UC-NRLF EXCHANGE , BIOLOGY LIBRARY 6 The Occurrence of Aluminium, and its Absorp tion from Food, in Dogs DISSERTATION Submitted in Partial Fulfillment of the Re- quirements for the Degree of Doctor of Philosophy in the Faculty of Pure Science of fcolumbia University By ARNOLD K. BALLS Highland Democrat Print Peekskill, N.Y. May, 1920 The Occurrence of Aluminium, and its Absorp tion from Food, in Dogs DISSERTATION Submitted in Partial Fulfillment of the Re- quirements for the Degree of Doctor of Philosophy in the Faculty of Pure Science of Columbia University By ARNOLD K. BALLS Highland Democrat Print Peekskill, N.Y. May, 1920 BIOLOGY R G OUTLINE I. Introduction. II. Method. III. Is the aluminium of aluminized bread soluble in the digestive juices ? IV. Does aluminium occur in the tissues of normal dogs? V. Is aluminium absorbed from food containing it in po- tentially soluble form? VI. Is aluminium absorbed by dogs from aluminized bread? I. Introduction This paper describes an investigation of the general subject of the occurrence of aluminium in the animal body, both under normal conditions, and on diets containing aluminium. The problem is of considerable theoretical interest. The occurrence of aluminium in the bodies of normal animals has never been specially investigated, but large amounts are certainly not present, or they would have been easily recognized ere this. This condition is in marked contrast to the abundance of aluminium in the plant and mineral worlds. That aluminium, like silicon, exists in the animal body in small amounts, could even be assumed on the ground of accidental contamination, unless some specific eliminative selection takes places, which would point biologically to an incompatibility between the metal and the organism. Ab- sorption, and the retention of aluminium, if either or both occur, bear directly upon the possibility of substituting one element for another in the body. Information concerning such replacements consequently means an increase in our knowledge of the functions of the substances involved. From the practical standpoint, the wide use of alum baking powders, especially in the South, makes it desirable that the effect of this aluminium should be determined. The problem was first attacked by Papillon 1 in 1870, who, after feeding aluminium phosphate to a young rat, found that the bones contained 6.95% of alumina. He concludes that his "researches show it possible to substitute a certain quantity of strontium, of magnesium, or of aluminium for the calcium normally occurring in the bones. H. Weiske 2 was unable to confirm these results with aluminium, although he verified them with regard to the other elements used. He recognizes the fact that such substitutions may not occur to a marked extent in an animal kept on a normal diet; but that by iPapillon: Comp. rend, de 1'Acad. de sciences (Paris), Ixxi, p. 372, (1870). 2 H. Weiske: Zeit. fur Biologie. vlii, 239, (1872), ibid, xxxi, 421, (1895) V 1 depriving the animal of a sufficient amount of the element to be replaced, (in this case calcium) substitution may be effected. J. Konig 3 first practically applied this principle of Weiske's, using food deficient tin calcium He was, however, unable to find any large amounts of aluminium, but states that he made no at- tempt to look for small amounts, and concedes the possibility of their being present. It was in this unsatisfactory condition that the problem was first taken up at this laboratory. In 1906 House and Gies 4 showed that a concentration of alu- minium greater than 1-65,536 moles per liter markedly inhibited the growth of seedlings. It was later found by Steel 5 , working with Gies, that in dogs, aluminium passed from the alimentary tract into the blood, when they were given food mixed with alum, and that when aluminium chloride was injected intravenously this pro- cess was reversed, aluminium passing into the intestines from the blood. Kahn 8 , also working in this laboratory, showed that aluminium is absorbed by dogs from food prepared with alum baking powder. On the other hand, the Referee Board of Consulting Scientific Experts of the United States Department of Agriculture 7 were unable to find aluminium in the blood of four men who "took one gram of aluminium a day for several days." The form in which the aluminium was taken, however, is not mentioned. More recently, Leary and She'ib 8 , using dogs and white rats, which were fed with aluminium hydroxide, identified aluminium in the livers of all the animals used. In our experiments we have used dogs, both because of their convenience, and since, of the animals available, their metabolism is nearest to that of human beings. 