THE AMERICAN JOURNAL OF PHYSIOLOGY. EDITED FOR Cl)c American Pjtmological ^ocict^ BY H. P. BOWDITCH, M.D., BOSTON freDERIC S. LEE, Ph.D.. New York R. H. CHITTENDEN, Ph.D., New Haven jacqueS LOEB, M.D.. Chicago W. H. HOWELL, M.D., Baltimore w. P. LOMBARD, .M.D., A^N Arbor W. T. PORTER, M.D., BOSTON CONTENTS. No. I, September i, 1901. Page A Study of the Metabolism ix Dogs with shortened Small In- testines. By Joseph Erlanger and Albion Walter Hewlett ... I Studies on Reactions to Stimuli in Unicellular Organisms. VII. — The Manner in which Bacteria react to Stimuli, especially to Chemical Stimuli. By H. S. Jennings and J. H. Crosby . . 31 The Formation of Allantoin from Uric Acid in the Animal Body. By Robert E. Swain 38 Some Decomposition Products of the Crystallized Vegetable Proteid Edestin. By p. a. Levene and Lafayette B. Mendel . . 48 No. II, October i, 1901. Do Spermatozoa contain Enzyme having the Power of Causing Development of Mature Ova? By William J. Gies 53 Concerning the Poisonous Effect of Pure Sodium Chloride Solutions upon the Nerve-Muscle Preparation. By Harvey Gushing ']^ Cerebral Pressure following Trauma. By IV. B. Cannon. ... 91 On the Analogy between the Effects of Loss of Water and Lowering of Temperature. By Arthur W. Greeley 122 Notes on Regeneration and Regulation in Planarians. By Frank R. Lillie 129 Artificial Parthenogenesis produced by Mechanical Agitation. By A. P. Mathews 142 No. Ill, November i, 1901. The COxMposition of Tendon Mucoid. By W. D. Ciitter and William y. Gies 155 Phlorhizin Diabetes in Cats. By Jtilitis F. Arteaga 173 On the Production of Artificial Parthenogenesis in Arbacia BY the Use of Sea-Water Concentrated by Evaporation. By S. J. Hunter 177 vi Contents. No. IV, Deckmber I, 1 90 1. Page An Analysis of the Influence of the Sodium. Potassium, and Calcium Salts of the Blood on the Automatic Contractions OF Heakt-Muscle. By W. H. Howell 181 The Action of Pilocarpine and Atropine on the Embryos of the Starfish and the Sea-Urchin. By Albert P. Mathews ... 207 The so-called Cross Fertilization of Asterias by Arbacia. By Albert P. Mathews 216 The Chemical Constituents of Tendinous Tissue. By Leo Buerger and William J. Gies 219 No. V, January i, 1902. Studies on Reactions to Stimuli in Unicellular Organisms. VIII.— On the Reactions of Infusoria to Carbonic and other Acids, with especial reference to the Causes of the Gatherings Spontaneously Formed. By H. S. Jennings and E. M. Moore 233 The Movements of the Intestines Studied by means of the Rontgen Rays. By IV. B. Cannon . 251 The Reflexes connected with Autotomy in the Hermit-Crab. By T. H. Morgan 278 A Physiological Study of the Pulmonary Circulation. By Horatio C. Wood, Jr 283 Artificial Parthenogenesis in Starfish produced by a Lowering of Temperature. By Arthur W. Greeley 296 On the Prolongation of the Life of the Unfertilized Eggs of Sea-Urchins by Potassium Cyanide. By Jacques Loeb and Warren H. Lewis 305 Contributions to the Physiology of the California Hagfish, POLISTOTREMA StOUTI. II. — ThE ABSENCE OF REGULATIVE NeRVES for the Systemic Heart. By Charles Wilsojt Greene .... 318 The Physiological Action of Formaldehyde. By Waldemar Koch 325 No. VI, February i, 1902. On the Relation of Lipase to Fat Metabolism — Lipogenesis. By A. S. Loevenhart 331 The Physiological Zero and the Index of Development for the Egg of the Domestic Fowl, Gallus Domesticus. By Charles Lincoln Edwards 3^1 The Excretion of Nitrogen during Nervous Excitement. By Francis Gano Benedict 308 Contents. vu Page Studies on the Physiological Effects of the Valency and POSSIBLY THE ELECTRICAL CHARGES OF lONS. I. — ThE TOXIC AND Antitoxic Effects of Ions as a function of their Valency AND POSSIBLY THEIR ELECTRICAL CHARGE. By Jacques Loeb . . 4II A Contribution to the Physiology of the Nervous System of THE Medusa Gonionemus Murbachii. Part I. — The Sensory Reactions of Gonionemus. Bj Robert M. Yerkes 434 The Liberation of Volatile Sulphide from Milk on Heating. By Leo F. Rettger 43° No. VII, March i, 1902. The Influence of Temperature, Odors, Light, and Contact on THE Movements of the Earth-Worm. By Amelia C. Smith . . 459 Proceedings of the American Physiological Society (Issued March i, 1902) ix-xxx INDEX xxxi PROCEEDINGS OF THE AMERICAN PHYSIO- LOGICAL SOCIETY. FOURTEENTH ANNUAL MEETING. University of Chicago, December 30 and 31, 1901. ^/ PROCEEDINGS OF THE AMERICAN PHYSIOLOGICAL SOCIETY. SOME NEW OBSERVATIONS ON BLOOD PLATES. By GEORGE T. KEMP and O. O. STANLEY. The experiments of Dutjen (Virchow's Archiv, 1901, clxiv, p. 239) were repeated and his observation of amoeboid movements of the blood plates on agar-agar containing sodium hexa-metaphosphate NagPgOjg were corroborated. In the blood of animals into whose circulation methylene blue had been injected, the plates showed blue granules scattered through the colorless part which was making amoeboid movements. These blue-stained granules resembled in staining properties the nuclei of leucocytes. Macallum's method for the detection of phosphorus in the cell showed the plates to be rich in phosphorus, the green color of the reaction being seen principally, if not entirely, in granules. From these experiments, together with the observation of Lilienfeld (corroborated by us), that the plates leave an insoluble residue after digestion with hydrochloric acid and pepsin, the authors conclude that the plates contain nucleo-proteid in the form of granules, and that they also contain a part (probably protoplasmic) which is capable of making amoeboid movements. The experiments of Arnold with blood in 10 per cent solution of KI were repeated and while the Manlbccr ■Siwd Stcchapfcl iorms, were constantly present, the complete budding off of the structures which he took to be plates was not always observed. Such separated por- tions as were seen were thought by the authors to be artefacts, and to be not identical with the blood plates. RELATION OF BLOOD PLATES TO THE INCREASE IN THE NUMBER OF RED CORPUSCLES AT HIGH ALTITUDES. By GEORGE T. KEMP. The red corpuscles and blood plates were counted first in Paris. The method employed is described in this Journal, 1901 , v, p. iv. The mean of five counts on five consecutive days was as follows : red corpuscles xii Fourteenth Annual Meeting, 4,800,000 and blood plates 457,000. The ratio of plates to red corpuscles was i : 10.5. A journey was then made to the Corner Grat, Switzerland, 10,290 feet above the sea-level. The journey was accomplished in forty-eight hours, and as the ascent of the mountain was made by the electric railway, the experiment was not complicated by physical exertion or fatigue. On the Corner Crat, seventy-two hours after the last blood count in Paris there were 7,000,000 red corpuscles (mean of two counts), and 1,206,900 plates. The ratio of plates to red corpuscles was i : 5.8. The number of plates had, therefore, increased more than one hun- dred per cent in seventy-two hours. The change in the character of the plates was most striking. The field was strewn with them and they were at least twice the average size of those seen at Paris. The number of small red corpuscles was very great. The large plates and the small red corpuscles overlapped in size. A careful examination failed to reveal any color in the plates. The small red corpuscles were always stouter than the plates and could not have been mistaken for them. Nothing that could be taken for a connecting form between them was observed. In spite of this, there was a marked resemblance to the crise hematoblastic described by Hayem. NOTES ON THE PHYSIOLOGY OF THE CIRCULATORY SYS- TEM OF THE CALIFORNIA HAGFISH, tPOLISTOTREMA STOUTI. By C. W. GREENE. Three distinct propelling organs are situated on the system of blood vessels in the hagfish. These are the hearts, which may be distin- guished as the systemic heart, the caudal heart in the tail, and the portal heart on the portal venous system. The systemic heart is of peculiar interest in that it is free from all nervous regulation. Stimu- lation of the vagus nerves, and of various portions of the central ner- vous system, has no influence on the rate and force of the heart, though other muscles are profoundly influenced. The portal heart was first observed to contract in Myxine by A. Retzius and J. Miiller. This organ contracts at a rate of from 40-50 per minute and main- tains its rhythm for many hours after isolation from the body. The portal heart is also without a special nervous regulative mechanism. Pj'ocecdings of the Amei^ican Physiological Society, xiii Studies on the relation of the activity of the systemic heart and of the portal heart to the composition of the blood have revealed that the osmotic pressure of the blood is much higher than in bony fishes. Tested by the freezing-point method the hagfish serum depresses the freezing-point 1.934° €-1.992° C, the equivalent of about 3 per cent sodium-chloride solution, /. e., practically the same as the sea-water, 1.945° C, of Monterey Bay. ON THE QUESTION WHETHER DEXTROSE ARISES FROM CELLULOSE IN DIGESTION. By GR.A.HAM LUSK. It was shown that neither twenty grams of cauliflower fed to a dog diabetic with phlorhizin, nor ten grams of paper fed to a phlorhizin- ized goat increased the dextrose in the urine. Since dextrose fed in phlorhizin diabetes is quantitatively eliminated in the urine, it follows that dextrose is not produced from cellulose in digestion. THE ACTION OF ALCOHOL ON MUSCLE. Bv FREDERIC S. LEE and WILLIAM SALANT. Ethyl alcohol in various percentages was injected into frogs in quantities proportional to the weight of the animals. In each ex- periment one gastrocnemius muscle was protected from the action of the drug, and compared with the alcoholized gastrocnemius of the opposite side. It was found that in small quantities (J minim of 10 percent alcohol per gram of frog), the drug has no effect. Larger quantities (i to 4 minims of 10 per cent alcohol per gram of frog) allow the muscle when stimulated to contract more quickly, relax more quickly, perform a greater number of contractions in a given time, and do more work than a muscle without alcohol, while the onset of fatigue is at the same time delayed. In some cases the increase in the amount of work performed is more than 10 per cent. Experiments in which curare and alcohol were used together make it for the present uncertain whether this seemingly beneficial result is to be attributed to the action of alcohol on the muscle protoplasm, or on the intramuscular nerve substance. Whether the alcohol exerts xiv Foui'teenth Aii7itial Meeting. its action by serving as a food, or in some other manner, is not decided. In larger quantities than the above the alcohol is detrimental to muscular action, diminishing the whole number of contractions, in- ducing early fatigue, and diminishing the amount of work that the muscle is capable of performing, even to the extent of doing away entirely with contractile power. In , such quantities the drug is distinctly poisonous. The after-effects of the administration of alcohol have not yet been studied. EXPERIMENTS ON THE RELATION BETWEEN THE SPLEEN AND THE PANCREAS." By LEO F. RETTGER (presented by LAFAYETTE B. MENDEL). In investigating the problem of the possible functional relation between the spleen and the pancreas, a considerable number of ob- servations on dogs have been made in repetition of the experiments of Herzen ^ and of Gachet and Pachon.^ This preliminary report brings a confirmation of the observed influence of spleen extracts in vitro (Herzen), and in vivo (Gachet), in increasing the proteolytic power of pancreatic extracts. The specific effect of intravenous injections of spleen extracts in augmenting the trypsin content of the pancreas of splenectomized dogs has been controlled by comparable injections of other fluids, c. g., physiological saline solution, boiled extracts of spleen, extracts of liver, pancreas, etc. In none of the latter cases was any specific *' trypsinogenic " effect manifested. While the positive results obtained were perhaps not in every instance as pronounced as those reported by earlier investigators, the effects of the spleen reactions are characteristic, as indicated by the protocols and demonstrations presented. EXPERIMENTS ON ALLANTOIN EXCRETION. By LAFAYETTE B. MENDEL. I. Thymus glands were finely divided, mixed with water, and in- XxQdiWZ&A per rectum into dogs. The absorption of the " purin " con- stituents was demonstrated by the characteristic excretion of allantoin ^ Herzen : Revue g^nerale des sciences pures et appliquees, June 15, 1895. 2 Gachet and Pachon : Archives de physiologic, 1898, xxx, p. 363. Proceedings of the American Physiological Society, xv in the urine eliminated during the succeeding day. The output of uric acid was also noticeably increased. The animals had previously been fed for a day or two on a diet of casein and lard, to exclude the influence of dietary nuclein compounds. The quantity of allantoin excreted was noticeable; and the masses of crystals separated directly from the concentrated urine were exhibited. The experiments were carried out by Mr. F. P. Underbill. 2. In an investigation of the physiological action of the nucleates and nucleic acid separated from wheat germs by the methods first described by T. B. Osborne, products containing about three per cent of phosphorus were fed to a cat in doses of five grams, with milk and cracker meal. The urine readily yielded crystals of allantoin, which were exhibited. These vegetable nucleates thus resemble the so-called nucleoproteids from animal tissues. The nucleic acid from wheat apparently shows a similar behavior in metabolism. A study of the action of the vegetable products when introduced directly into the circulation has been begun. The results already obtained are in part comparable with those recently recorded by Bang for the guanylic acid and the nucleoproteid of the pancreas. These experiments have been undertaken by Mr. G. B. White. THE ROLE OF THE CELL NUCLEUS IN OXIDATION AND SYNTHESIS. By R. S. LILLIE (presented by W. T. PORTER). Sections of fresh and alcoholic tissues of the frog were treated with an alkaline solution containing para-diamido-benzene and alpha- naphthol. This solution, at first colorless, turns violet on oxidation from the formation of indophenol by condensation. Different tissues produce a more or less rapid appearance of the color according to their greater or less oxidative activity. The chief results and conclusions were as follows, (i) In general the coloration is intensest and appears soonest in those organs and in those regions of organs which contain the most numerous and most densely aggregated nuclei, e.g. kidney, spleen, thymus, ventral gill-vestiges. Oxidations and their correlated syntheses appear there- fore to depend chiefly upon the nuclear matter of the cell. 2. The formation of the color is prevented by acid, and by strongly xvi Fcmrteenth Annual Meeting. reducing substances ; and is retarded by the presence of poisons which interfere with oxidation, such as hydrocyanic acid and its salts. 3. The formation of the color appears on microscopical exami- nation to take place chiefly at the surface of the nucleus. The oxi- dative and synthetic activity of such organs as the kidney and liver thus probably depends largely upon their great extent of nuclear surface. 4. The mucosa of the intestinal tract has a marked oxidative activity; the color appears chiefly at the inner ends of the epithelial cells near the nucleus. The layer of closely aggregated nuclei resulting from the columnar arrangement of the cells must be regarded as of great importance in effecting oxidations and syntheses during absorption. 5. Leucocytes and lymphoid structures in general have a marked oxidative activity. This is probably of importance in the activity of leucocytes in phagocytosis. 6. In the kidney-tubule oxidation is effected chiefly by the glandular region ; the Malpighian corpuscles are comparatively inactive. STUDIES IN DIURESIS. By J. T. HALSEY. Nussbaum's investigations on the circulation in the frog's kidney with their far-reaching importance in explaining the physiology of diuresis, failed to receive anatomical confirmation when Adami repeated these experiments. Recently Marcuse, while studying phlorhizin diabetes, confirmed Nussbaum unreservedly. In the hope of possibly finding an explanation for the contradictory results the matter was again taken up by myself, and the results of these new investigations form the basis of this communication. Every care was taken to carry out the investigation according to Nussbaum's directions. The only deviation made was that in addi- tion to ligaturing the renal and genito-urinary arteries, all branches of the abdominal aorta were ligatured, in order to prevent, if possible, blood reaching the kidneys through anastomotic arterial circulation. Canadian bullfrogs, male and female, females of several varieties of the small indigenous frogs, and female Rana Esculenta from Strassburg in Alsace, sixty-three frogs in all, were used in the course Proceedings of the American Physiological Society, xvii of the investigation. After completion of the operation and a short time before death, a solution of ammonium carmine, or a suspension of granules of vermilion or prussian blue, was injected. In every case, without exception, microscopic examination of the kidneys disclosed the coloring matter or the pigment inside the capillaries of a greater or less number of the glomeruli. A careful study of the sections, however, led to the belief that, as so few of the glomeruli were receiving blood, and as in these the blood stream appeared to be so sluggish, it was a justifiable conclusion that the glomeruli in the Nussbaum frog are a physiologically negligible factor. The study of the behavior of various substances injected into the veins of such frogs gave results in entire accordance with those of Nussbaum. Urea and indigo carmine are excreted by the kidney under these conditions, whereas dextrose, egg albumin, peptones (Witte's), and ammonium carmine are not. Under the influence of phlorizin, dextrose was secreted as Marcuse had found. When theobromine (diuretin) was injected simultaneously with dextrose, the urine reduced Fehling's fluid so strongly that the presence of sugar seemed established. After injecting NaCl, Na.^HPO^, and Na.2S04, these salts were regularly present in the urine. It seems impossible at present to explain the results of these experiments in any other way than by attributing to the tubular epithelium an active power of excreting urea and the three salts mentioned. Under the influence of phlorizin or theobromine, the tubular epithelium seems to acquire the power of excreting dextrose, a power which it does not possess under normal conditions. ON SALINE DIURESIS. By ARTHUR R. CUSHNY. Magnus has recently shown that the sulphate of sodium injected intravenously causes a more profuse diuresis than that produced by the chloride and that this difference is due to some local effect in the kidney and not to the changes in the circulation. He appears to consider that the sulphate stimulates the secretory elements of the kidney, a view also adopted by Sollmann. A simpler explanation may be given on Ludvvig's theory, namely that the sulphate is not so readily absorbed by the renal tubules as the chloride, and thus xviii Fourtee7ith Annual Meeting. retains a larger amount of water. On the injection of chloride and sulphate together intravenously, the chloride and sulphate increase in the urine along with the fluid. As the diuresis decreases, the chloride falls much more rapidly than the sulphate, which increases in percentage. This may be explained by the chloride and water being absorbed, while the sulphate permeates the tubules less readily. If this view is correct the chloride should decrease still more rapidly on accelerating the absorption while the sulphate should be less affected. This may be done by compressing the ureter during the injection, and in a number of experiments the ureter was narrowed until the pressure within it rose to 20-30 mm. mercury. The urine from this ureter was compared with that from the normal one of the other side, when it was found that the water from the former was less by G6 per cent, the chloride by 82 per cent, and the sulphate by 30 per cent. This can be explained only by absorption in the tubules, and the sulphate therefore penetrates the epithelial cells much less readily than the chloride. ON THE GLANDS OF THE OVIDUCT IN THE FOWL. By ARTHUR R. CUSHNY. The oviduct of the fowl is about 37-50 cm. in length and in the upper part possesses a thick wall with large villous projections. In the middle the wall is thinner and has fewer villi, while in the lowest 7.5- 10 cm. (uterus) the wall is again thick and villous. Histologically the oviduct consists of a layer of unstriated muscle externally (thin ex- cept in the uterus), a thin connective tissue submucosa, a mucosa consisting largely of glands with little connective tissue, and a lining epithelium. The glandular layer is entirely absent in the first part of the duct towards the fimbriated tube. The lining epithelium is of the pseudotransitional variety and is covered with cilia whose function appears to be the propulsion of the ovum downwards in a spiral. From above downwards the glands may be divided into albumen- secreting, mucous, membrane-secreting and lime-secreting. The albumen glands are very complicated, branching and coiling through the mucous coat. When they are not actively secreting, the cells are filled with fine granules which grow in size, coalesce, and form secretion vacuoles which are voided into the tubules and thence poured into the oviduct before the ovum passes. The lumen of the Proceedings of Ike American Physiological Society, xix gland is entirely closed before the secretion is formed, but is dilated with masses of albumen as the secretion vacuoles burst. The mucous part of the duct lies between the albumen and the membrane secret- ing glands and consists of typical mucous beaker cells with com- pressed ciliated cells between them. This part of the duct is 1-3 cm. long and secretes very actively, but the purpose of the secretion is unknown. The membrane-secreting part of the oviduct possesses glands of which the cells are filled with secretion-vacuoles or granules which coalesce and are excreted into the lumen. In the uterus the lime secreting glands are very complex, with few apertures. The gland cells are uniformly clear except towards the lumen, where they present granules staining strongly with eosin and with aniline blue. . EXPERIMENTS WITH ZYGADENUS VENENOSUS (POISON CAMASS).