox blood contains a sub- stance which, for a time at least, is able to maintain the tonus of the skin capillaries of the frog. Improved perfusion metliods. The technique of our preliminary perfusion experi- ments was rather crude, and we had to modify it sev- eral times before getting it quite satisfactory. I shall describe briefly our final technique, because it presents some points which are of interest also from a theo- retical point of view. It has been often discussed whether rhythmical per- fusion, imitating the action of the heart, is more effective than continuous. TTe mid that when the perfu- sion fluid contains blood corpuscles, rhythmic perfu- sion is necessary to prevent agglutination of cor- puscles. Our apparatus for rhythmical perfusion is as follows (Fig. 33) : Air (or any other gas or gas mix- ture) is supplied along the tube (1) at a suitable pres- sure. The cylinder (2), in which the tube (3) can be raised and lowered, constitutes a pressure regulator. HORMONAL CONTROL 147 the surplus of air supplied being blown off through the mercury (or water). (4) is a metal tap turned at a regular rate, corresponding to the pulse rate, by a small motor. When the tap is in the position shown, Fig. 33. Apparatus for rhythmical perfusion. the compressed air is admitted to a vessel (5) (in reality a series of vessels) and when the tap is turned through 180° the air is blown off through the tube (6) and the fluid resistance (7). We obtain at each cycle a definite systolic pressure, regulated by the resistance in (2), and a diastolic pressure, regulated by (7). The 148 CAPILLARIES rubber stopper of each perfusion vessel is perforated by three glass tubes: (8) is a T tube admitting the compressed air and serving further for the introduc- tion of perfusion fluid, (9) and (10) are screw clips. The tube (11) is provided with a glass bulb, closed at the top and containing air. At each pulse the perfusion fluid is driven up into the tube (11) and back again, thereby keeping the perfusion fluid constantly and completely mixed. The narrow tube (12) can be con- nected through suitable rubber tubing with the per- fusion cannula. "When a T tube is inserted just in front of the cannula and connected with a manometer the effective perfusion pressure can be measured. In all of the experiments we have perfused the hind legs of brown frogs (R.temporaria) narcotized with urethane through the femoral artery. In the first series of experiments one leg was perfused at a time, the femoral vein was widely opened and a strong ligature was tied around the proximal end of the perfused leg to prevent the perfusion fluid from entering the frog's body. In later experiments we perfused both legs simul- taneously with the two fluids to be compared, which, of course, greatly facilitated an accurate comparison. Improved dialysis methods. For a closer study of the active principle in mam- malian blood the first step was to devise a method by which we could remove the colloids from compara- tively large quantities of blood without, at the same time, destroying or removing the x-substance. "We made some tests with precipitation of the blood pro- teins, but, as the results were negative, they Avere soon abandoned, and we returned to dialysis as a means of increasing the quantity and concentration of the x-hormone obtainable from a given quantity of blood. I think it justifiable here to describe our dialysing HORMONAL CONTROL 149 technique in some detail, because it has a certain bear- ing upon the essential function of capillaries to be discussed in a following lecture : the exchange of dif- fusible substances across the capillary wall. I described in the first lecture a very peculiar arrangement of capillaries, a true "rete mirabile"' annexed to the gas gland in the swim bladder of fishes. This is, I believe, the most perfect dialysing apparatus in existence, and my problem was to copy it in a form which should lend itself to technical exe- cution. Let me recall to your memory by means of the diagram the capillary arrangement in question. We have two currents of fluids flowing in opposite directions in two sets of parallel tubes, regularly inter- calated, so that each tube carrying blood in one direc- tion is surrounded by tubes carrying blood in the opposite direction. Let us suppose that there is a dialysable substance in the gland which must, on no account, get out into the general circulation. When the blood from the gland passes along the system of par- allel capillaries the hypothetical substance, which we will call A, will diffuse over into the blood running in the opposite direction towards the gland. As the blood from the gland passes along, its content of A will be- come diminished, but, since the blood entering the afferent capillaries is assumed to be free from A, the diffusion can go on, until the whole quantity of A has been transferred from the efferent to the afferent capil- laries. The actual function of the rete belonging to the gas gland is unknown, but it is obvious that it could act in the way now described as a counter-current dialysis apparatus, which must be of a very high degree of perfection, since every cubic centimeter of blood present in the efferent capillaries presents a sur- face for dialysis of about 4200 cm. 2 Directly to imitate technically the anatomical struc- 150 CAPILLARIES ture described is obviously an impossibility, but the counter-current principle can be utilized if, instead of a large number of parallel channels, we arrange only two, one in each direction, eacli separated from the Fig. 34. Diagram of dialysing apparatus. other by a membrane, suitable for the dialysis con- templated. To make up for the large number of chan- nels and obtain a sufficient surface area we must, ob- viously, make our channels as long as possible. The apparatus based on these principles is constructed as follows. In circular discs 14.5 cm. in diameter and 9 HORMONAL CONTROL 151 mm. thick, cast out of pure tin to make them as indiffer- ent chemically as possible, a shallow spiral groove is formed on each side from a point near the centre almost to the circumference. The length of each spiral groove in our apparatus is 250 cm. ; it is 4 mm. broad and 1.5 mm. deep, thus occupying 100 cm. 2 of the total area of the disc (165 cm. 2 ) and having a capacity of 15 cc. The grooves can be connected up by the tubing shown in the diagram, Fig. 34, and when, for instance, four dialysing membranes are put between five discs and the whole clamped firmly together, we have an apparatus in which blood or any other suitable fluid can be admitted at one end of a channel 10 meters long and separated by a dialysing membrane of 400 cm. 2 from an exactly similar channel in which another fluid, e.g., saline, is running in the opposite direction. The dialysing surface of this apparatus is 6.7 cm. 2 per cc. of blood. This is by no means bad, but a comparison of the imitation with the original in the swim bladder of the eel cannot but inspire a deep sense of humility before the wonders of the creation. The preparation of dialysing membranes. Mr. Rasmussen and I have made a large number of tests to find the best type of dialysing membrane. We can conceive a dialysing membrane as being porous. When all the pores are below a certain size, molecules of that size or larger will be unable to pass the mem- brane. Smaller molecules can pass through, and a separation by dialysis can be effected. The rate of dialysis of a given substance depends upon the number of pores available, and it is clear that, to obtain the best results, the membrane must contain the largest number of pores of a uniform size, which should be just sufficient to keep back those substances the pres- ence of which in the dialvsate is not desired. 152 CAPILLARIES In our case we desired to keep back the true colloids, especially the proteins, and several methods were available by which membranes could be prepared which were just impermeable to proteins. We have tried connective tissue membranes, several kinds of parchment paper membranes and finally collo- dion membranes-, prepared according to the methods of Brown (1915) and Walpole (1915), respectively. The connective tissue and commercial parchment paper membranes were not uniform or reliable enough for our purpose, but we found that we could prepare excel- lent parchment paper of the desired degree of per- meability by dipping hardened niters for a certain time in sulphuric acid of suitable strength. Collodion membranes are prepared according to Brown's method by pouring out a solution of collodion in ether-alcohol over a plane horizontal surface and leaving it to dry completely by evaporation of the ether and alcohol. The dry membrane is practically imper- meable even to water, but Brown has found that any desired degree of permeability can be imparted to it by soaking it for some time in a mixture of alcohol and water. The higher the percentage of alcohol the more permeable will it become. According to Walpole 's method, finally, the collodion solution is not allowed to dry out completely, but, when the ether has evaporated, while the membrane still contains a certain proportion of alcohol, it is put into water and the rest of the alcohol removed by diffusion. By the three methods described we could prepare membranes which possessed the desired degree of absolute permeability, being just impermeable to protein. To estimate correctly their value as dialysing mem- branes the substance which it is desired to get through g capacity cc. < 0.05 0.11 2.7 22 ) HORMONAL CONTROL 153 should be available, or at least a substance possessing molecules of about the same size, but, failing this, a useful provisional test can be obtained by measuring the rate at which water will filter through the mem- brane at a definite head of pressure. We take as the filtering capacity the quantity of water passing through 100 cm. 2 of the membrane per minute at one atmos- phere's ( = 10 m. water) pressure. We find for the three membranes having the same degree of permeability: Filterin Brown's collodion. Our parchment, Walpole 's collodion, (Hardened filter paper, The figure for filter paper has been put in for the sake of comparison. It is seen that the filtering capac- ity of the Walpole collodion is many times higher than that of the other membranes. For the purpose of removing the x-substance from the blood we employed a dialysing apparatus as de- scribed with four Walpole collodion membranes and led a current of blood (usually defibrinated) through one channel at the rate of 1.2 liters in twenty-four hours. Against this was led a current of mammalian Ringer at the rate of 0.5 liters in twenty-four hours. The operation was always performed in a cold room at an average temperature of 3° to 6°. The activity of blood and dialysate from different animals. By means of the apparatus and technique described, we have tested the activity of blood and dialysate from a few mammals. Ox blood and dialysate, which was tested very often, was always active, but the 154 CAPILLARIES activity was rather variable. Whole defibrinated blood was not quite as active as blood diluted with two parts of frog's Ringer to five parts of blood, and the dialy- sate from a sample of blood was sometimes more active than the original sample. With the best speci- men of dialysate tested, the circulation in the perfused leg remained good for two and a half hours, but on an average a good circulation in fairly narrow capillaries could be kept up for only about one hour. It turned out later that in all these tests the perfusion pressure was higher than the normal blood pressure of the frog. Defibrinated pigs' blood, which was tested twice, showed only a slight effect, and the dialysate from one of these samples, none at all, but the serum from a sample of horse blood was found to be very active. Defibrinated rabbits' blood showed also normal activ- ity, keeping the capillaries narrow for one hour, but, when the animal's loss of blood, amounting to about 60 per cent, was made good by gum saline and the animal was bled again after twenty-four hours, this second sample showed only a slight activity, maintain- ing the capillary circulation for thirty-five minutes. In another experiment, hirudinized rabbits' blood also showed only slight activity. Attempts to isolate the active substance. We found that the dialysate from ox blood could be preserved on ice for several days without losing any of its activity, when it remained clear. In all cases where turbidity was brought about by the growth of bacteria the activity was destroyed. The neutral active dialy- sate could be heated for a few minutes to boiling with- out completely losing its activity, though it was in all cases weakened by the process. The active dialysate could further be evaporated in a vacuum to dryness, a process which took about three hours at a temperature HORMONAL CONTROL 155 of 25° to 30° for a quantity of 300 cc. When the residue was taken up in distilled water (and some C0 2 added to redissolve it completely) the fluid retained some, at least, of its activity, though the results were rather variable. By evaporating the dialysate to about half its original volume we could obtain a distinct increase in activity, but further progress in this direction was debarred by the simultaneous increase in concentra- tion of all the salts present. We attempted to extract the dry residue with acetone, methyl and aethyl alcohol and ether. These extracts were evaporated to dryness and taken up in Ringer's solutions, but showed no ac- tivity. In the case of one aethyl alcohol extraction we found the alcohol-insoluble residue distinctly active after solution in water, and we may conclude, there- fore, that the x-substance is insoluble in sethyl alcohol. We tried several precipitating reagents, but the re- sults were not encouraging. When, however, the pro- tein free dialysate was acidified with sulphuric acid and a small amount of phosphotungstic acid added before the evaporation in vacuum, a precipitate was formed, which appeared to contain the active sub- stance though it was extremely difficult to get it back into solution without destroying it altogether. By very cautious treatment of the precipitate with baryta and precipitation of the excess of baryta by means of CO;, we have at least twice obtained a solution which showed distinct activity after addition of the necessary salts to make up a Ringer solution. Since, in these cases, x should have been concentrated between ten and twenty times, if nothing had been lost, the result can only be described as a poor one, and we felt very disappointed when we had to admit that all our at- tempts at concentration of x had failed. Work on this line had to be put on one side for half a year, during which time preparations were made for 156 CAPILLAKIES an attempt to separate x from the salts of the blood by double dialysis. We would dialyse it out of the blood in the way with which we were now familiar, and the dialysate should be taken through another dialysing apparatus provided with much less permeable mem- branes, which should allow the simple inorganic salts to pass out while holding back the larger molecules of organic substances to which x would possibly belong. These experiments, whose chances were admittedly slight, were never carried out, however, because the problem suddenly took on an entirely new aspect. The influence of the pituitary gland on the tonus of capillaries. I had sometimes thought of testing a number of extracts from glands known to possess internal secre- tion, in the hope of finding x in one of them, but I have always had a strong dislike for experimentation at random, and no tests of the kind were carried out. My assistant, Mr. Eehberg, noticed, however, a paper by Pohle (1920; dealing with a development of cutaneous oedema in frogs after operative removal of the pituitary gland. Xow, oedema is very often closely associated with, in fact dependent on, a state of dila- tation of the capillaries, and Mr. Eehberg at once sug- gested that we should extirpate the pituitary on some frogs and study the effect on the capillary circulation in the skin and web. The suggestion was promptly car- ried into execution by Mr. Eehberg himself and the results were extremely striking. For the first few hours after the operation there is no apparent change in the circulation, but thereafter the capillaries in the skin and web begin to dilate, and after twenty-four hours they are usually strongly dilated. At the same time another change takes place, which has nothing to do with the circulation, but which HORMONAL CONTROL 157 has proved very helpful in the study of pituitary function. The frog (B. temporaria) becomes much lighter in colour, in fact, quite pale. The microscope reveals the fact that this change is due to the black pigment cells in the skin, which are usually expanded and greatly ramified, but which, in the operated ani- mals, contract into small balls, intensely black, but occupying only a minute fraction of the skm area. The colour of a frog, in which the whole of the pitui- tary gland has been removed, remains pale, but the state of the capillaries undergoes some remarkable changes. After a variable period, generally one or two weeks, the capillaries regain their ability to contract, but the cutaneous circulation is now characterized by its want of equiUbrium, just as may be the case in a definite area after section of the corresponding nerves. States of extreme contraction may change abruptly into pronounced or even maximal dilatation, and a circulation which can be accepted as normal is ob- served only occasionally for short periods. It might be supposed that at this stage cutting of the sciatic would abolish completely the control of the animal over the circulation in the web. This does not seem to be the case, however, but the point has not as yet been sufficiently investigated. The frogs from which the hypophysis has been re- moved usually survive the operation only for two to four weeks. We have not observed the oedema, which, according to Pohle, should result from the removal, but a special study of this point has not been made and we have used another species of frog. It is well known that the hypophysis is not a single organ with one well-defined function, but consists of several portions which differ greatly in their histo- logical structure. Distinction is usually made between a glandular portion (often termed the pars anterior), 158 CAPILLARIES a nervous portion (pars posterior) and an interme- diate portion. The terms anterior and posterior are rather unfortunate, because in the frog the hindmost part is the one having a glandular structure and cor- responding, therefore, to the pars anterior in mammals. 1 While the complete removal of the hypophysis of the frog requires some operative dexterity, the re- moval of the glandular portion alone is quite easy, and this operation has been performed repeatedly. We find that the initial effects on the cutaneous circula- tion and pigmentation are the same as those resulting from removal of the whole gland, but in a week or less the colour begins to darken and a normal capillary . circulation is restored. The animals are able to survive this operation for an indefinite period. We conclude, therefore, that the glandular portion is not the one responsible for the production of the capillary hor- mone, though its removal disturbs for a time the func- tion of the really active tissue, which must be located either in the nervous or in the intermediate portion. The effects of pituitary extracts. If a hormone which increases the contractile tonus of capillaries is normally produced by the portio ner- vosa or intermedia of the pituitary gland we must expect to find it in the pituitary extracts which are commercially obtainable. We have tested thoroughly the extract prepared by Parke, Davis & Co. from the "posterior" portion of the pituitary and sold under the name of pituitrine. This extract is, in reality, pre- pared from the nervous and intermediate portions of the hypophysis of cows, and 1 cc. of the extract is i A very complete account of the morphology and physiology of the pituitary gland is to be found in the book of Houssay (1918). HORMONAL CONTROL 159 stated to correspond to 0.2 g. of the fresh pituitary substance. Injection into the web of a frog of a minute drop of this extract brings about the contraction of both arte- ries and capillaries, while the veins remain unaffected. When a highly diluted extract is injected the action on the arteries may be absent, but the capillaries are still strongly affected and may contract even to oblit- eration. The intact epidermis is impermeable to the active substance, but when the web is treated for a few minutes with 5 to 10 per cent veronal sodium, it becomes permeable, and the subsequent application of 1 per cent pituitary extract causes definite contrac- tion of the capillaries, a number of which may even become closed. Perfusion experiments with 'pituitary extracts. It is evident that the pituitary extract has a specific action upon the capillaries in the frog's skin, but if it is to be accepted as containing a hormone normally present in the blood, it must be shown to be still active in much smaller concentrations. To investigate this point we again took up the perfusion experiments on the hind legs of brown frogs. Our first perfusion fluid was made up from Ringer solution, to which was added 0.1 per cent glucose, 1/3 volume of washed red cor- puscles from the ox and 1/10000 co mm ercial pituitrine. It had the effect of promptly stopping the circulation by causing complete contraction of both arteries and capillaries. The following experiments with weaker concentra- tions of pituitrine showed that the addition to the Ringer of 1 in 50000 to 1 in a million parts pituitrine generally had the desired effect of maintaining the tonus of the capillaries without causing complete con- traction and without having any visible influence upon 160 CAPILLARIES the arteries. The addition of 1 part pituitrine to 5 million parts perfusion fluid had no perceptible influ- ence in the few cases in which it was tried. I shall reproduce one protocol in some detail as an example. Femoral artery clamped 3 50 . Perfusion with 3 per cent gum Ringer -f- pituitrine 1 : 1000000 begun 3 52 . We found it advisable always to begin the perfusion with a solution free from eoipuscles to wash out the frog's own blood and make sure by the disappearance of all corpuscles that no blood got in by collateral circulation from the frog. Perfusion with the same fluid -|- 1/3 vol. washed ox cor- puscles begun 4 0<> . The circulation is, on the whole, in a good condition. Capillary network rather dilated. Stasis in a few capillaries. 4 13 Capillaries become distinctly narrower. 4 15 A large number of capillaries very narrow. The transient contraction of capillaries at i 15 may very probably have been due to innervation. We found it advisable in later experiments to cut the sciatic some minutes before beginning a perfusion experiment. 4 25 Capillary network about normal. 4 28 Changed to a perfusion fluid free from pituitrine. 4 35 Some dilatation of capillaries. Sciatic nerve cut. 5 23 Capillaries much dilated. Stasis developing in several places. 5 25 Changed to gum Ringer 4- pituitrine 1 : 1000000. 5 2G Changed to gum Ringer -+- corpuscles -+- pituitrine. The capillary network now began to contract and the fol- lowing measurements were made of the diameters of seven capillaries selected for the purpose. The measurements are in arbitrary scale divisions. 1.5 1.0 1.2 1.0 1.7 Total 9.7 1.7 0.8 1.3 1.2 1.6 " 8.7 1.3 0.8 0.5 0.8 1.0 " 6.5 1.5 0.6 0.4 0.7 0.8 " 6.1 7 00 Stasis has developed in a large number of capillaries and many small bleedings have occurred from these (the 5 32 1.5 1.8 5 41 1.1 1.0 5 47 1.1 1.0 5" 1.0 1.1 HORMONAL CONTROL 161 web had been allowed to dry up somewhat). In those capil- laries which are still open the circulation is good, and these capillaries are, on the whole, narrow, some of them very narrow. The problem concerning the causation and preven- tion of stasis in the capillaries is a complicated one, and its theoretical aspects can be more conveniently dealt with in a later lecture, but certain facts must be pointed out here, because they are of great prac- tical importance in testing the tonus of capillaries by means of perfusion experiments. The development of stasis depends upon the rate at which fluid leaves the capillaries through the endo- thelial wall. If this rate is so rapid that the corpuscle emulsion becomes concentrated beyond a certain point the corpuscles stick together and block the passage. This increases the pressure and a vicious circle is thus set up, rapidly leading up to the capillaries being filled with a densely packed mass of corpuscles. When the perfusion fluid does not contain colloids filtration takes place, and when the capillaries have become dilated beyond a certain point the rate of filtra- tion is increased and stasis often develops with aston- ishing rapidity. When the perfusion fluid contains a suitable colloid, such as 3 per cent gum acacia, the filtration of water through the normal capillary wall is practically inhib- ited, and stasis develops only when the capillary wall becomes permeable to the gum molecules. The causes bringing about this change are not understood in their entirety, but one important condition may certainly be a blood pressure in excess of the normal. It appears that the capillaries in the web of a frog are able to withstand an abnormally high pressure for some time, but finally they give way and become 162 CAPILLARIES relaxed to such an extent that the colloidal particles can pass out. The presence of a substance like pitui- trine counteracts the tendency to relax under an abnor- mally high pressure, but seems to be able only to delay the final relaxation. The influence of pituitrine is brought out very clearly in perfusion experiments without gum, where stasis occurs in the absence of pituitrine as soon as the capillaries are relaxed. With the addition of gum and a low pressure, that is, with the conditions as nearly physiological as possible, the circulation can be maintained sometimes for hours without the addition of pituitrine, but after the first half hour or so nearly all the capillaries become strongly dilated and remain so, while with pituitrine practically all the capillaries remain narrow. As an example of this typical difference, I give the following protocol of an experiment in which both legs of a frog were perfused simultaneously. Left leg Right leg "With pituitrine Without pituitrine ll 15 Sciatic nerve cut. II 16 Sciatic nerve cut. 11 36 Femoral artery opened. II 42 Femoral artery opened. 11 37 Perfusion with gum ll 43 Perfusion with gum Ringer -}- 1 : 50000 pitui- Ringer begun. trine begun. 11 38 Perfusion with gum ll 45 Perfusion with gum Ringer -j- 1/3 ox cor- Ringer -f- 1/3 ox corpus- puscles + 1:5000000 cles + 1:5000000 ace- acetyl-choline -f- 1 : 50- tyl-choline. 000 pituitrine. Acetyl-choline added to both perfusion fluids to maintain the arteries in a constant state of dilatation. It proves excel- lent for the purpose. Perfusion pressure measured by a mercury manometer just behind each cannula. Mean pressure ten mm. mercury in both systems. HORMONAL CONTROL 163 ll 45 Slight dilatation. No sta- Somewhat more dilated. No sis. stasis. 