3 J. Konig-: Zeit. fur Biologie.. x, 69, (1874). 4 House and Gies: Amer. Jour. Physiol. xv, (1906). Proc. p. xix. 5 Steel: Amer. Jour. Physiol. xxviii, 94, (1911). 6 Kahn: Biochem. Bull, i, 235, (1911). 7 Bull. 103, U. S. Dept. of Agriculture. 8 Leary and Sheib: Jour. Amer. Chem. Soc. xxxix, 1066, (1917). II. Method Tor description of the analytical methods employed, see Balls: Biochemical Bull V, pp. 195-202. (1916) III. Is the Aluminium of Aluminized Bread Soluble in the Digestive Juices? As a preliminary to animal investigation, it was thought de- sirable to find out if the aluminium of aluminized bread was soluble in artificial mixtures, made to represent, as nearly as possible, those juices with which it would come in contact during the process of digestion. To this end a sample of bread dough was made 1 and divided into three parts. The first of these was baked very thoroughly, the second could be called without exaggeration "well baked," while the third was poorly baked, the interior being quite doughy, and closely resembling ordinary bread. The difference in degree of baking is shown by the appended summary of data for the mois- ture content, determined on one-gram samples by drying to con- stant weight in a vacuum oven at 75. C. Table 1. Moisture content of bread samples. (% H 2 O) Very well baked. Well baked. Poorly baked. 5.84% 7.02% 16.14% 5.84% 7.16% 16.04% These different grades of bread were mixed with the extrac- tion media used, in the proportion of 5 grams per liter. The mixtures were kept in an incubator at body temperature for twen- ty-four hours, and shaken ten times during that period. The solid material remaining was then filtered out, and the aluminium in the solution determined according to the following procedure : Approximately of this composition: Flour, 550. grams. Sodium Chloride, 8. grams. Sugar, 5. grams. Baking Powder, 16. grams Water, 400. grams. The baking powder used was "Parrot and Monkey Brand," an alum baking powder purchased in the open market. It contained, according to the statement on the label, 31.% anhydrous soduim alum. This was found to be correct. 3 425707 The solution, first partially evaporated in a porcelain dish, was taken to dryness in a platinum one. The material was then gently ignited in an electric muffle until completely ashed. This ash was fused with approximately six times its weight of a mixture of potassium and sodium carbonates, and the melt dissolved in hydro- chloric acid solution. Repeated evaporations with HC1 were made to dehydrate the silica. The residue was finally moistened with cold concentrated acid; after standing some minutes diluted with water, warmed, and the silica, always small in amount, filtered off. The aluminium present in the acid solution was determined by the procedure of Schmidt and Hoagland 2 . This method depends in principle upon the precipitation of aluminium as phosphate from A solution containing phosphate, whose acidity has been reduced b> adding, first, ammonium thiosulphate, then ammonium acetate and acetic acid. Reprecipitation under similar conditions is neces- sary, otherwise the precipitate will be contaminated by iron and calcium. The precipitate is first heated gently to drive off the accompanying sulphur, then igniated and weighed as A1PO 4 . This method gives low results, but its advantages lie in afford- ing a direct determination of aluminium in the presence of the iron and phosphate always found in biological material. The method, as applied to small amounts of aluminium occurring in biological substances, was found by Howe 3 and others in this laboratory to give satisfactory results. Mellor* claims that water decomposes aluminium phosphate slightly into a soluble aoid phosphate and an insoluble basic phosphate, thus giving low results. This is probably merely an empirical way of describing a hydrolytic process. By washing the precipitates with dilute ammonium phosphate solution instead of with hot water as directed by Schmidt and Hoagland, we found, as one would naturally expect from the common ion effect, that less material was lost. A final washing with ammonium nitrate solution, however, is advisable, as the presence of much excess of ammonium phosphate in the precipitate requires very high and prolonged ignition to remove. The bread samples were extracted in the following four ways : I. With distilled water. TI. With a 0.2% solution of hydrochloric acid. (Represent- 2 Schmidt and Hoag-land: Jour. Biol. Chem. xi, 387, (1912). See also: Biochem. Bull, v, (1916). 