^ By REID hunt. By following the methods ordinarily used for the isolation of plant alkaloids a substance having the following properties was obtained from the alcoholic extract of Zygadenus ven. : as left from the evaporation of its solution in chloroform it was hard, glassy, trans- parent, slightly yellow in color, alkaline to litmus, easily soluble in alcohol and acidulated water, insoluble in distilled water. When dis- solved in alcohol and the alcohol slowly evaporated, a semi-crystalline mass was left. When this substance was treated with sulphuric acid a yellow color appeared; on warming, the color became violet, then orange and cherry red. The solution showed a green-yellow fluorescence. Warmed with hydrochloric acid a pink-red color was produced ; the solution became fluorescent on the addition of acetic acid. When a little of the substance was warmed with oxalic acid a blood-red color resulted. The substance had an intensely acrid burning taste which was very persistent. Dissolved in alcohol or chloroform and applied to the ^ These experiments were performed last July in the chemical laboratory of the Montana State College of Agriculture. Most of the work was done in connection with Mr. V. K. Chesnut of the U. S. Department of Agriculture. xx Fo2irtee7ith Animal Meetiup;. skin it caused a burning sensation which soon became painful ; the pain continued long after the solution was washed off. Later the spot became almost anaesthetic. The dust of the dried plant as well as the alkaloid caused intense irritation and sneezing when applied to the nose. The substance is very poisonous, five or six milligrams pro kilo body-weight being fatal to rabbits. The immediate cause of death in animals poisoned by this substance was paralysis of the respiration ; there was also a marked lowering of the blood pressure and a slowing of the heart. The latter occurred after the administra- tion of atropine and was probably due to a direct action upon the cardiac muscle. Convulsions (probably due to asphyxia) occurred in all cases ; the convulsions were followed by periods of great muscular weakness. The relaxation of the frog's muscle was greatly prolonged in animals poisoned by this substance. Thus both the chemical and physiological properties of the active principle of Zygadenus venenosus agree very closely with those of the mixture of alkaloids known as " veratrine " ; ^ experiments to determine the exact character of the alkaloid (or alkaloids) present are now being made. The alkaloid is excreted very rapidly in the urine, while it is absorbed rather slowly from the stomach. By the use of diuretics (diuretin or caffeine) the drug may be removed from the circulation so rapidly that few poisonous symptoms may result ; rabbits and sheep invariably recovered from fatal doses of zygadenus when treated in this manner. Strychnine and atropine seemed to intensify rather than diminish the effects of the poison. THE EXCRETION OF LITHIUM. By C. a. good. HiJFNER states that on giving 25 mgm. of lithium carbonate by the stomach, none of the metal could be found in the urine, although the spectroscopic method he employed permitted the recognition of 0.00003 mgm. When lithium was given in 50 mgm. doses the spec- 1 After the completion of the work here reported Vejux-Tyrode published a note (Journal of medical research, Nov., 1901) in which he states that Pfaff has isolated from Zygadenus ven. a "white, crystalline, neutral body" which causes a change in tlie form of the contraction of the frog muscle similar to that pro- duced by vcratrine. Vejux-Tyrode obtained a similar body from Zygadenus frumentii. Proceedings of the AfneiHcan Physiological Society, xxi trum was obtained from the urine. No quantitative estimations have hitherto been made. In a number of experiments on cats in which lithium chloride was injected hypodermically, the lithium was estimated by taking advantage of the insolubility of the phosphate in ammonia solution. The dose of 0.5 gram was found fatal, the animal dying after about a week. When large quantities (1-2 grams) were injected hypodermically, very considerable amounts were obtained from the stomach (by lavage), and from the bowel, and the saliva also contained appreciable quantities. In fatal poisoning more was found in the stomach and bowel contents than in the urine. In experiments in which small doses were administered repeatedly, more lithium was excreted in the urine than by the alimentary tract. AN ATTEMPT TO OBTAIN REGENERATION OF THE SPINAL CORD. By PERCY M. DAWSOX and EDWIN N. RIGGINS. A -FEW years ago experiments were performed in this laboratory which led to the conclusion that regeneration of the dorsal-root fibres into the cord will take place under the proper conditions. The authors then stated the opinion that with proper technic a severed spinal cord might be made to regenerate its broken tracts both ascending and descending. ^ Acting on this hope the following experiment was performed. In a young bitch anaesthetized with morphia and ether, the spinal cord was exposed and divided with a very sharp knife at the level of the lowest dorsal vertebra. The animal was then nursed with the great- est care for a period of one hundred and twelve days. During this time it was under constant observation. At autopsy the site of operation presented a very satisfactory appearance. The dura was adherent above to a mass of scar tissue but otherwise appeared quite normal. The cord observed by transmitted light, showed only a fine white line at the plane of section. Although the animal had remained in excellent condition and the operation had been so successful from a surgical point of view, there was never any conclusive evidence of conscious sensation or of volun- tary motion in the " hind dog'' 1 Regeneration of the dorsal root fibres of the second cervical nerve within the spinal cord : Baer, Dawson and Marshall. Journal of experimental medicine, 1899, IV, p. 29. XX 11 Fourteenth Annual Meeting. A STUDY OF METABOLISM IN A CASE OF LYMPHATIC LEUKEMIA. By YANDELL HENDERSON and G. H. EDWARDS.. By analyses of the urine of a case exhibiting all the characteristics of lymphatic leukaemia, the uric acid and phosphates were found rather below than above the normal. From this the conclusion was drawn in conformity with the views of Milroy and Malcomb that the nuclein metabolism of leukaemia, of this form at least, differs from the leucocytosis produced by injecting nuclein, in being due, not to an increased formation of leucocytes, but to a diminished katabolism. This may be due to an arrested development in the small leucocytes. The investigation is still in progress. A NEW INSrRUMENT FOR DETERMINING SYSTOLIC AND DIASTOLIC BLOOD-PRESSURE IN MAN. By JOSEPH ERLANGER. Since the work of Howell and Brush has shown decisively that instruments which make use of the principle of Marey — viz., the pressure under which an artery will give its maximum pulsations is the mean pressure in that artery — give, not mean pressures, but mini- mum or diastolic pressures, and since the same authors have shown that the systolic and diastolic pres- sure may actually be affected in opposite directions, the importance of a sphygmomanometer that will give parallel records of the variations in systolic and diastolic pressures becomes evident. Therefore the following instrument has been devised. An arm-piece like that employed in the Hill- Barnard sphygmometer is fastened about the arm above the elbow. Its rubber bag (A) communicates with a mercury manometer (B) and with a Pulitzer bag (C) for vary- ing the pressure. The maximum pressure is obtained by noting the Proceedings of the American Physiological Society, xxiii pressure required to obliterate the pulse distal to the point of appli- cation of the pressure. With the rest of the apparatus the diastolic pressure is obtained. When the stopcock (D) is open the pressure in the apparatus is also transmitted to the rubber bag (E) enclosed in the glass bulb (F). The air space between these communicates with the exterior through the stopcock (G) and with the tambour (H) the lever of which writes upon a slowly revolving drum. The tambour is perforated by a minute opening. To determine the minimum pressure, the stopcock (G) being open, the pressure is quickly raised above the expected minimum. The three-way stop- cock (I) is then closed and the valve (G) is closed. The lever will now record the pulsations. After, say, twenty pulsations have been recorded, the stopcock (I) is turned so as to communicate with the capillary (K). The pressure falls slowly and is stopped after a fall of 5 mm. Hg. The lever will quickly return to its original level on account of the equalization of pressure permitted through its minute perforation, and the pulsations under the diminished pressure are recorded without interruption. This manipulation is repeated until the pulsations begin to diminish in size. A series of records upon a straight line will thus be obtained at various pressures. They will show a distinct maximum. The pressure at which this maximum record was obtained is the minimum or diastolic pressure in the artery. A NEW FORM OF ERGOGRAPH. By WINFIELD S. HALL. The principal objection raised to the classic form of the Mosso ergo- graph is that the muscle soon becomes so far fatigued that it cannot lift the weight at all, and, therefore, is not credited with accomplishing work, yet it is exerting not a small amount of unmeasured energy. If a spring were used a small amount of work would be shown by a small extension of the spring. Again, if the observer wishes to study the degree and variations of tension in the " isometric " muscular contractions, the spring ergographs seem necessary. Two other ob- jections have been urged against the weight ergograph : (i) The highest tension required in moving a weight is that which overcomes its inertia and starts the weight with the other movable parts of the apparatus into motion. Once it is started, a sudden stopping of the xxiv Fourteenth Annual Meeting. motive force (the finger) is not followed by an instantaneous stop- ping of the weight, pulley, and cord. Thus the graphic record may be distorted. (2) All work with the weight ergograph shows only " isotonic " contractions, while it is possible with a spring ergograph to study both isotonic and isometric contractions. The writer has attempted to construct a weight ergograph in which the various imperfections of the classic ergograph are, in a measure at least, remedied. The principal feature of this ergograph is that the muscle works only during contraction. This is accom- plished through the device of a differential pulley. The arrangement accomplishes the same thing physiologically as Pick's Arbeit- sammler. Next in importance is the fact that the weight moves so slowly that quick movements of the finger never result in the inertia of the weight affecting the curve. The objection that the nearly exhausted finger cannot start the weight, and will, therefore, make no record and do no mechanical work, is not a valid objection from the standpoint of everyday work. Every object that the laborer lifts requires a greater expenditure of energy to start it moving than to keep it moving. DEMONSTRATION OF APPARATUS. By warren p. LOMBARD. I. Ergograph for tJie Index-finger xq^oxX.q.^ for Thomas A. Storey, Assistant in Physiology, Leland Stanford Jr. University. The instru- ment was devised by Mr. Storey in the summer of 1900, in the Physiological Laboratory of the University of Michigan. It permits of an accurate record of the angular movement about the metacarpo- phalangeal joint of the index finger. It may be employed both by vol- untary and by electrical excitation of the abductor indicis. It enables the muscle to work against a weight or spring, and isotonically or isometrically. It permits movements of the finger to be recorded directly on a horizontal or vertical drum, or indirectly by means of a distant lever or other appropriate device. It allows the use of vari- ous forms of " Arbeitsammler " in connection with it. It permits the attachment of weights so that the strain on the finger shall (except for inertia effects) be constant, and the effect of the throw of the weight be minimized. The essential element of the device is a lever, bearing on its axis Proceedings of the Amei^ican Physiological Society, xxv pulleys of various diameters for the attachment of cords to be con- nected with the weights or springs to be moved, or the apparatus employed to record the movement of the finger. A direct record of the movement of the lever can be taken from a celluloid pointer at the extremity. The centre of the metacarpo-phalangeal joint of the index finger is brought into line with the axis of rotation of the lever, and the end of the first phalange is fastened in the clamp at the end of the lever arm. The rest of the hand and thumb are fixed on suitable supports. A photograph and working drawings of the instrument, together with curves obtained by it, were shown, and a working model, differing slightly from the original, was exhibited and its use demonstrated. 2. The following apparatus intended for the use of students were demonstrated: A platinum mercury key; an upright rheocord ; a graphite rheostat — to demonstrate the effect of rate of change of strength of a battery current upon excitation of nerve; a short con- tact key — to demonstrate the effects of duration of a battery current upon the excitation of nerve; a modified form of Fitz's belt pneumograph ; a ball and socket joint to connect the disk on a tambour with a lever ; a celluloid lever; a spring support for a glass slide ; plaster of Paris electrodes and a method of moistening the skin for excitation of human nerve and muscle. DEMONSTRATION OF APPARATUS. By W. T. porter. The following pieces of the Harvard Physiological Apparatus were shown: Adjustable plate, aortic cannula, circulation scheme, femur clamp, platinum electrodes, capillary electrometer, artificial eye, frog board and clips, inductorium, rocking key, short-circuiting key, simple key, optical lantern, light muscle lever, heavy muscle lever, moist chamber, square rheochord, signal magnet, sphygmograph tambour. The Mechanism of Fibrii.lar Contraci'iox of the Heart. By W. T. Porter. Further Experiments on the Importance of Sodium for the Heart- beat. By D. J. LiNGLE. B y invitation. XX vi Fourteenth An7iual Meeting. On the Prolongation of the Life of Unfertilized Eggs of the Sea- Urchin BY Potassium Cyanide. By J. Loeb and W. H. Lewis. This journal^ 1902, vi, pp. 305-317. The Physiological Effects of the Electrical Charge of Ions and THE Electrical Character of Life Phenomena. By J. Loeb. Compare this journal, 1902, vi, pp. 411-433. The Nature of Nerve Stimulation and Alterations of Irritability. By A. P. Mathews. Effects of Potassium Cyanide and of Lack of Oxygen on the Development of Sea-urchin Eggs. By E. P. Lyon. By invitation. The Formula for Determining the Weight of the Central Nervous System in Frogs of Different Sizes. By H. H. Donaldson. The Chemical Analysis of the Brain. By W. Koch. By invitation. The Mode of Action of Certain Substances on the Colored Blood- corpuscles, with Special Reference to the Relation between so-called Vital Processes and the Physico-chemical Structure OF Cells. By G. N. Stewart. A Convenient Rabbit- Holder. By G. Lusk. On the Surface Action of Metals. By F. G. Now. An Arterial Cannula and other new Physiological Apparatus. By G. P. Dreyer. Glycocoll in Gelatoses. By P. A. Levene. Read by title. A Contribution to the Physiology of the Thyroid Gland. By L. Breisacher. Read by title. On the Nucleic Acid of the Suprarenal. By W. Jones. Read by title. On Gluco-phosphoric Acid. By P. A. Levene. Read by title. Embryochemical Studies. II. — The Presence of Mono-amido-acids IN THE Developing Egg. By P. A. Levene. Read by title. Proceedings of the A^nerican Physiological Society, xxvii The_ Frontal Lobes (cerebral) and the Formation and Retention of Associations. By S. I. Franz. Read by title. The Movements of the Intestines Studied by Means of the Ront- gen Rays. By W. B. Cannon. This journal, 1892, vi, pp. 251-277. Read by title. The Relation of the Parathyroid to the Thyroid Gland. By W. S. Carter. Read by title. THE COMPOSITION AND CHEMICAL QUALITIES OF THE ALBUMOID [N BONE. By p. B. hawk and WILLIAM J. GIES (reported by W. J. GIES). (Read by title.) In the first report to this society of the discovery of osseomucoid attention was drawn to the fact that the method used for the prepara- tion of the glucoproteid would also favor a study of the albuminoid constituents of osseous tissue. The collaginous residue remaining after extraction of osseomucoid from ossein yields an insoluble, elastin-jike substance on boiling in water. This substance is neither the elastin of Smith nor the keratin of Brosicke, but appears to be almost or quite identical with Morner's chondroalbumoid. Although our product is digestible in pepsin-hydrochloric acid, it appears to be somewhat more soluble in dilute acid and alkali than chondro- albumoid. Unlike the latter body, however, it does not contain loosely bound sulphur. We have prepared a number of samples of osseoalbumoid from ossein by the method Morner used for the preparation of the albumoid substance in cartilage. The chief difficulty in this work has been the removal of phosphates and the preparation of ash-free products. Our analyses thus far indicate the average elementary composition given in the summary below, where comparison is also made with keratin and elastin. C H N S O Osseoalbumoid . . . 50.03 6 85 15.93 0.55 26.64 Ligament elastin . . 54.08 7.20 16.85 30 21.57 Hair keratin .... 50.65 6.36 17.14 5.00 20.85 xxviii Fourteenth Annual Meeting. Osseoalbumoid does not contain phosphorus. Unfortunately, analytic comparisons with chondroalbumoid are not now possible, as Morner made no analyses of that body, although he found that the nitrogen content (three determinations) of albuminates made from it varied between 15 and 16 per cent. We have obtained larger propor- tions of this residual substance from bone than from cartilage. It is our purpose to study chondroalbumoid in this connection also. A COMPAR.^TIVE STUDY OF THE REACTIONS OF VARIOUS MUCOIDS. By L. D. mead and WILLIAM J. GIES (reported by W. J. GIES). (Read by tide.) Comparative studies of many of the precipitation reactions of osseomucoid, chondromucoid and tendomucoid have shown thus far a very striking sameness in result. Each of these glucoproteids also is digested in pepsin-hydrochloric acid, with a formation of proteoses and peptones and the separation of nitrogen-containing substance rich in reducing material, probably chondroitin sulphuric acid or essentially the same body in each case. The microscopic appearance of the phenylosazone bodies obtained from each is the same as that of dextrosazone, indicating glucosamine among the products of acid hydration. All these compound proteids contain sulphur obtainable as sulphate and as sulphide. They are acid to litmus, neutralize alkali, have essentially the same elementary composition and yield practically the same amount of heat on combustion. In physical appearance the sub- stances whether dry, freshly precipitated, or in solution, are practically identical. Attempts to obtain crystalline mucoid, by the methods which recently have given such fruitful results in other connections, have thus far been without success. When the electric current is passed through neutral or alkaline mucoid solutions (consisting of sodium or calcium salts of mucoids) turbidity results within a short time, and flocks eventually form and can be filtered off. Our studies in this general connection have not been completed. We are convinced, however, that the connective tissue mucoids are practically identical substances. Proceedings of the American Physiological Society, xxix ARE PROTEIDS WHICH ARE PREPARED BY THE USUAL METHODS COMBINED WITH FAT OR FATTY ACID? By E. R. POSNER and WILLIAM J. GIES (reported by W. J. GIES). (Read by title.) Chemical analysis of the glucoproteids has resulted in wide varia- tions in the figures for elementary composition, not only for bodies from different sources, but for products of similar origin. Such variation has been attributed to admixture of impurities, particularly of non-nitrogenous character. Nerking's recent experiments with mucins, ovomucoid, and various simple animal and vegetable proteids indicate that possibly the mucin substances, and other proteids as they are commonly prepared, are admixed or combined with fat or fatty acid. In order thoroughly to test this matter we have analyzed numerous samples of "chemically pure" connective tissue mucoids and albuminoids. Using Dormeyer's method on quantities of proteid from 2 to 13 grams in weight, and following Nerking's procedure, our extractive results were always entirely negative. We are convinced, therefore, that the mucoids and albuminoids as they are prepared to-day are not " fat-proteid compounds." ON THE TOXICOLOGY OF SELENIUM AND ITS COMPOUNDS. By L O. woodruff and WILLIAM J. GIES (reported by W. J. GIES). (Read by title.) The researches of Tunnicliffe and Rosenheim indicate that the numerous cases of " arsenical poisoning " in England recently may have been due in part to selenium. Through the kindness of Prof. Victor Lenher our studies are being made with absolutely chemically pure preparations. Thus far our results on dogs confirm most of the general observations of Rabuteau, and of Czapek and Weil. We are unable, however, to discover Rabuteau's crystals in the blood of the heart after death, or to agree with him that death results from mechanical interference with the circulation. Selenium is very much more toxic than tellurium, although its poisonous effects are qualitatively much the same. The expired XXX Fourteenth Annual Meeting, methyl compound of selenium is produced in much less quantity than that of tellurium under similar conditions. Injection of four milligrams of selenite or selenate per kilo under the skin of dogs usually results in death in a few minutes. Speedy death follows the introduction of like amounts per os or rectum. Four grams of the finely powdered metal, when taken into the stomach, manifested no toxicity whatever, and passed out in the, faeces. The introduction of soluble salts is quickly followed by elimination of selenium in the urine and the breath. After subcutaneous injections, the distribu- tion of selenium to the organs is similar to that found by us recently for tellurium. Selenium, although chemically related to sulphur, is very much like arsenic in its toxic properties. INDEX TO VOL VI. A BSORPTION, fat, 17, 33t. -^^ , proteid, 22. Albumoid, bone, composition, xxvii. Alcohol, action on muscle, xiii. Allantoin, estimation, 39. , excretion, xiv. , formation from uric acid, 39. Apparatus, physiological, xxiv, xxv. Arteaga, J. F. Phlorhizin diabetes in cats, ^73- Associations, formation and retention, xxvii. Atropine, action, 207. Autotomy, relation to reflexes in hermit- crab, 278. "DACTERIA, reaction to stimuli, 31. -*-' Benedict, F. G. The excretion of nitrogen during nervous excitement, 398. Blood-corpuscles, acted upon by certain substances, xxvi. Blood plates, constituents, xi. , relation to red corpuscles at high al- titudes, xi. Blood-pressure, determined in man, xxii. Blood-salts influence contractions of heart, 181. Brain, chemical analysis, xxvi. Breisacher, L. A contribution to the physiology of the thyroid gland, xxvi. Buerger, L. and W. J. Gies. The chemi- cal constituents of tendinous tissue, 219. BuLLARD, W. N. See Cannon, 91. /^ANNON, W. B. Cerebral pressure ^-^ following trauma, 91. Cannon, W. B. The movements of the in- testines studied by means of the Rontgen rays, 251, xxvii. Cannula, arterial, xxvi. Carter, W. S. The relation of the para- thyroid to the thyroid gland, xxvii. Cell nucleus, role in oxidation and synthesis, XV. Cell, vital processes and physico-chemical structure, xxvi. Cellulose, digestion, xiii. Cerebral pressure following trauma, 91. Chemotaxis, T/Z- Crosby, J. H. See Jennings and Crosby, 31- Cross-fertilization, 216. CusHiNG, H. Concerning the poisonous effect of pure sodium chloride solutions upon the nerve-muscle preparation, 77. CusHNY, A. R. On the glands of the oviduct in the fowl, xviii. CusHNY, A. R. On saline diuresis, xvii. Cutter, W. D., and W. J. Gies. The composition of tendon mucoid, 155. T^AWSON, P. M., and E. N. Riggins *~^ An attempt to obtain regeneration of the spinal cord, xxi. Defecation, 269. Development, index for egg of domestic fowl, 351. of sea-urchin eggs affected by potas- sium cyanide and lack of oxygen, xxvi. Dextrose, origin from cellulose in digestion, xiii. Diabetes, 173. Diuresis, xvi. Donaldson, H. H. The formula for de- termining the weight of the central ner- vous system in frogs of different sizes, xxvi. Dreyer, G. p. An arterial cannula and other new physiological apparatus, xxvi. "Tj'ARTHWORM, movements influenced -*--' by temperature, odors, light, and contact, 458. Edestin, decomposition products, 48. Edwards, C. L. The physiological zero and the index of development for the egg of the domestic fowl, Gallus domesticus. A contribution to the subject of the in- fluence of temperature on growth, 351. XXXll Index. Edwards. G. H. See Henderson and Edwards, xxii. Egg, presence of mono-amino-acids during development, xxvi. Electrical charge of ions influences physi- ological action, 411. Embryo, normal measurements, 351. Enzyme, spermatozoa, 53. Ergograph, xxiii. Eklanger, J. A new instrument for deter- mining systolic and diastolic blood-pres- sure in man, xxii. Erlanger, J., and A. W. Hewlett. A study of the metabolism in dogs with shortened small intestines, i. "PAT absorption, 17, 331. -*- Formaldehyde, action, 325. Franz, S. I. The frontal lobes (cerebral) and the formation and retention of asso- ciations, xxvii. Frontal lobes (cerebral), associations, xxvii. r^ ELATOSES, glycocoll, xxvi. ^-^ GiES, W. J. Do spermatozoa con- tain enzyme having the power of causing development of mature ova ? 