12 00 Contracted again. About Dilated. normal. 12 16 Completely normal. A little better, but distinctly dilated. Some narrow capil- laries, however. I 06 Perfusion pressure raised to about 20 mm. in both sys- tems, with no immediate effect in either. I 20 Capillaries have con- Capillaries greatly dilated, tracted somewhat. Many of them are now very Melanophores maximally con- narrow. Melanophores tracted. spread out, colour of leg nearly normal. Leg very pale. 1" No change. Perfusion Numerous stases and bleed- fluid (25 cc.) used up. ings. 2 0S Perfusion fluid (25 cc.) used up. The last observation, that the commercial pituitrine causes expansion of the black pigment cells in just the same way as does the circulating hormone from the frog's own hypophysis, has been repeated very often, and it can be concluded that the hormone circulates in the frog's blood in concentrations which show consid- erable variations, but must correspond usually to some- thing between 1:100000 and 1:1000000 pituitrine (Parke-Davis). In the experiments mentioned in a former lecture in which we had cut the femoral artery of one leg of a frog to get rid of the nerve plexus which might be present in its wall, we were able to follow the develop- ment of the collateral circulation by the change in colour of the leg. So long as the circulation was imper- fect the leg was perceptibly paler in colour than the other leg, which was normally provided with blood. 164 CAPILLARIES We have made a few perfusion experiments with suitable concentrations of adrenaline iustead of pitui- trine, but have failed to find any tonic influence of this substance upon the capillaries. Is the x-substance in mammalian blood identical iritJi the pituitary hormone? We have now demonstrated the existence of a pitui- tary hormone acting in very great dilution specifically on the contractile elements of certain capillaries, in- creasing their contractile tonus. We have shown, also, that a substance x, having a similar action on the same capillaries, is a normal constituent of the blood of cer- tain mammals. Is it legitimate to conclude that the two substances are identical? Can we put the solution of the problem in the following suggestive mathematical form ? x = pituitary hormone According to the literature about pituitary extracts (Biedl, 1916), the "pressor principle" contained in such extracts is dialysable (p. 131). It will stand boiling for a short period (p. 131). It is insoluble in sethyl alcohol and ether (p. 131), but soluble in methyl alcohol (pp. 691, 695). It is precipitated by phosphotungstic acid (p. 691), and When present in solution with proteins it is partially destroyed if the proteins are removed by heating to the boiling point in an acid solution (p. 131). According to the experience gained in our unsuccess- ful attempts to isolate x from the blood, this substance is similar to the pituitary "pressor principle" with regard to the following reactions : It is dialysable. It will stand boiling for a short period. HORMONAL CONTROL 165 It is insoluble in aethyl alcohol. It is precipitated by phosphotungstic acid. In the following reactions there is some apparent dissimilarity. X is completely destroyed when the proteins of the blood are removed by heating in acid solution. We have not succeeded in extracting x by methyl alcohol from the evaporated residue of active dialy- sate. These differences may probably be accounted for by the fact that the solutions on which we have worked are several thousand (probably about a hundred thou- sand) times more dilute tban the extracts on which the reactions given above have been worked out. r Very definite evidence for the identity of x with the pituitary hormone is furnished finally by the follow- ing observation: When the two legs of a brown frog are perfused simultaneously, one with dialysate from ox blood and the other with Ringer, the Ringer leg- will turn pale, while the dialysate leg will retain its dark colour more or less completely, and there is a definite parallelism between the power of a dialysate to maintain the tonus of the capillaries and its power to maintain the colour of the leg by expanding the black pigment cells. I think, therefore, that it is possible to assert with about as much confidence as one can have in an experi- mental science, where the pitfalls are numerous and often extremely well concealed, that a hormone, pro- duced by the pituitary gland and acting both on the contractile element of the frog's cutaneous capillaries and on the melanophores of the frog's skin, is nor- mally present in mammalian blood. *Ihe concentration of this hormone in the blood is variable. It corresponds to something between one ten-thousandth and one millionth part of the concentration in Parke-Davis 166 CAPILLARIES pituitrine, probably on an average to about one hun- dred-thousandth part. The concentration of the active substance in the com- mercial extract is probably very low. Judging by analogy from the concentration of adrenaline in the adrenal glands and thyroxine in the thyroid, the con- centration of the active substance in pituitary extract is not at all likely to be above one to one thousand, and it may be considerably lower. Y The concentration of the active substance itself in blood is, therefore, perhaps of the order of one part in a hundred million, but it may very well be much lower; With regard to the normal function of this substance in the body, say of man, we are still profoundly igno- rant, and I would warn emphatically against drawing any hasty conclusions. Capillaries in different animals and in different organs differ a great deal too much for any conclusion by analogy to be at all safe. Even in the frog there are very notable differences between the susceptibility of different capillaries to pituitrine. We have observed repeatedly in our perfusion ex- periments that the venous ends of the capillaries in the web are not prevented from dilating by a concentration which will keep the greater part of the capillary net- work in a state of contraction. The muscle capillaries of the frog are able to main- tain their tonus at least for a couple of hours without pituitrine and, on the other hand, an injection into the lymph spaces of 1 cc. of 1 per cent pituitrine, which brings about a general contraction of the cutane- ous capillaries, does not seem to affect the muscle Qapillaries. It is evident that a great deal of work will have to be done to clear up the reactions of even the most essential capillary systems to pituitrine, but I have no doubt that such work as is now bein°- undertaken HORMONAL CONTROL 167 by Mr. Reliberg will bear ample fruit, both theo- retically, with regard to the normal and pathological physiology of capillaries, and practically, by opening up therapeutical applications of pituitrine. It appears to me very suggestive that pituitrine has, according to Biedl, been successfully employed in cases showing shockdike depressions of the arterial blood pressure, since an important factor in such cases, to which I shall come back in a later lecture, is a dilatation of capillary blood vessels. LECTURE VIII THE MECHANISM OF SOME CAPILLARY REAC- TIONS, ESPECIALLY IN THE SKIN OF MAN XT will be convenient at this stage to make some applications of the information gained and to describe and analyse certain capillary reactions to be observed especially in the skin of man. Microscopic observation of human skin capillaries. The method by which the human skin capillaries can be made accessible to microscopic observation was dis- covered in 1912 by Lombard. It consists simply in plac- ing a drop of highly refractive transparent oil on the skin. The oil will produce a smooth surface, and, by replacing the air in the surface layers of the epider- mis, make this so far transparent that the papilla? of the cutis and their capillary loops may become visible when illuminated by an oblique beam of strong light. The application of a cover slip (Schur, 1920) may sometimes prove advantageous. In most fields it is possible only to see the tips of the capillary loops (Fig. 36), but at the base of the finger nails the papilla? are seen in profile, and a number of loops can be observed for their whole length. In 1916 Weiss published the first of an extensive series of articles from the clinic of Otfr. Muller, Tubingen, on the microscopic appearance of the skin capillaries in man. This study has been carried on mainly from a clinical point of view, and the changes in the appearance of the capillaries — especially those CUTANEOUS REACTIONS 169 at the base of the nail — in a number of diseases has been described. I do not propose to enter upon a de- tailed discussion of this work, the more so as I have some doubt whether sufficient attention has been paid to the variations of the capillaries in normal subjects and in response to simple stimuli. - inn i\0 Fig. 35. Arterioles, capillaries and venules at the base of the human nail. After E. B. Carrier. The reactions and behaviour of normal cutaneous capillaries have been studied by W. Hagen (1921) and in my laboratory by Miss Carrier (1922). By the use of the concentrated light from a small arc lamp, suitably filtered, so as to remove the rays from both ends of the spectrum, which may influence the capil- laries, and at the same time to increase the contrast between the blood and the tissue, Miss Carrier has succeeded in seeing in favourable places the connec- tions of the capillaries with the arterioles as well as 170 CAPILLARIES with the venules. The figure (35) shows that the arte- rioles are normally very narrow. It should be remem- bered, however, that the walls are invisible, and it is only the axial current of red corpuscles that can be ob- served. The diameter of the arterioles is, therefore, somewhat greater than is indicated by the figure. The floiv of blood in the capillaries at the base of the nail. At the base of the nail the most striking feature at first glance is the variety in size and shape of the capillary loops. At one extreme are vessels so very fine as to allow the passage of only a single file of red cells. At the other extreme are large vessels, several hundredths of a millimeter in width, which may have the form of a single loop or that of a figure eight, or may be greatly twisted and folded on themselves, espe- cially if there has been a little injury to the skin. The length of the capillary loops varies between wide limits. It is largely influenced by the type of work for which the hands are used and the freedom with which the skin grows over the nail. The venous branch is generally broader, and sometimes a great deal broader, than the arterial. While the general shape and relative size of capil- laries remain practically unaltered for a long time, and usually all the capillaries present are open and admit a current of corpuscles, there are many minor spontaneous alterations in the diameter of individual capillaries and in the rate of flow through them, as described and figured by Hagen. When the stream of corpuscles is rapid it appears almost homogeneous, but in a slow stream the leucocytes show up as little white spots. Often there is a slow stream, which looks granu- lar, due to agglutination of the red corpuscles, espe- CUTANEOUS REACTIONS 1T1 cially under the influence of cold. This granular stream has been noticed by many writers, and much impor- tance has been attached to it by some. The agglutination has probably something to do with a phenomenon which has been described as peristaltic contractions of capillaries (Thaller and Draga, 1917), and which can sometimes be observed, though prob- ably not in all individuals, when the current is slow: the stream slows and then starts forward again with a gap, about a tenth of a millimeter in length, inserted in the col umn of blood. The capillary walls are not visi- ble, so whether this represents a true contraction of the capillary or a peculiarity in the distribution of the cells cannot be decided by simple inspection. "What is seen is a space free of red blood cells. It has been explained as a contraction wave, passing over the capillary and helping thus to forward the blood stream. Hagen has made the observation that, when the flow stops altogether, the cells in a capillary draw together towards the tip of the loop, leaving a gap filled with plasma at each end, and this observation seems sufficient to explain the gaps which appear with a slow, jerky stream. It is occasionally possible, how- ever, to see gaps in a stream moving at a fairly rapid rate. That these are not due to capillary contractions is indicated by the following observations : In the first place, the gaps, when seen, travel at very variable and often very high rates, as would be expected if they are just gaps in the moving column of corpuscles, but not if they are contractions pushing the corpuscles before them, and in the second place, when the stream is watched in two capillary loops, which come off at the bifurcation of one arteriole, it may some- times be observed that gaps will travel simultaneously over both capillaries. There is, therefore, certainly no definite evidence of the existence of peristaltic waves 172 CAPILLARIES along capillaries, and the conception itself is inherently extremely improbable. The shifting of open capillaries in the skin and other tissues. On the back of the hand it is usually possible to see only the tips of the capillaries, but occasionally the Fig. 36. Field of capillary loops limited by subpapillary veins, from the back of the human hand. Figs. AD have been drawn at three-minute intervals. In E all the capillaries present have opened after mechanical stimulation. After E. B. Carrier. venous branch may be traced all the way to the richly anastomosing network of venules below, and we can see this horizontal plexus receive the blood from the capillary loops at regular intervals. When a single small field (Fig. 36, A to E) is watched over a period of several minutes, it is found CUTANEOUS REACTIONS 173 that here, unlike the conditions at the base of the nail, all the capillaries are not open at a given moment. They are opening and closing continuously, as seen in the figure, where drawings made every three minutes have been reproduced, together with a representation of all the capillaries present in this field, and opened up by the stimulus of a light pressure. Similar shifting of the open capillaries over the field has been observed again and again in the human skin when in a completely unstimulated condition. Hagen has seen corresponding changes in a flat capillary network in the rabbit's ear and given a diagrammatic representation of a number of changes taking place during a few minutes. They also regularly occur, though they are not so conspicu- ous, in the web of frogs, as I have described (1921), and I have seen them repeatedly in frogs' muscles. Quite recently a shifting of the blood flow from one capillary to another and from one glomerulus to another has been described by Richards (1922) for the frog's kidney. I intend to apply cinematographic methods to make a closer and more extended study of these shiftings of open capillaries than is practicable by visual methods, because I believe them to represent the very essence of capillary circulation regulation. A very small number of open capillaries (compared with the total number existing) is sufficient, as we have seen, to supply the wants of a resting tissue, but if the open capillaries were always the same the distri- bution of the substances supplied by them would be very unequal. The tissue elements nearest to the open capillaries would get more than they required, and the more distant would be starved. When the position of the open capillaries is continuously changing, a cell, starving at one moment, will get all it wants when the capillary nearest to it is opened up. In this way an 174 CAPILLARIES adequate, though intermittent, supply can he secured at every single point, while, at the same time, the blood is utilized with the utmost economy, the necessary minimum only being present at any one time in any tissue. A regulation of this kind may possibly be brought about by metabolic products accumulating at points where the distance through which they have to diffuse in order to reach an open capillary is too great, and by their action causing the opening of a nearer vessel. It may be conceived as simply due to oxygen lack — in some tissues at least — and it may be due, perhaps, to the pituitary hormone, which gradually disappears from a capillary when it is closed to the passage of blood. It is at present impossible to make a choice between these and other possibilities. Several factors may work together, and all I can say is that the prob- lem requires and will undoubtedly repay the most careful experimental investigation. In the skin of man and, very noticeably, in the palm of the hands, there is a further shifting of open capil- laries (and venules), in so far as fields of a few milli- meters diameter in which there are only a few open capillaries, while the underlying vessels are indistinct, alternate with others in Avhich a large number of capil- laries are open and the underlying vessels are promi- nent. The difference is often quite distinct when seen with the naked eye (Ebbecke, 1917, p. 51), and it can be observed how the white and red spots may be con- tinually changing in form and location. The local vasomotor reactions to 'mechanical stimuli. I have described in the second lecture the local white and red reactions which appear, after a short latency, on the human skin, when this is stroked lightly (white reaction) or more heavily (red reaction) Avith a blunt CUTANEOUS REACTIONS 175 instrument. I here reproduce, from a paper by L. R. Muller (1913), Figures 37 to 39, which show these reactions very distinctly. In Fig. 37 the skin has been stroked lightly and the white reaction only has developed. In Fig. 38 the strok- ing has been heavy, without being definitely painful, since there are no traces of a reflex erythema in the left part of the figures. To the right the two lower red stripes are seen to be bordered and prolonged by white. ■P^-JII ■Ps. 3,"*f3wM «EBC K*"* v ^B m ^E£->BSk r lMxwB&&/f*' W*. %x&wwmWi$3£- ■: ■■ | ■' ,^tik :: 'j^^Es ^^1 Vg&mm?**!?- ■'•*■■ --afSsif wL V Wa^ ; '"^'W fl mm*'' BE* '' .,« i ■§%?- Fig. 37. White reaction after stroking the human skin. After L. E. Muller. Miss Carrier has now made a microscopic study of these reactions in order to elucidate, if possible, their mechanism. When a light pressure is applied with a fine pointed glass needle under the microscope, an opening up and dilatation of capillaries takes place almost instan- taneously over the entire field. This would appear to be analogous to the reaction taking place in the rab- bit 's ear after the same stimulus, though there appears &' * m- ^ m-j . j A ■ ..- f j '-fl l§W : '* ! f- 4S ■& l| fi«.' i: - .- ' jh^H ^b r I'i 1 Fig. 38. Red reaction after stroking the human skin. After L. B, Miiller. Fig. 3£). Left: Bed reaction after stroking with blunt instrument. Right: Red reaction and re- flex erythema after scratching with a needle. After L. R. Miiller. CUTANEOUS REACTIONS 177 to be the difference that in the rabbit's ear the capil- laries directly pressed dilate first, while there is a definite spreading (through nerve fibrils) to the sur- rounding field. In the skin of man the spreading ap- pears to be absent and the capillaries dilate only in so far as they are directly mechanically affected. After the initial flush, which lasts for a variable period up to one-half minute, the field may become paler than before, but this reaction has not been constantly observed. The effect of stroking the skin, as in the usual der- matographic reactions, was then watched under the microscope. When a blunt point was passed lightly over a microscopic field the initial reaction was the same as that obtained with a light pressure. The capil- laries first opened up so that there Avere many more in the field than at the beginning of the experiment. Then, after about fifteen seconds, they began to close, and the white reaction came. One of the most striking fea- tures is that the underlying plexus of venules disap- pears also and the tissue background becomes more white and clear. The colour reaction as it appears to the naked eye depends much more on the changes in the venules than on the capillaries proper. When the stimulus is so strong that the resulting re- action is a central red line with a white band on either side, the microscopic picture is at first the same as above. The capillaries open up both in the immediate path of the stimulus and to either side. In the course of fifteen seconds, however, the central part, which has been, of course, more severely stimulated, becomes more flushed. The capillaries are fuller, the underlying vessels more prominent and the tissue background more red. But in the field to either side the capillaries and other vessels begin at the same time to close and very soon both the central red and the surrounding 178 CAPILLARIES white reactions are well developed. It can sometimes be seen without magnification, on the arm of a fair- skinned person, that there is a general flush, out of which the red and white reactions develop. Both the white and the red reactions certainly in- volve, respectively, contraction and relaxation of capil- laries and venules. It has not been possible to see if the arterioles Contract in the white reaction, but they certainly relax in the red, as Miss Carrier has found by the following experiment. On an arm which is cyanotic with cold, where it is impossible to obtain the white reaction because the capillaries have lost the power to contract, the red reaction can easily be ob- tained and has a brilliant arterial colour. Miiller (1913) found that the local vasomotor reac- tions could be obtained just as well in cases where the segment of the spinal cord corresponding to the skin area had been destroyed and the possibility of a reflex thereby excluded. Ebbecke (1917) found, further, that he could obtain these reactions after damage to the cor- responding peripheral nerves, even in cases where these had degenerated, and that local anesthesia with novocaine did not prevent their occurrence. The ex- periments with local anaesthesia have been repeated and varied by Miss Carrier, who confirms them entirely. It is impossible to abolish these reactions by local anaesthesia. These results rule out the possi- bility that the red reaction can be due to a local axon reflex in sensory nerve endings and point to the smooth muscle cells themselves as being directly relaxed by a mechanical stimulus. When the stimulus is strong they relax more and more during a period of a minute or more ; with a weak stimulus they recover after fifteen to thirty seconds, and a more or less complete contrac- tion develops and gives the white reaction. This con- traction might conceivably be due to a hormonal effect CUTANEOUS REACTIONS - 179 of the increased blood supply, but this possible expla- nation is unsatisfactory, as it was shown by Cotton, Slade and Lewis (1917) and confirmed by Miss Carrier that the white reaction could be obtained equally well on an arm in which the arterial supply had been previously blocked, even after ten minutes, when the arm was distinctly cyanotic. The white reaction may, according to the evidence at present available, be due either to a direct effect of a weak mechanical stimulus on the contractile elements or to an axon reflex in sympathetic fibres. This latter possibility cannot be excluded, since a certain number of the sympathetic fibres supplying the vessels reach them by way of the arteries. These escape degenera- tion when the spinal nerve is cut, and they may be resistant to the action of novocaine or cocaine. Ebbecke mentions the fact that in scars, whether fresh and red or old, the local vasomotor reactions can- not be obtained, and it is worthy of note that the capillaries (and venules) of scars appear to react dif- ferently in several respects from the vessels of the nor- mal skin. They are usually more contracted — the scars appear white — but, when exposed to cold, they relax at an earlier stage and make the scars stand out with a distinct blue colour against the otherwise pale skin (Ebbecke). Bier (1897, p. 452) mentions the case of a very seasick man: his face was deathly pale, but the numerous scars from his student duels stood out with a blue colour. We can infer that the capillaries and venules in the scars had failed to contract along with the arteries. The dilatation may have been a conse- quence of the oxygen lack. On the ear of a rabbit a red reaction can be easily \ obtained microscopically, but I have been unable so far to get the white, though I have seen contraction of arterioles on mechanical stimulation. This contraction 180 CAPILLARIES spreads to such a distance along the artery (2 mm. in both directions) that I believe nerve fibres (probably sympathetic) must be involved. Ebbecke has observed the red reaction on certain internal organs, notably the liver, spleen and kidney of several mammals. On the surface of the liver it can be produced with the greatest ease by an extremely weak stimulus, such as a stroke with a soft brush. On the kidney the white reaction can be very clearly ob- tained, even after removal of the organ from the body. In other internal organs and in muscles definite local vasomotor reactions appear to be absent. The local vasomotor reactions of capillaries on mechanical stimulation are of interest in that they emphasize the difference in behaviour between differ- ent capillaries. Their significance from the point of view of the economy of the organism is unknown and their mechanisms, in spite of the work spent upon them, is not clearly understood. The 'paradoxical reactions to venous blood. A number of experiments have been made on the human skin by Rehberg and Miss Carrier and on the rabbit 's ear by Rehberg and myself in order to verify and to explain certain results obtained by Bier (1897, 1898) and by Zak (1921), which appeared to be funda- mentally at variance with some of our own and, indeed, with generally accepted ideas. Bier has studied the mechanism of the reactive hyperemia, which manifests itself in the limbs of ani- mals and man when the arterial supply of blood has been cut off for some time, and he contends that this hyperemia can be brought about only by arterial blood, while the capillaries react by contraction to venous blood; they have what Bier has termed " Blutgef iihl " : a sensitivity to the quality of the blood. CUTANEOUS REACTIONS 181 This theory, which does not a priori commend itself to a physiologist of a mechanistic disposition, is sup- ported by experiments which cannot, however, he lightly dismissed, and which appeared to us to deserve the most careful consideration. I shall begin by quot- ing some of these experiments. In experiment 74 (1898) Bier, with a bandage, sud- denly cuts off the blood supply to the arm of a man who is lying in the horizontal position. He states that he obtains in this way a limb filled with about the nor- mal amount of blood. When the arm is now allowed to hang down the veins very rapidly swell, so that they are soon quite full. Experiment 