3 Howe: Biochem. Bull, v, 158, (1916). *Mellor: Treatise in Inorgunic Analysis, p. 608, (1913). 4 ing the acid concentration found in the gastric juice.) III. With a 0.2% solution of hydrochloric acid, to which was added 0.1% of a commercial preparation of pepsin. This mixture is the so-called ''artificial gastric juice," and its action will represent fairly completely the digestive pro- cesses in the stomach. IV. After allowing the bread to digest for 24 hours in arti- ficial gastric juice, sufficient sodium carbonate was added to furnish an excess of 5 grams per liter 5 , 0.2% of a commercial preparation of trypsin was then added, and the mixture allowed to incubate for another 24 hours. Each determination was made in duplicate with the results shown in the following table. Table II. Data showing the amounts of aluminium extracted from "alu- minized bread" by different solvents, and the relation of these amounts to the total quantity of aluminium in the bread. Under Column I. is given a description of the sample of bread used, and under Column II. is a description of the extraction medium. The w r eight A1PO 4 in milligrams, representing the aluminium extracted from 2.5 grams of bread by 500. c.c. of the extraction medium is given under Column III., while Column IV. gives the mean of the duplicate determinations, expressed in terms of the percentage of the total aluminium present, extracted by the extraction me- Al. extracted dium. (= x 100.) Total Al I. II. III. IV. Very well Distilled water 0.2 mg. baked bread. 0.2 2.0% 0.2% HC1 9.9 10.5 86.0% 0.2% HC1 + 0.1% pepsin 11.6 10.8 93.0% 0.2% HC1 + 0.1% pepsin 1.1 followed by 0.5% Na 2 CO 3 + 0.2% trypsin 1.4 11.0% 0.2% HC1. Al determined 9.1 in dialysate, instead of in filtrate. 9.8 79.0% ^Representing the usual alkalinity of the duodenal juices. 5 0.2% HCI + 0.1% pepsin 10.2 Al determined in dialysate. 9.6 83.0% Total aluminium in bread 12.0 sample. (2.5 grams) 12.0 (100.0%) Well baked Distilled water, 0.9 mg- bread 0.8 5.0% 0.2% HCI 10.9 10.5 93.0% 0,2% HCI 4- 0.1% pepsin 11.4 11.5 100.0% 0.2% HCI + 0.1% pepsin 0.8' followed by 0.5% Na 2 CO 3 +0.2% trypsin, 0.7 7.0% Total aluminium in bread 11.5 sample. (2.5 grams) 11.3 (100.0%) Poorly baked Distilled water 1.0 mg- bread 0,8 8.0% 0.2% HCI. 9.4 8.8 85.0% 0.2% HCI 4- 0.1% pepsin 10.0 10.5 96,0% 0,2% HCI 4- 0.1% pepsin 1.0 followed by 0.5% Na 2 CO 3 4- 0.2% trypsin. 1.2 10.0% .Total aluminium in bread 10.5 sample. (2,5 grams) 10.9 (100.0%) Pooly baked 0.2% HCI + 0,1% pepsin 10.2 mg. bread + an 10.4 96.0% equal weight 0.2% HCI 4-0.1% pepsin 7.4 of lean meat, Al determined in dialysate. 7.4 69.0% 0.2% HCI 4- 0.1% pepsin 1.0 followed by 0.5% Na 2 CO 3 4- 0.2% trypsin. 0.7 8.0% Inspection of this table will show that but little aluminium of the bread is extracted by water, the solubility increasing, how- ever, with decreasing thoroughness of the baking. In the alkali- trypsin mixture the amount dissolved is also very small, although larger than with water. The same relationship also exists between 6 solubility and degree of baking. With hydrochloric acid the ex- traction is about two-thirds of the total, and the artificial gastric juice dissolves out practically all the aluminium present, the amount dissolved out being, in both of these cases, independent of the degree of baking which the material had undergone. With the acidic extraction media the aluminium is dissolved in great part, but not entirely, to form a true solution. This can be seen by comparing, in Table II, the amount of aluminium that passes through' the ordinary filter paper used in filtering off the insoluble bread residue, with that dialysed from a similar mixture through a collodion membrane. This dialysis increases the manip- ulative error