53. GiES, W. J. See Buerger and Gies, 219. GiES, W. J. See Cutter and Gies, 155. Gies, W. J. See Hawk and Gies, xxvii. Gies, W. J. See Mead and Gies, xxviii. Gies, W. J. See Posner and Gies, xxix. Gies, W. J. See Woodruff and Gies, xxix. Gluco-phos])horic acid, xxvi. Glycocoll in gelatoses, xxvi. Gonionemus, sensory reactions, 434. Good, C. A. The excretion of lithium, xx. Greeley, A. W. On the analogy between the effects of loss of water and lowering of temperature, 122. Greeley, A. W. Artificial parthenogene- sis in starfish produced by a lowering of temperature, 296. Greene, C. W. Contributions to the physiology of the California hagfish, Polistotrema stouti. — W. The absence of regulative nerves for the systemic heart, 318. Greene, C. W. Notes on the physiology of the circulatory system of the California hagfish, Polistotrema stouti, xii. Growth, influence of temperature, 351. T T AG-FISH, circulatory system, xii. ^ ■*■ Hag-fish, innervation of heart, 318. Hall, W. S. A new form of ergograph, xxiii. Halsey, J. T. Studies in diuresis, xvi. Hawk, P. B., and W. J. Gies. The com- position and chemical qualities of the albumoid in bone, xxvii. Heart, fibrillar contraction, xxv. , hag-fish, innervation, 318. Heart-beat, influenced by sodium, xxv. Heart-muscle, influence of salts on auto- matic contractions, 181. Henderson, Y., and G. H. Edwards. A study of metabolism in a case of lym- phatic leukaemia, xxii. Hewlett, A. W. See Erla.nger and Hewlett, r. Howell, W. H. An analysis of the in- fluence of the sodium, potassium, and calcium salts of the blood on the auto- matic contractions of heart-muscle, 181. Hunt, R. Experiments with Zygadenus venenosus (poison camass), xi.x. Hunter, S. J. On the production of arti- ficial parthenogenesis in arbacia by the use of sea-water concentrated by evapora- tion, 177. T LEOC/ECAL valve, competence, 264. -*■ Infusoria, reaction to acids, 233. Intestine, course of food in, 262. , shortened in dogs, i. Intestines, movements, 251. Ions, relation of electrical charge to physi- ological action, 411. JENNINGS, H. S., and J. H. Crosby. J Studies on reactions to stimuli in uni- cellular organisms. — VII. The manner in which bacteria react to stimuli, espe- cially to chemical stimuli, 31. Jennings, H. S., and E. M. Moore. Stud- ies on reactions to stimuli in unicellular organisms. — VIII. On the reactions of infusoria to carbonic and other acids, with especial reference to the causes of the gatherings spontaneously formed, 233. Jones, W. On the nucleic acid of the suprarenal, xxvi. T7' EMP, G. T. Relation of blood plates -*-^ to the increase in the number of red corpuscles at high altitudes, xi. Kemp, G. T., and O. O. Stanley. Some new observations on blood plates, xi. Koch, W. The chemical analysis of the brain, xxvi. Koch, W. The physiological action of formaldehyde, 325. Index. XXXI 11 T EE, F. S., and W. Salant. The action "^^ of alcohol on muscle, xiii. Leukaemia, metabolism, xxii. Levene, p. a. On gluco-phosphoric acid, xxvi. Levene, P. A. Embryo-chemical studies, IL — The presence of mono-amido-acids in the developing egg, xxvi. Levene, P. A. GlycocoU in gelatoses, xxvi. Levene, P. A., and L. B. Mendel. Some decomposition products of the crystallized vegetable i^roteid edestin, 48. Lewis, W. H. See Loek and Lewis, 305, xxvi. Life, prolonged in unfertilized eggs of sea- urchins by potassium cyanide, 305. LiLLlE, F. R. Notes on regeneration and regulation in planarians (continued), 129. LiLLiE, R. S. The role of the cell nucleus in oxidation and synthesis, xv. LiNGLE, D. J. On further experiments on the importance of sodium for the heart- beat, XXV. Lipase, fat metabolism, 331. Lithium, excretion, xx. LOE15, J. Studies on the physiological effects of the valency and possibly the electrical charges of ions. I. — The toxic and antitoxic effects of ions as a function of their valency and possibly their electrical charge, 411. LoEB, J., and W. H. Lewis. On the pro- longation of the life of the unfertilized eggs of sea-urchins by potassium cyanide, 305, xxvi. LoEB, J. The physiological effects of the electrical charge of ions and the electri- cal character of life phenomena, xxvi. LoEVENHART, A. S. On the relation of lipase to fat metabolism — lipogenesis, 331- Lombard, W. P. Demonstration of ap- paratus, xxiv. LuSK, G. A convenient rabbit-holder, xxvi. LusK, G. On the question whether dex- trose arises from cellulose in digestion, xiii. Lyon, E. P. Effects of potassium cyanide and of lack of oxygen on the development of sea-urchin eggs, xxvi. TX/TATHEWS, A. P. Artificial parthe- ■'■*-*- nogenesis produced by mechanical agitation, 142. Mathews, A. P. The action of pilocarpine and atropine on the embryos of the star- fish and the sea-urchin, 207. Mathews, A. P. The so-called cross fer- tilization of Asterias by Arbacia, 216. Mathews, A. P. The nature of nerve stimulation and alterations of irritability, xxvi. Mead, L. D., and W. J. Gies. A com- parative study of the reactions of various mucoids, xxviii. Mechanical agitation, parthenogenesis, 142. Mendel, L. B. New experiments on allan- toin excretion, xiv. Mendel, L. B. See Levene and Men- del, 48. Metabolism, fat, 331. , influenced by nervous excitement, 39S. , in dogs with shortened intestine, i. , leukaemia, xxii. , nitrogen, 39S. Metals, surface action, xxvi. Milk, liberation of sulphide on heating, 450. Moore, E. M. See Jennings and Moore, 23.1- Morgan, T. H. The reflexes connected with autotomy in the hermit-crab, 278. Mucoids, reactions, xxviii. "^T ERVE stimulation and irritability, xxvi. ■'- ^ Nervous system of gonionemus, 434. Nervous system, weight, xxvi. NovY, F. G. On the surface action of metals, xxvi. Nucleic acid, suprarenal gland, xxvi. /'^^'A, development, 53. ^-^ Oviduct, glands in fowl, xviii. "DANCREAS, relation to spleen, xiv. ■^ Parathyroid, relation to thyroid, xxvii. Parthenogenesis, produced by concentrated sea-water, 177. , produced by lowering the temperature, 296. , produced by mechanical agitation, 142. Peristalsis, intestine, 260. Phlorhizin diabetes, 173. Physiological zero for egg of domestic fowl, 351- Pilocarpine, action, 207. Planarians, regeneration, 129. Porter, W. T. Demonstration 01 appara- tus, XXV. Porter, W. T. The mechanism of fibrillar contraction of the heart, xxv. Posner, E. R., and \V. J. Gies. Are pro- teids which are prepared by the usual XXXIV Index. methods combined with fat or fatty acid ? xxix. Proceedings of the American Physiological Society, xi. Proteid absorption, 22. Proteids, combined with fat or fatty acid, xxix. Pulmonary circulation, 283. "P ABBIT-HOLDER, xxvi. -*■*- Regeneration in planarians, 129. Rettger, L. F. The liberation of volatile sulphide from milk on heating, 450. Rettger, L. F. Experiments on the rela- tion between the spleen and the pancreas, xiv. RiGGiNS, E. N. See Dawson and Riggins, xxi. C ALANT, W. See Lee and Salant, *^ xiii. Secretion, kidneys, xvi, xvii. , nerve-control, 207. Selenium, toxicology, xxix. Smith, A. C. The influence of tempera- ture, odors, light and contact on the movements of the earthworm, 458. Sodium chloride, poisons nerve and muscle, 77- Spermatozoa, enzyme, 53. Spinal cord, regeneration, xxi. Spleen, relation to pancreas, xiv. Stanley, O. O. See Kemp and Stanley, xi. Stewart, G. N. The mode of action of certain substances on the colored blood- corpuscles, with special reference to the relation between so-called vital processes and the physico-chemical structure of cells, xxvi. Surface action, metals, xxvi. Suprarenal gland, nucleic acid, xxvi. Sw.\iN, R. E. The formation of allantoin from uric acid in the animal body, 38. T^ENDINOUS tissue, chemical constit- ■^ uents, 219. Tendon mucoid, 155. Thyroid gland, physiology, xxvi. Thyroid, relation to parathyroid, xxvii. T TNICELLULAR organisms, reaction ^ to stimuli, 31, 233. V ALENCV, physiological effects, 411. WATER loss, effects analogous to those produced by lowering the tem- perature, 122. Wood, H. C, Jr. A physiological study of the pulmonary circulation, 283. Woodruff, I. O., and W. J. Gies. On the toxicology of selenium and its com- pounds, xxix. A/ERKES, R. M. A contribution to the ^ physiology of the nervous system of the medusa Goriionemus Murbachii. Part I. — The sensory reactions of Gonionemus, 434. '7 YGADENUS venenosus^ death camass, THE American Journal of Physiology. VOL. VI. SEPTEMBER i, 1901. NO. I. A STUDY OF THE METABOLISM IN DOGS WITH SHORTENED SMALL INTESTINES. By JOSEPH ERLANGER and ALBION WALTER HEWLETT. [From the Physiological Laboratory of the Johns Hopkins University ?^ CONTENTS. Page Introduction 1 General remarks on dogs with shortened intestines 5 Diet and methods 6 The urine of dogs with shortened smr'll intestines 9 The faeces ' 15 Summary 25 Conclusions 26 Introduction. THE effect of excision of various parts of the gastro-intestinal canal upon metabolism and upon the nutritive value of the food-stuifs has been a subject of considerable study by numer- ous observers. Such studies possess a double interest. On the one hand the physiologist looks to this method of excision for the light that it may throw on the normal function of the various parts of the intestinal tract, and, on the other hand, the operation is of considerable practical interest to the physician and to the surgeon. For the surgeon it points out the limits, or, more correctly, enlarges the field of his encroachment on the gastro- intestinal canal ; and for the physician it serves as a guide to the proper nutrition of patients when parts of the gastro-intestinal canal have been removed from functional activity either by the hand of the surgeon or by the inroad of disease. 2 Joseph Erlaiiger and Albion Walter Hewlett. The effect of excision of the stomach has been more or less care- fully studied in a few cases, which have been collected by Deganello.^ They include three cases in man and two series of experiments on dogs. In a few of these the nutritive value of the food-stufFs was investigated. The disturbances in absorption were but slight. In general, nitrogen absorption was affected shortly after the operation, but after the lapse of a certain interval of time, four or more months, digestion seemed to proceed almost normally. The ratio of the ethereal to the alkaline sulphates was increased in Deganello's case, but even this disturbance tended to disappear in the course of time. It is interesting to note in connection with Deganello's case that although the ratio above mentioned was increased there was no absolute increase in the amount of ethereal sulphates eliminated in the urine. The effect of throwing the pylorus out of function has been made the subject of careful work by Rosenberg.^ In a certain number of dogs upon which the operation of gastro-enterostomy had been per- formed, he found that more than the normal amount of nitrogen, fat, and carbohydrates escaped absorption. He believes that the explana- tion of these disturbances lies in the altered relative action of the digestive juices. Under normal conditions the chyme passes into the intestine in small portions. When the controlling action of the pylorus is removed large quantities of acid gastric juice are poured into the intestine at short intervals and the inorganic acid in excess probably alters the action of the intestinal juices. A considerable part of the large intestine has been excised in man by Treves.^ The patient recovered rapidly from the effects of the operation. The composition of the faeces was not studied. A simi- lar operation has been performed on dogs by Harley.^ He subjected his animals to a thorough series of feeding experiments and this makes his work in many ways more complete than any of this nature that has heretofore been done in the same field. As we shall often refer to his observations throughout this paper, an account of his results would here be out of place. Complete excision of the small intestine has never been a success, so far as we are aware. The difficulties of such an operation are prac- ^ Deganello: Archives italiennes de biologic, 1900, xxxiii, p. T18. '^ Rosexberg: Arcliiv fiir die gesammte Physiologie, 1S9S, Ixiii, p. 403. ^ Treves: Lancet, London, 189S, No. i, p. 276. ■* Harley: Proceedings of the Royal Society, London, 1899, Ixiv, p. 255. Metabolisin in Dogs with Shor levied Small Intes tidies. 3 tically insurmountable. However, in dogs a very large part of the small intestine has been removed successfully. Senn,^ as a result of his work on dogs and cats, believed that the removal of more than one third of the small intestine results fatally sooner or later. Trzebicky- believes that the removal of one half of the ileum of dogs is fatal, and that the danger increases the nearer to the stomach the excision is made. Monari ^ states that he has successfully removed from a dog seven eighths of the small intestine. Monari's dog is of especial in- terest to us because so far as we know it is the only dog upon which metabolism experiments have been performed after excision of the small intestine. At the autopsy on this dog only 28 cm. of small intestine were found. De Filippi,* who made the metabolism experi- ments, found that the animal was capable of nourishing itself almost normally. The only change referable to the operation seemed to be an incomplete absorption of fat (19 per cent appeared in the faeces). Carbohydrates were completely used and no more nitrogen escaped absorption than in a normal dog used for comparison. Records of extensive resection of the small intestine of man are not numerous. In no case does the relative length of bowel resected approach that of Monari's dog nor that of the dogs that have been the subjects of our experiments. Dreesmann,^ in 1899, was able to collect twenty-six operations on men in which more than one metre was removed. Of those that survived the operation only four had symptoms clearly referable to the intestine and these symptoms con- sisted only of a tendency to mild diarrhoea. In one of the four cases three metres and in the others somewhat more than two metres of bowel were resected.'' The largest resection was 330 cm. It is very interesting to note that this patient showed no subjective symptoms referable to the operation. The ability to absorb food-stuflfs was not investigated. To these cases of Dreesmann we are able to add one reported by Schlatter,' and one unreported case of Mitchell. Mitchell's patient died as a result of the operation, but the case is ^ Senx : Experimentaler Beitrag zur Darmchirurgie, cited by Dreesmann ; Berliner klinische Wochensclirift, 1899, xxxvi. p. 337. - Trzebicky : Archiv fiir klinische Chirurgie, 1894, xlviii, p. 54. ^ MONARi : Beitrage zur klinischen Chirurgie, 1896, xvi, p. 479. * De Filippi : Archives italiennes de biologic, 1894, xxi, p. 445. ^ Dreesmann: Berliner klinische Wochenschrift, 1899, xxxvi, p. 337. ^ Fantino: Gazetta medica di Torino, 1896, xlvii, p. iSi; abstract in Central- blatt fiir Chirurgie, 1896, xxiii, p. 614. ■^ Schlatter: Correspondenzblatt fiir schweizerische Aerzte, 1899, xxix, p. 417. 4 Joseph Erlanger and Albion Walter Hewlett. interesting from the fact that a larger proportion of bowel was removed than in any other operation on record. Approximately ten feet of small intestine were removed on account of gangrene from infarction. A faecal fistula was established by suturing both ends of the small intestine to the abdominal wall. The patient gradually sank without any apparent cause and died on the tenth day. At the autopsy it was found that all of the small intestine had been removed except about six inches of the ileum in the neighborhood of the ileo-c?ecal valve and about one foot of intestine adjacent to the stomach. In only three cases has the metabolism of these patients been studied. An investigation of the metabolism of Schlatter's patient was made by Dr. Plaut one month after the operation. The patient, aged twenty-three, was allowed to eat ad libitum. Great quantities of food were consumed, the average for nine days containing 3 1 .8 grams of ni- trogen and 109 grams of fat per day. 10.47 P^"" cent of the nitrogen ingested and 13.91 per cent of the fat appeared in the faeces. The nitrogen loss was at the upper limit of normal, the loss of fat consider- ably above the normal under such circumstances (4 to 6 per cent). Although the patient did not suffer from diarrhoea he never regained his former activity, and eight months after the operation he could walk but slowly and with rests, and could eat only bouillon, soups, and meat. Riva Rocci ^ studied Fantino's patient, from whom 310 cm. of small intestine had been removed. The stools were more frequent than normal and contained a larger percentage of fats. On a diet containing 14.6 grams of nitrogen and 36 grams of fat, 29 per cent of the nitrogen and 23 per cent of the fat appeared in the faeces. As can be seen from our results these losses are the more remark- able because the diet was so restricted. The patient, although sixty years old, prevented loss of weight by taking a very ample diet. Giovanni Sagini^ investigated the exchange of materials in Ruggi's patient. The motor function of the intestine was normal. In one series of experiments the loss of nitrogen was 5.9 per cent of that ingested, the loss of fat 12.1 per cent. In a second series of experiments the nitrogen loss was 13.2 per cent of the ingesta, while of the fat 15.3 per cent escaped absorption. ^ Riva Rocci : Gazetta medica di Torino, [896, xlvii, p. 121. '■^ RUGGI : II policlinico, sezione chirurgica, 1896, iii, p. 49. Metabolism m Dogs with Shortened Small Intestines. General Remarks on Dogs with Shortened Intestines. In the early part of 1900 Flint and Rand ^ conducted a series of experiments in the Anatomical Laboratory of the Johns Hopkins University to determine the limits to which intestinal resection could be carried in dogs and to study the anatomical results of such resections. Of the dogs operated upon by them three were alive in the autumn of 1900, and through the kindness of Mr. Flint and Mr. Rand the authors of this paper were allowed to use these dogs in the metabolism experiments which will be described below. For convenience we shall call these dogs No. i, No. 2, and No. 3, which numbers correspond to No. 2, No. 12, and No. 11 respectively of Flint and Rand's series. All the measurements of length of intestine were made at the time of operation before the resection was begun. In dog No. I two separate operations were performed. The total length of the ileum and jejunum as measured at the first operation was 232 cm. Of this 80 cm. were removed at the first operation and 84 cm. at the second, making a total of 164 cm. Thus, about 70 per cent of the combined jejunum and ileum were removed. How- ever, this percentage may be considered as only approximately cor- rect, for it was found at the second operation that the remaining small intestine had lengthened somewhat after the first operation. This dog was studied by us in December 1900, eight months after the second operation. In dog No. 2, 238 cm. were removed at one operation from a total of 289.6 cm. of combined jejunum and ileum, this being 82 per cent of the movable^ small intestine. This dog was studied by us in November 1900 about seven months after the operation. In dog No. 3, 298 cm. were removed at one operation from a total of 357.5 cm., being an exsection of 83 per cent of the movable small intestine. Dogs from which larger amounts of intes- tine were removed died. Following the operation each of the dogs developed diarrhoea and lost weight. In dog No. i, and dog No. 2, the nutrition gradually improved, the lost weight was recovered, and at the time of our experiments they appeared to be well nourished. They showed, however, a marked tendency to diarrhoea. This diarrhoea mani- ^ Flint and Ranu : results not yet published. 2 By movable .small intestine is meant that portion which is supphed with mesentery. 6 Joseph Erlanger and Albion Walter Hewlett. fested itself when the dogs were placed on the ordinary diet of scraps of meat from the kitchen which food contained much indiges- tible matter. When placed on the easily digestible diet used in our experiments the diarrhoea ceased. Dog No. 3 differed from the two other dogs in that it never re- gained its former state of nutrition after the operation. It looked lean, and was constantly below its normal weight. It had a vora- cious appetite and at the same time had an almost constant diarrhoea. The malnutrition appeared to be due simply to a failure to absorb a sufficient quantity of nourishment from its intestines. This dog was the only one of the three that seemed much affected by the extensive resection of small intestine. Unfortunately, we were pre- vented from making exact feeding experiments upon this dog. We began our work with it and in our inexperience placed it upon a diet which was insufficient to keep up its nutrition. The dog became emaciated rapidly. Finding that its condition was becoming serious we returned it to the former diet of refuse from the kitchen. The dog then developed a most severe diarrhoea and despite a voracious appe- tite and unquenchable thirst continued to lose rapidly in weight. Four days after the return to this diet the animal died. At autopsy the intestinal wound was found perfectly healed and no cause for death other than malnutrition was discovered. Thus, we were un- able to get any exact results from this dog which seemed to be suffering most as a result of the exsection of small intestine. In conjunction with the work on these dogs with shortened in- testines a similar series of experiments was performed upon a normal dog to be used for comparison. This normal dog can best be com- pared with dog No. 2, for both were of the same breed, of the same size and approximately of the same weight. Diet and Methods. General remarks. — In our study of the effects of exsection of the small intestine upon dogs No. i and No. 2 and our normal dog we have followed closely the work of Harley.