75 is of a similar nature, except that the arm is held up and the blood forced out before the bandage is placed around it. At first the veins are hardly visible, but, after the arm has hung down for about five minutes, they stand out clearly, filled with blood. In experiment 76, on the arm of a fair-skinned man, Bier produces a slight hyperemia by venous stasis, so that the arm is slightly blue, and then completely occludes the blood supply with an Esmarch bandage. The arm is allowed to hang down before a hot air reg- ister. At first it is a uniform blue, but soon becomes spotted with white, and after fifteen minutes the white spots predominate on the upper arm. At this time the forearm and hand are still largely blue, but white spots do not fail to appear even on the finger tips. An examination of the blue spots with a hand lens shows that there are many little white islands in them. A similar experiment to this has been made also on a pig (No. 77), in which the axillary artery and vein were dissected out and the whole of the leg, except these vessels, tied off. The artery is closed for ten minutes. On opening it a very pronounced hypersemia develops rapidly. When it is closed again the colour 182 CAPILLARIES changes to blue, but goes back after two minutes to the natural colour and thereupon to a pronounced pale- ness. The toes on the hanging leg remain blue, how- ever. This experiment succeeds also, but the paleness develops more slowly when the brachial plexus has been removed (No. 79). A reactive hyperemia after artificial ischsemia is produced in the hind leg of a pig (Bier, 1897, No. 27). The animal is asphyxiated and it is noticed that when the skin gets blue the hyperemia quickly fades away, to develop again when the animal is allowed to breathe. These and a number of other experiments are ex- plained by Bier as due to contraction of capillaries. In experiments 74 and 75 this contraction drives the blood into the veins, while in experiments 76 and 77 the pressure is so high that only some capillary fields are strong enough to contract against it. From Zak's experiments I quote the following: The blood supply to the arm is stopped. When the hand is thereupon energetically opened and closed about thirty times all five fingers become white. The whitening does not change once it has appeared, even when the hand is at rest. If the artery is now opened a hypersemia spreads rapidly over the forearm and more slowly out on the fingers, taking up to ten seconds to fill the blanched area completely. Zak explains this as a contraction of the capillaries which will not toler- ate venous blood. When products of metabolism are piled up by exercise, the capillaries are stimulated to contraction. In another series of observations Zak deals with the appearance of the well-known reflex erythema, pro- duced in the skin by painful stimuli (as shown above), in combination with a cutting off of the arterial supply of blood. He finds that the want of arterial blood pre- vents the appearance of the reflex erythema, though CUTANEOUS REACTIONS 183 the direct red reaction comes about in the normal fashion. The erythema appears, however, when arte- rial blood is admitted and stands out clearly when the initial flush has subsided. A cutting off of the arte- rial supply during the latent period after the stimulus prevents the appearance of the reflex erythema, and even when the supply is cut off after the erythema has developed, white islands appear in it after one to two minutes and it soon completely disappears. "We, my collaborators and myself (Eehberg and Car- rier, 1922), can confirm, on the whole, the correctness of the observations here quoted, but, though they seem to furnish strong evidence in favour of Bier's view, we think we can show that they are to be interpreted on quite different lines. The blanching of the fingers in Zak's exercise experi- ment is due, not to any piling up of metabolic products, which, indeed, should take place in the forearm where the muscular' movements are performed, but to simple mechanical squeezing out of the blood by the pressure exerted in bending the fingers. The capillaries are not contracted at all, as they will fill up slowly from the veins when the cuff shutting off the supply of arterial blood remains on. When the cuff is allowed to remain on for twenty minutes the hand is a deep blue, and pres- sure on any spot of the skin gives the impression of touching a membrane laid over a shallow lake of ink, the colour flows back so easily and promptly over the spot made momentarily white by the pressure. Even now, blanching of the fingers can be produced by exer- cise, but the blue returns again almost immediately. Zak's observations on the failure of the reflex ery- thema to appear when the supply of blood is shut off by a pressure cuff on the arm are not entirely con- firmed by Eehberg and Miss Carrier. They find that a blue flush can be obtained instead of a red one when 184 CAPILLARIES the arm is held down in such a position that the relaxed capillaries can be filled with blood by gravity. At the height of the heart the available pressure is insufficient. The experiments of Zak, when rightly interpreted, lend no support to the view that capillaries are able to close against venous blood. The white spots described by Bier are more difficult to explain. These spots show best, owing to the greater contrast of colour, when a hyperemia is produced be- fore a pressure cuff is applied to shut off the supply of blood. In this case white spots begin to show on the blue arm after five to ten minutes. They grow in size until they are one or more centimeters in diameter or merge into larger patches, covering a considerable por- tion of the arm. Previous to these spots on the arm there also appear light red spots, the colour of arterial blood, on the hand, where true white spots are seen only occasionally. Microscopic observation of the back of the hand in an area remaining blue while the flow is stopped shows the following changes. During the first minute or two the open capillaries shift, as they do normally, but gradually more and more become visible in the field, until, at the end of two to three minutes, all the capil- laries are open and remain so for the twenty to twenty- five minutes ' duration of the experiment. The underly- ing vessels, however, are only just visible. The blood becomes a deep blue colour. On releasing, the colour almost instantly changes to arterial, the background becomes red also and the venous plexus very full and prominent. This is quite transitory, however, and in a minute some of the capillaries begin to disappear, the background is less brilliant and two or three minutes later the skin looks quite normal. If a red spot is observed under the microscope it cannot be distin- CUTANEOUS REACTIONS 185 guished from a field with a normal circulation. A white spot on the back of the hand shows microscopically a complete disappearance of underlying vessels and the tips of a very few capillaries only are visible. When the pressure is released the white spot also becomes hyperasmic but not so intensely as the surrounding blue skin, and it returns even more rapidly to a nor- mal condition. In the arm a definite distinction between "red spots" and "white spots" could not be arrived at by means of the microscope, because the normal colour of the skin is white and only a few capillaries are nor- mally open. The evidence presented, however, justifies the conclusion that there are two kinds of spots. The first are arterial spots, that is, areas containing arte- rial blood, fed probably through the collateral circu- lation by the medullary branch of the superior pro- funda artery, which runs for a distance through the humerus, so as not to be compressed by the cuff, and then sends a branch to anastomose with the recurrent radial. It was observed that if the finger was pressed over such a spot, so that the blood was pressed out of it and the surrounding tissue, when the pressure was removed the blood leaked into the surrounding tissue from the periphery, but the arterial spot itself filled immediately from below. It is a rather remarkable fact that the small supply of arterial blood which gets into the occluded arm through anastomoses in the bone can be located in definite spots on the surface, usually to be found again and again in the same places. The steady leaking through of arterial blood explains the gradual filling up of the veins in the hanging arm in Bier's experiments 74 and 75. The real white spots, in which the capillaries and venules contract and drive out venous blood, can be shown to be due to cooling. As we have seen, cold is, 186 CAPILLARIES within certain limits, a very powerful stimulus for contraction of capillaries in the human skin. An arm which is supplied with a minimum of blood cools rap- idly, and it has been found that, when the cooling is increased artificially, the white spots become larger and more numerous. When the arm was kept in water at 35° comparatively few "white spots" appeared; these were pinkish in tint, and it was found that when the blood was pressed out of them it returned rapidly from below. In other words, they were arterial spots. The reaction of the capillaries in the rabbit's ear towards venous blood is demonstrated by the following experiments : An area of about 1 sq. cm. in the ear is compressed by means of a Roy and Brown chamber until it is completely ansemic. When the pressure is reduced after several minutes the field becomes very hyperemia A clamp is arranged near the base of the ear so as to cut off the supply of blood. The blood becomes venous in colour but the hyperemia of the area is maintained. The area is compressed for ten minutes, the ear is clamped and the pressure thereupon reduced. In spite of the very low blood pressure resulting from the clamping, the blood pours into the area, which becomes intensely hypersemic and remains so when the blood becomes venous. Finally, the area is first compressed for fifteen minutes. A clamp is put on and kept on for fifteen minutes, during which time the blood in the clamped ear becomes very venous and the capillaries dilate. Alter the thirty minutes the area is decompressed, with the result that the venous blood flows in slowly from all sides and fills the capillaries, which cannot become much dilated, however, because the supply of blood is insufficient. Opening the clamp for a moment produces an intense hyperaernia. CUTANEOUS REACTIONS 187 I have described in a preceding lecture the effects on the circulation in the rabbit's ear when the animal breathes air with a low and declining oxygen per- centage. It was shown that along with the cyanosis a very pronounced capillary hypersemia of the ear re- sulted. A moment comes, however, when the hyper- semia is suddenly reversed, the arteries, capillaries and venules contract very strongly and the ear, which was, just before, extremely hypersemic and deep blue, becomes strongly anaemic and pale, both macroscopi- cally and microscopically. It is an extremely striking observation when the ear is watched under the micro- scope to see all the small blood vessels contract and pour the blood into the veins. The macroscopic veins themselves are not visibly affected. This change, which corresponds to Bier's observation (Xo. 27), comes, generally, when the pulse has become irregular and the respiration is reduced to single spasms at compara- tively long intervals. To find the mechanism of this striking reaction we have, in several experiments, cut the nerves to one of the ears before the respiration experiment. Cutting the anterior and posterior auricular nerves, which are mainly sensory, has no effect whatever, but section of the cervical sympathetic either abolishes the contrac- tion reaction altogether or diminishes its intensity to a very marked extent. The latter result is probably due to the sympathetic fibres reaching the ear by way of the third cervical nerve (Fletcher, 1898). In any case the observed contraction is due to a powerful stimulus sent out, probably from bulbar centres, along sympathetic fibres. TVe can now summarize our analysis of Bier's and Zak's experiments by the statement that venous blood never induces any local contraction of capillaries, but in all the cases investigated shows a distinct, often 188 CAPILLARIES very pronounced, dilator action, which may, however, in certain circumstances, be overcome by strong con- strictive stimuli, acting directly or through the inter- mediary of sympathetic nerve fibres on the liouget cells. The last reaction examined is evidently responsible for the well-known "paleness of death," which may often be so conspicuous, also, in human beings. The observations made by Hooker (1921) on cats which were suddenly killed by an overdose of ether show that asphyxia is not a necessary condition for the develop- ment of a praemortal contraction of the smallest blood vessels. Hooker noticed that in animals in a state of histamine shock the capillaries remained dilated at death — an observation which has its parallels also in human pathology. Hooker's conclusion that the central nervous system is not involved in the reaction is scarcely valid. It was arrived at from the observation that section of the cervical sympathetic does not pre- vent the contraction, but, as I have shown above, there are other nerve paths through which sympathetic stimuli may reach the ears. Hemorrhagic paleness. It is well known that a large loss of blood brings about a very characteristic paleness of the skin and mucous membranes. Though this paleness is probably partly due to the dilution of the blood which occurs rapidly after haemorrhage, it seems very likely that an active contraction of capillaries and venules may also be responsible for it. In experiments in which dogs were bled to a very large extent Meek and Eyster (1921) observed a sudden contraction of the microscopic vessels of the ear, but their description points rather to a praemortal con- traction than to a compensatory reaction. They point CUTANEOUS REACTIONS 189 out, also, that the quantity of blood in the skin is so small (2 to 3 per cent of the whole) that the contrac- tion would have very little effect if it were confined to the cutaneous vessels. From, his experiments on artificial plethora in dogs, in which quantities of dogs' blood amounting to dou- ble the normal blood volume were injected, "Worm Muller (1873) concluded that the large extra amounts of blood which, as he could show, would remain for a considerable time within the vascular system without causing any appreciable increase in the arterial blood pressure, must be, to a large extent, stored in capil- laries which were normally empty. Xo real proof of this contention has been brought forward, however. I presented several years ago (1912) a great deal of evidence to show that compensation to haemorrhage and plethora is mainly brought about by means of the portal system, but this does not exclude, of course, the existence of other compensatory mechanisms and it would certainly be worth while to investigate the reac- tions of capillaries generally to haemorrhage and plethora. If such reactions do occur, as I expect to be the case, their significance for the organism is obvious, but the mechanism necessary to put them into effect appears to me to be very obscure. The vasoneurotic constitution. I shall say, finally, a few words about the so-called vasoneurotic constitution as described by Parrisius (1921) and adopted by Hagen (1922). It is charac- terized by the great lability or downright instability of the innervation of the vascular system which mani- fests itself in the capillaries as well as in the arteries. Frequent changes in the innervation occur either "spontaneously" or from comparatively trivial causes. In the case histories given by Parrisius it is mentioned 190 CAPILLARIES how the fingers of certain patients become sometimes quite white (simultaneous contraction of arteries, capillaries and venules), sometimes strongly red (dila- tation of all vessels) or deep blue (contraction of arte- rioles with dilatation of capillaries and venules). Fig. 40. Capillary loops at the base of the nail. Case of vasoneurosis. After Parrisius. The innervation varies not only from time to time but also from place to place to a very considerable extent. Capillaries located side by side at the base of the nail, or elsewhere, may show quite astonishing dif- ferences, as seen in Fig. 40, where the narrow loops are nearly normal, while some are enormously dilated. Even within the same capillary vessel there may be dilated portions between others of normal bore, true capillary aneurisms, Fig. 41. All these symptoms demonstrate the instability of the (sympathetic) innervation and, without commit- CUTANEOUS REACTIONS 191 ting myself to any opinion, I would j)oint to the sug- gestive analogy between these observations and those which have been made on the frog a certain time after section of the nerves to one of the legs or after oper- ative removal of the pituitary gland, referred to in preceding lectures. Fig. 41. Capillary aneurysms. After Parrisius. Since the innervation of the skin vessels in the nor- mal subject is by no means constant, there cannot, of course, be a sharp distinction between the normal and the vasoneurotic constitution ; and, according to Hagen, who has examined the capillary circulation in a large number of children and young people, persons who show some deviation from the normal in this direc- tion are by no means rare. In the severest cases of vasoneurosis, local arterial spasms constitute the most pronounced symptoms. 192 CAPILLARIES These lead to cyanosis, often with very considerable dilatation of the corresponding capillaries and sub- papillary venules, and even to Kaynaud's gangrene. Parrisius and Erben (1918) are of the opinion that a spastic contraction of the subcutaneous veins is largely responsible for the cyanosis in these cases, and Parrisius describes a phenomenon which he believes to be explicable only on the assumption that the sub- cutaneous veins are closed by contraction. It consists in the fact that when a round spot on the cyanosed skin is pressed with a finger it becomes quite white and the blue colour creeps in from all sides like the closing of an iris diaphragm, but not at all from below. This observation is easily explained by the fact that the subcutaneous veins are provided with valves which will not allow the return of blood from them, while there are no valves in the small cutaneous veins. There is no acceptable evidence for a spastic contraction of sub- cutaneous veins, which must be considered as a priori rather improbable. LECTURE IX THE EXCHANGE OF SUBSTANCES THROUGH THE CAPILLARY WALL IN the preceding lectures I have been dealing chiefly with the physiology of the contractile elements in the capillary wall, and I have purposely left out, for the time being, everything pertaining to that which constitutes, after all, the chief function of the capil- laries : the exchange of substances between the blood and the tissues, or tissue fluids, taking place through the capillary wall. Apparently, at least, this function of exchange is a very complex one: gases, water, inorganic salts, or- ganic crystalloids of the most varied description, and, in certain tissues, even colloids are constantly passing through the capillary endothelium, and not infre- quently the direction of the passage changes. "After bleeding, the total blood volume in the body is very rapidly recovered. The capillary walls seem to take up the liquid and solid material required, and this mate- rial is at the same time reconstituted so as to produce blood plasma of normal composition. ... If blood is transfused from one animal to another the liquid part of the injected blood is rapidly eliminated." In view of such facts as these it is, no doubt, tempting to ascribe to the capillaries themselves the power of regu- lating the passage of substances through their walls, and the statement just quoted from my friend and eminent predecessor as Silliman lecturer, Dr. Hal- dane (1917, p. 81), seems to imply the belief in such a 194 CAPILLARIES regulation, which is to him the central object for study in physiological science. I venture to think, however, that the essential aims of physiology are better served by an attempt, however hopeless it may appear, to find causal explanations, to find out by what forces the exchange of substances is brought about, to determine as exactly as possible what the capillary endothelium is capable of doing and what it is not. What we have to do is, therefore, to see how far the known physical and chemical forces will carry us in attempting to explain the exchange of each single sub- stance through the capillary wall, which we assume to be an inactive permeable membrane. When these forces fail, Ave have to determine the extent and circum- stances of the failure, to discover which substance or substances can be actively transported through the endothelium, in what direction and, if possible, at what rate. Instead of giving, at this stage, an abstract defini- tion of what I mean by the phrases "known physical and chemical forces" and "active transport," respec- tively, I think it a better plan to illustrate it by means of examples, chosen from the very domain with which we are dealing. The exchange of gases in the tissues. As a case in which the physical forces are certainly sufficient to explain the exchange, I can select nothing better than the supply of oxygen to muscles, the prob- lem which I mentioned in my first lecture by way of introduction to the physiology of capillaries in general. Oxygen is soluble in water and watery fluids, like blood and tissue fluids generally. In the dissolved state, as well as in the free gaseous state, it will spread, by what is termed diffusion, from any point where its concentration happens to be high to all other points EXCHANGE OF SUBSTANCES 195 where the oxygen concentration is lower. The rate of diffusion depends upon the concentration difference. The concentration of dissolved oxygen can he meas- ured by the pressure of the oxygen in an atmosphere with which the dissolved oxygen is in equilibrium, and I define as the diffusion constant for oxygen the num- ber of cc. of the gas which will in one minute diffuse through the distance of \\>. (0.001 mm.) over an area of 1 cm. 2 , when the concentration difference corre- Fig. 42. Apparatus for measuring diffusion through tissue membranes. sponds to an oxygen pressure of one atmosphere. These somewhat arbitrary units have been chosen be- cause they are convenient in physiological applica- tions. Hufner (1897) measured the rate of diffusion of oxygen through water, and the diffusion constant, cal- culated from his determinations, is 0.34. I have meas- ured the diffusion through some animal tissues and I shall briefly describe one of the methods used (1919), because it illustrates the principle and is applicable also to other soluble substances, the diffusion rates of which it would be very useful to know. 196 CAPILLARIES The apparatus (Fig. 42) consists of two metal cham- bers, A and B, of 1.5 cc. and about 50 cc, respectively. A piece of suitable membranous tissue, as, for instance, a thin flat muscle, is arranged as a partition wall between the two chambers, which are filled with blood. The mixing screws, 7 and 8, are arranged to secure the completest possible renewal of the fluid along both surfaces of the membrane. The blood in B is saturated with oxygen at one atmosphere pres- sure, while that in A is oxygen free. The oxygen pass- ing through the membrane enters at once into combina- tion with the hasmoglobin, and its quantity can be determined after a suitable period of diffusion. When the area and thickness of membrane employed is also determined the diffusion constant can be calculated. From a number of determinations by this and an- other very different method, I have found the following diffusion constants for oxygen at 20° : In Water 0.34 (Hufner) Gelatine 15 per cent 0.28 (Krogh) Muscle 0.14 " Connective tissue 0.115 ' ' Chitine 0.013 India rubber 0.077 ' ' These experiments show that the animal tissues are permeable to oxygen but it is seen that they offer a much greater resistance to the passage of oxygen mole- cules than does water or even strong gelatine. The diffusion constant for animal tissues has been found to increase with increasing temperature about 1 per cent per degree between 0° and 40°, taking the rate at 20° as unity. By means of the determinations of the oxygen diffu- sion in muscle, combined with measurements given in the second lecture of the distribution and dimensions of muscle capillaries, it becomes possible to calculate EXCHANGE OF SUBSTANCES 197 the oxygen pressure head necessary to provide a muscle with the amount of oxygen required by it in different conditions. It was shown in that lecture that in the cross sec- tion of a striated muscle we find the open capillaries distributed with conspicuous regularity among the muscle fibres. In a resting muscle only a few are open and the distances between them are considerable in consequence. In a working muscle they are very close together. In either case we can, without committing any serious error, suppose each capillary to supply oxygen independently of all the others to a cylinder of tissue surrounding it. In a transverse section such a cylinder is represented by an area which can be taken as cir- cular, and the average area belonging to each capillary can be calculated by counting the number of capillaries in a transverse section and by division of the total area bv the number found. Fig. 43. Supposing Fig. 43 to represent the cross-section of a capillary (r) with the cylinder of tissue (R) supplied by it, we shall have oxygen molecules constantly leav- ing the capillary through the wall and entering the sur- rounding tissue, where they will be used up at a rate determined by the metabolic processes. The oxygen pressure difference between the inside of the capillary wall and a point at the distance x from the centre of the 198 CAPILLARIES capillary must be proportional to the oxygen-metabo- lism p and inversely proportional to the diffusion rate d; and when these are known, together with the radii r and R of the capillary and cylinder, respectively, and the distance x, it becomes possible to establish a mathe- matical formula from which the pressure difference can be calculated. Not being much of a mathematician myself, I have asked my friend the Danish mathemati- cian Mr. Erlang to Avork out such a formula for me. It runs 10 4 p/ , x x'-r*\ in which T and T* are the oxygen pressures (in atmospheres) in the capillary and the point x, respec- tively, d is the diffusion constant as defined above, and p is the number of cc. of oxygen used up per minute by 1 cc. of muscle ; the distances r, x and R are measured in cm. Putting x = R and substituting ordinary loga- rithms for the natural we get the formula 10 4 p / R W - r T n -T R = ^(1.15R 2 W which will give us the maximum pressure difference necessary to supply the muscle. If this difference is found to be smaller than the oxygen pressure of the venous blood leaving the muscle it follows that every point of the muscle tissue can be supplied with oxygen by diffusion alone: diffusion is