and was therefore not applied to the trypsin-alkali and the water extracts, where the quantities of available alumi- nium were much smaller. By adding to the bread, in repetitions of the experiment described, an equal weight of lean meat, finely divided, the presence of a quantity of meta-protein, proposes, and peptones from digested meat was found not to affect the results obtained from the bread alone, except to increase the amount of aluminium not in true solution and therefore not dialysable, but still capable of passing through a filter paper. This difference is more strikingly indicated in subsequent results on aluminium solu- tions containing large quantities of peptone, where fully half of aluminium in "solution" could not be dialysed out. In those cases where the solubility is small, the percentages dissolved are, of course, useful only for comparison. The weights of material dissolved probably represent the values, under the con- ditions of the experiment, for a saturated solution of the alumi- nium compound in question (whatever that may be). The total amount dissolved is therefore merely a function of the dilution as long as any solid remains. A series of experiments similar to those above, were also made on a solution of- aluminium phosphate, since the phosphate is the form in which previously ionic aluminium would be found after entering the intestine, and it was accordingly decided to use this substance in feeding some of our animals. One gram of freshly precipitated, thoroughly washed, moist aluminium phosphate 8 , was dissolved in a liter of 0.2% HC1. On the addition of 20. grams of Witte peptone no precipitation oc- curred, indicating that proteoses and peptones would not interfere with the solubility of aluminium in the stomach. The aluminium 6 The material contained 84.03% H 2 O. An amount equal to one gram of the dried material was employed. 7 was only partially in true solution, as shown by the fact that only about half of it was dialysable. The aluminium was again pre- cipitated when an excess of alkali was added equal to the percentage of alkali found in the intestines. The addition of the alkali as above, but in the presence of bile salts (0.2% sodium taurocholate was used) was productive of no difference in results. The fol- lowing table summarizes these findings. Table III. In each case the solution used consisted of one gram of A1PO, dissolved in a liter of dilute HC1, to which the peptone was added as described above. Under Column I is a description of the sub- sequent treatment of this solution, of which 50. c.c. were used in the determinations under 1., and 100. c.c. portions in the deter- minations under 2. and 3. Column -II. expresses the weight in milligrams of A1PO 4 in aliquot parts of the solutions after the speci- fied treatment. Under Column III. is given the percentage of the Al. dissolved total A1PO 4 , which remains in solution. (= x 100.) Total Al. I. II. III. 1. Dialysed through 24.3 mg. collodion. Dialysate used. 24.3 49.0% 2. 0.5% Na 2 CO 3 added in 0.6 excess of amount necessary to neutralize acid. Filtrate used 0.6 0.6% 3. 0.2% Na taurocholate -f- 0.9 excess of 0.5% Na 2 CO 3 added. Filtrate used. 0.7 0.8% We conclude, then, that aluminium is undoubtedly rendered soluble and absorbable, both from aluminized bread and from alu- minium phosphate by the gastric juice, and that such soluble aluminium is not entirely rendered insoluble by the alkaline intes- tinal fluids. The locality of this absorptive process also becomes evident. Some absorption is to be expected from the stomach during the digestion of aluminized food, for the gastric juice will contain at that time, a high concentration of aluminium. Since the stomach is not particularly well adapted to absorptive processes however, most of the soluble aluminium will find its way into the intestine, where a great part will again be rendered insoluble. This process, however, cannot be instantaneous. The upper duo- 8 denum is known to be frequently acid, and here absorption may be rapid, while in the lower portions of the intestines the amount of metal taken up by the body is probably negligible. IV. Does Aluminium Occur in the Tissues of Normal Dogs? No observations of the occurrence of aluminium in normal animal tissues have been recorded. It is probable that while any