^ We have often used his results as a standard of comparison with our own. In preparing our diagrams we have also used his figures in the construction of some of the curves, believing that we could thus show more strik- ^ Harley : Loc. cit. Metabolism in Do^s with SJiortened Small Intestines. v> ingly the differences in the results following the removal of small intestine as compared with those following the removal of the large intestine. Diet. — The diets used in the following determinations consisted of 150 grams of lean beef and 100 grams of dry soda biscuit to which a varying amount of olive oil was added. In order to secure uniformity in the meat moderately large quantities were well ground and mixed, then weighed out into 150 gram lots and placed in flasks. The flasks were sterilized in the autoclave for fifteen minutes at 120° C. and from each lot of meat the contents of two flasks were analyzed separately for nitrogen and ether extract and the results were averaged. The biscuits were weighed into lots of 100 grams, which were then wrapped and used as required. Two packages were analyzed for nitrogen and ether extract. The first diet for each dog consisted of 150 grams of the beef and 100 grams of soda biscuit. For convenience we shall call this diet A. The second, diet B, con- tained 150 grams of beef, 100 grams of biscuit, and 25 c.c. of olive oil (specific gravity 0.912). Diet C consisted of 150 grams of beef, 100 grams of biscuit, and 100 c.c. of olive oil. No limit was placed on the amount of water taken. Some was given in the food and in addition the dogs were allowed to drink as much as they desired about twice a day. Before beginning the analyses the dogs were kept at least three days on the diet to be used in order to approximate a condition of equilibrium. No difficulty was experienced in obtain- ing a complete consumption of the food in the case of dog No 2 and of our normal animal. Dog No. i, however, refused to eat all of its food while on the last diet containing 100 c.c. of olive oil. In order not to lose this series the nitrogen and ether extract was determined in the residue. This being subtracted from the amounts of nitrogen and fat in the total food, the remainder represents the total taken by the animal. Methods. — During the experiments the dogs were kept in zinc- lined cages. The urine was obtained from dog No. 2 and from our control normal dog by means of catheterization so far as this was possible. This operation is not difficult in the case of sluts but we were disappointed as to its value ; for even though we catheterized four times a day the animals would void into the cage at times. The method had the value, however, of sharply separating one day's urine from the next. Dog No. i was allowed to void into the cage entirely, for he could not be catheterized nor could he be trained to void when 8 Joseph Erlanger and Albion Walter Hewlett. taken from the cage. It will be noticed that the amount of nitrogen in the urine of this dog is somewhat below that voided by our other dogs. Such a loss is inevitable when special methods of collecting the urine cannot be employed. But as not less than four days were used in computing the average, these averages are probably relatively correct, although the absolute values are only approximate. The faeces were taken from the floor of the cages. In order to determine, accurately the faeces belonging to the period of experiment the dogs were given lampblack on the day preceding the period and again on the last day of the period. We then began with the faeces first appearing after the blackened faeces and ended with the last of the second lot of blackened faeces. In this way the total amount of faeces obtained was correctly marked off, although the amount obtained from day to day varied considerably. The total nitrogen was determined by the Kjeldahl method as modified by Argutinsky.-^ The amount of water in the faeces was determined by drying at ioo° to I io° C. to constant weight. The ether extract was obtained by the method described by E. Voit.^ It consists of a twenty-four hour extraction of the acidified and dried material in a Soxhlet apparatus with ethyl ether, evaporation of the ether, re-extraction with petroleum ether, and weighing. As is done in most metabolism experiments we shall call this ethereal extract, fat. The alkaline and ethereal sulphates were determined by the gravi- metric method.^ The determination of carbohydrates in the faeces was not systema- tically carried out during the course of the experiments because we found that they were entirely absent in dog No. 2 and in our normal dog when tested for by the method described by Hoppe-Seyler,* During the final series of experiments on dog No. i Pavy's method of testing for carbohydrates^ was applied to the faeces after the fat had been extracted. About 10 grams of faeces were used. A very well ^ Argutinsky : Archiv fur die gesammte Physiologie, 1890, xlvi, p. 581. 2 E. VoiT: Zeitschrift fiir Biologic, 1897, xxxv, p. 555. ^ Salkowski: Virchow's Archiv fiir pathologische Anatomie, 1880, Ixxix, p. 551. * Hoppe-Seyler : Handbuch der physiologisch- und pathologisch-cliemischen Analyse, Berlin, 1883, 5te Auflage, p. 504. ^ Pavy: Physiology of the carbohydrates, London, 1894, p. 61. Metabolism in Dogs with Shortened Small Intestines. 9 marked reduction of Fehling's solution resulted after boiling with sulphuric acid, although when tested for by the method of Hoppe- Seyler carbohydrates were apparently completely absent. The reduc- ing substances thus indicated must have been formed during the treatment of the original substance with the reagents employed in Pavy's method. They may have been derived either from the normal mucin of the faeces,^ from the proteids or from the cellulose and allied substances (hemi-cellulose and dextrane). A small amount of the proteids ^ and of the cellulose may have come from the food-stufifs directly. But a larger amount may have come from the large number of bacteria present in the faeces.'^ It is quite impossible in our pres- ent state of knowledge to decide these questions. This same uncer- tainty is reflected in the literature. While it is generally agreed that no starch granules can be found in the faeces of an animal placed on an easily absorbable diet,** statements differ as to the chemi- cal findings. For instance Harley ^ found that when his dogs were on a mixed, easily absorbable diet the faeces contained no carbohy- drates even after resection of the large bowel. He does not state his method. On the other hand, Tsuboi^ experimenting on dogs with practically the same amount of carbohydrates found 0.57 gram per day in the faeces. Tsuboi's method of determining the carbohydrates consisted in treating the faeces with dilute acid and then applying Allihn's method. It is interesting to note in this con- nection that in the faeces of starving animals Tsuboi found no sub- stances capable of reducing copper sulphate. The Urine of Dogs with Shortened Small Intestine. Amount.— In the interpretation of the following figures only the most obvious results will be noticed. Such a series of experiments contains so many sources of error that finer deductions from the figures hardly seem to be justified. The urine of dogs deprived of a large part of the small intestine 1 VoN Jaksch : Klinische Diagnostik, 1896, 5te Auflage, p. 277; Pfeiffer : Journ. Landw. 1885, xxxiii, p. 535, cited by Atwater and Langworthv : Bulletin 45, United States Department of Agriculture, 1898, p. 381. 2 Kermauner : Zeitschrift fiir Biologie, 1897, xxxv, p. 316. 3 Sahli : Lelirbuch der klinischen Untersuchungsmethoden, 1899, p. 463. ^ Moeller: Zeitschrift fiir Biologie, 1897, xxxv, p. 291. ^ Harley: Loc. cit. ^ Tsuboi: Zeitschrift fiir Biologie, 1897, xxxv, p. 68. lo Joseph Erlanger and Albion Walter Hewlett. differs but little from that of the normal animal and others have shown that the amount of urine to diminish as fat is added to the diet, and that w amount there is a corresponding rise in specific ever, is not an invariable result of the additio Harley further showed that the same holds removal of the entire large intestine. , We have dogfs after extensive resection of small intestine Harley,^ Pugliese''^ in normal dogs tends ith this diminution in gravity. This, how- n of fat to the diet.^ true for dogs after found it true also for (Table I). TABLE I. Showing the excretion of water by way of the urine and faeces. Diet. Normal Dog. Dog No. I. Dog No. II. Urine. Water Urine. Water Urine. Water Amt. c.c. Sp. gr. faeces. Gms. Amt. c.c. Sp. gr. faeces. Gms. Amt. c.c. Sp. gr. faeces. Gms. A. (10-13 gms. fat) 268 1.040 24.3 183 1.038 15.0 268 1.031 33.0 B. (33-36 gms. fat) 175 1.040 41.6 LSI 1.039 25.0 241 1.043 36.6 C. (83-104 gmj .fat) 173 1.046 423 99 1.042 36.2 130 1.048 93.7 As the dogs were allowed to drink freely the coincident increase in the total amount of water in the fsces could hardly be said to be the cause of the diminution in the urine. As may be seen by the accompanying table the increased amount of water in the faeces could not account for the much larger diminution in the quantity of urine. It must be attributed to some effect of the oil in altering the metabol- ism of the body, an effect probably associated with a diminution in the amount of nitrogen excreted. Nitrogen. — The amount of nitrogen excreted by way of the urine in dogs deprived of small intestine shows no great deviation from that excreted by normal dogs. Harley showed that as fat was added to the diet the amount of nitrogen in the urine diminished both in normal dogs and in those deprived of the large intestine. And the same fact has been demonstrated in normal sheep by Wicke and Weiske.* We 1 Harley: Loc. cit. ^ Pugliese : Archiv fiir Physiologic, 1897, p. 473. ^ Laas : Zeitschrift fiir physiologische Chemie, 1895, xx, p. 233. * WiCKE and Weiske: Zeitschrift fiir physiologische Chemie, 1895. xxi, p 42 ; 1896, xxii, p. 137. Metabolis7n in Do •£ ^ H = en Ratio of ethereal to alkaline. A (10-13 fat) B (33-36 fat) C (83-104 fat) a4484 0.3824 03825 0.0535 0.0426 0.0472 1 : 7.4 1:7.9 1:7.4 0.3697 0.3275 0.2614 0.0577 0.0665 0.0440 1:5.7 1:4.3 1:4.8 0.3970 0.5686 0.3249 0.0629 0.1034 0.0432 1:5.4 1:4.6 1:6.5 Average 0.4044 0.0477 1:7.6 0.3195 1 0.0557 1:4.9 0.4302 0.0698 1:5.2 There is, therefore, an increase in intestinal putrefactive processes after the removal of a large part of the small intestine. This may be explained on the supposition that more food material reaches the large intestine and that it resides longer in the large intestine than in the normal dog. Consequently, the putrefactive changes there are greater in amount than is normal. In normal animals it has been noted by Laas (Joe. cit.') that the addition of fat to the diet is practically without constant result; that the addition of fat to the diet does not increase or diminish the amount of intestinal putrefaction. But the addition of fat to the diet of dogs with resected large intestine showed a decided tendency to diminish putrefaction. In our observations the diarrhoea caused by adding large quantities of fat to the diet obscures this point. If we consider only the results obtained when our animals were on diets A and B containing relatively small amounts of fat, a tendency of fat Metabolism in Dogs with Shortened Small Intestines. 15 to increase the total amount of ethereal sulphates eliminated be- comes evident. (Fig. i.) It is not unreasonable to suppose that this is brought about by a tendency of the fat either to hurry the food through the shortened small intestine which is already on the border of functional insufBciency or to retard the absorptive processes in the small intestine. The effect would be to permit bacteria of the large intestine to act on the larger amount of unabsorbed food material. This fact probably explains in part the effect of an increase of fat in the diet of dogs with shortened intestine in increasing rather than in diminishing the total elimination of nitrogen. (See Table III.) The F/ECes. The faeces in our dogs after large intestinal resections varied greatly in character. When on a mixed diet largely made up of kitchen refuse dog No. 3 had an almost constant severe diarrhoea, with yellowish fluid stools. The other dogs also suffered to a lesser extent from diarrhoea on this diet. When, however, dog No. 3 was placed on an easily absorbable and, as was afterwards seen, insuffi- cient diet, the diarrhoea was quickly checked so that the dog had only four stools in nine days and these were very hard. Likewise, when dogs No. i and No. 2 were placed on an easily absorbable diet the diarrhoea was quickly checked and the dogs showed some ten- dency to constipation. When a larger amount of oil was added to the food the faeces became softer and the stools more frequent. Dog No. I had a single, somewhat firm, well formed stool each day when on diets A and B. Diet C made the stools very soft and two or three per day were not at all unusual. Dog No. 2 had profuse fluid stools after being given a large amount of oil. The fluidity of these stools, as the analyses show, was not due to the high percentage of water present but rather to the large percentage of oil, which amounted to one fourth of the weight of the dry fasces. The faeces possessed a characteristic glistening silky appearance described as due to the presence of fat. The faeces in the control dog changed but little in character as fat was added to the diet. They became somewhat more copious and also somewhat softer but presented none of the marked changes seen in the faeces of the animals that had been operated upon. Microscopical examination. — Microscopically, there seems to be very little qualitative difference between the stools of normal dogs 1 6 Joseph Erlanger and Albion Walter Hewlett. and those of dogs with shortened intestines. In both when on the same diet and on different diets muscle fibres in various stages of disintegration, fat globules and fatty acid crystals occur. We ex- amined simultaneously the faeces of our normal dog when on diet A, and the faices of dog No. 2 when on diet B. Both contained the same elements as mentioned above and it was difficult to decide that these elements were more abundant in the one than in the other. Reaction. — The reaction of the faeces to litmus was not constant in dogs with large resections of small intestines. It was more fre- quently acid than alkaline, but the acidity or alkalinity was never marked. The reaction of the faeces seemed to be entirely without significance. Amount of water, — The percentage of water present in the stools varied considerably at different times in the experiment. The averages of our dogs with intestines removed did not differ materially from the averages of normal dogs. It may be noticed in the follow- ing table (Table IV), that the percentage of water falls somewhat as the fat in the diet is increased, although owing to the increased weight of the fasces the absolute amount of water present is in- creased. (See Table I.) TABLE IV. Showing the percentage of water in the faeces of a normal dog and of dogs with shortened small intestines. Diet. Normal Dog. Dog No. I. Dog No. II. Per cent. A (10-13 grams fat) B (33-36 grams fat) C (83-104 grams fat) 64 SO 71 64 64 61 75 69 70 Average 72 63 71 The greater variations in the normal dogs are due possibly to the few days taken to make up some of the averages. The average of figures for Harley's control animals is 65.8 per cent. It is interest- ing to compare with these results the average percentage of water in the faeces of dogs whose large intestines have been removed. Harley's figures are uniformly high, being 77.9 per cent for one dog Metabolism in Dos[s with Shortened Small Intestines. 17 and 81.3 per cent for the other, a marked increase in the amount of water in the stools. The accompanying diagram (Fig. 2) shows the curves for the percentages of water in the faeces on varying diets. On account of its irregularity, the curve from our control animal is omitted and the curves for Harley's normal dogs have been used instead. These figures show the importance of the large intestine in the absorption of water in the fsces. When the large intestine Fat in Diet in Grams ft, i 10 20 30 40 50 60 70 80 90 100 90 85 80 75 70 65 60 55 ^- j» , - ^' »•'' , .- ■ , *^* • "^ '■.^ ^ ^ ""^^ "-~-~^ •""i 1 . .-. •-^ L~^ ^^-___^ -1 ■llir~" --. __j____ — ,'" — — • Figure 2. — Showing the percentages of water in the faeces : — of normal dogs (Harley). of dogs after removal of large intestine (Harley). of dogs after extensive resection of small intestine (upper is Dog No. II; lower is Dog No. I). is removed the water in the faeces is increased, but when present, the percentage of the water in the faeces varies between 60 per cent and 75 per cent, even though a large amount of the small in- testine has been removed. The removal of the large intestine, therefore, allows more water to pass into the f.neces than the removal of a much longer piece of small intestine. Absorption of fat. — It is generally held that most of the fat in the diet is absorbed in the small intestine.^ The recent work of Ham- ^ Harley: Loc. cit.: Czerxv: Virchow's Archiv fiir pathologische Anatomie, 1874, lix, p. i6r; Deucher : Deutsches Archiv fiir klinische Medicin, 1S96, Iviii, p. 210. 1 8 Joseph Erlanger and Albion Walter Hewlett. burger,^ however, has shown that the large intestine is capable of absorbing fat with a degree of rapidity and completeness not ex- ceeded in the small intestine. Of especial interest, therefore, is the question of the behavior of animals with a large amount of small intestine removed toward increasing quantities of fat in their diet, De Filippi^ states that there was some increase in the fat of the faeces of his dog after removal of seven eighths of the small intestine. Plaut found in Schlatter's patient ^(192 cm. of small intestine removed) on a diet containing on an average 31.8 grams of nitrogen and 100.9 grams of fat, 13.91 per cent of the fat ingested in the fjeces. In a normal man only about 4-6 per cent should appear in the faeces on this diet. Riva Rocci working on Fantino's patient* (310 cm. removed) with a diet containing 14.6 grams of nitrogen and 36 grams of fat found that 23 per cent of the fat appeared in the faeces. Ruggi's patient failed to absorb from 12 to 15 per cent of the fat ingested. Before considering the amount of fat in the faeces of dogs from which a large percentage of the small intestine has been resected it seems best to consider the effect which increasing the amount of fat in the diet has upon the fat in the stools of a normal animal. Harley'' and Laas ^ have shown in dogs and Kayser'^ in men that when the fat in the diet is increased the absolute amount of fat in the stools may increase slightly, but that there is a greater relative absorption of fat by the intestines, or, in other words, that a smaller percentage of the fat taken appears in the faeces. A curve for one of Harley's dogs is represented in Fig. 3, which shows the downward direction of the curve for the percentage of unabsorbed fat when the amount of fat in the diet is increased. The figures for our normal dog correspond closely to Harley's figures. Upon the addition of a moderate amount of fat to the diet, there was a marked drop in the percentage of un- absorbed fat. When, however, there was a very large amount of fat in the diet the percentage of unabsorbed fat rose very slightly above the percentage obtained upon a moderate addition of fat to the diet. Thus, on diet A (13.4 grams fat) 8.9 per cent of the fat taken was unabsorbed. On diet B (33.8 grams fat) 3.4 per cent was unab- 1 Hamburger: Archiv fiir Physiologic, 1900, p. 433. - De Filippi: Loc. cit. ^ Harley: Loc. cit. 3 Schlatter : Loc, cit. * Laas : Loc. cit. * Fantino : Loc. cit. ' Kayser : Archiv fiir Physiologic, 1893, p. 371. Metabolism in Dogs with Shortened Small Intestines. 19 sorbed and on diet C (102. 3 grams of fat) 4.4 per cent was unabsorbed. The curve obtained from these figures is shown in Fig. 3. Where the large intestine has been removed the fat absorption does not differ materially from that in the normal dog. The direction of the curve is still downward and the percentages absorbed are approxi- mately the same. This apparent diminution of fat escaping absorption is probably the result of the fact that a certain amount of the ether extract in the faeces is derived from an excretion into the intestinal canal.^ Thus the ether extract of the faeces of starvation may amount to 0.67 gram per day 2 and Harley " has shown that after extirpation of the pancreas more fat is found in the small intestine than was ingested. It is impossible to say how much of the fat in the faeces of an animal on a mixed diet is excretory and how much has been derived from unab- sorbed fat. The amount of fat in starvation faeces is probably relatively greater than that in normal faeces, just as the nitrogen of starvation fasces is relatively greater.* The same argument that we shall offer to explain the event of an increase of nitrogen in the faeces of animals with shortened small intestine applies to the fat in the faeces. Should the amount of fat in the faeces of such an animal be in excess of the normal amount, there is every reason for believing that such an excess represents fat that has escaped absorption. After removal of the small intestine, our dogs showed an inability to absorb fat normally if a large quantity of it was present in the diet. Dog No. I with 70 per cent of the movable small intestine removed gave a curve of fat non-absorption which nearly approached the curve in the normal dog (Fig. 3). With a very small amount of fat in the diet the amount of fat in the faeces was approximately equal to that in the normal dog, about 8.5 per cent (Table V). When a moderate amount of fat was added to the diet there was a drop in the percentage of fat eliminated in the faeces similar to the drop which occurs in normal dogs. The fall, however, was not so pronounced, the percentage eliminated being 5.6 per cent as compared with 3.4 per cent for our normal animal. When the fat in the diet was still further increased, the percentage eliminated in the faeces rose rapidly 1 Fr.Vott: Zeitsclirift fiir Biologic, 1892, xxix, p. 325; Prausnitz: Zeitschrift fiir Biologie, 1S97, xxxv. p. 335. 2 TsuBOi : Loc. cit. ^ Harley : Journal of physiology, 1895, xviii, p. i. * C. VoiT : Hermann's Lehrbuch der Pbysiologie, iSSi. 20 Joseph Erlanger and Albion Walter Hewlett. and with an average of 83.4 grams of fat in the diet 15.9 per cent passed through the intestine unabsorbed, while our control animal on a larger fat diet absorbed all except 4.4 per cent of the fat ingested. Dog No. 2 with 82 per cent of the movable small intestine removed showed an inability to absorb fat normally even when very small amounts were present in the diet. On a diet containing 13.4 grams Fat inBiet in Grams 63 to ■si .1) 10 20 30 40 50 60 70 80 90 100 Z5% 20% 15% 10% 5% • .' .' ,' /■ y / .' ,'' ^ ^p . .' ,' ' ^. '' -•' y ' ,' « ^ '' •, \ ^^ ^^ ' < "^ .' ' V \, "■•' ^ s s^ -• ^1 ^ • ' — ■^ Figure 3- — Showing the percentages of the fat in the diet which appears in the faeces of : — 1. Normal dogs (upper, ours ; lower, Harley's). 2. Dogs after removal of large intestine. 3. Dogs after e.xtensive resection of small intestine. of fat 1 1.3 per cent was unabsorbed in dog No. 2 as compared with 8.9 per cent in the control animal. With a moderate amount of fat in the diet this inability to absorb is shown more clearly. Instead of the percentage of unabsorbed fat sinking as it does in a normal animal (3.4 per cent) there is a slight rise to 1 1.5 per cent (Fig. 3). When the fat was increased still further to 104.6 grams the amount Metabolism in Dogs with Shortened Small Intestines. 21 which was not absorbed by dog No. 2 made up 26 per cent of the amount ingested, whereas in the normal animal, only 4.4 per cent passed through unabsorbed. We may say, therefore, that in a normal dog practically all the fat in the diet is absorbed, even though very large amounts be ingested. After resection of from 70 to 83 per cent of the movable small intestine, however, a dog absorbs fat properly only when the fat is present in very small amounts in the diet. When large amounts of fat are present, from one sixth to one fourth of the amount ingested passes through unabsorbed (Table V). TABLE V. Showing the amount of fat in f^ces and the percentage of fat not absorbed. Diet. Normal Dog. Dog No. 1. Dog No. 2. Amount. Grams. Not absorb'd. Per cent. Amount. Grams. Not absorb'd. Per cent. Amount. Grams. Not absorb'd. Per cent. A (10-13 grams fat) B (33-36 grams fat) C (83-104 grams fat) 1.193 1.137 4.539 8.9 3.4 4.4 0.915 1.907 13.342 8.4 5.6 15.9 1.515 4.176 27.207 11.3 11.5 26.0 To recapitulate, dogs after removal of the large intestine absorb just as much fat as normal dogs. Increasing the fat in the diet of normal dogs and of dogs deprived of large intestine diminishes the percentage of fat which appears in the faeces. Small amounts of fat are absorbed by dogs with shortened small intestine in nearly normal per- centages. When large amounts of fat are added to the food, abnormally large percentages of fat escape absorption, even though the faeces be evacuated but once per day. When we consider that the length of the sm.all intestine removed was 164 to 298 cm., while the length of the large intestine was only about 60 cm.,^ we are not justified in saying from our figures that a given length of small intestine absorbs fat more completely than the same length of large intestine. On the other hand, looking at the small and large intestines as wholes, we can say that the small intestine is necessary for proper fat absorption, whereas, the large intestine is not necessary for fat absorption, although under exceptional circumstances such as in our dogs it doubtless assists in the absorption of fat. 1 Ellenberger and Baum : Anatomie des Hundes, 1891. 