quantitatively sufficient to cover the oxygen requirements of the muscle. The pressure heads necessary have been calculated for a small number of typical instances from guinea pig muscles, in which the capillaries and their average distances have been measured. The oxygen consump- tions given in the table are more or less arbitrary, since no determinations could be made in the circnm- EXCHANGE OF SUBSTANCES 199 stances. They have been assumed in accordance with the results of determinations by Barcroft and Kato (1915) on dogs' muscles. The two lowest values are to be expected after several hours' complete rest — prefer- ably with the nerves cut. The highest assumed during rest (3 per cent) was observed by Barcroft and Kato on a muscle which had been active under artificial stimulation a short time before the sample was taken. The maximum assumed for work is a little higher than the highest figure actually measured by Barcroft and Kato (13 per cent). a b c d e f g 9 a ' o^.ij En t^ ro w q, co ,9 a H S " -K ^ r °.S 3.2 13 a -2 2 s a . u »" lis ^ 33a 3^ 0.5 31* 100 3.0 6.3 3 0.02 Rest, -J 1.0 85 61 3.0 3.4 8 0.06 3 270 34 3.8 2.5 32 0.3 5 1400 15 4.6 0.6 200 2.8 Work, 10 2500 11 5.0 0.4 390 5.5 Maximum circulation 15 3000 10 S 0.25 750 15 The oxygen pressure of the venous blood leaving the muscles is probably between 4 and 5 per cent. When a tension difference of 6.3 per cent is necessary to supply the whole of the muscle there will be oxygen lack in those places which are at the greatest distance from open capillaries. The other capillary numbers * This figure has been calculated from an estimation (necessarily some- what uncertain) of the distance between open capillaries (2B) on a liv- ing animal. The corresponding value for the diameter of the open capil- laries (tr) is taken from measurements on a vitally injected preparation with 85 open capillaries per mm; 200 CAPILLARIES counted during rest give, when combined with the oxy- gen consumptions assumed, a low positive oxygen pressure everywhere in the muscle. In a series of experiments undertaken in Barcroft's laboratory, Verzar (1912) has shown that the oxygen consumption of resting muscles depends to a certain extent on the supply of oxygen and decreases when this is diminished. Similar results have been obtained by Gaarder (1918) in my laboratory in determinations of the respiratory exchange of fishes, which have an extremely small number of open capillaries in their muscles. Verzar and Gaarder concluded that the oxy- gen pressure in parts of the tissues must normally be zero, and it is evident that the countings given in the above table are not at all inconsistent with their re- sult. It is, indeed, probable that the capillary circula- tion in resting muscles is regulated so as to maintain the pressure at about zero in certain parts which must, however, be constantly changing with the shifting posi- tions of the open capillaries. After massage and during work the oxygen pressure in the muscular tissue becomes practically equal to \that of the blood. This would hold even if higher rates of oxygen consumption were assumed, and it is evi- dent that in these cases the circulation takes place through a much larger number of capillaries than would be necessary to insure the supply of oxygen. It is, therefore, rather probable that the increase in num- ber of open capillaries serves to meet other require- ments of the muscle. In column / I have calculated the total surface of the open capillaries in 1 cc. of muscle. This is the area primarily available for diffusion and exchange of sub- stances of any kind whatever between the blood and the tissue. When many capillaries are open this area is seen to be enormously increased. EXCHANGE OF SUBSTANCES 201 When diffusion of oxygen through the capillary wall and the tissue cells is more than sufficient to cover the needs of muscles during the heaviest work, it follows that it will probably be sufficient in all tissues, since the available capillary network is, in most active tis- sues, even closer than in muscles, and it follows a for- tiori that the carbon dioxide produced in the tissues can always be eliminated by diffusion into the capil- laries, since the diffusion constant for C0 2 in tissues is some thirty times higher than for oxygen. The C0 2 pressure difference between any point in the tissue and the blood must, moreover, in all circumstances, be an absolutely negligible quantity; and, in so far as the hydrogen ion concentration is determined by the COj pressure, there can be no measurable difference be- tween the tissue cells and the blood. The exchange of crystalloids through the capillary wall. Although there are a few other substances, such as urea, according to the investigations of Gad-Andresen (1921), which are probably able to diffuse freely through most of the cells of the body, the gases are the only ones about which we can say with anything approaching certainty that their only means of trans- port in the body is simple physical diffusion, and for a number of substances we can point out definite in- stances where the physical forces are clearly insuffi- cient, where molecules are brought steadily and at a considerable rate from a place where their concentra- tion is low to another where the concentration remains considerably higher. As an example, I would mention the problem, raised by Heidenhain in 1891, and much discussed in the following years, of the transport of calcium from the blood to the milk in a cow. The cow produces milk at the rate of 25 1. a day, with a calcium 202 CAPILLARIES content of 42.5 g. or 1.7 g. per 1. This quantity of cal- cium is derived from the blood, in which the calcium content does not exceed 0.18 g. per 1. The calcium ions cannot possibly move spontaneously from the blood to the alveoli of the mammary gland. Their transport must involve the expenditure of energy, must involve work performed in living cells by means of some spe- cial "machinery," so far utterly unknown, adapted for that purpose; must involve, in short, secretion of calcium. The point I wish to emphasize in connection with this and the many other well-known examples of glandular secretion, is that there is no need to assume, indeed, there is no cogent reason to suppose, that any part of the secretory work is done by the capillary endothelium. "We have the gland cells for that, large, robust cells with a complicated protoplasmic structure, a structure which we can, however dimly, recognize as having something to do with the secretory work the cells are called upon to perform ; and all that is re- quired of the capillary endothelium is that it be per- meable to the substances which the gland cells take up and secrete. When, in the above example from the mammary gland, the gland cells absorb calcium ions so quickly that their concentration outside the endothe- lium is kept definitely lower than in the blood, there can be no doubt that, with the enormous endothelial surface available, the endothelium need to be only mod- erately permeable to calcium ions to provide all the calcium required by the gland. Still another point comes out clearly from this dis- cussion : The glands are not well suited for a study of the properties of capillary endothelium, since we can- not experimentally separate the endothelial functions from those of the gland cells. For the purposes of such a study, tissues must be selected where we can bring EXCHANGE OF SUBSTANCES 203 a fluid in direct contact with the outside of capillaries and investigate the exchange of substances between such a fluid and the blood. The subcutaneous tissue is one of the places where such an interchange of sub- stances can be conveniently studied. It has been made use of in a very large number of experiments and is being used many thousand times every day by physi- cians for the introduction of the most various sub- stances into the blood of patients. The general result of all the experiments, consciously and unconsciously made, is that all crystalloid substances pass freely through the capillary endothelium in both directions. Usually the injected substances appear in the blood after such a short interval of time that absorption by way of the lymph vessels can be excluded. In a small number of cases it has been ascertained by special pre- cautions that the substances in question were, in fact, taken up directly through the capillary wall. In many experiments a passage of substances has been recorded from the blood to an artificially injected fluid or a natural oedema. In those cases where the point was directly studied concentration equilibrium was ob- tained, or at least approached, for the substances involved between the blood and the outside fluid. Results similar to those obtained by means of sub- cutaneous fluids have been obtained from injections into the abdominal cavity or pathological cases of ascites, in spite of the fact that in this case the diffus- ing substances must traverse the peritoneal epithelium in addition to the capillary endothelium. In a number of experiments made in the nineties by Heidenhain and his school and by Cohnstein, diffusible substances were injected into the blood and compari- sons were made between their concentrations in the blood and in the lymph from the thoracic duct at differ- ent periods after the injection. In many of these experi- 204 CAPILLARIES ments the concentrations were found to be somewhat higher in the lymph than in the simultaneous sample of blood, and in some it was even higher in one of the lymph samples than in any of the blood samples. This was assumed by Heidenhain and his school to be due to active secretion from the capillary blood into the lymph, though it was admitted that diffusion took place also. I believe it is generally agreed now that such a conclusion cannot be binding, in view of the unsurmountable difficulties in the way of determining the real correspondence in point of time between the samples of blood and lymph respectively and the fur- ther difficulty of hitting the real concentration maxi- mum in each fluid by means of a comparatively small number of samples taken at ten- or twenty-minute intervals. Authors who believe in the secretory powers of the capillary endothelium have mentioned several cases in which they consider that active transport of sub- stances must have taken place {e.g., Volhard, 1917). It would serve no useful purpose to consider these in detail, because the conditions are generally, as in the above example, too obscure to allow any valid conclu- sion being reached, and when I review all the facts that have come to my notice I have no hesitation in saying that there is no trustworthy evidence of the capillaries having any power of hindering or favour- ing the passage by diffusion of all kinds of crystalloids through the endothelium. The rates at Avhich different substances will diffuse through the capillary wall seem to be closely related to their rates of free diffusion in water or gelatine. For some inorganic salts and for glucose Clark (1921) has arrived at this conclusion by an interesting series of experiments, to which I shall refer in some detail at a later stage of my argument. EXCHANGE OF SUBSTANCES 205 In a very important contribution to our subject Schulemann (1917) has studied the rates at which organic dyes, subcutaneously injected into mice or rabbits, are distributed over the whole organism and taken up by certain cells. He compares these rates with the rates at which the same dyes will diffuse through gelatine and finds a general and close corre- spondence. Since the dyes are distributed by the blood, after passing into the capillaries at the place of injec- tion, and stain the cells by which they are taken up, after renewed passage through the endothelium, Schulemann 's results show that the passage takes place by diffusion and that the relative rates of diffusi- bility are the same through the capillary endothelium as through gelatine. The general significance of the permeability of the capillary endothelium to crystalloids is obvious : it se- cures the supply of all the substances, which may be required for the intracellular metabolic processes, at the very surface of each cell. If any substance is taken up by a cell for storage, conversion or transport, its concentration in the tissue fluid at the surface of that- cell is lowered, and the lowering leads automatically to a diffusion of that particular substance towards the surface where it is required. The impermeability of the capillary wall to colloids. Schulemann has found that dyes which are unable to diffuse from a watery solution into gelatine will not, when injected subcutaneously into an animal, produce any general vital staining, but remain at the place of injection, where they may be taken up by certain cells. These dyes are clearly unable to diffuse through the capillary endothelium, and the same appears to be the case with regard to colloids generally : they are unable to diffuse through the normal capillary endothelium in 206 CAPILLARIES most organs. Certain exceptions to this rule will be dealt with in the next lecture. From our physiological point of view the greatest importance attaches to the impermeability of capil- laries to the normal colloids of the blood — the proteins. When a protein solution is injected subcutaneously the absorption is extremely slow and apparently does not take place at all through the capillary endothelium. Lewis (1921) has made subcutaneous injections of serum which he was able to recognize by a "comple- ment binding" reaction, sensitive to one part in a mil- lion. He found the specific protein in the thoracic duct after forty minutes, but it could be detected in the blood only after three and a half hours. Protein-free fluids, injected subcutaneously, remain protein free until they are absorbed, and pathological oedema fluids may remain practically protein free for an indefinite period, as shown, for instance, by a case of ascites mentioned by Volhard (p. 1478). The exchange of water through the capillary wall. The impermeability of the capillary wall for colloids forms" the basis of the mechanism for absorbing iso- tonic solutions of crystalloids into the circulation, de- scribed in 1896 in the classical paper by Starling: "On the absorption of fluids from the connective tissue spaces." To demonstrate the absorption of a salt solution, iso- tonic with the blood, from the tissue spaces directly into the capillaries, each of the surviving hind limbs of a dog was perfused with the dog's own defibrinated blood, which was made to circulate regularly through the leg. One of the legs was first made cedematous by the injection of 1 per cent NaCl solution, and it was found that, while the blood circulating through the nor- mal leg remained practically unaltered, the blood cir- EXCHANGE OF SUBSTANCES 207 culating through the cedematous leg became gradually more dilute by taking up the fluid from the oedema. This absorption seemed very puzzling, since the initially higher osmotic pressure of the outside fluid must cause water to pass from the blood into the tis- sue spaces, while the constantly higher hydrostatic pressure of the blood in the capillaries must set up a filtration of water and salts in the same direction. Starling showed that the explanation of the observed absorption lay in the osmotic pressure of the blood colloids. It is unnecessary here to go into the question about the exact nature of osmotic pressure, which is still a debated problem, and I need only remind you that the term osmotic pressure expresses the attraction of dis- solved substances for the solvent fluid; that its exist- ence can be demonstrated when the (watery) solution is separated from pure water by a membrane, per- meable to water, but not to the dissolved substance, in which case the pure water will continually pass through the membrane into the solution, which in- creases in volume and becomes more dilute. When the solution is put under pressure a filtration of water will take place in the opposite direction and, when the pres- sure is increased to such a height that the filtration just balances the osmotic current of water and the volume of the solution neither increases nor decreases, the filtration pressure set up is equal to and can be used as a measure of the osmotic pressure of the solution. The osmotic pressure of a solution depends upon the number of molecules, ions or other particles present, quite irrespective of their kind or size, and for one gram molecule of an undissociated substance (180 g. glucose, for instance) dissolved in one liter of water it is equal to 22.4 atmospheres. 208 CAPILLARIES The osmotic pressure of human blood amounts to about 6.5 atmospheres. Of this pressure by far the greater part is due to the inorganic salts dissolved in the plasma, and most of the rest to organic crystal- loids. The blood sugar, for instance, amounting to about 1 g. per 1., exercises an osmotic pressure of 0.125 atmosphere or 1.3 m. water pressure. Starling showed, however, that even the proteins, which make up practically the whole of the colloids of the blood, have a definite though small osmotic pressure, which can be exercised and measured when the blood is sepa- rated from a protein-free solution, containing the blood salts, by a membrane which is impermeable to proteins. Starling constructed osmometers from small glass bells provided near the top with two vertical tubulures. Over the mouth of the bell was tied a peritoneal mem- brane, which was rendered absolutely watertight by being soaked in 10 per cent gelatine for some minutes after it had been tied on. The membrane was prevented from bulging by fixing it over a perforated silver plate. One of the tubulures was connected either with a long, narrow, vertical tube or with a small mercurial ma- nometer. These osmometers were filled with serum and their lower ends allowed to clip into a salt solution, which was generally chosen so as to be slightly hyper- tonic. In these circumstances the fluid in the vertical tube would sink a little at first, but in all the experi- ments it began to rise within two or three hours and rose steadily for three or four days, the final height varying from 30 to 40 mm. of mercury or 400 to 550 mm. of water. When the osmometers are started with a pressure higher than this, filtration of salt solution takes place until the same point of equilibrium is reached. In the capillary blood vessels we have, just as in the EXCHANGE OF SUBSTANCES 209 osmometer, a membrane which is permeable to crys- talloids and impermeable to colloids. An absorption of isotonic salt solution can, therefore, take place, and, indeed, must take place, when the hydrostatic pressure in the vessels — the capillary blood pressure — is lower than the osmotic pressure of the proteins. Since the surface of the capillaries available for the osmosis is, as we have seen, very large, a rapid absorption of salt solution can be effected by this mechanism, but to study the exchange of water more closely we must obtain more detailed information about the osmotic pressure of the colloids of the blood and about the filtration pressure of the blood in different capillaries. Since Starling's publication the osmometers for col- loids have been repeatedly improved and more accu- rate determinations of the osmotic pressure of the blood colloids made, but nothing of a very essential nature has been added to Starling's explanation. The fractional osmotic pressure of the blood. tie- in connection with a study of the physiological and pathological variations in capillary permeability we have in my laboratory — Mrs. Krogh, Mr. Rasmussen and myself — undertaken some determinations of what I shall call the fractional osmotic pressure of blood. It is well known that no sharp distinction can be made between crystalloids and colloids and that xerj large differences exist between the sizes of particles of dif- ferent colloids. It appeared to us probable that the particles responsible for the colloid osmotic pressure of blood might be of very different size, that certain capillaries were permeable to particles up to a certain size and that the effective osmotic pressure in such capillaries might, therefore, be lower than in others which were less permeable. I shall give a brief account 210 CAPILLARIES of the preliminary results of this research, which is still in progress. The method adopted is based on Sorensen's deter- 3. 4. n Fig. 44. Micro-osmometer. Natural size. EXCHANGE OF SUBSTANCES 211 ruinations of the osmotic pressure of pure proteins. It is best described by reference to Fig. 44. In Fig. 44, 1 is a capillary glass tube, connected be- low by means of a short piece of rubber tubing with the collodion tube 2. This collodion tube and part of the capillary tube is filled with the blood to be determined up to a sharply defined surface near the top of 1. A piece of narrow rubber tubing 3 is closed with the clip 4. The outside fluid — generally Ringer's solution — is put in a glass tube 5 just wide enough to let the collo- dion tube go in. The whole is mounted in a small test tube 6 with some mercury 7 to lift the tube 5 and keep it in position relative to the collodion tube 2. The appa- ratus is mounted in front of a horizontal microscope, through which the meniscus in 1 can be focused. The microscope is provided with a vertical micrometer eye- piece. In experiments on the blood of warm-blooded animals the tube 6 is placed in a water bath at 37°. When the clip 4 is kept closed an exchange of water and diffusible substances takes place through the collodion membrane. The meniscus rises in 1 and the air above it becomes compressed until equilibrium is reached between the osmotic pressure and the filtration pressure. To measure the pressure the rubber tube 3 is connected with a manometer and air pressure ar- rangement and the pressure adjusted until the menis- cus remains stationary. The reading of the manometer (corrected for the height of fluid in the capillary 1) then gives the osmotic pressure. The collodion tubes are prepared over a capillary glass tube by dipping it in a solution of collodion in ether-alcohol, allowing it to dry in air for a certain time, removing the alcohol still remaining by putting the tube in water, whereupon the collodion tube can be pushed off from the glass. The gradations in permeability have been obtained 212 CAPILLARIES by the use of collodion solutions with different propor- tions of ether and alcohol and by varying the time of drying in air before the immersion in water. The more alcohol the collodion contains at the moment of immer- sion the more permeable it will become. We have used 8 per cent collodion in three different mixtures of alcohol and ether, A with eighty parts of alcohol to twenty ether, B with equal parts and C with twenty alcohol to eighty ether. The time of drying has been varied from one-quarter to twenty minutes at 20°, and the tubes are designated by the mixture and the time of drying, 4B, for instance, meaning a tube of the B mixture dried for four minutes before immer- sion in water. After numerous tests we have done the actual osmometric determinations with only four kinds of tubes, y±A, 4B, 6C and IOC. When correctly made y±A is just impermeable to gum acacia. A gum solution will show the same osmotic pressure in such a tube as in 4B, for instance. Rather often, however, it happens that a y±A tube proves to be permeable to gum, in which case the osmotic pres- sure, when tested, falls slowly to 0. The difficulty in making these very permeable tubes quite uniform is probably responsible for the somewhat irregular re- sults which they have given with blood. 4B is practically impermeable to blood proteins. After an osmotic test with blood, lasting twenty-four hours or more, the outside fluid will give a just visible albumin reaction with Spiegler's reagent. For 6G and IOC we have not yet succeeded in finding substances to characterize what I shall term their absolute permeability, but we have found that glucose diffuses very slowly through IOC ; the osmotic pres- sure of a 0.1 per cent solution of glucose will take sev- eral hours to fall to 0. We suppose, therefore, that the EXCHANGE OF SUBSTANCES 213 tube IOC is decidedly less permeable tban any normal capillaries. The tubes have been tested further by measuring the filtration rate of water through them. The following results have been obtained on tubes from different batches at room temperature. Filtering capacity C'c. filtered per minute throu gh 100 cm.! at 1 atm. pressure VU IB GC 10C 6.5 0.94 0.21 0.06 6.3 0.85 0.19 0.03 1.2 0.84 0.21 0.04 1.7 0.86 0.15 2..'; 1.04 3.6 0.72 The results show, again, that uniform tubes of the y^A type have not been obtained, while the others are sufficiently good for our purposes. The dimensions of our osmometers are much smaller than any hitherto used. The collodion tubes are about 30 mm. long and 4 mm. in diameter, holding, when mounted, only 0.32 cc. The outside fluid is reduced to about 0.1 cc. This has not been done primarily to econo- mize the material but in order to obtain constant read- ings with as little delay as possible. The essential point is to allow those molecules which can just pass through a certain membrane sufficient time to get into equi- librium on both sides of the membrane. This is best attained by making the area of the membrane as large as possible in relation to the volume of the tube and, further, by reducing the volume of the outside fluids as far as possible. We have per cm.- of membrane 0.1 cc. internal and 0.03 cc. external fluid. Even with these favourable conditions it may take many hours to reach equilibrium, especially with the tight membrane 10C 214 CAPILLARIES and with the membrane y±A, which is permeable to protein. With the first of these, equilibrium has gen- erally been reached in twelve to sixteen hours, the pressure remaining constant for twenty-four hours afterwards ; with the second it has also generally taken about twelve hours, but in some cases equilibrium has not been secured at all, the protein continuing to dif- fuse through at a very slow rate. The chief difficulty in these experiments is to insure absolute sterility of the osmometer and fluids. If bac- teria are present they will multiply rapidly, especially in the outer fluid, with the result that the manometer readings cannot become constant but generally show a tendency to fall at an increasing rate. The principal determinations so far made have been put together in the following table : Treatment of blood Tp. ioc Osmotic mm. ec pressure water UB %A Protein Osm. press. per c ]c protein mm. water Frog's Blood Hirudinized, 17° 125 50 55 13 2.1 26 " 12° 125 60 60 45 1.5 40 Average, 125 55 57 29 S3 Babbit's Blood Hirudinized 37° 340 325 325 285 5.6 58 " 37° 325 265 225 210 4.8 47 Average, P '', " ? 