considerable quantities, if present, would have been discovered, at the same time traces might have been as easily overlooked, es- pecially when the attention of investigators was not particularly directed to this end. Before determining whether aluminium is absorbed by the dog from food containing it in potentially soluble form, it was neces- sary, therefore, to ascertain whether aluminium exists normally in dog tissus, and if so, to what extent. Accordingly, dogs were selected which had been fed for a period of about two months on a normal diet. This consisted of meat, cracker meal, lard, and bone ash 1 , samples of which con- stituents, when extracted with artificial gastric juice, according to the methods described in the previous sections, failed to give a filtrate containing aluminium. From 200. to 500. grams of blood were removed from each dog through a femoral artery, under local anaesthesia, and the dogs afterwards maintained on the same diet. Three weeks later two of the dogs were bled to death, the operation being conducted as before. After exsanguination was nearly complete, physiological salt solution was admitted to a femoral vein, and the perfusion continued until practically all of the blood had been washed out of the body The washings were discarded. This perfusion is necessary, for were it not done, and the blood later found to con- fain aluminium, then all the other tissues would contain aluminium in proportion to their blood contents (which is not accurately de- terminable) in addition to any stored by the tissue itself. Various parts of these two animals were then selected and prepared for analysis. The parts selected were those in which aluminium might be expected, for physiological reasons, to occur. They are also intended for direct comparison with those dealt with 1 Hashed lean beef 15. grams per kilo of body weight. Cracker meal 2. grams per kilo of body weight. I ard 3. grams per kilo of body weight. Bone ash 1. gram per kilo of body weight. ill section V, where these selections will be found to have greater biological significance. Here we are merely concerned with the- presence or absence of aluminium. The tissues were freed from organic matter in one of the fol- lowing ways: 1, Decomposition of the material according to> Neumann, with nitric and sulphuric acids ; removal of any nitric- acid by boiling with water, and evaporation of the excess of sul- phuric acid from a silica or platinum dish. 2. Ignition in plat- inum, and subsequent fusion of the ash with sodium and potassium carbonates. The residues were then treated precisely as those obtained from the bread extracts, previously described, and the aluminium determined as phosphate. In the case of bone, however, the fusion with alkali carbonate is not necessary. Here also the procedure of Schmidt and Hoagland must be modified. The large quantity of calcium phosphate present makes necessary a greater dilution of the solution for the first precipitation. Comparatively small samples of the material must also be used, even at the sacrifice of accuracy. Working with 2.5 grams of bone ash, representing about four times that weight of fresh bone, the solution in which aluminium phosphate is first precipitated should measure about 800, c.c., and the quantities of reagents be regulated accordingly. Otherwise calcium phosphate will contaminate any precipitate of aluminium phosphate. The following results were obtained. TABLE IV. Data showing the aluminium content of normal dog tissues Column I. designates the animal used; Column IL the part of the animal used; Column III. the weight of the part in grams; Column IV, the method of decomposition of the tissues; Column V, the weight of A1PO 4 found 1 in milligrams ; Column VI. the aluminium calculated to A1 2 O 3 , and Column VII. the aluminium calculated to milligrams of A1 2 O 3 per 100, grams of the original material. x Six blank determinations, using the same reagents and glassware were run parallel to those recorded here. They gave the following results in milligrams of A1PO ; Neumann Method. Ignition Method 0.2 mg. 0.3 mg. 0.3 0.3 0.2 0.3 The proper corrections have accordingly been subtracted from each Weight recorded in this column. 10 I. c. D. a H. E. II. . III. Blood 588. 513. 330. 