2 2 Joseph Erlanger and Albion Walter Hewlett. Absorption of proteids. — The percentage of the nitrogenous mate- rial of the diet which appears in the faeces of normal dogs may be taken as from 7 to 13. We shall speak of the nitrogenous contents of the fcxces as consisting of the unabsorbed nitrogen of the food and in the percentages shall calculate it as such. As is well known ^ a large amount if not all,^ of the nitrogen in the faeces resulting from a mixed easily absorbable diet consists of material excreted into the intestinal tract. When, however, we find the nitrogenous contents of the fasces increased as a result of removal of intestine such an increase, in the absence of an inflammation of the mucous membrane, comes in large part, we believe, from the unabsorbed nitrogen of the food. We must, however, take into consideration in this connection that the bacteria have an indirect influence on the utilization of food in that the material that the bacteria incorporate into themselves is lost to absorption, the bacteria being, so to speak, filtered off by the intestinal mucosa.^ Therefore with the thrift of the micro-organisms there ought to be a corresponding increase in the amount of nitrogen in the faeces. If the constant corresponding to the nitrogen excreted could be eliminated, the amount of nitrogen lost by our oper- ated dogs as compared with that lost by our normal dog would be relatively greater than that ratio as we have calculated it. To take an example, if on a certain diet a normal dog eliminated 0.8 gram of nitrogen in the faeces as compared with 1.6 grams in an- other dog with a large amount of small intestine removed, the latter would lose twice as much as the former. If, however, 0.4 gram (a low estimate) in each case had been excreted by the intestines the amounts of diet nitrogen lost would be 0.4 gram and 1.2 grams respectively, that is, in the dog after operation the loss in diet ni- trogen is three times that lost by the normal dog. As we have no means of computing what proportion of the nitrogen is ex- cretory we shall disregard this factor altogether. At the same time it is necessary to recognize that our figures err on the side of making the nitrogen losses in operated dogs relatively small when compared with those in normal dogs.* As we have already said, pre- cisely this same reasoning holds good in regard to fat losses. ^ TsuBOi : Loc. cit. 2 Prausnitz : Zeitschrift fiir Biologic, 1897, xxxv, p. 335. ^ Prausnitz : Loc. cit. * In this connection it is of interest to examine the tables obtained from our dog No. 3 from which 83 per cent of the small intestine had been removed. This Metabolism in Dogs with Shortened Small Intestines. 23 As we have said, the percentage of the nitrogenous material in the diet which appears in the faeces of normal dogs may be taken as from 7 to 13 per cent. As the fat in the diet is increased, the nitrogen of the diet remaining constant, the percentage of nitrogenous mate- rial absorbed remains approximately constant (Fig. 4). When the large intestine has been removed, the percentage of nitrogenous material eliminated in the faeces by Harley's dogs is high, 15,5 to 16.5 per cent on a diet containing but little fat. As the fat was increased Harley's two dogs varied somewhat. In one there dog was on a small, in fact, starvation diet consisting of 100 grams of lean beef and 50 grams of corn-meal and containing 4.45 grams of nitrogen and 5.24 grams of ether extract. The average daily elimination of nitrogen in the faeces was 0.494 gram, which amount possibly approximates the amount of nitrogen excreted from the intestines of our operated dogs. The rise in the daily excretion of nitrogen by way of the urine which was very marked toward the end is to be interpreted as a pre- mortal increase (Schultz : Archiv fiir die gesammte Physiologie, 1899, Ixxvi, p. 381). It is, however, more marked and extends over a longer period than is usual in such a rise. The dog died four days after returning to a sufficient diet. The urine was not examined for sugar. DOG No. III. Eighty-three per cent of movable small intestine removed. Diet insufficient to preserve equilibrium. Day. Weight. Kgms. Diet. Excretion of N. N. Grams. Fat. Grams. Fasces. Grams. Urine. Grams. 9 11.22 4.4.S 5.24 none 6.652 10 4.45 5.24 1.745 6.896 11 4.45 5.24 0.944 7.769 12 4.45 5.24 none lost 13 4.45 5.24 none lost 14 4 45 5.24 none 8.868 15 4.45 5.24 none 9.160 16 4.45 5.24 1.089 10.050 17 9.76 4.45 5.24 664 13.712 Averages . . 4.45 5.24 0.494 24 Joseph Erlanger and Albion Walter Hewlett. was a marked drop in the percentage eliminated in the faeces, while in the other the percentage remained persistently high (Fig. 4). The amount of nitrogen in the fjeces after excision of a large part of the small intestine in man has been studied by Plant/ by Riva Rocci,^and by Giovanni Sagini.^ De Filippi working on dogs found no change in the amount of nitrogen excreted. Plant in Schlatter's case found that the faeces contained 10.47 percent of the nitrogen ingested. Riva Rocci found that the faeces of Fantini's patient contained 29 per cent of the nitrogen ingested. Ruggi's patient showed in two series of experiments 5.9 per cent and 12. i per cent of the nitro- gen ingested in his faeces. In our dogs after resection of a large Fat in Diet in Grams =1 § ■S SJ -pi ?5% 10 20 30 40 50 60 70 80 90 100 zo 15% 10% 5% ^ ,j» • ^, --'' ' ,.- .--■ ,' , ■" - '■ ,'' -'' •, ,-' tr ,'' < \ " -~ -H 1 .'' •■' ,-'\ s a '' .'N — -. ''' S ,■ • • ■— ■ Figure 4. — Showing the percentages of the nitrogen in the diet which appear in the fasces : — of normal dogs (upper, our dog; lower, Harley's dog). of dogs after exsection of large intestine (Harley). of dogs with shortened small intestine. part of the small intestine there was not a great change in the per- centage of nitrogen eliminated in the faeces provided the dog was on a diet low in fat. In dog No. i with 70 per cent of the movable small intestine removed 9.9 per cent of the nitrogenous material 1 Plaut : Loc. cit. '^ Riva Rocci : Loc, cit. 3 RuGGi : Loc. cit. Metabolism in Dogs with SJiortcned Small Intestines. 25 ingested appeared in the faeces. In dog No. 2 with 82 per cent of movable small intestine removed 14 per cent of the diet nitro- gen appeared in the faeces, which is above the percentage in the normal animal and yet not so high as in dogs whose large intes- tines have been removed. As fat is added to the diet the dog with a shortened small intestine shows its inability to absorb food properly, for not only the fat itself, but the nitrogenous material as well, passes through the alimentary tract in larger amounts than in normal animals. Thus, on a diet containing a moderate amount of fat the percentage of nitrogen eliminated by our normal dog was 12 per cent, while of our other dogs, dog No. i eliminated 14 per cent and dog No. 2 17 per cent. On a diet containing a very large amount of fat 12.9 per cent of the nitrogen ingested was eliminated by our normal dog, 24.5 per cent by dog No. i and 24.4 per cent by dog No. 2. The curves obtained from the above figures show well the effects of excision of a large portion of the small intestine upon the absorption of nitrogenous material. (Fig. 4.) When there is a small amount of fat in the diet the amount of nitrogen in the faeces is about equal for normal dogs and for dogs with small intestines removed, but it is high for dogs with large intestines removed. As fat is added to the diet the amount of nitrogen in the faeces remains about the same or even grows less in the case of normal dogs and of dogs with large intestines removed. In dogs with the small intestines removed, however, there is an increase in the amount of nitrogen in the faeces proportionate to the increase of fat in the diet. With large amounts of fat in the diet one fourth of the nitrogen ingested passes out unabsorbed, a loss twice as great as in the normal animal. Summary. A dog with a large amount of small intestine removed behaves very much like a normal dog so long as it is on an easily absorbable diet which contains only a small amount of fat. The elimination of fat and nitrogenous material in the faeces may not exceed the elimination in the normal dog (dog No. i) or it may do so to a slight extent (dog No. 2). When, however, the fat in the diet is increased the insufficiency of the shortened intestine as an organ for absorption becomes very apparent. The amount of fat in the faeces may be 25 per cent of that ingested, while the nitrogen in the faeces may also amount to 25 per cent of the diet nitrogen. The 26 Joseph Erlanger and Albion Walter Hewlett. loss of fat is in itself of slight importance, for a large amount of fat is still absorbed and as a matter of fact our dogs increased in weight on a diet rich in fat. The diminution in fat absorption and espe- cially the loss of nitrogenous material expresses an interference in the power for absorption in such an animal. This lack may mani- fest itself under other conditions of diet. For example, dog No. 2 after its feeding experiments were over was placed on an unlimited diet of kitchen refuse. It developed a severe diarrhoea, became emaciated and only recovered when great care was taken with its further diet. The whole appearance of dog No. 3 was that of an animal suffering from malnutrition in consequence of its inability to absorb nourish- ment properly. When living on a diet of kitchen refuse it ate ravenously, had profuse diarrhoea and remained emaciated. It seems to us that the inert non-digestible material in such a diet might easily increase peristalsis so as to carry through a large proportion of the unabsorbable food into the large intestine. This may be quickly evacuated with a consequent loss of absorbable material to the animal or it may reside for a longer time in the large intestine ex- posed to the wasteful action of bacteria. Such an animal might starve to death, owing to its inability to cope with the indigestible material. Conclusions. (i) Dogs from which from 70 to 83 per cent of the combined jejunum and ileum have been removed may live indefinitely after recovery from the operation. Their nutrition may appear to be perfectly normal, or it may be so poor that even when eating rave- nously they do not seem to be able to keep well nourished. (2) Such dogs are peculiarly liable to be affected with diarrhoea which may be caused by a diet too rich in fat or one containing too much inert non-digestible material. This diarrhoea is of very serious moment to such a dog and may cause its death. (3) The urine of such dogs shows no great variation in quantity, specific gravity or nitrogenous contents from that of normal animals. (4) The conjugated sulphates in the urine are increased absolutely and relatively to the alkaline sulphates indicating an excess of in- testinal putrefaction. (5) The quantity of faeces varied in our two animals. In the dog from which 70 per cent of the movable small intestine had been removed there was no marked increase in the amount of fseces ; in Metabolis7n in Dogs with Shortened Small Intestines. 27 the dog from which 82 per cent had been removed the amount of faeces was increased. (6) The percentage of water in the faeces of dogs deprived of large amounts of small intestine may equal or only slightly exceed the percentage of water in the fseces of normal dogs. This is in con- trast with the increased percentage of water in the faeces of animals deprived of the large intestine. (7) On a diet poor in fat the dog with a shortened small intestine absorbs the fat as well or almost as well as a normal dog. As the fat in the diet is increased the fall in the percentage eliminated in the faeces which occurs in the normal animal may either occur to a lesser extent or may not occur at all in dogs deprived of small in- testine. With large amounts of fat in the diet 25 per cent of that ingested may appear in the faeces, whereas, in our normal dog only about 4.5 per cent appeared. (8) The addition of fat to the diet of a normal dog does not greatly affect the amount of nitrogenous material eliminated by the faeces. The addition of fat to the diet of dogs deprived of small intestine causes an increased elimination of nitrogenous material in the faeces. On a diet rich in fat the amount of nitrogen eliminated in the faeces may be double that eliminated by a normal dog, although on a diet poor in fat there is no great difference between the two. 28 Joseph Erlaiiger aiid Albion Walter Hewlett. r- 1 ^ rt n -1- -t- -t; 'i- VO O ro Ov t LO 1 , t^ -t- •2 "■■= t^ r-i l--^ t>l t^ CO CO t^ 1^ t^ 1>^ t^ -*^ — rt 03 fl> CS o-t: S Ov ^ • -t- ^ i^; o j^ 00 ro I rh 1^ )- . t^ CO d d ^ d d d d d d d d 0) I- ■•-; ■s "1^ - oi p til . VO t^ > u K— o o O o p O O VO ro O VO -t oi (>a CO t-i CO CV) d vd (NJ a^ . 1>; (Nl (Nl ro ro ro ro (SJ (Nl '"' ro ro o o G 75 )2; rt O 1— ( t-H (M , . . rt LO VO i^ LO VO 1^^ ,-H (Nl ro P ^^ ' ' Cvl (Nl (Nl Metabolism in Dogs ivith SJiortened Small Intestines. 29 1 io of eal t line. ^ -1- ,^ ^ nO ro CO (M (SI Cv CO \d -f LO -h -t -f ^ u^ LO -f ro -f ^ a S ro s ro iy~. (>a ro (^ ro LO -^ ro <^ -1- ^" p q p P p p p P P p p p C/3 i—i r-H ^ i-H i-H '"' '"' '"' '"' ^ .-H '^ ^ ^ ^ ^ ^ CM ^ CO ^. vO vC 00 ON On u ^ Tt- -+- t^ OD !>. ro ^ ro LO ro NO ^- -f d •i- ro LO NO (^ ON 00 d 06 d c^ o (SI (^ Q (N N (SI (SI 30 Joseph Erlanger and Albion Walter Hewlett. O ~^^ "oZ ^ ,—1 (» t^ ON -t- CO CN] CO CNl (M NO LO lO LO Ratio therea alkali •o LO 'l- ■+ • W-; ro LO ro »+- NO rl- NO NO NO ^" ^" ^_^ __^ • I— 1 ^ 1 ^ ^ ^ t-H __l ^_, ^ 1 . LO ro NO 'l- CNJ (M (NJ l^ t^ t^ CM CM ro NO •f- ON ro O) tX) ro ^' jj £ to NO NO -i- ON t^ O -t- ro ■i- 3 o O o o o o f—i O o o O O 2 C/3 w"^ o o o o o o o o o o O o o c:> "re c o CO ro -1- -i- ro ON ro u-j o CO ON ON NO CO NO CO ON s o 2 •+• NO CO >-o IT", ^ LO NO NO ro 00 (M co ro p H'O o O O O O o o O o o o O o O ui ir-j O LO t^ i>"j NO (Nl „ o o ^J- t^ O l^J ro IZ § i/~- O O On o NO On J—l ro t^ ro lO CO ON O l^ o (N t^ On ON 1^ W-; NO o uo -1- W-) -^- •i- LO -+- LO lO •+ -J- 'J- -^ Tf •^ ^ ^ co t^ nT) o t^ ro ro >^ t^ ro o CO yj r- NO CNl <6 .,__, ,__( ,^_, o (M ,__( - ^ (M -1- LO ro Tt- 'l- ON t^ . LO O ;?'l • • -t- o CM fe • tn O ro ro On rO rO ^ (X> -1- ,_! I^ t^J CO -1- NO s! ■^ ON NO . t^ NO NO NO t^ t^ ^ " ^ !« O o o o o O o o CO O o o o t^ On S £ ro ? ro LO t^ -t- I/-; o yj~j ij^j o ro CO ON ro . o o o o o o '^ • • •^ CNJ • (M NO NO l) ro • ro NO ro • NO ro s s c n O NO NO : NO NO l-H NO bog t^ On l^ (M W^ 'l- LO -1- ro U-, -t- CO k;^ O • On On On ON ON ON >, vO t>. 00 On o »^. NO t^ CO On LT. NO ^^ r— * l-H .-H CM CNJ ^~* STUDIES ON REACTIONS TO STIMULI IN UNICELLU- LAR ORGANISMS.— VII. THE MANNER IN WHICH BACTERIA REACT TO STIMULI, ESPECIALLY TO CHEMICAL STIMULI. [From the Zoological Laboratory of the Unmersity of Michigan, Jacob Reighard, Director.] By H. S. JENNINGS and J. H. CROSBY. IN earlier numbers of this series of studies the manner in which ciliate and flagellate infusoria react to various stimuli has been described. It has been shown, especially in the second,^ fourth,^ and fifth '^ of these papers, that the so-called tactic phenomena of these organisms are not as a rule due to an orientation, or direct turning of the organism to or from the source of stimulus, as has often been assumed to be the case. On the contrary, the phenomena are due to a definite movement or reflex action, produced by the stimulating agent, and always taking place in essentially the same manner. The organisms when stimulated by a chemical, by heat, by cold, by mechanical shock or other similar agent, swim backward and turn toward a structurally defined side. To this simple reaction is due the collecting of the organisms in certain regions and their apparent avoidance of other regions, — the so-called positive and negative chemotaxis, thermotaxis, etc. (The reaction to the electric current is complicated by other factors). Early in the progress of the work, it was incidentally observed in a number of cases that the bacteria have an analogous method of re- action. At that time opposition had developed in certain quarters to the account given in these studies of the method by which the so- called tactic phenomena take place in the infusoria. In view of this fact, and the further fact that Pfeffer in his classical studies on the reactions of unicellular organisms had distinctly asserted ^ that the 1 This journal, 1899, ii, pp. 311-341. ^ Ibid., 1900, iii, pp. 229-260. •^ Ibid., pp. 355-379- * " Dass die Ansammlungen niclit etvva nur zu Stande kommen, weil die zufal- lig einschwarmenden Organismen beim Versuch des Entfliehens zuriick.schrecken, lehrt die direkte Beobachtung." — Pfeffer, Untersuchiingen aus dem botanischen Institut, Tiibingen, 18S8, ii, 648, note. See also ibid., 1884, i, 464. 31 32 H. S. Jennings and J. H. Crosby. reactions of the bacteria were not of the nature which our observa- tions showed them to be, it was not deemed worth while to publish an account of these reactions of the Bacteria until the description given of the reactions of the infusoria had been confirmed. This description has now been confirmed and extended to other organisms by various observers. A careful study of the movements and re- actions of certain bacteria was therefore undertaken during the past winter, with the intention of publishing the results of the investi- gation. This work was entirely finished when the valuable paper of Rothert,^ dealing partly with the same subject, appeared. This paper, so far as it covers the same ground, agrees throughout with our own observations and clearly establishes the fact that the re- action method of the bacteria to chemicals is not by an orientation, but is analogous to that of the infusoria, as described in these studies. As the subject is of much interest, and as our work was done from a different standpoint from that of Rothert, with dififerent meth- ods and, to a certain extent, different organisms, a brief account of our own observations will not be superfluous even after the publi- cation of Rothert's paper. In a field where so much uncertainty and disagreement exists, the mutual confirmation of two investigators working independently is of importance. For a discussion of the literature and of the general bearing of the results, reference should be made to the paper of Rothert. The bacteria studied by us were those occurring in cultures of hay and of aquatic plants decaying in water. Two species of Spirillum, apparently Spirillum volutans and S. undula, were selected for special investigation. The gross features of the reactions of bacteria to chemicals, as usually shown, are well known through the work of Engelmann,^ Massart,'^ Verworn,^ and others. Particularly striking is the reaction to oxygen. The bacteria, mounted on a slide and covered with a cover glass, collect (i) about bubbles of air; (2) about the edge of the cover glass, next to the air; (3) about green plant cells, diatoms, desmids, etc., which are giving off oxygen through the action of 1 Rothert: Flora, 1901, Ixxxviii, pp. 371-421. 2 Engelmann : Archiv fiir die gesammte Physiologic, 1881, xxv, pp. 2S5-292. 8 Massart, J.: Bulletin de I'Academie royale de medecine de Belgique, 1891 (3), xxii, pp. 148-167. ■• Vkkworn : Psychophysiologische Protistenstudien, 1889, pp. 103-106. Reaction of Bacteria to Stimuli. 33 the chlorophyll. (See in Verworn's General Physiology, Figs. 211 and 214, and Davenport's Experimental Morphology, Part i, Figs. 3-5^.) Such collections may be observed in any preparation con- taining the ordinary bacteria of decay. How are these collections formed .-• Do the bacteria turn and direct their course toward the centre of diffusion of the oxygen, — proceeding directly toward the region of greatest oxygen density? Or are the collections brought about more indirectly, in a manner similar to that by which Paramecium collects in regions containing an acid .'' Before describing the observations by which this question is answered, it will be necessary to give a brief account of the form and usual movements of the organism Spirillum. Spirillum volutans forms an elongated rod, very slender, with a length of from 15 to 50 /u-. It is curved into the shape of a spiral of from two to six turns, so that it resembles a corkscrew in form. At each end of the spiral are found one or two fiagella. There is no observ- able difference between the two ends of the organism, nor is there any inarked differentiation of two sides, such as distinguishes oral and aboral sides in Paramecium, for example. Movements. — Spirillum swims in the direction of the long axis of the spiral by means of its fiagella. At the same time it revolves, the revolution following the direction of the spiral, and being therefore (usually at least) from left over to right, if one faces in the direction in which the organism is swimming. At intervals the movement is reversed, the organism swimming then with the opposite end in advance, and revolving in the opposite direction. As a rule, neither end seems to be preferred as the anterior one, Spirillum swimming indifferently in either direction. Given individuals are observed, how- ever, at times to swim for long periods with a certain end in advance, the reversals lasting but a moment. Occasionally an individual may be seen revolving on its long axis without progressing in either direction, while in other cases there is a rapid whirling on a tranverse axis. But these methods of movement are rare. "Chemotaxis."' — If the Spirilla are mounted on a slide beneath a cover-glass, in company with some desmids or other green algal cells, after a time they will be observed to have formed collections about the algse, as illustrated in the figures referred to above. How does this occur? 34 H. S. Jennings and J. H. Crosby. Careful observation shows the course of events to be as follows : — At the beginning the bacteria are scattered uniformly throughout the preparation. They are swimming rapidly in all directions. At first they pass close to the green plant cells without any reaction whatever. The algae begin, in the light, to give off oxygen, so that after some time each desmid or other alga must be conceived as sur- rounded by a zone of water impregnated with oxygen. Now begins the collection of the Spirilla about the algae. The bacteria surrounding the algae do not change their direction of motion and swim toward the centre of diffusion of the oxygen. On the contrary, all continue to swim in the same direction as before. A Spirillum passing close to the alga into the oxygenated zone does not at first change its movement in the least. It swims across the zone till it reaches the other side. It is here that the reaction occurs ; the organism reverses its movement and swims in the opposite direc- tion till it reaches the opposite boundary of the oxygenated area. It then reverses again, and this is continued, — the direction of move- ment being reversed as often as the organism comes to the boundary of the zone of oxygen. The Spirillum therefore remains within the area, which thus acts like a trap. Other Spirilla, swimming at random, enter the area in the same way, react at the outer edges in the same manner, and remain. In the course of time therefore the zone of oxygen swarms with Spirilla. There is thus no orientation shown either by the organisms within the area or by those outside. Within, the Spirilla are swimming in all directions, crossing each other's paths at every angle, and agree- ing only in the fact that the movement is reversed on coming to the boundary of the zone of oxygen. A single individual may be seen to oscillate back and forth from one side of the area to the other an indefinite number of times. Without, movements are occurring absolutely without relation to the position of the alga and its oxygen zone. Many Spirilla pass close to the edge of the zone, but do not enter unless their original course carries them directly into it. Many of the bacteria therefore remain scattered throughout the preparation, not gathering about the alg.