295 275 247 53 Human Blood Oxalate, 37° 490 515 <50 7.2 72 Hirudinized 37° 480 450 S.5 53 Oxalate, 37 = 710 460 405 <90 7.0 58 Average, (710) 477 457 61 We find for the frog a colloid osmotic pressure which is, on the whole, very low. In the tubes 4B and 6C it is practically the same, and we may conclude that there EXCHANGE OF SUBSTANCES 215 are no colloid particles present of such a size that they can pass through the pores of 4B but are held back by 6C. There are, however, a certain number of smaller particles which are able to exercise an osmotic pres- sure of 70 mm. water in the collodion tube IOC. The pressure of 55 mm. found in 4B and 6C must be due mainly to the proteins, and we find that a fraction of these is held back by the more permeable membrane y±A and is still able to exercise an osmotic pressure. The size of these protein molecules or molecular aggre- gates must be larger than that of the rest. In the rabbit the osmotic pressure of the blood is much higher. There is a small difference in osmotic pressure between 4B and 6C, which may be real and indicate the presence of particles slightly smaller than protein molecules. It is very remarkable that almost the whole of the protein fraction possesses particles of such dimensions that they are held back even by the very permeable membrane y±A. In human blood the case seems to be different. Prac- tically the whole of the proteins present have diffused out through the most permeable membrane so far employed. This does not mean, of course, that the blood proteins cannot be separated into fractions with dif- ferent osmotic activity, but only that the membranes employed by us for the purpose have been too per- meable. On the basis of the assumption that we can consider the osmotic pressure observed in the tubes 4B as due exclusively to the proteins, an assumption which must be at least approximately correct, we can calculate the average osmotic pressure produced by 1 per cent pro- tein in the blood. "We find this to be considerably lower in the case of the frog than in the mammals and rather variable for each single species, so that a calculation of the colloid osmotic pressure from a determination 216 CAPILLARIES of the protein percentage of the blood can never give quantitatively reliable results. The osmotic determinations here given should be considered as preliminary only. They are to be con- tinued and extended to the blood of more animals, to tissue fluids and to the possible effect of certain drugs on the state of aggregation of the osmotically active particles. I believe we have some right to expect that, when thus extended and combined with precise deter- minations of capillary permeability, they will eventu- ally yield results of considerable interest. The remarkable differences in colloid osmotic pres- sure observed between the frog, the rabbit and man will find their explanation, I think, when we come to consider the other factor on which the exchange of water between the blood and the tissue spaces must primarily depend : the blood pressure in the capillaries. LECTURE X THE EXCHANGE OF SUBSTANCES THROUGH THE CAPILLARY WALL (CONTINUED) The capillary blood pressure. THE hydrostatic pressure of the blood in the capillaries determines, by its relation to the osmotic pressure of those colloids for which the capillary wall is impermeable, the direction and rate of exchange of water (isotonic salt solution) between the tissue spaces and the blood. When the capillary blood pressure is higher than the osmotic pressure an excess of water over that attracted osmotically will filter out through the capillary walls, and oedema will develop. When the capillary pressure is lower than the osmotic pressure any excess of fluid in the tissue spaces will become absorbed — as can be observed to take place in most normal tissues. The rate at which such an absorption can take place depends upon the difference between the osmotic pressure and the capil- lary blood pressure. The determinations of capillar}- blood pressure to be found in the literature are not very numerous and are very discordant, varying, for instance, in the case of the capillaries in the human skin between 750 mm. water pressure (v. Recklinghausen, 1906) and 70 to 170 mm. water (average 100 mm.) (Basler, 1914). I do not think it necessary to discuss all these results, since the meth- ods b} r which they have been obtained are all open to serious objections, and I am in the fortunate position 218 CAPILLARIES of having at my disposal a few measurements, which cannot be challenged on methodical grounds, from a study undertaken in my laboratory by Miss E. Carrier and Mr. P. B. Rehberg (1922). 1 During the numerous microscopical observations made by us a very definite, though necessarily quali- tative, impression was gained about the relations of velocity and pressure in the smallest blood vessels. The small arteries and especially the arterioles are normally extremely narrow when compared with the network of capillaries supplied by them. Sometimes the internal diameter of an arteriole does not much exceed that of a single one of the capillaries supplied by it; more often it is about double that of a normal open capillary. The corresponding observation is also made again and again that the arterial current is ex- tremely rapid in comparison with the capillary blood stream. The difference between the capillaries and the venules, on the other hand, is not at all pronounced, though the current becomes certainly more rapid in the veins than it is in the capillaries. The picture of the whole system differs in a very significant way from that presented by an injection preparation. One is often reminded of a relatively broad stream (running at first in a number of separate channels) supplied by a system of pipes — the arterioles — and it is impos- sible to doubt that the main resistance to be overcome lies in the arterioles where, therefore, the main fall in pressure must take place, while comparatively insig- nificant pressure differences must suffice for the cur- rent through capillaries and veins. The determinations by Miss Carrier and Rehberg were undertaken prima- rily to obtain a quantitative basis for this conception. i Special reference should be made, howeveT, to the papers by Leonard Hill (1920) in which the normally very low values for capillary blood pressure are strongly insisted upon. EXCHANGE OF SUBSTANCES 219 The pressure in the capillaries of the human skin was measured in the following manner : A piece of glass tubing about 4 mm. in diameter and 10 cm. long is drawn out at one end to a fine point, 0.01 to 0.02 mm. in outside diameter. A little normal saline is then run into the top of this glass tube and blown out to make sure the lumen is patent. The glass tubing is connected by a long piece of rubber tubing to the top of one arm of a water manometer. When the level- ing bulb is raised the water in the other arm also rises. A pressure equivalent to the height of this column of water is therefore exerted at the capillary tip of the glass tube, which is connected with the first arm of the manometer, but the pressure is held in check by the surface tension of the saline so long as the opening is surrounded by air. If now a favourable field for micro- scopic examination, such as the base of the finger nail or the back of the hand, is covered with oil and illu- minated with a strong light, it is possible, under the binocular microscope, to pierce the tip of a capillary loop with the glass capillary tube and hold the tip for a few seconds in the lumen of the capillary. If the pres- sure in the capillary is higher than that in the glass tube, blood will run up into the salt solution. If, how- ever, the reverse is the case, no blood will run in. If the blood does run in, the pressure is raised by several cm. and another capillary loop tried. If the blood here fails to run into the glass needle, two or three more loops are tried to confirm this result and the pressure is then lowered to an intermediate point. By narrowing down the limits between which the blood does or does not run up against the pressure, it is possible to come within % cm. of the pressure in the capillaries. Occa- sionally the pressure in the capillary is just equal to the pressure in the glass needle. When such is the case 220 CAPILLARIES the blood pulsates in the tip of the glass needle with every beat of the heart. This method has been controlled in several ways and found to be absolutely reliable. Unfortunately, it can be used only on rather wide capillaries which are firmly imbedded in a tissue which is not too soft. In other cases it is too difficult to pierce the capillary with the glass needle, and in any case it requires considerable dexterity of manipulation. The capillary pressure is greatly influenced by the vertical position of the point on the surface where the determination is made relative to the thoracic cavity. A series of determinations of capillary pressure on the hand of one of the subjects (A.R.) will make this clear. The position of the hand relative to the centre of the clavicle is given in the table. Positions above the clavicle are indicated by — , below by +. The pressures are given both in cm. water and (by division with 1.05) in cm. blood. Position, —20 + 1 +7 +8 +12 +19 +33.5 +36 blood, 4.3 4.3 4.3 5.7 9.5 16.2 27.5 30.5 I ' re; ' s,l,e '| water, 4.5 4.5 4.5 6 10 17 29 32 The pressure is seen to be constant and very low, from 7 cm. below the clavicle, upwards. Below that point it increases regularly with the increasing vertical distance. The results are shown graphically on the chart, Fig. 45. In another subject (P.R.) the neutral point was about 10 cm. below the clavicle, and the pressure meas- ured at or above this level was 7.5 cm. water. A few other subjects showed minimum capillary pressures in the hand between the limits 4.5 and 7.5 cm. water. To understand why the capillary pressure becomes constant above a certain level one must bear in mind the slightly negative pressure in the thorax and the EXCHANGE OF SUBSTANCES 221 peculiar collapsible structure of the veins connecting the capillaries with this cavity. The pressure is made up of two components : the hydrostatic pressure, due to the difference in level between the capillary system and the thoracic cavity, and the frictional resistance pressure, determined by the total cross-section of the veins and the velocity of flow. When the hands are lifted above a certain level the veins begin to collapse, with the result that the frictional resistance is in- 25 20 20 25 30 35 40 Fig. 45. Capillary pressure in hands at different levels above and below clavicle. Dotted line : Theo- retical hydrostatic pressure. After Behberg and Carrier. creased. As the walls of the veins are quite soft and yield to the slightest excess of pressure from outside, the negative hydrostatic pressure which would be set up in a rigid tube when lifted above the level at which it enters the chest is automatically compensated by the increase in frictional resistance in the venous system, which collapses more and more as the hand is raised. It is a simple physical consequence, therefore, of the conditions of flow in collapsible tubes that the capil- lary pressure must become constant at all levels above that at which the veins begin to collapse. 222 CAPILLARIES The venous blood pressure as affected by position. It will be noticed that in the chart, Fig. 45, the ob- served capillary pressure at the lowest level at which the hand could be held falls short a little of the hydro- static pressure to be expected. This has been noticed before in experiments on venous pressure and is very conspicuous in the foot of man (v. Recklinghausen, Hooker, 1911), and to investigate it Miss Carrier and Rehberg have made some venous pressure measure- ments. 20 30 -10 60 60 70 80 90 100 NO 120 130 MO 150 Fig. 46. Venous pressure in human foot at different levels below clavicle. The dotted line representing the theoretical hydrostatic pressure should have been drawn more to the right, probably from a point near 25 on the abscissa. They determined the venous pressure in a manner similar to that employed by v. Recklinghausen and by Hooker (1911), by measuring the outside pressure necessary to bring a vein to collapse. On well-filled veins this method works with an accuracy of ± 1 cm. and has no systematic errors. On veins which are EXCHANGE OF SUBSTANCES 223 already half collapsed the method is difficult to use and not very accurate. When compared with the capillary pressure the pres- sure in superficial veins in the hand or foot is generally lower by 2 to 3 cm. water pressure. On a subject in a natural standing position they find, like v. Recklinghausen and Hooker, that the venous pressure in the foot falls very considerably short of the hydrostatic pressure, reckoned from the lower end of the sternum about 25 cm. below the clavicle in their subject, and is, moreover, very variable. When the sub- ject stands on one leg on a low stool and the leg to be examined hangs down freely, the discrepancy is dimin- ished and constant results are obtained after standing for five minutes. The chart, Fig. 46, gives a number of determinations on the subject P.R., who was sitting or lying in a chair with the foot resting in different positions. In all the deeper positions the actual pres- sure is distinctly lower than the hydrostatic, which in the higher it exceeds slightly. The venous pump. This interesting phenomenon is due to slight mus- cular movements in the leg which the subject may feel but is unable to suppress completely, as was shown by Hooker (1911), who found, further, that when the legs were paralysed or the subject anaesthetized the pressure would rise up to the level to be expected theo- retically — slightly above the hydrostatic pressure cor- responding to the vertical distance below the chest. The details of the mechanism of this "venous pump" undoubtedly deserve further study, both from the anatomical and from the physiological point of view, but, generally speaking, the pumping action is due to compression of veins by muscular movements. Owing to the valves any compression of a venous seg- 224 CAPILLARIES rnent will, at least partially, empty that segment in the central direction, by which action the pressure is low- ered and the segment can fill up again from below. The efficacy of the venous pump is very considerable. As we have seen, even the slight involuntary move- ments made by a person standing in the erect position may be sufficient to reduce the pressure some 40 cm. In walking, the pressure in the veins of the foot is usually reduced to very near zero, as can be seen or felt on the veins of the foot of a walking person. In the arm of man the efficacy is much less, but even on the hand hanging down vertically the high pressure to be felt by testing the veins can be considerably reduced by rap- idly opening and clenching the hand a few times. An effective venous pump is certainly present in the legs of all animals of high stature. In the hoof of the horse a special arrangement of valved veins is de- scribed (Lungwitz, 1910) which are alternately com- pressed and dilated in stepping movements and pro- vide a pump which will reduce the pressure in the capillaries of the toe. The. liability to filtration oedema of the loiver parts of the body in large animals. When the venous pump is not acting, the capillary pressure in the foot becomes decidedly higher than the effective osmotic pressure of the blood, even if the capillaries are assumed to be as impermeable as is compatible with an adequate supply of such crystal- loids as sugar and certain amino acids, and the occur- rence of filtration cedema is unavoidable. In the experi- ment on P.R., referred to above, the subject was kept standing for fifteen minutes, during which time the circumference of the hanging foot increased by swelling from 27 to 29 cm. Some swelling of the hands is said to be frequent in soldiers having to march or EXCHANGE OE SUBSTANCES 225 stand with hanging arms for some time. Other exam- ples of this filtration oedema caused by the capillary pressure exceeding the effective osmotic pressure of the blood will be given later. It is evident that, in spite of the activity of the venous pump, which depends wholly on contractions of "voluntary" muscles, the capillaries in the lower parts of the body of a large animal (the udder of a cow, for instance) must often be exposed to rather high hydrostatic pressures, and it is from this point of view that the possession of blood with a high colloid osmotic pressure affords an important protection against filtration oedema, the more urgently required, the higher the level of the heart is above the ground. The fact that in most large animals the heart is placed at the lowest possible level in the chest is perhaps also of some significance in this connection. Its posi- tion is, in fact, generally so low that the elephant and the giraffe are (as far as I know) the only animals having their heart at a higher level than man. In the giraffe with a total height of 5 m. the heart is at a height of about 2.5 m., and it would be extremely interesting to know just how the giraffe avoids the development of filtration a?dema in its long legs. Unfor- tunately, we have not found it possible to obtain giraffe blood for determinations of the fractional osmotic pressure. I mentioned in my first lecture the fact that the functional capacity of an anatomical structure may depend on its size just as much as upon its form. I would like to draw your attention to the example of this rule which we have, I believe, in this case. It might be imagined that a circulatory sj^stem like that of a mammal might be reproduced in any desired dimen- sions, but it is at least not improbable that the giraffe is not very far removed from the limit at which, in an 226 CAPILLARIES animal living on land, the unavoidable increase in hydrostatic capillary pressure can be compensated by increasing the colloid osmotic pressure of the blood, which, in its turn, must be limited by the consequent increase in viscosity and perhaps by other factors. In aquatic animals, like the whales, the hydrostatic pressure at any one point will be the same inside and outside the capillaries, so that these animals will prob- ably require only a low colloid osmotic pressure in their blood. Speculations such as these, though admittedly loose, are sometimes very useful. Sooner or later an oppor- tunity offers of putting them to the test. It is, of course, very gratifying to find them confirmed, but generally they are even more useful when they turn out to be wrong, because, in that case, they serve to discover at what point the reasoning went astray and to guide it back into a channel which may possibly lead it on- wards. The problems of physiology are so complicated that, to put it tersely, one cannot expect to be able to reason correctly from the facts for more than five minutes at a stretch. The normal relations between capillary pressure and colloid osmotic pressure. We want very badly some determinations of capil- lary pressure in other organs than the skin, and we want especially a series of determinations in cases where the arteries are dilated, with and without simul- taneous dilatation of capillaries. In the latter case, especially, there must be a substantial increase in capillary pressure over that in the veins, but it seems to me rather doubtful if it can become high enough to produce filtration oedema, except in cases where the venous pressure is also appreciably increased. Normally, at any rate, there seems to be in almost all EXCHANGE OF SUBSTANCES 227 tissues a definite excess of the effective osmotic pres- sure over the capillary blood pressure. This appears to follow from the two facts that animals, and also man, in a normal condition are able to stand a considerable dilution of the blood, brought about, for instance, by haemorrhage and subsequent injection of saline, with- out developing oedema, and that salt solutions injected anywhere in the subcutaneous tissue, in muscles or in the peritoneal or pleural cavities, are rapidly absorbed directly into the blood. The dilution of the blood after hemorrhage. After haemorrhage the blood increases in volume and becomes diluted without any fluid being introduced into the organism from outside. Scott (1916) has shown that the fluid thus taken up from the tissues is probably a solution of salts with very little protein or perhaps none at all, as one would expect from the relations of osmotic pressure to capillary pressure and the normal permeability of the capillary wall. There remains, however, an extremely interesting problem from the point of view of the regulatory powers of the organism. By the haemorrhage the power of the blood to absorb such a fluid is not appreciably altered, though there may be a slight fall in capillary pressure, and there is, certainly, not normally any reserve of fluid in the tissue spaces which could be taken up. The only possibility appears to be that the haemorrhage induces by some mechanism a mobilization of water and salts from tissue cells into the tissue spaces. The nature of such a mechanism is entirely unknown and I should not like to venture even a guess regarding it. The exchange of water against diffusible substances. We have to consider briefly the effects upon the exchange of water caused by a change in the concen- 228 CAPILLARIES tration of diffusible substances either in the blood or in the tissue fluids. Just because the substances in question are diffus- ible, such a change can be only temporary, but while it lasts it is able to exercise a very considerable influ- ence upon the water exchange, since the rate of diffu- sion of most substances is slow when compared with that of water. As a striking example of this influence I would mention the pulmonary oedema produced ex- perimentally by Laqueur (1919) by the injection of 1 cc. of concentrated (50 per cent) glucose solution into the trachea of rabbits. The sugar attracts water osmoti- cally from the blood in the pulmonary circulation ; dif- fusible salts — NaCl in particular — pass out also at the same time that the sugar is passing slowly inwards. In less than an hour the quantity of fluid in the lungs has increased to 15 cc. or more and has become isotonic with the blood, whereupon the volume begins to de- crease by absorption into the blood. In numerous experiments by Brasol (1884), Leathes (1885), "White and Erlanger (1920) and others, glu- cose solutions have been injected into the blood of animals with the invariable result that water is drawn from the tissues into the blood at such a rate that the normal osmotic pressure (total osmotic pressure) is re-established in one-half to two minutes, whereupon there is a comparatively slow return to normal condi- tions, as the surplus sugar diffuses out into the tis- sues and is excreted by the kidneys. In an interesting series of experiments Clark (1921) has studied the absorption of nearly isotonic solutions of various substances into the blood from the peri- toneal cavity of rabbits. He finds that the rate of absorption of the fluid as a whole depends upon the diffusibility of the dissolved substance, becoming slower with diminishing diffusibility. Glucose diffuses EXCHANGE OF SUBSTANCES 229 so slowly, compared with the salts, that the amount of fluid in the peritoneum is actually increased during the first couple of hours, because salts will diffuse out from the blood and draw the necessary water to main- tain the isotonicity of the fluid. After three hours about 75 per cent of the glucose is absorbed, but, even at this time, the volume of fluid present may be in excess of the volume injected. Clark draws from his experiments the very inter- esting conclusion that the membrane in question (capil- lary endothelium + peritoneal epithelium) does not show any selective permeability. The relative rates at which the different substances pass through are, at least approximately, the same as the rates at which they diffuse through dead membranes and even through gelatine or water. The exchange of water against slowly diffusible sub- stances is probably to a large extent responsible for the production of lymph by exudation from the blood vessels in active organs. In their beautiful researches on the ' ' effects of func- tional activity in striated muscle and the submaxillary gland" (1915), Barcroft and Kato have measured the exudation from the blood in these organs in dogs by the ingeniously simple method of measuring the blood flow and comparing the hsemoglobine percentages in the arterial and venous blood respectively. An increase in Hb. percentage, brought about by the passage through an organ, means, of course, a corresponding concentration of the blood and thus measures the quan- tity of fluid given off to the organ. The production of lymph, which they found to be insignificant in the resting gastrocnemius muscles and nil in the gland, becomes very considerable during, and for hours after, activity. In muscle the enormous exudation of 5 cc. per 100 g. muscle per minute has even been recorded 230 CAPILLARIES for a short period, though in most cases it did not exceed 2 cc. Even this means that a volume of lymph equal to that of the organ is produced in less than one hour. The mechanism of this lymph production is prob- ably complex, but the fact that it is closely correlated quantitatively with the rate of oxidation in the active organ, is an indication that it may partly, at least, be caused by metabolic products, which diffuse sufficiently slowly to be osmotically active. A determination of the fractional osmotic pressure of the lymph from active organs will probably yield conclusive information on this point. Differences in capillary permeability. In the preceding lecture I have given the reasons which compel us to assume that in most tissues the capillaries are normally practically impermeable to protein. It is quite possible that in some places the perme- ability is even lower and is able to cause an appre- ciable increase in the effective osmotic pressure of the blood, but it remains to be seen if such a supposition can be substantiated by direct evidence. It has been shown conclusively, on the other hand, and notably by Starling's admirable investigations on the formation of lymph (1894), that in certain organs the capillaries are normally permeable to protein to such an extent that the effective osmotic pressure be- comes lower than the capillary blood pressure and a filtration of lymph through the capillary wall is con- stantly going on. The most permeable capillaries are those of the liver, which fact is naturally correlated with their very peculiar structure as referred to in my third lecture. The normal blood pressure in the liver capillaries is extremely low, as inferred by Bayliss and Starling EXCHANGE OF SUBSTANCES 231 (1894) from simultaneous determinations of pressure in the portal vein and the cava, and the flow of lymph is, therefore, not very great. Any rise of pressure in the liver capillaries will, however, bring about an in- crease in the lymph flow proportional to the pressure, and the composition of this lymph will approach so near to that of the blood plasma that it must be con- cluded that the capillaries are permeable to all the blood colloids. The filtration of the colloids is, however, undoubtedly slower than that of the crystalloids, and this will cause some dilution of the lymph, as compared with the plasma, and check the flow to a certain extent. The capillaries of the intestinal mucous membrane are also permeable to protein, but the protein content of the intestinal lymph is always lower than that of the the blood and in the light of the experiments on frac- tional osmotic pressure, described in the preceding lecture, it seems natural to assume that the capillaries in question are permeable