340. " 1st bleed'g 235. " 2d bleed'g 445. IV. Neumann V. VI VIL 0.1 mg. 0.04mg. 0.01 mg. Spleen Kidneys (2) Bile + Gall Bladder Muscle (Leg) Claws Bone (Femur) Liver Blood 1st bleeding 2d bleeding Spleen Kidneys (2) Bile+ Gall Bladder Muscle (Leg) Claws Bone (Femur) Ferratin, from liver Remainder of liver 13, 38. 15. 60, 3.5 10, 230. 320. 10. 35. 6. 60. 3. 10. 0.130 200. Ignition Neumann Ignition 1,1 0.0 0,0 (XI 0.0 0,1 0.0 0,0 0.1 0,0 0.1 0.1 0.0 0.1 0.0 0.0 0.1 0.1 0.46 0,09 determination lost determination lost Neumann 0.0 In but one instance was the amount of aluminium phosphate obtained large enough to exceed a reasonable error in making the weighings. In this case the blood undoubtedly contained alumi- nium, the presence of which was confirmed by the qualitative tests described later. Unfortunately, the animal died after the first bleeding and the body was inadvertently discarded before alumi- nium was known to be present in the blood It was therefore im- possible to continue work on the other tissues of this particular dog. We regard a contamination of the material during the course of analysis, and accidental errors in the manipulation, as very un- likely, but, of course, as possibilities. The suggestion is also to be considered that the animal, whose previous history is unknown, may have eaten aluminized food prior to the beginning of our ex- 11 periment. The fact that this single result is directly at variance with all the others, leads us to regard it as exceptional, to say the least. It is reasonable to suppose that most of the elements could be found in living matter if our means for their detection were deli- cate enough. From these results, however, we may conclude that aluminium is present in the normal dog in amounts that are too small usually for detection. V. Is Aluminium Absorbed from Food Containing it in Poten- tially Soluble Form? Our work has shown aluminium to be absorbable from bread, containing it supposedly as the hydroxide. Some light on the ab- sorption of aluminium from the digestive tract was obtained in section III. We wish, now, after referring to the findings of previous workers, to discuss the fate of aluminium after it has entered the body. For this purpose a dog (G) was selected, whose blood content of aluminium was previously found to be too small to be detected by our methods. The dog was then fed for one month on a diet similar to the one previously described, but with the substitution of freshly precipitated, moist aluminium phosphate for the bone ash of the normal diet. This diet is somewhat deficient in calcium. The fact that aluminium, even if administered in a more soluble form, would be changed to phosphate in the upper small intestine, led to the choice of the phosphate as the form in which to feed the aluminium. In addition, any chances for acidosis, or of a toxic effect of the anion attendant on feeding water soluble alu- minium was thus eliminated. The body weight of the dog at the beginning of the experiment was 14.0 kilos ; at the end 12.3 kilos. A dose of 10. grams of aluminium phosphate per day was found to be the maximum which could be administered without causing diarrhea. The health of the animal was apparently not particu- larly injured by the experiment, although a decrease in body weight and a loss of appetite were apparent, particularly toward the last. The dog was bled to death and perfused in the manner already explained. Analyses of the tissues were made for aluminium as before. The results are recorded in Table V. The presence of aluminium in these precipitates was shown in each case with Gop- pelsroeder's test. All these precipitates of A1PO 4 were white, in- 12 fusible, and, when tested for calcium by being moistened with hydrochloric acid, and introduced on the end of a platinum wire into the non-luminous flame of a bunsen burner, imparted no color to the flame. TABLE V. Amount of aluminium occurring in the tissues of dog G, fed on freshly precipitated A1PO 4 . The different columns and their numbering have the same significance as in Table IV. I. II. III. G. Blood before experiment 330. g. Blood, after experiment (Plasma) 625. Blood, after experiment (Corpuscles) 150. Bile + Gall Bladder 35. Spleen 30. Kidneys (2) 85. Muscle (Leg) 57. Bone (Femur) Freed from marrow 10. Bone marrow (Femur) 8. Claws 4. Ferratin, from liver 0.348 Remainder of liver 240. IV. V. VI VII. Neumann 0.0 mg. 0.00 mg. 0.00 mg. 1.4 1.5 1.5 0.0 3.2 0.2 0.59 0.63 0.63 0.00 1.34 0.08 0.09 0.42 1.80 0.00 1.58 0.15 Ignition 0.6 0.25 2.5 1.1 0.0 0.5 Neumann 0.4 0.46 0.00 5.8 0.0 0.21 60. 