-e, no matter how long the slide is allowed to stand. But through their continued movement in all directions, dense groups are soon formed about the algal cells. It is evident therefore that the collections of bacteria arise through the agency of a " motor reflex " essentially similar in character to that of the infusoria, described in previous numbers of these studies. Reaction of Bacteria to Stimuli. 35 This motor reflex consists in the bacteria in a reversal of the direction of movement, upon stimulation. The direct cause of the reaction is a change in the nature of the surrounding medium. In the cases already described, it is the change from water containing much oxygen to water containing little oxygen. The " boundary " of the oxygen zone, above spoken of, is of course merely the region where the change in oxygen density is sufficiently great to cause the reaction. The bacteria collect in exactly the same manner about air bubbles, and about the edge of the cover-glass, next to the air. In these cases the bacteria usually collect in a narrow zone a short distance from the air surface. If their movements be observed here, it will be found that the reversal of motion is brought about in two different regions, (i) The passage from the optimum zone of oxygen to a region having less oxygen pressure causes the reaction. (2) Passage from the optimum into a region having greater oxygen pressure, — next to the air surface, — causes the reaction with even greater precision than the opposite change. The Spirilla therefore remain in the narrow optimum zone a short distance from the bubble or the edge of the cover-glass. The above phenomena are cases of what has been spoken of as " positive chemotaxis." " Negative chemotaxis," or the avoidance of regions containing certain chemicals, takes place in the same manner, save that the reaction occurs when the organism comes, from the outside, against the outer boundary of the area in question. Thus, if a drop of a \ per cent solution of sodium chloride be introduced beneath the cover-glass by means of a capillary pipette, the following phe- nomena will be observed. The bacteria do not orient themselves and move in radial lines away from the centre of diffusion of the salt solution. On the contrary, all move in random directions, as before. But on coming against the outer boundary of the salt solution, the organism reacts by reversing the direction of its movement. Hence it does not enter the drop. As every Spirillum that comes in con- tact with the drop reacts in the same way, the drop remains empty. Solutions of most acids, alkalies, and salts act in the same manner, so that a drop of any of them (of sufficient concentration) remains empty when introduced beneath the cover-glass. In addition to the observations on Spirilla, the reactions of a num- ber of the other bacteria found in decaying vegetable matter were studied. In every case the reactions took place in the manner above described for Spirillum, so that there can be no doubt that this 36 H. S. Jennings and J. H. Crosby. method is of general occurrence among the bacteria. In this respect our results agree throughout with those of Rothert.^ The same method of reaction often occurs when the bacteria strike against a solid obstacle. The movement is reversed, the organism swimming in the opposite direction. This manner of reaction was first observed by Engelmann in the reaction of Bacterium (Chromatium) ,photometricum to light.''^ If a small circumscribed area on the slide is lighted from beneath, the bacteria, swimming at random, pass into this area in the same man- ner as described above for an area of oxygen. On attempting to pass from this light area into the dark, this organism, according to Engel- mann, suddenly reverses its movement and swims backward, — thus remaining in the lighted area. This reversal lasts in the case of Bacterium photometricum but a short time, the organism beginning soon to swim forward again. This is due to the fact that this bac- terium has flagella only at one end and normally swims with that end in advance. The reversal of the movement is therefore soon followed by a return to the original direction. This reaction was called by Engelmann a " Schreckbewegung; " it is clearly identical with the " motor reflex " described in these studies. Bacteria thus react to chemicals, to mechanical obstacles, and to light (or darkness) in the same way, — by a "motor reflex," com- parable to that of the ciliate infusoria. This method of reaction is denominated by Rothert apobatic taxis, in contradistinction to strophic taxis, which consists in a turning of the organism toward or from the source of stimulus, so as to bring the axis of the body into a definite orientation with respect to the stimulus. The bacteria would thus show apobatic chemotaxis and apobatic phototaxis, using this method of denomination. The really fundamental phenomenon in these cases is the definite reflex action produced by the stimulus ; whether aggregation or scattering of the organisms occurs and where it occurs, depend merely on what agencies produce this reflex. The " motor reflex" of the bacteria differs from that of the infusoria in the same way that the form and structure of the body differ in the two cases. In such bacteria as Spirillum there is no differentiation as between the two ends, or between the two sides of the organism. In correlation therewith, movement takes place indifferently in the ^ Loc. cit. " Engelmann : Archiv fiir die gesammte Physiologic, 1883, xxx, 95-124. Reaction of Bacteria to Stimuli. 2)1 direction of either end, and the motor reflex consists merely of a reversal of the direction of the movement, — without subsequent return to the original direction except as a response to a new stimulus. In the infusoria there is a differentiation both between the ends and between the sides of the animal. The movements reflect these differentiations. The organism swims normally with a certain end in advance, and usually swerves toward a certain side. The motor reflex consists in a reversal of the direction of movement, so as to swim toward the opposite end, together with a turning toward a definite side, and this is always followed soon by a return to the original motion with the anterior end in front. In the case of Engel- mann's Bacterium photometricum we have an interesting intermediate condition. Here there is a differentiation between the two ends of the organism, only one bearing flagella, while apparently all sides are alike. The reaction to a stimulus consists in a reversal of the direc- tion of movement, as in the other bacteria, but without any turning toward a certain structurally defined side, such as occurs in the infusoria. But the reaction is followed, as it is in the infusoria, by a return to the original direction of movement. The reactions thus give throughout, in their simplicity or complexity, a faithful reflection of the structure. THE FORMATION OF ALLANTOIN FROM URIC ACID IN THE ANIMAL BODY. By ROBERT E.' SWAIN. [From the Sheffield Laboratory of Physiological Chemistry, Yale U/iiversity.'] DESPITE the numerous investigations of recent years which have dealt with the synthesis and transformation of uric acid in the animal body, much remains to be ascertained before we can reach a clear conception of the processes determining the excretion of this substance. Any comprehensive consideration of the extent of uric acid production calls for knowledge regarding the fate of uric acid when introduced as such into the organism. The decomposi- tion products obtainable in the laboratory include such familiar com- pounds as urea, oxalic acid, and allantoin. The latter is readily formed by the gentle oxidation of uric acid.^ Wohler and Frerichs ^ conducted the earliest feeding experiments with uric acid. They searched in the urine for those products which Wohler had obtained by oxidation of uric acid with lead peroxide, viz., urea, oxalic acid, and allantoin. The latter substance could not be isolated from the urine of man, the dog, or the rabbit, after introduc- tion of uric acid into the system ; the investigators concluded that it was broken down further into compounds unknown to them, and thus escaped detection. Neubauer, ^ later, also failed to find allantoin in the urine of rabbits after feeding uric acid. A large increase in the urea output was observed, however. In 1876, Salkowski'^ first demonstrated allantoin excretion after feeding uric acid to dogs. Four grams were given to one animal on two days in succession ; i 42 grams of allantoin were separated from the urine by crystallization, and identified by analysis. This result has been confirmed by Minkowski ° in two cases. From one dog he 1 Claus: Berichte der deutschen chemischen Gesellschaft, 1874, vii, p. 226. 2 Wohler and Frerichs: Annalen der Chemie, 1848, Ixv, p. 335. 3 Neubauer : Annalen der Chemie, 1856, xcix, p. 217. * Salkowski : Berichte der deutschen chemischen Gesellschaft, 1876, ix, p. 719. 5 Minkowski : Archiv fiir experimentelle Pathologie und Pharmakologie, 1898, xli, p. 398 (foot-note 3). 38 Formation of Allantdin from Uric Acid. 39 obtained 0.763 gram of allantoin and o. 161 gram of uric acid after administration of five grams of uric acid with the diet. Mendel and Brown ^ likewise obtained a considerable yield of allantoin after feeding uric acid to cats. Recently, however, Poduschka^ has failed to obtain any allantoin after feeding uric acid to dogs. It is easy to understand the failure of the earlier investigators in view of the lack of any satisfactory method for the quantitative estimation and isolation of allantoin.^ But Poduschka's results were obtained after the application of a new and accurate quantitative method which will be discussed later. His analytical data may be briefly summarized here"^: — I. One gram of sodium urate, dissolved in water, was given to a dog with his ordinary diet. The large quantities of water required increased largely the subsequent excretion of urine. The output of urea considerably exceeded that corresponding to the nitrogen of the ingested urate. There was an insignificant increase in the allantoin excretion. II. A dog of 9.8 kilos received on the second day of fasting 100 c.c. of 0.8 per cent sodium sulphate solution subcutaneously, and 140 c.c. of water per OS. On the following morning the urine (70 c.c.) was removed by catheteriza- tion, and estimations of various constituents made. The animal then received two doses of one gram of sodium urate each. The 78 c.c. of urine obtained by catheterization during the following twenty-four hours were analyzed. The increase in urea-nitrogen demonstrated the absorption and decomposition of the uric acid. No noticeable production of allantoin was observed. The quantity of uric acid fed in the last experiment was sufficient to have yielded 1.03 grams of allantoin. From these data Poduschka concludes that the decomposition of uric acid in the dog does not involve the formation of allantoin. The experiments with which the present paper deals have been under- taken to explain the negative results recorded. Methods of allantoin estimation. — Before proceeding with a quanti- tative study of allantoin excretion it became necessary to consider the methods available for its estimation. The older methods of Meissner, which will be found described in various text-books, are ^ Mendel and Brown : This journal, 1900, iii, p. 267. 2 Poduschka : Archiv fiir experimentelle Pathologie unci Pharmakologie, 1900, xliv, p. 65. ^ Cf. Minkowski : Loc. cif., p. 396. * Poduschka -. Loc. cit., p. 65. 40 Robert E. Swain. somewhat unsatisfactory. Moscatelli ^ has outlined a method which the writer has applied in slightly modified form with excellent results. The fluid to be examined is treated with a solution of mercuric ni- trate, which precipitates the allantoin completely, together with other substances. The precipitate is washed with cold water and decom- posed with hydrogen sulphide. The filtrate is evaporated to a small volume after being made alkaline with' ammonia, and is again precipi- tated with an ammoniacal solution of silver nitrate. The silver pre- cipitate is filtered off after twenty-four hours and decomposed with hydrogen sulphide. From the solution thus obtained the allanto'in may be crystallized out. For analytical purposes it is sufficient to determine the nitrogen in the silver precipitate, from which the con- tent of allantoin can be calculated. The most serious criticism of this method applies to the solubility of allantoin -silver in an excess of ammonia. This fact has been recognized in the method recently devised by Loewi.- It likewise depends upon the quantitative pre- cipitation of allantoin with alkaline silver nitrate solution. The excess of ammonia is avoided by the substitution of magnesium oxide which may be added in excess without interfering with the complete precipitation of the allantoin-silver. In Loewi's method, the urine is first treated with mercurous nitrate which removes the other nitrogenous compounds without precipitat- ing allantoin. The mercury is removed from the filtrate by the use of hydrogen sulphide, and the solution is then treated with magnes- ium oxide and silver nitrate. The silver precipitate contains no nitrogenous compounds except allantoin. The latter may be esti- mated by determining the nitrogen in the washed precipitate, or by decomposition, reprecipitation with mercuric nitrate, the subsequent removal of the mercury with hydrogen sulphide, and the weighing of the allantoin after evaporation of the filtrate to dryness. Like Loewi, the writer has obtained very good results with this method. Great care must be taken, however, not to have the solution of mer- curous nitrate too strongly acid, and to prevent the formation of mercuric nitrate. Poduschka^ has lately recommended the following method. Fifty or one hundred cubic centimetres of urine are precipitated with ^ Moscatelli : Zeitschrift fiir physiologische Chemie, 1889, xiii, p. 203. 2 Loewi : Archiv fiir experimentelle Pathologic und Pharmakologie, 1900, xliv, p. 20. 3 Poduschka: Loc. cit., p. 61. Formation of Allaiitdin from Uric Acid. 41 the necessary amount of basic lead acetate, and the excess of lead is removed from a definite volume of the filtrate by means of concen- trated sodium sulphate solution. To a definite portion of the second filtrate, 5-10 per cent silver nitrate solution is added, the precipitate filtered ofif and rejected, and to the new filtrate dilute ammonia (i per cent) added drop by drop. The allantoi'n-silver is precipitated, and can be washed thoroughly with one per cent sodium sulphate solution until free from ammonia. The allantoi'n can then be esti- mated by a Kjeldahl determination of nitrogen in the precipitate. With this method Poduschka recovered from 93 to 95 per cent of the allantoi'n added to urine. Some results obtained in the course of a study of the relative ac- curacy of the methods outlined are detailed below. I. Meissner's method. — 0.332 gram of allantoi'n was dissolved in 500' c.c. of normal dog's urine which contained no allantoi'n when examined by Loewi's method. Allanto'in crystals recovered ■=■ 0.3012 gram. The crystals were somewhat discolored, therefore a Kjeldahl nitrogen determination was made. Total nitrogen found = 0.0906 gram = 0.2982 gram allanto'in = 89.8 per cent. II. Loewi's method. — (a) 0.3621 gram of allanto'in was dissolved in 500 c.c. of normal dog's urine as above, and an estimation of allanto'in made. Allantoi'n crystals recovered = 0.3524 gram = 97.3 per cent. (b) 0.2891 gram of allanto'in was dissolved in 500 c.c. of normal dog's urine, and an estimation made by determining the total nitrogen in the first precipitate of allanto'in-silver. Total nitrogen found = 0.0861 gram = 0.2834 gram allantoi'n = 98.0 per cent. III. Poduschka's method. — 0.3082 gram of allanto'in was dissolved in 500 c.c. of normal dog's urine. Total nitrogen found in the allantoin-silver precipitate = 0.0899 gram = 0.2961 gram allanto'in = 96.0 per cent. Experimental methods. — The present feeding experiments were made upon two dogs : one a large bitch of 17 kilos ; the other a very small young terrier of 5 kilos. Animals of such widely different size were purposely selected in order to demonstrate any possible difference in urine acid metabolism which might be attributable to this factor. They were confined in suitable metabolism cages, and 42 Robert E. Swain. in the case of the larger animal the urine was collected twice daily by catheterization. The diet consisted of casein (freshly precipitated from skimmed milk), lard, and cracker-dust. Casein was selected as a suitable proteid foodstuff instead of meat or similar products, since Salkowski ^ has demonstrated that allantoin excretion may occur in the dog after a meat diet. In the earlier experiments the daily diet was made up as follows : Food. Large dog. Small dog. Casein Cracker-dust Lard 200 grams. 50 " 50 " 150 grams. 50 " 25 " LARGE DOG (17 kilos). Experi- ment. Day. Body-weight. Kilos. Urine volume. c.c. Uric acid fed Grams. Allantoin in urine. Gram. I 1 2 3 16.9 17.0 17.1 140 200 280 1.0 - None. II 1 2 3 17.2 17.2 17.1 108 201 345 4.0 1- 0.430 1 J III 1 2 3 17.2 17.2 17.1 ISO 260 280 5.0 ■ 0.590 IV 1 2 3 17.3 17.2 17.2 200 240 190 6.0 - 0.620 ^ Salkowski : Berichte der deutschen chemischen Gesellschaft, 1878, xi, p. 500. Formation of Allantdm from Uric Acid. 43 The uric acid was administered as sodium urate with the daily diet. Water was given ad libitum. The allantoin was estimated by Loewi's method in all the experiments, during a period of three days after the uric acid feeding. Experience showed that the allantoin formed is entirely eliminated by the end of this time. The faeces obtained were usually examined roughly for uric acid by extraction with dilute sodium hydroxide solution, and subsequent acidification with hydrochloric acid. See protocols on pages 42 and 43. SMALL DOG (5 kilos). Experi- ment. Day. Body-weight. Kilos. Urine volume, c.c. Uric acid fed. Grams. Allantoin in urine. Gram. V 1 2 3 5.0 4.9 5.0 140 200 95 1.0 1 1 1- 0.240 1 J VI 1 2 3 5.2 53 5.2 160 110 85 4.0 0.620 1 1 0.140 1- 0.780 1 0.020 j VII 1 2 3 4.9 48 4.9 145 210 85 5.0 1 1 \ 0.935 Since the preceding experiments demonstrate that only a relatively small portion of the uric acid ingested reappears in the form of allantoin, it seemed desirable to obtain more definite quantitative data. The larger dog was therefore fed upon a diet which maintained nitrogenous equilibrium in this animal. After a preliminary period of three days, three grams of uric acid (in the form of sodium urate) were daily added to the food for three days ; this period was followed by an after-period of three days. The diet was carefully analyzed for nitrogen. The nitrogen content of the urine, faeces, and hair was ascertained, allantoin was estimated in the urine by Loewi's method, oxalic acid by Baldwin's modification of Dunlop's method,^ and uric acid by Hopkins' method. The diet consisted of — ^ Baldwin: Journal of experimental medicine, 1900, v, p. 30. 44 Robert E. Swain. 300 grams of casein containing . . 50 grams of cracker-dust containing 50 grams of lard containing . . . Total nitrogen of diet . 9.9 grams of nitrogen. 0.8 grams of nitrogen. 0.0 grams of nitrogen. 10.7 grams of nitrogen. The results of this experiment are given in tabular form. EXPERIMENT VIII. Date. 1901. March. Body Weight. Food. Urine. F.^CES AND Hair. Nitro- gen. Dietary. N. Uric acid. Vol. Nitro- gen. Uric acid. Oxalic acid. Allan- to'i'n. Kilos. Grams. c.c. Grams. 7 8 9 10 11 12 16.3 16.3 16.2 16.3 16.2 16.2 10.70 10.70 10.70 10.70 10.70 10.70 3.0 3.0 3.0 0.0 0.0 0.0 500 350 356 410 230 360 10.53 9.85 11.98 12.26 9.40 11.26 0.106 0.080 0.062 trace trace 0.0070 0.0053 0.0064 0.0032 0.0021 0.310 0.308 0.260 0.088 0.021 0.830 1.800 0000 0.112 1.261 0.260 Totals .... 64.20 9.0 =r 3.0 grams N. 65.28 .... 0.987 4.263 From these results it appears that the yield of allantoin excreted under these circumstances is far smaller than the quantity theoreti- cally obtainable from the amounts of uric acid fed. Thus from nine grams less than one gram of allantoin was recovered. Although the output of oxalic acid was slightly increased during the uric acid period, the increase is far too small to account for much of the meta- bolized material. The significance of oxalic acid as a metabolic pro- duct derived from purin groups such as are contained in nucleins has recently been emphasized by Salkowski ^ and his co-workers. The yield of allantoin obtained in the metabolism of varying quantities of uric acid is summarized in the following table : — 1 Salkowski: Uerliner klinische Wochenschrift, 1900, p. 494. Formation of Allantdin from Uric Acid. 45 SUMMARY OF ALLANTOIN EXCRETION. Uric acid fed. Grams. Allantoin excreted. Large dog. Small dog. Gram. 1 4 5 6 9' none 0.430 0.590 0.620 0.987 0.240 0.780 0.935 The difference in the allantoin output of the two animals is more striking when the effects are expressed in terms of units of body - weight, as indicated below: — Uric acid fed Allantoin output per kilo body-weight. per kilo body-weight. Large dog. Small dog. Milligrams. 59 none 200 48 235 25 294 35 353 37 529 58 800 156 1000 .... 187 In attempting to explain these differences, we recall the failure of various investigators- to demonstrate allantoin production in man 1 This was fed in the course of three days. '^ Cohn: Zeitschrift fiir physiologische Chemie, 1898, xxv, p. 509; Minkow- ski : Loc. cit., p. 398 ; Loewi : Loc. cit. , p. 22. 46 Robert E. Swain. after ingestion of uric acid-yielding food, such as thymus. Allan toin may be looked upon as an intermediate product in uric acid meta- bolism in the body. Ordinarily, where the conditions are favorable, uric acid and its antecedents are more completely oxidized in the system, and the nitrogen of these compounds presumably reappears in large part as urea. This might be looked upon as representing the ordinary fate of uric acid when introduced into the human organism in such doses as are permissible. Furthermore, allantoTn itself may be oxidized almost completely in the body.^ In the dog, also, comparable quantities of uric acid may apparently be meta- bolized beyond the stage where allantoin appears as an end-product — as Experiment I above and Poduschka's experiments show. When larger quantities (per kilo of body-weight) are fed, however, the sys- tem is apparently unable to bring about so complete a decomposition of the purin radical ; and under these conditions allantoin may appear as an end-product of the transformation of a part of the ingested uric acid precisely as it has repeatedly been shown to arise after nuclein,^ or even allantoin ^ feeding. The differences shown between man and other animals (dog and cat) thus appear as due to variations in the extent of metabolism in the organs involved, such as the liver, rather than to specific peculiarities of different animals. In regard to the importance of the liver, reference may be made to the occurrence of allantoin in cases of cirrhosis of the liver * and in diamid poison- ing.^ For the latter case, Borissow has already pointed out that allantoin occurs in the urine, presumably owing to the inhibition of normal processes of metabolism ; and Poduschka's experiments also suggest that hepatic changes may play an important role. Summary. The experiments demonstrate, in agreement with the observations of Salkovvski and Minkowski, that allantoin is excreted by the dog 1 Minkowski : Loc. cit., p. 399; Poduschka : Loc. cit., p. 64. 