to a certain fraction of the blood protein. This fraction appears to exceed one-half or perhaps two-thirds of the total protein by weight, and, since it must be the smaller molecules which can pass out, the resulting reduction in the effective os- motic pressure of the blood will be even greater. Bay- liss and Starling (1894) have found the pressure in the portal vein of medium-sized dogs to be about 100 mm. of water. The pressure in the intestinal capillaries must be higher, but the difference is probably small. The available information is consistent, however, with the assumption that there is normally in the in- testine some excess of capillary pressure over the effective osmotic pressure, which explains the observed constant production of lymph. Starling has, by his varied experiments, established the fact that any increase in pressure in the intestinal capillaries brings about a corresponding increase in 232 CAPILLARIES the flow of lymph from the intestine, and at the same time he has verified the older observation that, during any such increase, the percentage of solids (that is, of protein) in the lymph is diminished. This observation is of some theoretical importance. Generally speaking, the rate of lymph filtration must be determined by the rate of filtration of the slowest substance. If we suppose, for instance, that water is filtered at a more rapid rate than the salts, the excess of water in the filtrate would set up osmotic forces, counteracting the filtration of water and reducing its rate to that of the salts. Hence it follows that the pro- portion of crystalloids in the filtrate must be the same as in the blood. The osmotic force set up by a protein deficit in the lymph is so low, however (about 60 mm. water pressure for 1 per cent of protein, according to the determinations given in the preceding lecture) that it may be overcome by the filtration pressure, and this is obviously what is going on in the intestine when lymph is produced by filtration at a rapid rate. The pulmonary oedema produced by Laqueur (1919) by the introduction into the lungs of a small quantity of concentrated glucose was found to contain 1 to 2 per cent of protein, showing that the lung capillaries must also be (or can become) permeable to a fraction of the blood protein. The increase in capillary permeability due to dilata- tion. Apart from the differences in permeability between the capillaries of various organs, thus briefly described, it can be shown that the permeability of one and the same set of capillaries is variable and, especially, that it is normally increased by dilatation. This is, I be- lieve, a point of prime importance, which will go a long way to explain capillary reactions both in physio- EXCHANGE OF SUBSTANCES 233 logical and in pathological conditions, and I think it right, therefore, to illustrate and substantiate it by a number of examples. I have mentioned in a preceding lecture (VI) the effect of a small drop of a strong urethane solution applied to a single narrow capillary and arteriole in the mucous membrane of the frog's tongue. While the arteriole remains narrow the capillary dilates more and more as the blood flows into it and little or no blood flows out from the venous end, but at the same time, the concentration of the blood is visibly increased from moment to moment. The corpuscles become packed together, forming a solid lump near the venous end ; this lump grows into a long column by fresh cor- puscles being added to it. One is able to watch quite distinctly how the plasma disappears from the normal blood entering the capillary, while the corpuscles are piled up until, finally, the whole strongly dilated capil- lary is completely filled and we have the typical stasis. The entire plasma with all its colloids is filtered off through the capillary wall. Of what kind is the alteration which must take place in the capillary wall? Judging from the rapidity with which the fluid disappears, I was led at first to assume that real openings were formed in the capillary wall, and I formed the lrypothetical conception of the forma- tion of fissures between the endothelial cells which might occur, I thought, especially at points where the borders of three or more cells meet. This conception proved to be quite erroneous when put to the following- test. Diarysed and filtered India ink, the particles of which are submicroscopical and can be taken to be about 20(W. in diameter, is added to the blood so as to make the plasma in the capillaries of a distinct gray colour, while in the larger vessels it is almost black. 234 CAPILLARIES If microscopic openings were now formed in the capil- lary wall by the application of urethane the gray plasma would be filtered off, but the result of the actual experiment was that the India ink particles were held back quantitatively while the clear plasma disappeared as before. This experiment has been repeated and varied in several ways, always with the same result. A very striking experiment is the following : Perfusion of the hind limb of a frog is performed with India ink in Einger after the capillaries of the web have be- come strongly dilated by perfusion with pure Ringer. As the perfusion fluid has no measurable colloid osmotic pressure the water is filtered off through the capillary wall and an India ink stasis develops in the capillaries, which become jet black, while no trace of the substance can be found outside the vessels. We must conclude, therefore, that the capillary wall re- mains mechanically intact, while its permeability to colloids is increased pari passu with the dilatation. A number of other observations and experiments show that the increase in permeability cannot be due to any specific action of the urethane, but is a regular accompaniment of dilatation, quite irrespective of the manner in which the dilatation is brought about. Be- fore we go into these, a few words must be said, how- ever, about the ways in which an increase in capillary permeability manifests itself microscopically and macrosopically. Microscopically the development of stasis is the surest sign of increased permeability. Stasis is some- thing quite different from a simple retardation or stoppage of flow which leaves the quantitative relation between the corpuscles and the plasma unaltered, and has nothing to do, directly, with agglutination of cor- puscles. Agglutination is common when the flow is stopped and is easy to recognize by the spaces of clear EXCHANGE OF SUBSTANCES 235 plasma separating the lumps of corpuscles. In stasis the concentration of corpuscles is increased, and in the most typical stasis the red corpuscles become packed to the same extent to which they can become packed by centrifuging in a hematocrit. The fluid between them disappears completely and the column becomes trans- parent instead of being normally opaque. 1 When the stasis is able to reach this stage the capillary wall must, evidently, be permeable to all the plasma constituents. The blood flow in the capillary stops altogether and no further exudation of fluid through its walls can take place. When stasis develops rapidly the residting oedema is only slight. A. minor increase in permeability will not, as a rule, stop the circulation, but, when it is sufficient to upset the balance between nitration and osmotic attraction of water, a steady stream of fluid leaves the capillaries through their walls. The blood becomes concentrated during the passage and oedema may develop in the surrounding tissue spaces. Microscopically this stage is, in most cases, difficult to recognize, but the local oedema is often very conspicuous microscopically. It can be stated, as a general rule, that all those physical and chemical agencies which I have enumer- ated in the fifth and sixth lectures as giving rise to capillary dilatation Avill also, when applied in sufficient strength, cause oedema or stasis, or both. I may men- tion especially the case of histamine. Dale and Rich- ards found, as you may remember (Lecture II), that the local application for ten to twenty seconds of dilute histamine to the cat's pancreas caused capillary dila- tation. When the same solution was applied for a few i In normal blood a beam of light will undergo refraction at every passage from a corpuscle to plasma and vice versa. Through a mass of packed corpuscles it can pass in a straight line and suffers absorption only, as in a concentrated heemoglobine solution. 236 CAPILLARIES minutes the result was a pronounced local oedema, by which the single pancreatic lobules became separated as "if embedded in a colourless, transparent jelly." In their study of histamine shock Dale and Laidlaw (1918) obtained quantitative evidence of a progressive concentration of the blood, which can only be explained by the loss of fluid through the capillary walls. In some very interesting experiments Magnus (1899) has studied the effect of plethora induced by infusion of saline into the blood. He found that in normal ani- mals (dogs and rabbits) the blood can be considerably diluted without the production of subcutaneous oedema, showing the existence of a certain reserve of colloid osmotic pressure in the normal blood. In dead animals (one-half hour) the infusion produces a very pro- nounced cedema, and the same is the case in living animals which have been treated with capillary poisons (arsenic, phosphorus) or are deeply anaesthetized with chloroform, ether or chloral. These last-named results are of special importance, because they show how a dilatation which is too slight to give symptoms when the blood is normal, produces, nevertheless, an in- crease in permeability which gives rise to filtration when the colloid osmotic pressure of the blood is low- ered. Magnus mentions also some experiments by Cohnheim and Lichtheim, in which a very slight ex- perimental inflammation gave rise to a local cedema when the blood was diluted by infusion of saline. It has been observed in numerous experiments by different investigators that the capillary dilatation caused by local mechanical stimulation may give rise to considerable exudation of fluid. I have seen it some- times in the frog's tongue when a sharply localized dilatation was brought about by repeated very weak stimulation with a hair. Ebbecke (1917, p. 31) has seen an evanescent oedema of the liver surface in mammals EXCHANGE OF SUBSTANCES 237 accompany the red reaction produced by slight me- chanical stimulation, and in the skin of normal persons he has found it possible to raise a wheal by repeated light pricking of a small area with a needle. This is the way blisters are produced by manual labour in the hands of persons not accustomed to that kind of exertion. In our perfusion experiments with and without pitui- trine on the hind leg of the frog, Eehberg and I have often observed the gradual concentration of the "blood" and the increased lymph flow developing in the web when the capillaries became dilated, and Ave have found that dilatation beyond a certain point, cor- responding, no doubt, to a degree of permeability which allows the osmotically active gum molecules to leave the vessels, leads rapidly to a complete stasis. By long-continued stimulation of one glossopharyn- geal nerve in the frog, which causes, as described in the fourth lecture, dilatation of the corresponding capil- laries, Brack (1909) succeeded in producing a pro- nounced oedema in the part of the tongue innervated, and Ave have a counterpart of this reaction in the blis- ters formed in those areas of the human skin Avhere the vessels are dilated in consequence of irritative proc- esses in the corresponding spinal ganglia in the disease Herpes zoster. When I mention, finally, that in those cases Avhere capillary dilatation after chemical stimulation has been prevented by local anaesthesia (Krogh, 1920) the increase in permeability has also failed to appear, I think Ave can safely draw the general conclusion that no dilatation of capillaries involving mechanical stretching of the endothelium can take place Avithout being accompanied by an increase in the permeability, an increase AAmich runs, on the Avhole, parallel to the degree of dilatation and which allows all the normal 238 CAPILLARIES plasma colloids to filter off: rapidly when the capil- laries are strongly dilated. As things stand at present the above conclusion can be given only in qualitative terms. The next step re- quired is to obtain a quantitative formulation. I am afraid that it will be very difficult to make this obvi- ously necessary step. Attempts to measure the absolute permeability of capillaries. We have in my laboratory made some preliminary attempts to determine the absolute permeability of capillaries at different stages of dilatation by injec- tion of suitable dyes into the circulation while observ- ing the capillaries microscopically to see whether the dye would penetrate the capillary wall. It is very difficult, however, to observe a slight diffusion, and the experiments are complicated by the fact that the con- centration of the dye in the blood is rapidly reduced by absorption, especially in the liver. Somewhat bet- ter results can be obtained by perfusion of suitable organs (hind leg of the frog) with the dye solution, in which case the state of the capillaries can be regulated by the addition or omission of pituitrine. We have confirmed for a few dye substances the results of Schulemann, referred to in the preceding lecture, that the rates at which they penetrate the capillary wall are roughly proportional to the rates at which they diffuse through gelatine, but we have not, so far, succeeded in getting any substance which is really suitable for our purposes. The best results have been obtained with brilliant vital red and Chicago blue 6B. Both these dyes diffuse slowly through the normal capillary wall in the frog's tongue or web, but in all places where capillaries are somewhat dilated these become closely surrounded by a stained zone. EXCHANGE OF SUBSTANCES 239 The experiments are to be continued with other dye- stuffs and attempts are to be made to determine the size of particles which will be held back or let through the capillary wall in a normal and dilated state. A few experiments have been made with soluble starch, the particles of which are said to be about 5w in diameter. Starch is held back by normal capillaries. It passes out when they are strongly dilated and can be detected by the addition of a dilute iodine solution. The size of the pores in this latter case is, therefore, above 5^, while our experiments with India ink show that they are below 20CW*. Changes in capillary permeability not accompanied by corresponding changes in calibre. Though it appears, a priori, very probable that the permeability of the endothelial cells can become altered independently of any change in calibre, I have been able to find only very slight evidence of such altera- tions. Ebbeeke (1917, p. 65) and Parrisius (1921, p. 340) mention cases of Urticaria factitia in which the local cedema would develop after slight mechanical stimu- lation of the skin without any pronounced dilatation of the skin vessels, as judged by the hypersemic colour. It must be remembered, however, that some dilata- tion has undoubtedly taken place and that the cedema rapidly obscures the hyperemia and imparts a second- ary paleness to the surface. In the single case of Urti- caria factitia observed in my laboratory by Miss Car- rier the initial dilatation was very pronounced. L. Hess and H. Miiller (1915) have been able to pro- duce dropsy in rats, rabbits and dogs by subcutaneous injection of certain aromatic diamines (Toluylendia- mine). The dropsical fluid appears to be rich in pro- tein, indicating a considerable increase in capillary 240 CAPILLARIES permeability, while no indications of capillary dilata- tion are mentioned, which does not prove, of course, that it was absent. It has been repeatedly observed by a number of investigators that calcium salts, administered subcu- taneously or per os, may diminish exudation in inflam- mation (Chiari and Januschke, 1910), retard the ab- sorption of dyes from tissue spaces and their passage from the blood into the aqueous humour (Rosenow, 1916) and diminish exudation into the lungs after phosgene poisoning (Laqueur and Magnus, 1921). All these effects are commonly ascribed to an effect of the increased calcium content of the blood upon the capil- lary endothelium, which is supposed to be rendered less permeable. This is very probable, but the evidence so far brought forward can scarcely be accepted as conclusive. Quite recently some perfusion experiments have been made by R. Hamburger (1922) to study this problem, but, as far as I understand them, their results seem rather conflicting. It should be quite clear from the preceding account that the problems of capillary permeability are, in spite of the large amount of work spent upon them, very far from being satisfactorily solved. What is especially required is, I think, a quantitative formula- tion of the problems and quantitative, even if roughly approximate, determinations of the properties of the capillary wall. We want to know, as I have insisted, the absolute permeability of capillaries, the size of molecules which they will let through in different con- ditions, but we want just as badly to obtain some infor- mation about their filtering capacity. I have shown in a preceding lecture that artificial membranes of the same absolute permeability, which comes at least very close to that of the normal capil- lary wall, may show filtering capacities varying from EXCHANGE OF SUBSTANCES 241 less than 0.05 to 2.7. I will venture to guess that the filtering capacity of the capillary endothelium is a good deal higher than that of our best membrane, which has a thickness of about 200^ against less than 1/* for the endothelium, but it must be admitted that we have not a bit of real information on the point, and until we have that information we are unable to utilize properly what we know about osmotic pressure, capil- lary blood pressure and capillary surfaces. A very remarkable attempt to study quantitatively the exchange of water through the capillary walls has recently been made by Ellinger and Heymann (1921). These authors have perfused the hind leg of the frog with artificial solutions of known composition and de- termined the inflow as well as the outflow. The dif- ference between inflow and outflow gives them the exchange of water between the blood vessels and the tissues, which they have further controlled by weigh- ings of the perfused limbs. Their perfusion fluids have been partly Ringer solutions with various crystalloid additions, partly serum and mixtures of serum with crystalloid solutions. It would, no doubt, be possible to obtain by experiments of this type very valuable infor- mation, but, unfortunately, the authors have not paid any attention to the necessary condition of maintain- ing the capillaries in a normal or at least in a well- defined state of permeability. The perfusion experi- ments recorded in Lecture VII show that the artificial solutions rapidly produce an extreme dilatation of capillaries, Avhile the addition of mammalian serum may be able to keep them more or less normal for a time. Ellinger and Heymann find that the serum col- loids possess a power of retaining water in the circu- lation which is, they think, greatly in excess of their osmotic pressure, and this leads them to assume the existence of an "imbibition pressure of colloid soles" 242 CAPILLARIES different from and much greater than their osmotic pressure. They make no attempt to explain, however, why this "imbibition pressure" fails to appear when proteins or other colloids are studied in osmometers. While it must be admitted, I fear, that the actual results and conclusions of Ellinger and Heymann can- not be accepted, it is only fair to add that their method indicates a path along which progress can perhaps be made. LECTURE XI SOME APPLICATIONS OF THE PHYSIOLOGY OF CAPILLARIES TO COMPLEX PROCESSES IN HEALTH AND DISEASE IN the preceding lectures I have brought together and- arranged, to the best of my ability, the avail- able information on the qualities and reactions of capillaries. Nobody can realize, I think, more acutely than I do the fragmentary and altogether inadequate character of this information, but I believe, neverthe- less, that the attempt should be made to apply this information, such as it is, to a small number of physio- logical and pathological problems. Such an attempt will serve as a sort of test. If our information is, on the whole, sound — even if frag- mentary — it should not be in serious conflict with what is assumed to be known from other sources, and it ought to serve to throw some fresh light on certain aspects of some of the problems. Of the problems with which I propose to deal three are physiological: the negative pressure of the tho- racic cavity, the absorption of substances from the in- testine into the blood, and the interaction between the aqueous humour and the blood through the canal of Schlemm ; while four are pathological and comprise nothing less than the formidable problems of urticaria, inflammation, circulatory shock and oedema. The physi- ological problems have been selected partly on account of the peculiar difficulties which they present. 244 CAPILLARIES The "negative" pressure in the thoracic cavity. In the tissues generally and in such a cavity as the abdomen the pressure is everywhere and practically always very nearly atmospheric, and must be so, because the integuments give way very easily to any excess of pressure, whether positive or negative. The only significant exception to this rule is the pleural cavity, the walls of which have a sufficient power of resistance to allow the setting up of definite pressure differences. It is well known that the pressure in the pleural cavity is normally some 6 mm. of mercury or 80 mm. of water lower than the atmospheric. When a pneumothorax or a hydrothorax is established in one of the pleurae and the corresponding lung made to collapse, the air or fluid is gradually absorbed and the lung becomes expanded again in spite of its own tend- ency to elastic contraction. The absorption of the air of a pneumothorax into the blood is due to the fact that the total tension of dis- solved gases in the blood is lower by about 10 per cent than the atmospheric gas pressure, and I would sug- gest that the absorption of the fluid of a hydrothorax, as well as the maintenance of the normal negative pres- sure, is due to the excess of colloid osmotic pressure in the capillaries over that in the pleural cavity. There must be a slow but steady nitration of fluid from surrounding tissues into the pleural cavities, but this fluid is continually being taken up by the blood. The negative pressure cannot, of course, rise above the height necessary to expand the lungs and make them fill up every available space in the thoracic cavity. The absorption of dissolved substances from the small intestine into the blood. Through the intestinal epithelium of an average man there pass per day something like 400 grams IN HEALTH AND DISEASE 245 of sugar and 100 grams of amino acids. This quantity is dissolved in an amount of water which is not very exactly known but is usually estimated at about 5 liters, making up, therefore, a 10 per cent solution of sugars and amino acids. The solution contains also salts, and to judge from the salt content of the various digestive juices, which is, on an average, somewhat lower than that of the blood, and the amount of salts usually taken with the food, the absorbed solution probably possesses much the same salt concentration as the blood. In any case, its total osmotic pressure is greatly in excess of that of the blood. With the problem of the transport of this solution through the columnar epithelium we are not here con- cerned. We will deal only with the problem of its fate after it has entered the intestinal villi, and you will find that this problem, which is usually treated in a few lines in the text-books, presents very serious diffi- culties and will require renewed investigation. It is usually stated that the water and dissolved substances taken up in the intestinal villi are trans- ported practically exclusively through the blood, that there is only a slight increase in the flow of intestinal lymph after a meal, and that this lymph does not con- tain an increased amount of sugar or amino acids — or even, that the concentrations of these substances in the chyle are lower than in the portal blood. If these state- ments are correct they will, as far as I am able to see, amount to an assertion of secretory qualities on the part of the capillary endothelium in the villi, quali- ties which it is, as we have seen, unnecessary to assume for the ordinary capillaries of the body. When discussing the lymph flow from the intestine we were led to conclude that the endothelium of the intestinal capillaries is, like the capillary endothelium in the rest of the body, simply permeable to water and 246 CAPILLARIES crystalloids, and further — in this particular case — to a certain fraction of the blood proteins. If we try to pic- ture on this basis what will happen during absorption when a comparatively strong solution of sugars, amino acids and salts enters the pericapillary spaces from the epithelium, we cannot but conclude that the osmotically active substances of comparatively low diffusibility must draw water out from the capillaries. At the same time the diffusible substances will, of course, enter the blood stream through the capillary wall, and since the capillary surface is, as shown in the first lecture, very large and exceptionally permeable, while the blood in the capillaries is constantly and rapidly renewed, it is conceivable that a complete equilibrium may be reached, making the percentage concentration of the diffusible substances the same in the chyle as in a cor- responding sample of portal blood. When a complete equilibrium has been reached be- tween the crystalloids on both sides of the endothelium, it is conceivable further that part of the water drawn out from the blood by the initially concentrated solu- tion may be reabsorbed into the capillaries in conse- quence of the difference in colloid osmotic pressure between the blood and the chyle ; but it must be empha- sized that such osmotic reabsorption can take place only after complete equalization of the crystalloid con- centration, since a difference of less than 0.01 per cent sugar would be ample to counteract the possible excess of colloid osmotic pressure in the blood, and further, that the reabsorption stipulated must, in any case, re- main incomplete, in view of the fact that the normal relations between the hydrostatic and osmotic forces in the villi lead to a regular transudation of lymph. Reab- sorption of water is conceivable only up to the point where the chyle has reached the normal protein con- centration of the lymph during fasting. IN HEALTH AND DISEASE 247 We come, therefore, to the conclusion that, if the distribution of the absorbed water and crystalloids between the blood and the chyle is to be at all explica- ble as the result of simple osmotic processes, the con- centration of each substance in the chyle must never be lower than in the blood, while the quantity of