0.17 0.07 As was described in the footnote to Table IV., blanks were run and the appropriate corrections made. An element in the animal body can always be detected when present, by the use of appropriate analytical means. No matter how involved the chemical changes in which it may take part, it is always possible to find the element with proper methods. After absorption such an element would necessarily occur in the blood. The blood analyses would therefore furnish evidence for or against the passage of aluminium into the body. Since this metal is for- eign to the body there will be a tendency toward elimination, and 13 both the excretory organs and their excretions may be expected to contain the metal in comparatively large amount. For this reason the kidneys, the liver, and the contents of the gall bladder were also selected for analysis.. If, however, elimination is not so rapid as absorption, storage must occur, and the liver, from its tendency to store metals would be the most probable place for this accumulation. On the other hand there is always the possi- bility that some tissue may possess a particular retentive power for the metal in question, and to investigate this, samples were taken of muscle (from the leg), of a typical keratin (the claws), of bone (the femur), and of one of the glands of internal secre- tion (the spleen). Our results show most of the aluminium to be speedily elim- inated, through the kidneys and by the liver. Analyses of the urine of human beings fed with aluminized bread, show considerable aluminium and are in agreement with this finding 1 . There is, however, a storage of aluminium, at least temporary, which is not confined to any one particular tissue. The distribution is not uni- form, however, and seems to be roughly parallel to that of the iron. Iron is the only trivalent metal normally found in the body, and as far as we know, it occurs principally combined with proteins. Would it be possible for the closely related trivalent aluminium to replace part of the iron in such combinations, in much the same way, perhaps, as this occurs in the mineral world? To prove this absolutely would be difficult indeed, for tissues are like composite rocks rather than single minerals with definite empirical formulae, and any comparison involving the total amount of iron present would be utterly worthless. Evidence pointing strongly to such replacement, however, is not lacking. If the blood be mixed, at the time of collection, with sodium oxalate dissolved in physiological salt solution, clotting is prevented, and the corpuscles may be separated from the plasma by contrifuging. The concentration of aluminium in the corpus- cles was found to be much higher than in the iron-free plasma. Now in adult animals not suffering from anaemia, the red cor- puscles are synthesized in the marrow of the long bones. Ac- cordingly, the femur was taken from the dog's body, split, and the marrow, including as much of the spongy portion of the bones as possible, was removed with a steel instrument. Both the bone and the marrow contained aluminium but the latter did so in ^ ery large T Paper in process of publication. 14 quantities^ The presence of the metal in the red corpuscles, therefore, dates from the formation of the cell, and can be regard- ed as an original constituent rather than as a subsequently acquired "impurity." Furthermore, the erythrocytes are broken down in the liver and the iron formerly contained in the haemoglobin is stored there, in a great part as ferruginous proteins, the so-called "ferratins.