2 CoHN : Loc. cit., p. 507; Minkowski: Loc. cit., p. 393; Salkowski : Cen- tralblatt fiir die medicinische Wissenschaften, 1898, p. 929; Mendel and Brown : This journal, 1900, iii, p. 265 (Cat). ^ Minkowski : Loc. cit., p. 399; Poduschka : Loc. cit., p. 64. ^ MoscATELLi: Loc. cit., p. 202. 5 Borissow: Zeitschrift fiir physiologische Chemie, 1894, xix, p. 499: Po- duschka : Loc. cit., p. 59. Formation of Allaiitdin from Uric Acid. 47 after ingestion of uric acid. Considerable quantities of this substance may, however, be burned up beyond the allantoi'n stage. After uric acid feeding the output of uric acid in the urine is only slightly increased. In conclusion, the writer acknowledges his obligation to Professor Lafayette B. Mendel, not only for the suggestion of the subject of this investigation, but also for valuable assistance and criticism. SOME DECOMPOSITION PRODUCTS OF THE CRYSTAL- LIZED VEGETABLE PROTEID EDESTIN. By p. a. LEVENE and LAFAYETTE B. MENDEL. MORE than two years ago we undertook a study of the decom- position products of various albuminous substances and began the investigation by endeavoring to determine the proportion of nitrogen split off in different groups by the action of acids. Phos- photungstic acid was used to precipitate the basic compounds, and the estimations were made in somewhat the same way as those by Hausmann ^ in his research on the proteids. It was very soon found, however, that concordant and reliable results could not be obtained, and the method was given up as unsatisfactory. Similar experience led Henderson^ and Chittenden and Eustis^ to announce that Hausmann's method of determining the distribution of nitrogen in the proteid molecule is unreliable for quantitative purposes, and that results obtained by this method must be accepted with caution. Quite recently Kutscher^ has arrived at a similar conclusion, appar- ently without being aware of the observations just referred to. We have made a qualitative study of the hexon bases arising from the decomposition of the crystalline vegetable proteid edestin, and the earliest analyses together with demonstrations of the products separated were presented to the American Physiological Society in 1899.^ Edestin was selected for study at that time, because it could be obtained in large quantities as a relatively pure crystalline com- pound ; furthermore, the proteid is unusually rich in nitrogen (18.7 ^ Hausmann: Zeitschiift fiir physiologische Chemie, 1899, xxvii, p. 95 ; 1900, xxix, p. 136. 2 Henderson: Proceedings of the American Physiological Society, 1899; This journal, 1900, iii, p. xxx. Also, Zeitschrift fiir physiologische Chemie, 1899, xxix, p. 47. 3 Chittenden and Eustis : Proceedings of the American Physiological So- ciety, 1899; This journal, 1900, iii, p. xxxi. * KuTSCHER : Zeitschrift fiir physiologische Chemie, 1900, xxxi, p. 215. ^ Levene and Mendel: Proceedings of the American Pliysiological Society, 1899; Tiiis journal, 1900, iii, p. iv. 48 Some Decompositio7t Products of Edestin. 49 per cent), and the experience of Schulze ^ with the proteid from the seeds of Picea excelsa suggested the possibility that the edestin from the hemp-seed might likewise yield the hexon bases in abundance. Kossel and Kutscher^ have lately published the results of an ex- tensive investigation of the nitrogenous groups — especially the hexon bases — in various proteids. Edestin or similar crystalline vegetable proteids are not included in their research ; and since no reference is made to our earlier observations we have concluded to present a very brief account of these experiments. Additional in- terest is now attached to them in view of the specific differences which Kossel and Kutscher found among the proteids of wheat and corn (maize) which they studied. Thus the alcohol-soluble proteids, zein, gliadin, etc., yielded no lysin whatever, although it was readily obtained from the so-called gluten-casein of wheat. The edestin used in the present investigation was obtained by Osborne's method of extracting hemp-seed with warm 5 per cent sodium chloride solution.-' Large quantities of the carefully pre- pared and washed proteid were decomposed by heating with 20 per cent hydrochloric acid and stannous chloride for seventy-two hours. The tin was then removed from the diluted fluid with hydrogen sulphide; and after driving off the excess of this gas most ot the acid was removed by treatment with freshly precipitated lead hydroxide. The filtrate was again treated with hydrogen sulphide to remove the lead in solution, and the fluid concentrated. It re- acted slightly acid and still contained some hydrochloric acid. Nitric acid was added, and enough silver nitrate to combine with all the bases present. The quantity necessary for this purpose was ascer- tained by diluting 10 c.c. of the concentrated liquid to 25 c.c, and adding 5 per cent silver nitrate solution from a burette until a drop of the mixed liquids formed a brownish yellow trace when added to a solution of barium hydrate on a watch-glass. The cal- culated quantity of silver nitrate solution was then added to the entire acid solution, and the precipitate of silver chloride which formed was removed by filtration. From the resulting filtrate the ^ Schulze: Zeitschrift fur physiologische Chemie, 1898, xxiv, p. 276. 2 Kossel and Kutscher: Zeitschrift fur physiologische Chemie, 1900, xxxi, p. 165. ^ For the properties, etc., of edestin, cf. Osborne: American chemical jour- nal, 1893, xiv, p. 38; Chittenden and Mendel: Journal of physiology, 1894, xvii, p. 48. 50 P. A. Levene and Lafayette B. Mendel. three hexon bases were finally separated and identified by the methods of Kossel/ as follows : — Histidin. — To remove the histidin, the acid fluid was neutral- ized with barium hydrate solution. A slightly yellowish precipitate was formed and filtered off by suction. It was washed with half- saturated barium hydrate solution, then taken up with water and rendered slightly acid with sulphuric acid. The silver was next removed by means of hydrogen sulphide, the sulphuric acid with barium hydroxide, and the excess of the latter by means of carbonic acid. The filtrate from the barium carbonate was concentrated, made acid with hydrochloric acid, decolorized with charcoal, again concentrated with addition of more acid until crystallization began. This process was allowed to continue in a vacuum desiccator until crystals of histidin dichloride — some lo mm. long and about 6 mm. in thickness — were obtained. The mother liquor yielded more of the same crystals on further similar treatment. The material was washed with alcohol and ether, dried over sulphuric acid and analyzed. o. 157S gram yielded 0.1995 gram AgCl = 31.22 per cent CI. 0.1423 gram yielded by the Dumas method, 25 c.c, of N at 30° C and 747 mm. Hg pressure = 18.62 per cent N. , CfiHgNsO, . 2HCI . Calculated. Found. N . . . . 18.42 per cent 18.62 per cent CI ... . 31.14 per cent 31.22 per cent Arginin. — The filtrate from the histidin-silver compound was satu- rated with powdered barium hydrate. The white precipitate which formed rapidly turned brown. It was filtered off by suction, washed with concentrated barium hydrate solution, taken up with water and acidulated with sulphuric acid. The silver was next removed by means of hydrogen sulphide, the sulphuric acid with barium hydrox- ide, and the excess of the latter with carbon dioxide. The filtrate from the barium carbonate was concentrated to a thick syrup and transferred to a vacuum desiccator. A partly crystalline mass of arginin carbonate was gradually formed. The crystalline material was easily separated from the colored mother liquor and finally .a white mass obtained which was washed with alcohol and ether. ^ For references to the most recent methods, see Kossel and Kutscher : Zeitschrift fur physiologische Chemie, 1900, xxxi, p. 165. Some Decomposition Prodticts of Edestin. 5 i The bulk of the arginin remained in the mother liquor and wash- ings, from which more of it could be isolated. Only the white purified carbonate was used to obtain the nitrate and silver salts. The acid silver salt was analyzed.^ 0.250 gram yielded 0.06715 gram Ag = 26.86 per cent Ag. 0.100 gram yielded by the Dumas method 19.5 c.c. of N at 32° C. and 756 ram. Hg pressure = 20.74 per cent N. , C6H14N4O2 . HNO3 + AgNOs , Calculated. Found. Ag . . . . 26.54 per cent 26. 86 per cent N .... 20.64 psr cent 20.74 per cent Lysin. — The filtrate from the arginin-silver was treated with sul- phuric acid to remove the barium, then with hydrogen sulphide to eliminate the silver. After driving off the excess of gas by heat, the acid fluid was treated with phosphotungstic acid, and the precipitate formed was decomposed in the usual manner with barium hydrate.^ To the final solution containing the carbonate of lysin, and probably other bases, an alcoholic solution of picric acid was added and the material allowed to stand for twenty-four hours. The lysin picrate formed was redissolved several times in hot water and reprecipitated. It was dried in vacuo, then at 120° C. and analyzed. 0.1645 gram yielded, according to an analysis by the Kjeldahl-Gunning ^ method, 18.55 P^^ ^^"^ ^• C.2153 gram yielded on combustion, 0.3025 gram CO.j = 38.30 per cent C, and 0.1025 gram HoO = 5.28 per cent H. , CeHi.NoOo . C6H3N3O, . Calculated. Found. C . . . . 38.40 per cent 38-30 per cent H . . . . 4.53 per cent 5.28 per cent N . . . . 18.67 per cent 18.55 per cent In view of the high content of hydrogen found, the preparation was recrystallized and analyses obtained as follows : — I. C . . . 38.38 per cent H . . . 5.50 per cent II. C . . . 39.01 per cent H . . . 6.65 per cent ^ Cf. Hedin: Zeitschrift fiir physiologische Chemie, 1894, xx, p. 189. 2 Cf. KossEL: Zeitschrift fiir physiologische Chemie, 1899, xxvi, p. 586. 3 This analysis was made before the publication of Henderson's criticism of the 52 p. A. Levene and Lafayette B. Mendel. These results suggested that a partial decomposition of the salt had taken place during the successive recrystallizations. The picrate was therefore converted into the chloride by Kossel's method. ^ The material thus obtained was too small in quantity to furnish a satis- factory analysis. We believe, however, that the first analysis taken in connection with the mode of preparation and properties of the compound leaves little doubt as to the identity of the latter. The foregoing experiments have demonstrated that the crystal- lized vegetable proteid edestin resembles the ordinary animal proteids in yielding the three known hexon bases, arginin, lysin, and histidin, and differs from the alcohol soluble proteids of vegetable origin which fail to yield lysin among their decomposition products.'^ N-estimation in lysin by the Kjeldahl method: Zeitschrift fiir physiologische Chemie, 1900, xxix. p. 322. 1 Kossel: Zeitschrift fiir physiologische Chemie, 1899, xxvi. p. 587. 2 KossEL and Kutscher : Zeitschrift fiir physiologische Chemie, 1900. xxxi, p. 212. DO SPERMATOZOA CONTAIN ENZYME HAVING THE POWER OF CAUSING DEVELOPMENT OF MATURE OVA? By WILLIAM J. GIES. [From the Departmetit of Physiology in the Marine Biological Laboratory at Wood's //oil, Mass.^ CONTENTS. Page Historical 54 Experimental 56 Methods of procedure 56 Results with sperm extracts 59 Results with extracts of fertilized ova . . 70 Discussion of results 72 Summary of conclusions 75 OUR knowledge of the chemical properties of enzymes is very slight, and our understanding of the part they play in zymolysis anything but clear. Nevertheless, the great importance in biological events of these energy-transforming substances is generally recog- nized. The lack of precise information regarding the essential quali- ties of enzymes no doubt accounts for the current tendency to attribute indefinitely to ferment influence various processes of mor- phological or chemical character which are not satisfactorily compre- hended through ordinary experimental means, or which, in some cases, have not even been subjected to such investigation. A fundamental biological question has lately been put into this cate- gory. The process of segmentation in the fertilized egg has been ascribed in part, at least, to enzyme influence. With the advice and many helpful suggestions of Professor Loeb, I have attempted to ascertain whether any experimental justification can be found for recent statements that the spermatozoon carries substance into the ovum which effects proliferation by zymolysis. ^ I am indebted to the kindness of Professor Curtis for the use of the investi- gator's room at Wood's Holl, reserved for the Department of Physiology of Columbia University. 53 54 William J. Gies. Historical, Pieri,^ after some observations on Strongylocentrotiis lividus and Echinus esciilentus in the Marine Laboratory at Roscoff, in August, 1897, reported that he had extracted soluble sperm enzyme having power to bring about segmentation of the ovum. " Ovulase, " as he called it, was obtained by merely shaking the spermatozoa of these Echinoderms for a quarter of an hour in a flask with sea-water, or with distilled water. Microscopic examination of the filtrates showed that the spermatozoa which passed through the paper were without tails and immobile; " that is to say, dead." The fresh mature ova, well washed in sea-water, were placed in shallow dishes (size not stated) with the extract, immediately, or within ten hours, after its preparation. Segmentation proceeded slowly and reached the morula stage in about ten hours, with the usual phenomena of karyokinesis. Microscopic examination showed that there had been no penetration by spermatozoa. The "ovulase" in distilled water was less effective than that obtained in sea-water ; it produced only a few segmentations (greatest number not mentioned). At the end of his paper Fieri himself mentions two "objections" to his conclusions which it appears to the present writer destroy their force: (i) Only the spermatozoa in the distilled water (which extract he has distinctly indicated possessed the lesser, if any, seg- mental power) were always killed by the shaking process. He sug- gests that the spermatozoa might be eliminated, and pure " ovulase" obtained with the aid of the centrifuge or porcelain filter. (2) Some of the main supply of eggs in sea-water, from which those tested were taken, segmented (to what stage is not stated), " in spite of the pre- cautions taken." Fieri gives few details of his work, and no direct judgment can be passed on his methods. What proportion of the eggs developed .-' The few divisions caused by the distilled water extract can hardly be emphasized, for Fieri found that distilled water alone caused control eggs to become clear and fragmentary. Is it possible, in microscopic examination of myriads of such minute bodies as spermatozoa, to be certain that each individual can be seen } Is the apparent lack of ^ Pi^Ri : Archives de zoologie experimentale et generale, 1899, vii; Notes et revue, ix, p. xxix. Development of Mature Ova. 55 motility in those actually observed conclusive evidence of the death of all? Besides, not all of the fluid in use can be examined by means of the microscope. Further, what effect did boiling have on "ovulase"? Was it destroyed at that temperature, as all ferments are ? What means were taken to kill the spermatozoa which may have been present in the sea-water used to wash the eggs ? These important points Fieri has not considered. Shortly after Pieri's communication, Dubois ^ presented a brief note of a similar character. Dubois arrived at the conclusion that natural fertilization comes about through the action of a fecundative ferment. He claims that he was able to separate such a body, " d'une zymase fecundante," from the testicles of Echinus esailentus, but no experiments showing its qualities were reported by him. Dubois named the ferment (.-') " spermase " and credited it with the power of modifying a hypothetical substance pre-existent in the ovum, which he called " ovulose." As long as experimental evidence of the truth of such a conclusion is wanting, it must continue to remain an unsatisfying speculation. Winkler's^ experiments were made on SpliacrccJiimcs gninnlaris and Arbacia pustulosa. Every precaution was taken to prevent the action of live spermatozoa. Winkler made extracts of spermatozoa by shak- ing them for about half an hour with distilled water (quantities not stated). In order to prevent destructive action on the part of the distilled water, a precaution Fieri had not observed, Winkler added to the extract, before using it on the test ova, a sufficient quantity of evaporated sea-water to make the concentration of the extract the same as that of sea-water (*' ca. 4%"). Another kind of extract of sperm was made in the fluid obtained by evaporating 400 c.c. of sea-water to one fourth its volume.'^ The filtered extract was finally treated with enough distilled water to lower its concentration to that of normal sea-water, ^ Dubois : Comptes rendus hebdomadaire des seances de la Societe de Biologie, 1900, lii, p. 197. The author has not had access to the original paper and relies upon the review made of it by Winkler. (Ref. below.) ^ Winkler : Nachrichten von der konigliche Gesellschaft der Wissenschaften zu Gottingen. Mathematisch-physikalische Klasse, 1900, p. 187. * Winkler states that the sea-water he used contained " ca. 4% " of saline mat- ter and that by evaporating 400 cc to 100 c.c. he obtained a solution of "ca. 20%." The author fails to see how anything but a 16% solution was obtained if the process was conducted as described. Loeb's experiments have shown how necessary exact knowledge of concentration is in such work. 56 William J. Gies. In both kinds of extract the eggs showed some tendency to seg- ment, but only a few divided.^ Sometimes with the same extract the eggs of one individual " reacted," while the eggs of another did not. Finally, it is decidedly significant that the proliferation went at most only to the 4-cell stage, and that then separation of the cells occurred from the absence of retaining membrane, and " abnormal" forms re- sulted. In the control experiments these manifestations were not apparent. Winkler does not claim that the slight changes he observed were due to an enzyme. He states that he did not determine the effect of heat on the power of his extracts. The nature of the active sub- stance, he says, is completely unknown. It might be reasonable to assume that dissolved nucleoproteid had stimulated proliferation, but it seems much more probable that the initial segmentations Winkler observed were really due to increased concentration and the consequent osmotic conditions, not to ferment action or extractive influences. Errors in making up the saline solutions might of them- selves have accounted for all that was observed. A concentration very little above that of normal sea-water would produce the results.^ Further, it is well known that the eggs of sea-urchins are prone to divide into a few cells if they are allowed to remain undisturbed in normal sea-water for about a day.'^ Winkler's results are hardly positive enough, therefore, to permit of the deduction he draws ; they might, in fact, be used to show how unwarranted were Pieri's conclusions. Experimental. General methods of procedure. —The investigations recently done under Professor Loeb's supervision in this connection were con- ducted with Arhacia picnctiilata. In a few experiments, as will be pointed out, the testes of Strongyloccntrotus piirpiiratus were used. Males and females were kept together in a tank in running sea-water until they were needed. Immediately before they were used all extraneous matter was carefully washed off in an abundance of fresh water, which killed any adherent spermatozoa. The various instruments employed in the work were repeatedly washed in the same way. ^ " Nur ein nicht sehr grosser Theil." ■^ LoEB : This journal, T900, iii, pp. 436 and 437. ^ LoEB: Loc. cit. Development of Mature Ova. 57 The sea-water in these experiments was collected in a large stoppered bottle on one day for use upon the next. This insured the use of the same water for each set of experiments and the cor- responding controls. Gemmill ^ has shown experimentally that if free spermatozoa are kept in sea-water (in ''dilute mixture") for five hours they lose their ability to impregnate the ovum. Conse- quently our method rendered inert any spermatozoa which may have been alive in the water at the time of collection and made boiling unnecessary. Moreover, Loeb ^ has lately called attention to the fact that sea-urchins have practically died out in the immediate neighborhood of Wood's Holl, and that for this reason, even at the height of the spawning season, there is little or no danger that the supply of sea-water used in this laboratory contains any live spermatozoa of this animal. In procuring testes or ovaries the oral surface of the animal was cut away and the alimentary and vascular membranes carefully torn out. After thorough flushing in sea-water to eliminate body fluid and dissolved matter such as digestive enzyme, etc., the glands were transferred to perfectly clean vessels for appropriate treatment without delay. The ovaries, from which the eggs used as indicators were taken, were transferred directly to a shallow dish with just enough sea- water to cover them. In most cases the eggs from one animal were sufficient for a connected series of observations. As a rule the ovaries were full of eggs and mere shaking sufficed to liberate the latter into the surrounding fluid, where a comparatively thick layer quickly formed. A few drops of this sediment, containing thousands of eggs, were sufficient for each individual test. The ovaries were never taken from the animal until all other preparations had been completed, so that the eggs were perfectly fresh when employed. Only such unfertilized eggs as were found to be normal and mature were used. In each of the series of experiments to be described some of the ova were either fertilized directly with spermatozoa or were first subjected for an hour or two to the influence of solutions of higher osmotic pressure than sea-water (mixtures of 88 c.c. sea- water -f- 12 c.c. '^^ n KCl were usually made up for the purpose) and then were placed in sea-water to test their capacity for partheno- genetic division. In many experiments both methods were used. 1 Gemmill: Journal of anatomy and physiology, 1900, xxxiv, p. 170. ^ LoKB : Loc. cit., p. 450. 