chyle flowing from the intestine per unit time must be at least somewhat higher than the corresponding quan- tity of lymph flowing from the empty gut. This conclu- sion does not agree with the facts as usually stated, but there is, I think, some reason to believe that some of the observations made are incorrect. Hendrix and Sweet (1917) have analysed simulta- neous samples of blood and chyle during absorption of amino acids and glucose. Tbey found invariably that the amount of amino-nitrogen in the chyle rose con- siderably and became much higher than in simulta- neous samples of blood from the general circulation. The same was the case with the sugar, and in one — but only one — experiment they have compared sugar per- centages of the chyle with those of samples from the portal blood and found them to be practically identical. In some experiments, in which he injected large amounts of 0.3 per cent saline into the small intestine of fasting dogs, Heidenhain (1888) observed a consid- erable increase in the flow of intestinal lymph, the quantity of injected fluid transported through the lymph channels amounting on an average to about 1/10 of the quantity taken up by the blood, but in other experiments the flow of lymph was scarcely increased. Heidenhain appears to have thought that the intes- tinal villi might be so far contracted during absorption that the intercapillary spaces would be nearly obliter- ated and that almost all the fluid given off by the epi- thelial cells must enter the capillaries directly. This conception would account for the taking up of prac- 248 CAPILLARIES tically all the absorbed fluid into the blood, but it does not seem very probable. In his description of the blood vessels of the intes- tine, Mall (1887) mentions a very curious arrangement of small veins found in the submucosa. In injected specimens these veins appear to the naked eye as Fig. 47. Nests of small veins in intestinal submucosa of dog. After Mall. Highly magnified. "small coloured points" present in very large num- bers. When "highly magnified" they appear as shown in Fig. 47. Unfortunately the magnification is not stated. A large number of small veins branching off from the larger vessels coalesce to form a globoid or lenticular "rete." According to the description given by Mall of the lymph vessels of the intestine there can be no doubt that these veins are in close contact with IN HEALTH AND DISEASE 249 lymph vessels coining from the villi, and it is probable, therefore, that a further interchange of substances may take place, beyond that possible in the villi them- selves. It is impossible, however, to form any idea of the quantitative importance of such interchange until the number of these structures has been counted and their surface at least approximately estimated and compared with the capillary surface available in the villi. Everything considered, I think it most likely that the distribution of the substances absorbed from the intestine will turn out to be explicable on the basis of diffusion, but new experiments are urgently needed. The filtration of aqueous humour into the canal of Schlemm and the episcleral veins. It has been shown by Leber (1903) and recently con- firmed in a series of researches by Seidel (1921, 1922) that the aqueous humour is constantly filtered off from the anterior chamber of the eye through the canal of Schlemm, or ciliary plexus, into the episcleral veins. The canal of Schlemm is generally a circular plexus of small veins firmly imbedded in scleral tissue and separated from the anterior chamber by a layer of endothelium and the fine clefts known as the spaces of Fontana. According to the somewhat imperfect data given by Leber, the filtering surface of the canal can be esti- mated at something between 10 and 50 mm. 2 — say 30 mm. 2 The hydrostatic pressure in the eye is about 25 mm. Hg. and the quantity of aqueous humour filtering through is estimated at 6 mm. 3 per minute. When we figure out according to these data the filtering capacity of the canal as defined in a former lecture (VII), we find 60 cc, or about ten times as much as the filter- 250 CAPILLARIES ing capacity of the most permeable collodion mem- brane made. This agrees well with the fact that the protein-rich aqueous humour produced after punc- ture of the anterior chamber is easily filtered off and with the observation repeatedly made that India ink particles, the size of which is just submicroscopical, can also pass easily from the aqueous humour into the canal of Schlemm and the episcleral veins, but it raises a serious difficulty. I) I iSBH W. N, Ik gesi. / s ?^g C.r Fig. 48. Section through the border between cornea and sclera S. Cv. Canal of Schlemm. Lp. Spaces of Fontana. D. end of membrane of Descemet. After Leber. If the endothelium of the canal is permeable to pro- tein and even to India ink particles, a diffusion should take place in the opposite direction, and the aqueous humour ought normally to contain protein. The only explanation which I am able to suggest is that the endothelial cells themselves are impermeable, but that there are a number of extremely fine intercellular pas- sages through which there is a current of filtration which is sufficiently rapid to prevent the entrance of plasma from the blood. The canal of Schlemm in man is not a part of the regular venous system, in so far as no capillaries open IN HEALTH AND DISEASE 251 into it, but is a sort of diverticulum opening through numerous small passages into the ciliary veins. Seidel (1922) is of opinion that the whole of the canal does not normally contain blood but is filled with the fluid filtering off from the anterior chamber. In the following pathological conditions capillary reactions play an important and more or less conspicu- ous part : urticaria, inflammation, circulatory shock and oedema. I do not mean to convey any systematic idea by this enumeration, which only represents the order in which it is convenient from my point of view to discuss them. Urticaria, A large number of the most diverse stimuli may lead in the skin of man to local exudation of fluid which raises oh the spot affected a wheal or blister. The fluid has been examined in some cases by Torok and Vas (1900), who found a protein content of about 3 per cent. In a single case of a blister, due probably to the local action of heat (which had not, however, been noticed by the patient), we have measured a fractional osmotic pressure of the contained fluid amounting (in the collodion tube 4B) to 260 mm. water pressure. The increase in permeability of the capillary wall is, therefore, considerable. In the large majority of cases a dilatation of the capillaries has been demonstrated or can safely be assumed ; but, as referred to in the preceding lecture, this reaction may possibly be absent in some cases of urticaria. In other cases it may appear doubtful whether the dilatation observed is quantita- tively sufficient to account for the exudation. The mechanism of urticaria has been much discussed in dermatological literature and opinions appear to differ to an extraordinary degree, some authors hold- ing the complaint to be essentially of nervous origin, 03 IN HEALTH AND DISEASE 253 while others assume a primary alteration of the capil- lary wall and others again a change in the metabolic activity of tissue cells, the products of which will act both on vessels and on nerves. As far as I can see, the mechanism differs fundamentally in different cases. In the blisters of Herpes zoster (referred to in the preceding lecture) the mechanism is evidently purely nervous and it is quite possible that nervous reactions play some part also in many other cases of urticaria, though the reality of nervous intervention will often be very difficult to prove. Many capillary poisons produce urticarial eruptions when introduced into the system, and in some cases it has been shown (Biedl and Kraus, 1909) that we have to do with anaphylactic reactions by which vasodila- tory (capillario-dilatory) substances are produced within the organism. The difficulty in all these cases of general intoxication is not the capillary dilatation or exudation, but its special localization in certain groups of vessels in the skin of the patient. Urticaria can be produced locally by many different stimuli. Of these forms perhaps the most interesting is the Urticaria factitia produced in certain hypersen- sitive individuals by a mechanical stimulus which is so weak that it would, on normal persons, produce no more than a slight red reaction. Like the local red reac- tion the formation of a wheal after mechanical stimu- lation is independent of the nervous system, as shown by Ebbecke (1917, p. 25), who further emphasizes the fact that the capillaries, which are hypersensitive to mechanical stimuli, may show a normal reaction to poisons, like mustard, which possess an elective action on sensitive nerve endings. As the contractile cells of the cutaneous capillaries normally relax after mechanical stimuli of a certain strength and a local oedema can be produced every- 254 CAPILLARIES where on normal human skin by prolonged weak stimu- lation, I think we must accept the urticaria factitia as an excessive reaction of the Rouget cells to mechanical stimuli. W^^^^^^^R w^^ "- '13b ■w^^tmH^^^BB i". . ' JS : ' ■ 'V '■ ^'WrMki K '' " * ; f*i | p8fflfii 1 ' sr ; ' ' *sHr f '^B i'v. fi & '»■' '"*« 1 JB^ :■.?-. ll^B-V'lfl |; ';.- '' ; ''$>#g *■ ■■■i-lMsIB K ' %,, •< '■" |ffi| K b. ■ / JH K " isHliH *"'3IH m ■ i ' M i'4/-''.ipis%.4-|H 1 '"'^'ili brPwH^^H K- _,- -fi i^Hft&H jvM' ' ■ ''Jm • jK^anHgH Pigs. 50, 5]. Urticaria factitia. The figures were written with a blunt stick at five-minute intervals. In Fig. 50 the 6 had been written one minute before. The photograph Pig. 51 was taken five minutes later. Philippson (1900), Torok and Hari (1903) and Pilcher and Sollmann (1917) have studied the urti- carial reactions after local introduction into the skin of small quantities of various substances. Generally those substances which are found to be active in pro- ducing local exudation are either definite capillary poisons or belong to the heterogeneous group of "in- flammatory poisons, ' ' but with regard to some of them the action on the calibre of capillaries is unknown. This is the case especially regarding some drugs of the mor- IN HEALTH AND DISEASE 255 phine group, which are capable of raising wheals in concentrations of 1 to 0.1 per cent. Inflammation. In inflammation the circulatory phenomena are gen- erally very conspicuous and by a large school of pathologists they are regarded as the primary and essential symptoms to which all others can and should be referred. I am unable to agree with such a view. In my opinion, the vascular reactions in typical inflam- mation are, in the main, of a secondary character, though it must be admitted, of course, that they form, nevertheless, an element of prime importance in the complicated inflammatory processes. This opinion is mainly based upon the following considerations. In many cases of inflammation the vascular changes develop slowly and long after the application of tbe stimulus which is responsible for the inflammatory reaction. Typical examples are the inflammation pro- duced by strong chemical light and the inflammation clue to abrine, both of which have been briefly described in the fifth lecture. We must assume in these cases that neither the light nor the abrine acts primarily upon the capillaries or larger blood vessels, but that processes are set going in the tissue which ultimately lead to vascular reactions. In the light inflammation in the human skin a direct action of the light on the capillaries appears to be absent. The abrine causes, undoubtedly, some imme- diate reaction, but this is rather insignificant compared with the secondary effects. In many other cases of experimental inflammation the immediate vascular re- action is very pronounced, but even in a number of these it can very well be distinguished from the second- ary effects due to the processes set up in the tissue, and it is important to remember that quite as pro- 256 CAPILLARIES nounced immediate reactions can be produced by otber agencies which do not cause any subsequent inflam- mation. The emigration of leucocytes — one of the central, if not the central reaction in inflammation — is, no doubt, favoured by the dilatation of capillaries and perhaps also to some extent by the slowing of the current, but in the main, the process of emigration is independent of the vascular reaction and indicates the formation in the tissues of chemotactic substances. Recently Gessler (1921) has made some determina- tions of the metabolic oxygen consumption of small pieces of inflamed skin as compared with normal pieces from the same region. He reports increases in metabo- lism (at 37°) of 36 to 77 per cent. It seems natural to bring this metabolic activity into relation to the proc- esses of cell formation and tissue proliferation, which form such an important feature in inflammation. The attempt has been made by Ricker and Regendanz to explain all these reactions as simple consequences of the hyperemia, dilatation of capillaries and exudation of plasma, but the explanations offered cannot, I think, be regarded as convincing. In the main, I believe, we must agree with Ribbert (1909) in considering inflammation as a complex pro- tective and restorative reaction in which the vessels and the elements of the blood take their part along with the other tissue elements. I would emphasize the point, however, that though the reaction in its entirety is certainly beneficial to the organism, some of the partial reactions, at least of the vascular system, are often noxious. The development of complete stasis, for instance, in a large number of capillaries cannot but expose the tissues to grave dangers. It is, to say the least, difficult to see how the cedematous swelling of inflamed tissue can be beneficial. I would recall once IN HEALTH AND DISEASE 257 more the experience of Dreyer and Jansen that a light inflammation healed more quickly in that ear of a rab- bit the sympathetic nerve supply of which had been partly cut off, and the results of Bruce, Bardy and others that the diminution of the vascular reactions by ansesthetization may have a beneficial effect on the course of an experimental inflammation. It may very well be worth while to study the vascular reactions during inflammation with the object in view of getting them under control, of restraining them at the rjoints where they become harmful and of helping them on where they are beneficial. The calcium therapy inaugurated by Chiari and Januschke is, if rightly interpreted, to be considered as an attempt in this direction. Circulatory shock. The word shock has been used, and is often used, to denote very different pathological conditions, some of which have, perhaps, nothing whatever in common beyond the symptoms of collapse. Here we have to deal only with circulatory failure of the type described in the sixth lecture as resulting in certain animals from a large dose of histamine and there spoken of as hista- mine shock. You will remember that the analysis made by Dale and Laidlaw showed that the symptoms are due to a general capillary dilatation which takes up so much of the available blood that the return flow to the heart fails and the arterial blood pressure becomes gradually very low. A number of researches, among which should be es- pecially mentioned that splendid example of what can be achieved by hearty co-operation, the Reports on "Wound Shock and Haemorrhage by the Special Investi- gations Committee on Surgical Shock (1919), have shown that a condition essentiallv similar to histamine 258 CAPILLARIES shock can be brought about by severe traumatic inju- ries, and the work of the committee has proved con- clusively that traumatic shock is due primarily to the action of toxic substances formed in the injured tissue without the intervention of micro-organisms and distributed throughout the body by the circulating blood itself. A very essential feature in the aetiology of shock is the vicious circle which is set up by the poisoning of the capillaries. When the circulation begins to fail the blood supply to the tissues suffers and this in turn leads, by reason of oxygen lack or by reason of the diminished supply of tonic hormone, to still further dilatation. At more advanced stages the permeability of the capillary wall is so far increased that loss of plasma occurs, thus aggravating once more the failure of the circulation. Further important results of the committee 's inves- tigations are the demonstrations that anesthetization with ether and exposure to cold are apt to aggravate severely the conditions of shock. The effects of anaes- thetics have been discussed in Lecture VI and are easily understood, but the effects of cooling appear to be more complicated and a perfectly satisfactory explana- tion has not been found so far. The degrees of cold likely to occur should, in ordinary circumstances, give rise rather to capillary contraction, at least in the skin, but it is quite possible that the capillaries in the inter- nal organs react differently when the body tempera- ture becomes actually lowered. This point mil require further investigation. States of circulatory failure, similar in mechanism to traumatic shock in so far as they are mainly of toxemic origin, are, I believe, not at all infrequent. I am very imperfectly acquainted with the anaphylactic shock, the mechanism of which is perhaps very com- IN HEALTH AND DISEASE 259 plex, but which appears, according to Biecll and Kraus (1909), to have important features in common with traumatic shock; but I would draw attention to the symptoms in severe peritonitis and to those which often develop in patients after more or less extensive burns. In a recent paper, H. Olivecrona (1922) has shown very clearly that experimental peritonitis in rabbits, cats and dogs leads to typical circulatory shock show- ing the rapid pulse, the low blood pressure, the cya- notic pallor and the characteristic reduction in the volume of circulating blood. Evidence is presented to show that this form of shock is due to poisons, prob- ably disintegration products of proteins, liberated into the blood. It is well known that after burns, especially, perhaps, after more or less extensive scalding, a state of col- lapse, ending in the patient's death, may develop in one to two days in cases where the primary lesions are comparatively slight and do not involve any organ of vital importance. These cases, for which bacterial infec- tion could not possibly be made responsible, have, of course, aroused considerable interest and have been studied extensively, but so far as I am aware (I am not sufficiently familiar with clinical literature to feel sure), the only definite conclusion reached is that the symptoms are due to intoxication. When, however, these symptoms are studied in the light of the infor- mation obtained about circulatory shock, the similarity with the shock symptoms is, I think, very striking. 1 There is the typical fall of blood pressure, the rapid and very weak pulse, the concentration of the blood, i Cevario (1921) has made experiments on rats joined in pairs by- lateral cceliotomy. One of each pair underwent experimental scalding of a part of the skin. Both animals suffered to the same extent, which shows that toxic substances circulating in the blood must be responsible for the symptoms. 260 CAPILLARIES as shown by the increase in the corpuscle count. It has been noted (Helsted, 1905) that the intoxication from burns often leads to an increase in body temperature, a point in which there is apparently a difference between the toxic products from traumatized and from burnt tissue. According to a proposal originally made by Bayliss (1916), after consideration of the importance of main- taining the colloid osmotic pressure of the circulating fluids, circulatory shock is now very generally treated by intravenous injection either of blood or of an iso- tonic solution of gum acacia. injections of isotonic and hypertonic solutions of various crystalloids have been exhaustively tried, but found to bo useless, and it is easy to see that in cases where the permeability of the capillary wall is in- creased such solutions must leave the circulation very rapidly. When dissolved in normal saline, a solution of 6 per cent gum acacia has very nearly the same colloid os- motic pressure as human blood, 1 while its viscosity is slightly higher, and, when pure, it appears to be a per- fectly innocuous substance. 2 We have found the dif- fusibility of gum through collodion membranes to cor- respond very closely to the least diffusible fraction of the blood proteins, so that it will probably be held back in capillaries which have become permeable to the greater part of the normal plasma colloids. The commercial gum contains salts, and it has been stated by Bayliss (Med. Res. Com., 1919) that it is i According to experiments made in my laboratory the effective osmotic pressure of gum depends upon the salts of the solution. In pure water the osmotic pressure of gum is much higher. 2 Since writing the above I have learned of the recent observations showing bad effects from gum injections. A renewed investigation of the properties of gum and its reactions with human blood appears to bo necessary. IN HEALTH AND DISEASE 261 sufficient to dissolve the gum in 0.9 per cent sodium chloride, since the gum contains sufficient calcium and potassium salts. In a recent paper, Zondek (1921) has attempted to explain the effects of gum solutions as due to their calcium content, which he assumes to be very high. Determinations made by Mr. Rasmussen in my labo- ratory show, however, that the effective concentration of both calcium and potassium in a 6 per cent solution is only about double the concentration of the same sub- stances in Ringer, that is, the concentration of diffusible Ca and K ions correspond to 0.06 per cent CaCl. and 0.05 per cent KC1. The gum should, therefore, make an almost ideal provisional substitute for blood plasma. It should be clearly understood that in cases of shock the gam or blood injected can act only by reliev- ing the circulatory failure and breaking the vicious circle. If the capillaries go on dilating under the influ- ence of the poisons formed, the relief can be only tem- porary, and if the shock has reached such a stage that the capillaries have become permeable to gum, the injection will prove useless. It is, therefore, certainly worth attempting to obtain a curative effect by stimu- lating the tonus of the capillaries. Experiments to see if this can be done by means of pituitrine are planned in my laboratory. The formation and absorption of ozdema. The general problem as to the causes of oedema is one of extreme complexity, and in order to reach any valid conclusions we must establish certain distinctions between those types of oedema which are more or less dependent upon the state of capillaries and their pres- sure relations and those which are not. To these latter types all those cases of oedema should be referred in which the oedema is intracellular — is 262 CAPILLARIES brought about by a swelling of the tissue elements themselves, while the intercellular, porioapillary and lymph spaces do not contain more than the usual amount of fluid. A very instructive case of intracellular oedema has been described recently by F. Mendel (1!»22). It was characterized by a swelling of the cutis cells, which was present over the whole body but especially pro- nounced in the legs. It had brought about an increase in the weight of the patient from 77 to !). r > kg. There were no symptoms on the part of the kidneys or the heart. When the patient was deprived of sodium chloride in his food a rapid excretion of water began in a few days, and in a week the weight was reduced to normal. An experimental return to a diet containing salt brought back the (edematous symptoms. Mendel points out that, in ibis case, we have to do with an abnormal affinity of (be cutis cells for NaOl. They take up this substance from the blood so long as it is present in excess, or even in normal amount, in the blood; and, for obvious osmotic reasons, they must take up a corresponding amount of water. The state of the capillaries has nothing whatever to do with the proc- ess. They can neither hinder nor accelerate the absorp- tion of salt and wafer, both of which substances diffuse freely through the endothelium, whether it is normally or abnormally permeable. Without expressing any opinion regarding the rela- tive importance from a clinical point of view of intra- cellular and intercellular (edema, respectively, I would only emphasize the point that from the point of view of capillary physiology and pathology the cases of intracellular oedema should bo ruled out. They belong to an altogether different category. The cases of intercellular oedema with which we have to deal are themselves sufficiently complicated IN HEALTH AND DISEASE 263 and are controlled "by so many factors that great cau- tion must be exercised in drawing conclusions regard- ing them. The exudation and eventual reabsorption of fluid in the intercellular spaces will depend upon the capillary blood pressure, the colloid osmotic pressure of the blood, the permeability of the capillary wall, the efficiency of the lymph flow and the metabolic ac- tivities of the tissue cells. It cannot be surprising that the process resulting from the interaction of these factors is often difficult and sometimes impossible to disentangle. If we rule out to begin with the possible changes in capillary permeability and in cellular activity, we can put down as the simplest type of intercellular oedema the filtration oedema which will be produced in any tissue whenever the filtration pressure — the capillary blood pressure — exceeds the effective osmotic pressure of the blood. So long as the excess is slight the transu- date can probably be removed through the lymph chan- nels as rapidly as it is formed and no visible oedema results, but the rates at which removal can take place in different tissues have not been determined. A filtra- tion oedema can be brought about either by a sufficient decrease in the colloid osmotic pressure of the blood (a hydraemic plethora) or by an increase in the capillary pressure or by both factors acting in the same direc- tion. Examples of experimental filtration oedema have been mentioned in the preceding lectures. It is easily produced by the hydrostatic venous pressure in the human feet when these are allowed to hang down. In an interesting series of experiments Mende (1919) has measured the rate and degree of swelling in the arms of normal persons after the application of a rubber cuff in which a definite pressure was maintained. When the pressure was raised to 100 cm. water, an immediate 264 CAPILLARIES increase in volume of about 40 cc. was observed, due, of course, to the filling up of the vessels, especially the veins, with blood. During the next fifteen minutes a further increase of about 60 cc. took place, which can be shown to be due partly to dilatation of capil- laries and venules, partly to the formation of a filtra- tion oedema. This increase was continued, though more slowly, for a further period of nearly twenty minutes, when decompression took place after a total increase in volume of 120 cc. On decompression the volume went down about 70 cc. in a couple of minutes and there- after took thirty-five minutes to regain the original volume. Mende found that, by the application of 63 cm. water pressure for twenty -two hours, he could produce a definite oedema which would disappear in two to four hours after decompression, while with 50 cm. he ob- tained only passive hyperemia which would go back immediately on decompression. If the pressure applied could be taken to indicate the venous pressure reached in Mende 's experiments they would furnish an indication of the effective os- motic pressure in the blood of his patients. This is not the case, however, and all we can say is that the venous pressure was certainly considerably lower, and probably some 20 cm. lower, indicating an effective osmotic pressure of something between 30 and 40 cm. You may remember that in osmometers just imper- meable to protein we found an average osmotic pres- sure of human blood amounting to 46 cm., with indi- vidual variations from 40 to 51 cm. The lower pressure to be deduced from Mende 's experiments is probably due to an increase in capillary permeability brought about by the dilatation. The comparatively slow dis- appearance of the cedema after 63 cm. pressure points to the probability that the fluid contained some protein. A simple filtration cedema produced from capil- IN HEALTH AND DISEASE 265 laries which are normally impermeable to protein must be approximately protein free. 1 Such an oedema, of which the case of Ascites observed by Volhard and mentioned in a preceding lecture is a typical example, must be produced whenever the capillary pressure is raised above the colloid osmotic pressure of the blood and will be completely absorbed when the balance is turned in favour of the colloid osmotic pressure. 