*' Ferratin was accordingly prepared from the liver. The method used was that of Wohlgemuth 2 , which is essentially the same as the original proceedure of Schmiedeberg. :! The fresh, finely divided liver was covered with water, and the temperature gradually increased to boiling, when the aqueous extract was removed. Subsequent extractions with boiling water were then made. The protein was precipitated from the extract with dilute acetic acid and purified by dissolving in ammonia and reprecipitating with acid. The preparation was finally washed with absolute alcohol, with absolute ether, and dried. It was of a light yellow color, and contained on analysis, 4.3% Fe 2 O 3 , resem- bling the ferratin prepared for comparison from the liver of nor- mal dog F. (Section III.), which, however, contained only 2.4% Fe 2 O 3 . The aluminium content of the former preparation, (from dog G.), while not very accurately determined on account of the small sample of material available, was at any rate very large. Aluminium can therefore be seen to follow the iron from the synthesis of haemoglobin in the bone marrow to its disintegration in the liver and the storage of the iron as ferratin. That there is an "aluminium circulation," comparable to the iron circulation well known to physiologists, is then certain. Were this associa- tion of aluminium with the iron merely a physical one, it is not reasonable to believe that it would persist through such profound chemical changes. The replacement of protein-combined iron by aluminium, and the formation therefore of "aluminium-proteins/' such as we undoubtedly obtained in the case of our specimen of ferratin, seems to be the only logical explanation of the facts. The presence of aluminium in the bone itself is interesting as recalling the observation of Papillon, already mentioned. The quantity found by him is enormous in comparison, however. It will be remembered, that Papillon's experiments were made on a young animal. The diet used by us would probably not affect 2 Wohlgemuth: Zeit. Physiol. Chem xxxvii, 475, (1903). 3 Schmiedeberg: Archiv fur Exp. Path. u. Pharm. xxxiii, 101, (1894). 15 greatly the composition of the bones of an adult animal, whose skeleton had already fully developed. It was thought advisable therefore, to repeat the work on a puppy, where this lack of bone- building substances would be more marked, owing to the absence of milk from the food, and where the bones were still in the process of rapid growth. This experiment is still in progress (May 12, 1917). The principle of aiding substitution by creating a deficiency of the material to be replaced in the diet, was evidently not correctly applied by Koenig 4 . Since the aluminium replaces the iron rather than the calcium of the body, a diet deficient in iron, not calcium, is indicated, and experiments on an iron-poor diet would make an interesting chapter to the investigation. We may conclude then that aluminium is absorbed from the digestive tract of the dog in considerable quantities when fed as aluminium phosphate. Much of this aluminium is speedily eliminated in the bile and the urine. Some, however, is retained in the body, in all probabil- ity replacing the iron of ferruginous proteins, and accompanies this iron in its cycle : Bone Marrow Red Corpuscles Ferratin Bone Marrow. Still another portion remains in the bones, and, although to a much smaller extent than found by Papillon, con- firms his results. VI. Is Aluminium Absorbed by Dogs from Aluminized Bread? The experiments described here were undertaken to determine whether or not aluminium is absorbed in dogs fed on a diet con- taining "aluminized food." It was planned to repeat, in a general way, experiments previously conducted in this laboratory 1 , but to use, in the determination of aluminium, the method devised by Schmidt and Hoagland 2 . The experiments were performed on two dogs (A and B), which were fed aluminized food for a period of about three months. The dogs had been previously maintained on a definite diet of hashed lean beef, cracker meal, infusorial earth and water. During the experiment "baking-powder biscuits," made from alum baking-powder, according to Kahn 3 , were substituted for cracker *J. Konig-: loc. cit. l Kahn: loc. cit. 2 Gies: Biochem Bull, v, (1916). 3 Kahn: loc. cit. The ingredients used were the same as used in preparing the bread used in Section II. 16 meal, and in such quantity as to give each dog, daily, about 20. mg. of aluminium as A1 2 O 3 , per kilo of body weight. Representative data, showing the general condition of the dogs at weekly intervals throughout the experiments, are recorded in Table VI. TABLE VI, Data showing the weight and condition of the urine of dogs A and B during the experiments. 4 DOG A DOG B X G &** G Ill /> AJ4/1 Illfttfi - M c fe ^JMM^T* "-QL 7 LD 21-100m-8,'34 U : UNIVERSITY OF CALIFORNIA LIBRARY