2 Theoretically and, to some extent, also practically, to be sharply distinguished from the cases of simple filtration cedema, we have another group of cases in which the filtration is due to an increased permeability of the capillary wall without any change in capillary pressure or the colloid osmotic pressure of the blood. In such cases the transudate cannot possibly be reab- sorbed into the blood. So long as the increased capil- lary permeability persists the transudation will go on, and when the capillaries return to normal they become impermeable to the protein in the transudate, the osmotic pressure of which will prevent also the reab- sorption of the water and crystalloids. For the absorption of oedema fluids containing pro- tein two different mechanisms are at present recog- nizable. One is the return to the blood by way of the lymphatics, and the experiments of Lewis (1921), re- ferred to in the preceding lecture, show conclusively i It must always be borne in mind that the cedema fluid is in constant contact both with capillaries and with tissue cells, which latter elements may have some influence upon its composition. 2 In a case of nephrosis studied by Hagedorn, Rasmussen and Eehberg after the above was written (October, 1922), there was a general cedema. The oedema fluid was protein free. The capillary blood pressure was somewhat increased (about 15 cm. water), but the colloid osmotic pres- sure of the patient's blood, which contained 5 per cent protein, was only about 100 mm. (as compared with the normal 450 mm.), fully explaining the cedema as due to simple filtration. The urine contained 2.8 per cent protein with a colloid osmotic pressure of 240 mm. In this case the smaller and osmotieally most active protein molecules were preferentially eliminated by the kidneys. 266 CAPILLARIES that protein is transported along this route. The other has been pointed out in a recent paper by Landsberg (1921), who observed that the protein in a pleuritic exudate was gradually broken down by enzymes, pro- duced, no doubt, in adjoining cells. The cleavage prod- ucts would be able to diffuse into the blood and a direct reabsorption of the oedema thus becomes possible. In the clinical literature numerous cases are re- corded, of which characteristic examples are given by Volhard (pp. 1163, 1255, 1274), showing rapid absorp- tion of oedema fluids after diuretic drugs — caffeine, theobromine, calomel and others. The mechanism of these cures is at present unknown, and I would venture to remark that it appears idle to speculate upon it until the facts of the cases have been more completely ascertained. Ever since the splendid work of Claude Bernard (1852) inaugurated the study of vasomotor mechan- isms, that branch of physiological science which deals with the regulation of the circulation has been a favourite subject of research and has exercised a pro- found influence upon physiological and pathological thought generally. Until quite recently the word vasomotor was used as synonymous with arteriomotor. We have realized now that beyond the well-known arteriomotor we have certain capillariomotor mechanisms and, though the systematic study of the capillaries and their reactions is still in its early infancy, it is quite safe to predict that it will show a vigorous growth both as a branch of pure science and with regard to its direct and indi- rect applications for the benefit of mankind, and be recognized ultimately as of equal importance with the study of the heart and the arterial system. It is my hope that these lectures may help in some IN HEALTH AND DISEASE 267 measure to promote the growth of this young branch of physiology and to arouse an active interest in a study which, by revealing more and more the beautiful cor- relation of activities in the organism, will further the purpose for which the Silliman Memorial Lectures were established. BIBLIOGRAPHY Abderhalden, E., und Gellhorn, E. (1921). Weitere Untersuchungen iiber die von einzelnen Organen hervorgebrachten Substanzen mit specifischer Wirkung. VI. Mitteilung. Pfl. Arch. 193, 47. Barbour, H. (1921). The heat regulating mechanism of the body. Physiol. Reviews, 1, 295. Barcropt, J. (1908). Mechanism of vasodilatation in cat's submaxillary gland. Proc. Physiol. Soc, Journ. Physiol. 36. Barcropt, J. (1914). The respiratory function of the blood. Cambridge Univ. Press. Barcroft, J., and Kato, T. (1915). Effects of functional activity in striated muscle and the submaxillary gland. Phil. Trans. Roy. Soc. London. Ser. B. 207, 149. Bardy, H. (1918). Ueber Hemmung inflammatorischer Symptome. Skand. Arch. Physiol. 32, 198. Bargee, S., and Dale, H. H. (1911). /J-iminazolylethylamine, a depressor constituent of intestinal mucosa. Journ. Physiol. 41, 499. Basler, A. (1914). Untersuchimgen iiber den Druck in den kleinsten Blutgefassen der menschlichen Haut. Pfl. Arch. 157, 345. Bayliss, W. M. (1901). On the origin from the spinal cord of the vaso- dilator fibres of the hind limb and on the nature of these fibres. Journ. Physiol. 26, 173. Bayliss, W. M. (1902). Further researches on antidromic nerve impulses. Journ. Physiol. 28, 220. Bayliss, W. M. (1916). Viscosity and intravenous injection of saline solutions. Journ. Physiol. 50, 23. Bayliss, W. M. (1919). The action of gum acacia on the circulation. Journ. Pharm. and Exper. Ther. 15, 29. Bayliss, W. M., and Starling, E. H. (1894). Observations on venous pressures and their relationship to capillary pressures. Journ. Physiol. 16, 159. Beale, L. P. (I860). On the distribution of nerves to the elementary fibres of striped muscle. Phil. Trans. Roy. Soc. London, 1860, 611. Bernard, Claude (1852). Sur les effects de la section de la portion cephalique du grand sympathique. C. R. Soc. Biol. 4, 168. Biedl, A. (1916). Innere Sekretion. Dritte Auflage Bd. II. Biedl und Kraus (1909). Experimentelle Studien iiber Anaphylaxie. Wien. Win. Woch. 1900, No. 11. Bier, A. (1897). Die Entstehung des Collateralkreislaufs, I. Arch. f. pathol. Anat. und Physiol. 147. BIBLIOGRAPHY 269 Bier, A. (1898). Die Entstehung des Collateralkreislaufs, II. Arch, f. pathol. Anat. und Physiol. IBS, 256, 444. Brasol, L. v. (1884). Wie entledigt sich das Blut von einem Ueberscbuss von Tranbenzucker. Arch. (Anat. u.) Physiol. 1884, 211. Bremer (1882). Die Nerven der Kapillaren. Arch. f. mikr. Anat. SI. Breslauer, F. (1919). Die Pathogenese der trophischen Gewebsschiiden naeh der Nervenverletzung. Deutsche Zeitschr. f. Chirurgie, 150, 50. Brodie, T. G. (1914). Croonian Lecture. Proc. Eoy. Soc. 87, 571. Browicz, T. (1899). Erniihrungswege in der Leberzelle nebst einem Resume iiber die Eesultate der seit 1897 in den Publikationen der Akademie veroffentlichten Untersuchungen des Verfassers iiber die Leberzelle. Anz. Akad. Wiss. Krakau, 1899, 359-365. Brown, W. (1915). On the preparation of collodion membranes of dif- ferential permeability. Biochem. Journ. 9, 591. Bruce, A. N. (1910). tfber die Beziehung der sensiblen Nervenendig- ungen zum Entziindungsvorgang. Arch. f. exper. Path. u. Pharm. 63, 424. Bruck, C. (1909). Experimentelle Beitrage zur Aetiologie und Pathoge- nese der Urticaria. Arch. f. Dermat. u. Syphilis, 96, 241. Bruns und Konig (1920). Ueber die Strbmung in den Blutkapillaren der mensehlichen Haut bei kalten und warmen Biidern und iiber die Beaktion in und nach kiihlen Wasser- und Kohlensaurebadern. Zeit- schr. f. Physiol, und diatet. Therapie, 24, 1. Burn, J. H. (1922). The relation of nerve supply and blood flow to sweating produced by pilocarpine. Journ. Physiol. 56, 232. Carrier, E. B. (1922). The reaction of the human skin capillaries to drugs and other stimuli. Amer. Journ. Physiol. Carrier, E. B., and Eehberg, P. B. (1922). Capillary and venous pres- sure in man. Skand. Arch. Physiol. Cevario, L. (1921). Sulla patogenesi della morte per ustione. Pathologica, IS. Chauveau, A., et Kaufmann, M. (1887). Experiences pour la deter- mination du coefficient de l'activite nutritive et respiratoire des muscles en repos et en travail. C. K. 104, 1126. Chiari, R., und Januschke (1910). Hemmung von Transsudat- und Exsudatbildung durch Kalziumsalze. Wien. klin. Woch. 23, No. 12. Clark, A. J. (1921). Absorption from the peritoneal cavity. Journ. Pharm. and exper. Therap. 16, 415. Cohnheim, J. (1867). Ueber Entziindung und Eiterung. Arch. f. path. Anat. 40, 1. Cohnheim, J. (1872). Untersuchungen iiber die embolische Processe. Berlin. Cohnstein und Zuntz (1888). Untersuchungen iiber den Pliissigkeitsaus- tausch zwischen Blut und Geweben. Pfl. Arch. 42, 303. Cotton, T. P., Slade, J. S., and Lewis, T. (1917). Observations upon dermatographism with special reference to the contractile power of the capillaries. Heart, 6, 227. 270 CAPILLARIES Dale, H. H., and Laidlaw, P. P. (1910). The physiological action of ff-iminazolylethylamine. Journ. Physiol. 41, 318. Dale, H. H., and Laidlaw, P. P. (1911). Further observations on the action of »-iminazolylethylamine. Journ. Physiol. 43, 182. Dale, H. II., and Laidlaw, P. P. (1919). Histamine shock. Journ. Physiol. 52, 355. Dale, H. H., and Richards, A. N. (1918). The vasodilator action of histamine and of some other substances. Journ. Physiol. 52, 110. Doi, Y. (1920). On the existence of antidromic fibres in the frog and their influence on the capillaries. Journ. Physiol. 54, 227. Dreyer, G., und Jansen, H. (1905). Ueber den Einfluss des Lichtes auf tierischen Geweben. Mitteilungen aus Pinsens Lichtinstitut, 9. Ebbecke, TJ. (1917). Die locale vasomotorische Beaktion (L.V.E.) der Haut und der inneren Organe. Pfl. Arch. 169, 1-81. Ellikger, A., und Heymann, P. (1921). Die treibende Krafte fur den Plussigkeitsstrom im Organismus. Arch. f. exper. Path. u. Pharm. 90, 336. Erben (1918). Ueber vasomotorische Stbrungen. Wien. klin. Woch. 1918, No. 2. Eugling, M. (1908). Untersuchungen iiber den peripheren Tonus der Blutgefasse. Pfl. Arch. 121, 275. Faust (1904). Ueber das Faulnisgift Sepsin. Arch. f. exper. Path. u. Pharm. 51, 248. Finsen, N. E. (1900). Neue Untersuchungen fiber die Einwirkung des Lichtes auf die Haut. Mitteilungen aus Finsen 's Lichtinstitut, 1. Fleisch, A. (1921). Die Wasserstoflionenkonzentration als peripher regu- latorisches Agens der Blutversorgung. Zeitschr. f. allg. Physiol. 19, 269. Fletcher, W. (1918). The vaso-constrictor fibres of the great auricular in the rabbit. Journ. Physiol. 22, 259. Gaarder, T. (1918). Ueber den Einfluss des Sauerstoffdruckes auf den Stoffwechsel. Biochem. Zeitschr. S9, 94. Gad-Andresen, K. S. (1921). Die Verteilung des Harnstoffes im Or- ganismus. Biochem. Zeitschr. 116, 266. Gasser, H. P., Erlanger and Meek (1919). Studies in secondary traumatic shock. IV. The blood volume changes and the effect of gum acacia on their development. Amer. Journ. Physiol. 50, 31-52. Gessler, II. (1921). Ueber die Gewebsatmung bei der Entzfindung. Arch. f. exper. Path. u. Pharm. 91, 366. Glaser, W. (1920). Innervation der Blutgefasse in L. B. Mfiller: Das vegetative Nervensystem. Berlin, Springer. Hagen, W. (1921). Die Schwankungen im Capillarkreislauf. Zeitschr. f. d. ges. exper. Med. 11, 364. Hagen, W. (1922). Periodische, konstitutionelle und pathologische Schwankungen im Verhalten der Blutcapillaren. Virchow's Arch. Haldanb, J. S. (1917). Organism and environment as illustrated by the physiology of breathing. Yale University Press. BIBLIOGRAPHY 271 Halpert, A. (1920). Ueber Milrokapillarbeobaehtungen bei einem Fall von Raynaudseher Krankheit. Zeitschr. f. d. ges. exper. Med. 11, 125-137. Hamburger, K. (1922). Ueber die Bedeutung der Kalium- und Calciumio- nen fiir das kiinstliehe Oedeni und fur die Gefassweite. Biochem. Zeitschr. 129, 153. Heidenhain, B. (1888). Beitrage zur Histologie und Fhysiologie der Bunndarmschleimhaut. Pfl. Arch. 43, Supplementheft. Heidenhain, B. (1891). Versuche und Fragen zur Lymphbildung. Pfl. Arch. 49, 209. Helsted, A. (1905). Bidrag til Lauren om Bodsaarsag ved Forbramding. Banish Bissertation. Henderson, Y. (1908). Acapnia and shock. I. Carbon dioxide as a factor in the regulation of the heart rate. (With the collaboration of M. M. Scarborough, P. P. Chillingworth and J. B. Coffey.) Amer. Journ. Physiol. 21, 120. Henderson, Y. (1909). Acapnia and shock. II. A principle underlying the normal variations in the volume of the blood stream and the deviation from this principle in shock. Amer. Journ. Phvsiol. 23, 345-374. Henderson, Y. (1910). Acapnia and shock. VII. Failure of the circula- tion. Amer. Journ. Physiol. 27, 152-170. Henderson, Y., and Harvey, S. C. (1918). Acapnia and shock. VIII. The veno-pressor mechanism. Amer. Journ. Physiol. 46, 33-53. IIendrix, B. M., and Sweet (1917). Amino-nitrogen and glucose in lymph and blood before and after the injection of nutrient solutions in the intestine. Journ. Biol. Chem. 32, 299. Herring, P. F., and Simpson, S. (1906). On the relation of the liver cells to the blood vessels and lymphatics. Proc. Boy. Soc. London, 78, 455. Hess, L., und Muller, H. (1915). Beitrage zur Pathologie des Oedems. Zeitschr. f. exper. Path. u. Therap. 17, 189. Heubner, W. (1907). Ueber Vergiftung der Blutkapillaren. Arch. f. exper. Path. u. Pharm. 56, 370. Heymann, P. (1921). Ueber die Wirkung kleinster Saure- und Alkali- niengen auf die Gefiisse. Arch. f. exper. Path. u. Pharm. 90, 27-76. Hill, Leonard (1920). Capillary pressure (I, II). Proc. Physiol. Soc, Journ. Physiol. 54. Hill, Leonard (1921). The pressure iu the small arteries, veins and capillaries of the bat 's wing. Proc. Physiol. Soc, Journ. Physiol. 54. Hofman, F. B. (1907). Histologische Pmtersuchungen iiber die Innerva- tion der glatten Muskulatur. Arch. f. mikr. Anat. 70, 361-413. Hooker, B. B. (1911). The effect of exercise upon the venous blood pres- sure. Amer. Journ. Physiol. 28, 235. Hooker, B. B. (1920). The functional activity of the capillaries and venules. Amer. Journ. Physiol. 54, 30. Houssay, B. A. (1918). La aceion fisiologica de los extractos hipofisia- rios. Buenos Aires. 272 CAPILLARIES Hoyee, H. (1877). Ueber unmittelbare Einmiindung kleinster Arterien in Gefassaste venosen Charakters. Arch. f. mikr. Anat. IS, 603, Tafel 38-39. Hiji'NER, G. (1897). Ueber die Bestimmung der Diffusionseoefficienten einiger Gase fiir Wasser. "Wiedemann's Annalen. N. F. 60, 134-168. Jacobj, W. (1920). Beobachtungen am peripheren Gefassapparat unter lokaler Beeinflussung desselben durch pharmakologische Agentien. Arch. f. exper. Path. u. Pharm. 86, 49. Jacobj, W. (1921). Pharmakologische Wirkungen am peripheren Gefass- apparat und ihre Beeinflussung auf Grund einer spezifischen Veran- derung der Permeabilitiit der Zellmembranen durch Hydroxylionen. Arch. f. exper. Path. u. Pharm. 88, 333. Jansen, H. (1906). Experimentelle Studier over Finsen-Behandlingens Virkemaade. Danish Dissertation. Krimke, A. (1884). Die Nerven der Kapillaren. Dissertation. Miinchen. Keogh, A. (1912). The regulation of the supply of blood to the right heart. Skand. Arch. Physiol. 27, 227. Kkogh, A. (1918). Vfevenes Porsyning med lit og Kapillferkredslobets Eegulering K. D. Vid. Selsk. Biol. Medd. 1, No. 6. Krogh, A. (1919). The rate of diffusion of gases through animal tis- sues, with some remarks on the coefficient of invasion. Journ. Physiol. 52, 391. Krogh, A. (1919). The number and distribution of capillaries in muscles with calculations of the oxygen pressure head necessary for supplying the tissue. Journ. Physiol. 52, 405. Krogh, A. (1919). The supply of oxygen to the tissues and the regula- tion of the capillary circulation. Journ. Physiol. 52, 457. Krogh, A. (1920). Studies on the physiology of capillaries. I. The reac- tion to stimuli and the innervation of the blood vessels in the tongue of the frog. Journ. Physiol. 53, 399. Krogh, A. (1921). Fortsatte Studier over Kapilherernes Fysiologi. (Danish). K. D. Vid. Selsk. Biol. Medd. S, No. 3. Krogh, A. (1921). Studies on the physiology of capillaries. II. The reactions to local stimuli of the blood vessels in the skin and web of the frog. Journ. Physiol. 55, 412. Krogh, A., and Harrop, G. A. (1921). On the substance responsible for capillary tonus. Proc. Physiol. Soc, Journ. Physiol. 5-1. Krogh, A., and Harrop, G. A. (1921). Some observations on stasis and oedema. Proc. Physiol. Soc, Journ. Physiol. 54. Krogh, A., Harrop, G. A., and Kehberg, P. B. (1922). Studies on the physiology of capillaries. III. The innervation of the blood vessels in the hind legs of the. frog. Journ. Physiol. 56, 179. Krogh, M. (1915). The diffusion of gases through the lungs of man. Journ. Physiol. 40, 271. Kupfeer, C. v. (1876). Ueber Sternzellen in der Leber. Arch. f. mikr. Anat. 12, 352-358. BIBLIOGRAPHY 273 Kupffer, C. V. (1899). Ueljer die sogenannten Sternzellen der Saugetier- leber. Arch. f. mikr. Anat. 54, 254-288. Lanb-sberg, M. (1921). Studien iiber den Chemismus der Resorption d. pleuritisehen Exsudate Wien. Arch. f. inn. Med. 3, 467. Langley, J. N. (1921). The autonomic nervous system. Part I. Heffer and Sons, Cambridge. Laqueur, E. (1919). Ueber kiinstlich erzeugtes (osmotisches) Lungenbdem und iiber Resorption in der Lunge, Miinchen. med. Woch. 1919, 1721. Laqueur, E., und Magnus, R. (1921). Ueber Kampfgasvergiftungen. V. Experinientelle und theoietisehe Gruudlagen zur Therapie der Phos- generkrankung. Zeitschr. f. d. ges. exper. Med. 13, 200. Leathes, J. B. (1895). Some experiments on the exchange of fluid be- tween the blood and the tissues. Journ. Physiol. 19, 1. Leber, Th. (1903). Lie Cirkulations- und Ernahungsverhaltnisse des Auges. Leipzig. Leriche, R., et Policard, A. (1920). Etat des capillaires pendant l'ex- citation du sympathique periarteriel chez 1 'homme. C. R. Soc. Biol. 20, Nov., 1920. Lewis, J. H. (1921). The rate and route of absorption of subcutaneously injected serum in relation to the occurrence of sudden death after injection of antitoxic horse serum. Journ. Amer. Med. Assoc. 76, 1342. Lister (1858). The early stages of inflammation. Phil. Trans. 148, 645. Lombard, W. P. (1912). The blood pressure in the arterioles, capillaries and small veins of the human skin. Amer. Journ. Physiol. 29, 335. Loven, Cm:. (1S66). Ueber die Erweiterung von Arterien in Polge einer Neivenerregung. Verhandl. d. sachss, Ges. d. Wis., M. ph. CI. 1S66, 85. Lungvyitz (1910). Der Fuss des Pferdes, p. 157. Magnus, R. (1899). Ueber die Entstehung der Ilautbdeme bei experi- menteller hydramischer Plethora. Arch. f. exper. Path. u. Pharm. 42, 250. Mall, J. P. (18S7). Die Blut und Lymphwege im Diinndarm des Ilundes. Abh. d. Sachs. Ges. d. \Yiss. M." ph. CI. 14, 153. Mayer, S. (1902). Die Muskularisierung der capillaren Blutgefasse. Anat. Anz. 21, 442. Medical Research Committee (1919). Wound-Shock and Hasmorrhage. Spec. Rep. Series, No. 25. Meek, W. J., and Eyster, J. A. E. (1921). Reactions to liEemorrhage. Amer. Journ. Physiol. 56, 1-15. Mende (1919). Ueber Hyperamie und Oedem bei der Hemmung des Riickflusses des venbsen Blutes durch die Staubinde. Deutsche Zeit- schr. f. C'hirurgie, 150, 379. Mendel (1922). Oedema cutis proprium. Klinische Wochenschr. 1, 1502. Miller and Carlton (1895). The relation of the cortex of the cat's kid- ney to the volume of the kidney and an estimation of the number of glomeruli. Trans. Wisconsin Aead. Science, 10, 525. 274 CAPILLARIES Muller, E. (1920). Beitrage zur Kenntnis des antonomen Nervensys- tems. I. Ueber die Entwickelung des Syrnpathicus und des Vagus bei den Selachiern. Areh. f. mikr. Anat. 94, 208-247. Muller, L. R. (1921). Studien iiber den Dermographismus. Deutsche Zeitschr. f. Nervenheilkunde, 47-48, 413. Natus, M. (1910). Beitrage zur Lehre von der Stase. Arch. f. path. Anat. u. Physiol. 199, 1. Olivecrona, H. (1922). An experimental study of the circulatory failure in peritonitis. Acta Chir. Scand. 54. Parrisius, W. (1921). Kapillarstudien bei Vasoneurosen. Deutsche Zeitschr. f. Nervenheilkunde, 72, 310. Philippson, M. M. (1903). Ueber das fliichtige Reizodem der Haut. Arch. f. Dennat. u. Syph. 65, 387. Piloher, .1. D., and Sollman, J. (1917). Endermic reactions. I. Journ. Pharni. and exper. Ther. 9, 309. Pohle, E. (1920). Der Einfluss des Nervensystems auf die Osmoregula- tion der Amphibien. Pfl. Arch. 182, 215. Recklinghausen, H. v. (1906). Unblutige Blutdruckmessung. Arch. f. exper. Path. u. Pharm. 55, 376. Rehberg, P. B., and Carrier, E. B. (1922). Concerning the reaction of the human skin capillaries to venous blood. Skand. Arch. Physiol. Ribbert, II. (1909). Wesen der Krankheiteu. (Quoted from Handworter- bueh d. Naturwissenschaften VII, 543.) Rich, A. R. (1921). Condition of the capillaries in histamine shock. Journ. exper. Med. 33, 287. Richards, A. N. (1922). Kidney function. Amer. Journ. Med. Sci. 163, 1. RlCKER und Natus (1910). Versuch einer Theorie der chronischen Entzimdimgen. Virchow 's Arch. 202. Richer, S., und Regendanz, P. (1921). Beitrage zur Kenntniss der ortlichen Kreislauf sstbrungen. Virchow 's Arch. 231, 1. Rosenotv, G. (1916). Der Einfluss parenteraler Calciumzufuhr auf die Durchlassigkeit der Gefassvfand. Zeitschr. f. d. ges. exp. Med. 4, 427. Rothlin, E. (1920). Exper. Studien iiber allg. und. spez. Eigenschaften iiberleb. Gefasse unter Amvendung der chem. Reizmethode. Biochem. Zeitschr. Ill, 219. Rouget, Ch. (1873). Memoire sur le developpement de la tunique con- tractile des vaisseaux. C. R. 79, 559. Rouget, Ch. (1879). Sur la contractilite des capillaires sanguins. C. R. 88, 916. Roy, Ch., and Graham Brown (1879). The blood pressure and its varia- tions in the arterioles, capillaries and veins. Journ. Physiol. 2, 323. Schaeer, E. (1902). On the existence within the liver cells of channels which can be directly injected from the blood vessels. Proc. Roy. Soc. Edinburgh, 24, 65. Schulemann, W. (1917). Vitale Eiirbungen mit sauren Farbstoffen. Biochem. Zeitschr. 80, J. BIBLIOGRAPHY 275 Schur, II. (1920). Ilaut und Hauteapillaren im rnikro-episkopischen Bilde. Zeitschr. f. angew. Anat. und Konstitutionslehre, 5, 193. Scott, F. II. (1910). The mechanism of fluid absorption from tissue spaces. Journ. Physiol. 50, 157. Seidel, E. (1921). Exp. Untersuch. iiber intraokulare Saftstrornung. IX. Ueber den Abfluss von Kammerwasser. Arch. f. Ophthal. 104, 357. Seidel, E. (1922). Weitere exp. Untersuch liber die Quelle und der Yer- lauf der intraokularen Saftstrornung. XVI. Mitteilung. Arch. f. Ophthal. 108, 285. Sherrington, C. S. (1920). The integrative action of the nervous sys- tem. 6'th printing, Yale University Press. Smith, II. P., Arnold, II. R., and Whipple, S. H. (1921). Blood volume studies. VII. Amer. Journ. Physiol. .56', 330. Spalteholz, W. (1888). Pie Vertheilung der Blutgefiisse im Muskel. Abhd. d. sachs. Ges. d. Wiss. M.-Ph. CI. 14, 509. Spalteholz, W. (1893). Pie Verteilung der Blutgefiisse in der Ilaut. Arch. f. Anat. (u. Physiol.) 1893, 1. Starling, E. If. (1894). The influence of mechanical factors on lymph production. Journ. Physiol. 16, 224. Starling, E. II. (1896). On the absorption of fluid from the connective tissue spaces. Journ. Physiol. /.'», 312-327. Steinach, E., und Kahn, R. II. (1903). Echte Contractilitiit und moto- rische Innervation der Blutcapillaren. Pfl. Arch. 07, 105. Stewart, G. X. (1911). Studies on the circulation in man. I. Heart, 3, 33. Strigker, S. (1805). Studien liber Bau u. Leben der capillaren Blutge- fasse. Sitzungsber. d. Wiener Akad. d. Wiss. M.-N. Kb 5S, 2 Abt. .379. Strigker, S. (1879). Untersuchungen iiber die Contractilitiit der Capil- laren. Sitzungsber. d. Wiener Akad. d. Wiss. M.-N. Kb 71, 3 Abt. 313. Thaller und Praga (1917). Pie Bewegungen der Hautkapillaren. Wien. klin. Woch. 30. Tigerstedt, R. (1920). Pie Stromung des Blutes in die Kapillaren und Venen. Ergebn. d. Physiol. 18, 1. Torok, L., und Hari, P. (1903). Experimentelle Untersuchungen iiber die Pathogenese der Urticaria. Arch. f. Derm, und Syph. 65, 21. Torok und Vas (1900). Ueber den Eiweissgehalt des Inhaltes von Haut- blasen. Festschrift iur Ehren von Moriz Kaposi (quoted from Ebbecke). Verzar, P. (1912). The influence of lack of oxygen on tissue respiration. Journ. Physiol. 45, 39. Vimtrup, Bj. (1922). Beitrage zur Anatom'ie der Kapillaren. I. Ueber contractile Elemente in der Gefasswand der Blutkapiilaren. Zeitschr. f. d. ges. Anat. 65, 150. Volhard, A. (1917). Die doppelseitigen hiimatogenen Xierenerkrankun- gen (Brightsehe Krankheit). Ilandb. d. inneren Med. Bd. IIP, 1148-1722. 276 CAPILLARIES Walpole, G. S. (1915). Notes on collodion membranes for ultrafiltration and pressure dialysis. Biochem. Jouin. .9, 284. Weiss, E. (1916). Beobachtung und mikrophot, Darstellung der Haut- capillaren am lebenden Mensehen. Deutsche Arch. f. klin. Med. 119. Wernoe, Th. B. (1920). Aestesioscopia abdominalis (Danish) Ugeskrift for La?ger. 82, 1415. White and Eelakger (1920). Blood analysis following acacia-glucose injections. Amer. Journ. Physiol. 54, 1-29. With, C. (1920). Studier over Lysets Virkning ved Vitiligo (Danish) Hospitalstid. 1920. Woodlaxd, W. (1911). On the structure and function of the gas glands and retia mirabilia ... of some teleostean fishes. Proc. Zool. Soc. London, 1911, 183-248. Worm-Muller, J. (1873). Die Abhangigkeit des arteriellen Blutdruckes von der Blutmenge. Verhandl. d. sachs. Ges. d. Wiss. M.-Ph. Kl. 1S7S, 172. Zak, E. (1920). Ueber viszero-vasomotorisehe Zonen. Wiener klin. Woch. 1920, No. 25. Zak, E. (1921). Tiber den Gefiisskrampf bei intermittierenden Ilinken und iiber gewisse kapillomotorische Erscheinungen Wien. Arch. f. inn. Med. 2, 405. Zondek, H. (1921). Die Bedeutung kolloidaler Nahrlosungen fiir die Funktion des normalen, erschbpjften und vergifteten Herzens. Bio- chem. Zeitschr. 116, 246. PRINTED IN THE UNITED STATES OF AMERICA