58 William J. Gies. Under these test conditions the eggs employed were always found to develop into swimming larvse within twenty-four hours. These facts are not specially noted in the records given below because of their uniformity throughout. The " control " tests mentioned with each series refer to the eggs which had been placed only in normal sea-water for comparison with ova treated by special processes. In each of the following series of experiments the volume of sea- water in each test was, as a rule, 100 c.c. (Note exceptions farther on.) It was increased only by the addition of portions of extract as specified under each series and by the few drops of sea-water carrying the eggs, in pipette, from the main supply. The sea-water was contained in small bowls of uniform size, making the depth of the fluid (about an inch) practically the same for all of the experiments. Throughout each series the bowls were kept covered with glass plates. The air space above the fluid was about an inch in depth, thus insuring abundant supply of oxygen. Occasionally, as will be noted, eggs were placed in quantities of the extract alone, held in smaller vessels. These were also kept covered. The temperature of the room varied between 18-20° C. The amount of evaporation, as indicated by sensible condensation on the under side of the cover- plates, was comparatively slight during twenty-four hours, so that no material concentration occurred during the interval. The extracts of the spermatozoa were made directly from the testes. It was not thought necessary to attempt separation of the non- spermatic tissue elements. The testes were always thoroughly ground to a thick paste in a mortar with dry sand which had been heated above 100° C. for from fifteen to twenty minutes. Water and saline extracts were used within a few hours. Fluids containing preservatives, however, were given more time for extraction, as will be noted below. The extractions were made in bottles to permit of frequent and vigorous shaking. Clear filtrates were obtained in each case without special difficulty. In each series of experiments carefully measured quantities of ex- tract were added to sea-water, and the mixtures stirred to prevent inequalities of concentration. The eggs were distributed after the mixtures of sea-water and extract had been made. The experiments were begun in the morning. At intervals of an hour or two until late at night, samples of eggs were quickly removed with pipettes from the bowls to watch glasses for observation under the micro- scope. Hundreds were examined carefully each time. None were Development of Mature Ova. 59 ever returned to the main supplies. The eggs in each series were always under observation for from at least twenty to twenty-four hours, seldom longer than that, and unless otherwise stated the " results " recorded below are for periods of that length. Experimental Data. Our experiments are described here briefly, though in some detail, so that whatever value they may possess may be accurately estimated. The first series of extracts were made with spring water. Fresh water extract. — Fresh testes. — I. The glands from one animal were extracted in 15 c.c. H.^O for i hr., 30 mins. Three tests were made as follows : • — (i) Control (2) Extract — 4 c.c. (3) Fresh HjO — 4 c.c.^ Result: No segmentation. II. The glands from one animal were extracted in 15 c.c. HoO for 3 hrs. (i) Control (2) Extract — 2 c.c. (3) Fresh HoO — 2 c.c. Result : No segmentation. III. Glands from two animals in 10 c.c. HoO for 4 hrs. A. Control. B. Extract : (a) i c.c. (unfiltered), (b) 4 c.c, (c) 0.05 c.c, (d) eggs in 3 c.c. + equal volume of y ;/ NaCl. C. Some of (d) into sea-water after 2 hrs. Result : Irregular parthenogenctic forms in a very small propor- tion of (a), (b), and (c) after 4 hrs. A few groups of 8 and one or two of 16 cells from individual eggs, in 24 hrs., in (b). None beyond the 4-cell stage in (a) and (c). A {q^^ parthenogenctic in C as far as the 8-cell stage. No moruloe in any. No segmentations in the control. The results of the third series encouraged the belief that enzyme action was demonstrable, although we did not lose sight of the fact that perhaps increased concentrations, induced by unobserved cir- cumstances, or other unknown conditions, would account for the proliferations noted. In the fourth and fifth series the effects of fresh were compared with those of boiled extract. IV. Five sets of testes extracted in 60 c.c. HoO for 3 hrs. One half was boiled in an Erlenmeyer flask 10-15 '"riins. An appreciable concentra- ^ It will be understood from what was stated on page 58 that this abbreviated reference to the three tests means that besides being under normal conditions (in 100 c.c. sea-water alone), eggs were subjected to the influence of both 4 c.c. of extract in 100 c.c. of sea water and 4 c.c. of fresh HoO in the same large quantity of sea-water. This system will be adopted throughout for brevity's sake. 6o William J. Gies. tion resulted, but of course no approximation to the specific gravity of sea-water was effected. A. Control. B. Fresh extract : (a) lo c.c, (b) eggs in 8 c.c. ex- tract alone. C. Boiled extract: (c) lo c.c, (d) eggs in 8 c.c. extract alone. D, Samples of B and C in loo c.c. sea- water after I hr., 30 mins. Result: During the first 12 hrs. there was no segmentation in any of B and C. An occasional kidney-shaped cell was found in the control and D after 5 hrs. At the end of 24 hrs. there were a few 4 to 8 cell divisions in the eggs of (a) and (c) which had been trans- ferred to sea-water. Only a few 2 to 4 cell groups were found in the control at the end of the same period. V. Testicles from 15 animals extracted in 85 c.c. H^O for 3 hrs. One half was boiled as in the preceding series. A. Control. B. Fresh extract: (a) 20 c.c, (b) 10 c.c, (c) eggs in ID c.c extract alone. C. Boiled extract: (d) 10 c.c, (e) 8 c.c. D. Eggs in B and C transferred to normal sea-water after 1 hr. Result : Not a single segmentation could be detected. A very few of the eggs of (d) and (e) which had been transferred to sea- water were kidney-shaped as though in an initial parthenogenetic stage. The results of the first five series were indecisive, but, where positive, they strongly suggested initial osmotic parthenogenesis, caused probably by conditions beyond control, rather than zymolytic influences. On the assumption that the concentration of the extracts was somewhat lower than sea-water in spite of the salts and proteids dissolved from the testes, and that variations in effects occurred as a consequence, the sixth series was arranged to overcome this difificulty. VI. Fourteen sets of glands were extracted in 35 c.c. H2O for 3 hrs. Just before the filtered extract was used it was mixed with an equal volume of normal NaCl, making approximately a | « NaCl mixture (sea-water is equivalent to about f « NaCl). A. Control. B. Extract : 20 c.c, to c.c, i c.c, eggs in 10 c.c. ex- tract alone. C. Eggs in each of B transferred to 100 c.c. sea-water at the end of 2 hrs. Result : No divisions or irregular forms. The generally negative results of the preceding experiments made it seem desirable to resort to other means before abandoning the study of fresh water extracts. Various enzymes are more easily extracted after the containing cells have been dried and thoroughly broken up. This expedient was tried, therefore. Development of Mature Ova. 6i Dry testes. — The glands from each animal were macerated and spread out separately in a thin layer on watch glasses. These were placed in desiccators over concentrated sulphuric acid or calcium chloride. Drying was accomplished within eighteen hours. When desired for use the dry substance was scraped into a mortar, and ground up thoroughly with sand and extracted as in the previous experiments. VII, The dry substance of four sets of glands was extracted in 30 c.c. HoO for 3 hrs. A. Control. B. Extract: (a) 5 c.c. (unfiltered), (b) 10 c.c, (c) 5 c.c, (d) I c.c, (e) eggs in extract + equal volume ^^" 11 NaCl. Result : Within 12 hrs. no change. At the end of 24 hrs. a very few were in initial parthenogenetic stages, 2 to 4 cell groups, in all except (a). They could be found only after careful search and there were as many in the control as in any of the others. VIII. Eight sets of dried testes in 25 c.c. HjO for 4 hrs. Filtrate mixed with an equal quantity of ^/ n NaCl before using. A. Control. B. Extract : 7 c.c, eggs in extract alone. C. Some of the eggs in B were transferred to 100 c.c. sea-water after i hr., 45 mins. Result : No segmentations or parthenogenetic forms in any. It seemed necessary to conclude at this point that fresh water extracts of spermatozoa do not contain substance of zymolytic power or else that the conditions attending their use are unfavorable to such manifestation. Enzymes which are soluble in water are also soluble in solutions of electrolytes, so that attempts were next made with the latter. Salt water extract. — A common method of extracting enzymes in- cludes treatment of the tissue with ordinary salt solution. Sea-water itself furnishes such a dilute solution, but is not so favorable to rapid destruction of spermatozoa as fresh water or stronger salt solution. Since spermatozoa pass through ordinary filter paper, however often they may be subjected to filtration, it was necessary in using fresh testes to give particular attention to killing the spermatozoa by mechanical means. Prolonged grinding in a mortar with fine sand, as had been done previously, followed by continuous shaking for several hours, accomplished this. Rres/i testes. IX. Twelve sets of glands were extracted in 50 c.c. sea-water for 4 hrs. 62 William J. Gies. A. Controls (2). B. Extract : 20 c.c, 10 c.c, 5 c.c, i c.c, 0.25 c.c. Result : Not a single division could be found. The very greatest care is necessary, in this connection, in the use of solutions of electrolytes, because of the ready osmo-parthenogenetic response the eggs make to slightly increased concentration. There is little reason for believing that an enzyme is present in spermatozoa which is insoluble in dilute, but soluble in strong salt solution. Therefore it seemed unnecessary to try the effect of more concen- trated extractive. The tenth series shows the result of an effort to make the best of saline extraction of fresh testes, however, in a way somewhat different than that of the preceding. X. Eight sets of testes in 40 c.c. 1 71 NaCl for 2 hrs. One half was warmed to 35-40 C. 15-20 minutes. A. Controls (2). B. Extract (unwarmed) : (a) 5 c.c, (b) eggs in 5 c.c. extract alone. C. Extract (warmed) : (c) 5 c.c, (d) eggs in 5 c.c. extract alone. D. Some eggs of B and C in 100 c.c. normal sea-water after 2 hrs. Result : No segmentation within 6 hrs. In 12-24 hrs. a very few 2-cell groups were found with difficulty in (a), (b), and (c) and in one of the controls. Dry testes. The preliminary process of drying was also resorted to in this connection. XI. Dry material from three animals was extracted in 5 c.c. sea-water for 2 hrs. A. Control. B. Extract: 2 c.c. (unfiltered), i c.c, 0.25 c.c. Result: Not a sign of segmentation. Do the extracts possess poisonous qualities? — One condition that may appear to be against the action of an enzyme in the extracts used in these experiments is the possible presence of poisonous substances in the extract. This question now required a definite answer. We had varied the quantities of extract considerably, between all reasonable extremes, in the belief that the most favorable amount might be indicated, but it will be observed from the foregoing account of results that no such relation was suggested. The eggs which had been subjected to the extracts alone, and those placed in sea-water with the greater proportions of extract, usually showed abnormalities after a few hours, such as the development of enclosing membrane or transparent periphery (thicker and not comparable to the " vitel- line " membrane after fertilization), swelling, disintegration, discol- Development of Mature Ova. 63 oration, agglomeration of pigment, etc., but none of these changes were constant so far as their relation to observed conditions could be determined. The sperm extracts contained salts and dissolved proteids, of course, and it would be reasonable to assume that these bodies were present in larger proportion, in some of these experi- ments at least, than they ever are under normal conditions of fecundation. This important matter was definitely tested several times. The following results of two experiments are cited to show the facts in the case : A. Five sets of fresh testes were ground in the usual way and extracted for 2 hrs. in 30 c.c. fresh water. An equal quantity of '^' 71 NaCl was added to the filtrate. The eggs were placed in this mixture and samples transferred at intervals of an hour to 100 c.c. sea-water, to which fresh spermatozoa had been added. Results of examination at the end of 24 hours, the numerals indicat- ing the number of hours the eggs were kept in the extract: (i) Swimming gas- trul?e. (2) Blastulje (none alive). (3) A few dead blastulae, mostly morul?e. (4) Many unsegmented, none beyond the 32-cell stage. (5) About the same as those after the 4-hr. treatment. (6) Very few went so far as the 32-cell stage, many were in the 4 to 8 cell groups. There were no segmentations in the eggs kept for 24 hrs. in the extract. B. Six sets of fresh glands were extracted in 30 c.c. sea-water, 3 hrs. Eggs from one animal were placed in the filtered extract and also into an equal quan- tity of sea-water (as control). At intervals eggs were withdrawn from each supply and transferred to 100 c.c. sea-water containing perfectly fresh sperma- tozoa. Results at the end of 36 hours from the time of the first transferral, the numerals again indicating the number of hours the eggs were under the direct influence of the extract or the normal sea-water: (t) Plutei in each. (2) Advanced gastrute in each. (3) Gastrulse in each. (4) Many gastrute in the control ; hardly any live ones, mostly morul?e, among those treated with the extract. (7) A large number of blastul?e were present in the control, but no divisions beyond the 32 cell stage could be found among the eggs which had been in the extract ; most of the ova were unsegmented. There were no proliferations in the eggs retained in the extract itself. In the earlier tests the proportion of unsegmented cells was uniformly greater in the control than in the other series, whereas the living larvae were relatively more numerous in the latter. The extract seemed at first to stimulate, and later to inhibit karyokine- sis. Possibly, however, the accumulation of bacteria in the bowls containing extract was responsible for the latter effect. It is clear, from the foregoing, that the dissolved substances of our extracts have not prevented the eggs from segmenting. From this 64 William J. Gies. we may safely conclude that they doubtless would not interfere with zymolysis if such were demonstrable. The results of all the preceding series seemed to point in the same general direction and to indicate no mitotic action. Before accept- ing this negative conclusion, however, we proceeded to employ various other familiar methods for the separation of enzymes in the hope of eventually extracting and demonstrating the presence of such a sub- stance. Extract of spermatozoa which had been treated with, and preserved in alcohol Enzymes may readily be extracted from tissues hardened in alcohol. In fact they are frequently isolated by such preliminary treatment, which brings about disintegration of the cellular proto- plasm as well as coagulation of soluble proteid, and thus diminishes the proportion of undesirable extraneous material in the final extract. Through the kindness of Professor Loeb, I was enabled to make ex- tracts of the spermatozoa of Strongylocentrotus piirpurat/is, which had been preserved in an excess of 95% alcohol. The testes were taken from animals collected on the Pacific Coast about a year ago, while Professor Loeb was engaged there in his classical researches on artifi- cial parthenogenesis. In these experiments, with Arbacia as well as Strongylocentrotus, the alcoholic sperm mixture was filtered. Both the solid and fluid portions were transferred to shallow dishes and dried in the air. The liquid soon evaporated and left an oily residue which dissolved to a milky fluid when mixed with water. Strongylocentrotus purpiiratiis. XII. Three grams of the dry sperm res- idue were thoroughly ground with sand and 30 c.c. fresh H2O. After an hour an equal volume of y n NaCl was added. Extraction in this mixture was continued an hour. A. Control. B. Extract: 17 c.c, 7 c.c, and eggs in 8 c.c of extract alone. C. Some of the eggs in each of B were transferred to 100 c.c, sea-water after 3 hrs. Result: Not the slightest trace of segmentation. XIII. Two grams of the finely divided dry substance were extracted in 40 c.c sea-water for 3 hrs. A. Control. B. Extract: (a) 12 c.c, (b) eggs in 10 c.c. ex- tract alone. C. Eggs from B transferred to 100 c.c. normal sea- water after 2 hrs., 15 mins. Result: Only a few forms in initial parthenogenesis in the con- trol and in (a). These were found only after very careful search. Entirely negative results in the others. Development of Mature Ova. 65 It did not seem very likely that the alcoholic filtrate would contain a mitotic enzyme, if such a substance could not be extracted from the portion insoluble in alcohol. Yet, since some enzymes are soluble in diluted alcohol, the following experiments were made in order to ascertain definitely. XIV. Half the residue of evaporated alcoholic extract was dissolved in 40 c.c. sea-water and filtered. A. Control. B. Extract: (a) 15 c.c, (b) 5 c.c, (c) eggs in 20 c.c. of the extract alone. C. Eggs from (c) were transferred to 100 c.c. sea-water after i hr., 30 mins. Result : Within 6 hrs. no perceptible effect. At the end of i8 hrs. a number of irregular parthenogenetic forms and some groups of 4 and 8 cells in C. No traces of segmentation in any of the others. XV. The result in the preceding series seemed to be due to increased con- centration caused by the accumulated salts of the original alcoholic extract. If this assumption were correct, dilution of the extract should prevent the effect noticed above. Only a fourth of the residue was next dissolved in 50 c.c. sea-water. A. Control. B. Extract : (a) 20 c.c, (b) eggs in 20 c.c. extract alone. C. Samples of B were transferred to 100 c.c. sea- water after 2 hrs. Result : Only a very few irregular shapes in the control and the transferred eggs of (b). One 4-cell group was found among thou- sands in the control ; none among the others even after prolonged search. XVI. A third experiment was made with the alcoholic residue. The solution was made more concentrated again. The remaining portion (one fourth) of the evaporated extract was dissolved in 15 c.c. sea-water. A. Control. B. Extract: (a) 8 c.c, (b) eggs in 5 c.c. extract alone. C. Samples of each of B transferred to ico c.c. sea-water after 3 hrs. Result : Parthenogenetic groups of small cells in the transferred eggs of (b), but nothing of the sort in any other. The results of the last three series emphasize the necessity of pre- venting material change in the composition of the sea-water and suggest how easy it might be, in cases of slightly increased concen- tration to mistake ion parthenogenesis for enzyme proliferation. Arbacia. Twenty-one sets of testes were treated with 500 c.c. 95% alcohol. After remaining in contact with the latter for two days the solid substance was collected on a filter. 66 William J. Gies. XVII. The dry solid matter was thoroughly extracted in loo c.c. sea-water for 12 hrs- A. Control. B. Extract : (a) 25 c.c, (b) 15 c.c, (c) 10 c.c, (d) 5 c.c, (e) I c.c, (f) 0.5 c.c C. Samples of B transferred to 100 c.c. sea-water after 2 hrs. Result : A very small percentage of 2-cell groups was found in the control, in (b) and among those of (d) which had been transferred to normal sea-water. One 2-cell segmentation had been found among the normal eggs immediately after they had been taken from the ovaries. XVIII. In 24 hours the alcoholic filtrate (500 c.c.) had evaporated to 30 c.c. Practically all the alcohol had disappeared. The residue was made up to 100 c.c. with sea-water and filtered. A. Control. B. Extract: 25 c.c, 15 c.c, 10 c.c, 5 c.c, 1 c.c C. Samples of B transferred to 100 c.c. sea- water after 2 hrs. Result : An occasional 2 to 4 cell group in practically all in- cluding the control — less than 2 per 100. Glycerine extract. — Glycerine in water seems to be one of the best of enzyme extractors. Extracts of fresh Arbacia sperm were made by the previous general process in mixtures of equal parts of glycerine and water. It has been assumed, of course, that the glycerine in such extracts would exert specific deleterious effects and naturally careful control experiments were made to ascertain its influence in the quantities used in this series. These preliminary control tests de- termined the influence of glycerine under three general conditions: {a) its direct effect on the eggs, {b) its influence on normal fecunda- tion, {c) its action on artificial parthenogenesis. An abundant supply of equal parts of glycerine and sea-water was made for use in all these experiments. Normal eggs were found to remain unseg- mented in all proportions of this glycerine solution with sea-water, although a few irregular parthenogenetic forms were produced by 15 c.c. in 100 c.c. noruial sea-water. Quantities of this glycerine solution greater than 5 c.c. in 1 00 c.c. of sea-water prevented the normal segmentation by spermatozoa, but many swimming larvae formed in the presence of 2 c.c. of the glycerine solution per 100 c.c. sea-water. Even 15 c.c of the glycerine solution in 100 c.c. of sea-water did not, however, entirely prevent proliferation in ova which had pre- viously been kept for 2 hrs. in 88 c c. sea-water -)- 12 c.c. --§ 71 KCl, yet none of tlie segmentations under these conditions went beyond the 8 to 16 cell stage. With smaller quantities, swimming larvre were obtained. With these facts established the result of the following exi)eriments are not without significance. Development of Ma hire Ova. 67 XIX. Seventeen sets of testes in 75 c.c. of the above glycerine solution for 48 hrs. A. Control. B. Extract: (a) 15 c.c, (b) 5 c.c, (c) 2 c.c. C. Samples of each of B transferred to 100 c.c. sea-water after I hr. Result : Here and there a kidney-shaped cell was found among those of (a) which had been transferred to normal sea-water. No distinct segmentations. XX. Same glycerine extract after having been shaken with the tissue 24 hrs. longer. A. Controls (2). B. Extract : 5 c.c, 2 c.c, 0.5 c.c, 0.25 c.c. C. Some of each of B transferred to 100 c.c sea-water after i hr. Result : Not the slightest suggestion of segmentation. XXI. Twenty sets of testes were extracted in 80 c.c of the glycerine solution four days. The filtrate was poured into a litre of 95% alcohol. A bulky, though light, white flocculent precipitate formed at once. After 24 hrs. this precipitate was filtered off, treated with 25 c.c. of sea-water for several hours and the filtrate used in the following experiment : A. Control. B. Extract: (a) 10 c.c, (b) 5 c.c, (c) 2 c.c, (d) I c.c, (e) 0.25 c.c. C. Samples of each lot of B transferred to normal sea-water after 2 hrs. Result: One or two irregular parthenogenetic forms were found in (c) and among those of (a) which had been transferred to normal sea- water. The number of such was less than 5 per 1000. Ether extract. — Substances which cause the death of the cell or which appreciably lessen its vitality are known to favor solution of enzyme into the surrounding medium. Small quantities of alcohol or ether effect such results. Mathews^ has recently shown that expos_ ure of the unfertilized eggs of Arbacia to a saturated solution of ether in sea-water for ten to fifteen minutes leads to karyokinetic division of nearly all the eggs. In the use of ether in these experiments the greatest care was taken, therefore, to ascertain the influence of ether in the small quantities employed. A solution for general use in this connection was made by mixing sea-water and ether in the proportion of 100 c.c. of the former and 7 c.c of the latter. This amount seemed sufficient for any extractive usefulness ether might possess here. Intimate solution resulted. The odor of ether from the solution was still quite distinct at the conclusion of the experiments,, though not strong at ^ Mathews : This journal, 1900, iv, p. 345. 68 William J. Gies. any time. In three control experiments, similar to those outlined under the head of glycerine extract, it was found that as much as 15 c.c. of this ether solution failed to effect parthenogenesis, although after eighteen hours a few 2 -cell groups and irregular forms suggesting an initial stage of mitosis were found. As these were also present in the control, however, no importance could be attached to the result. After the usual treatment with sea-water plus \*' 71 KCl, swimming larvae developed whe^ the eggs were transferred to 100 c.c. of sea- water containing as much as 25 c.c. of the ether solution. The same result was obtained, with as much ether solution present, when spermatozoa were added to the eggs in 100 c.c. of sea-water. XXII. Ten sets of fresh testes were extracted in 60 c.c. of the ether solu- tion for 3 days. A. Control. B. Extract: 25 c.c, 15 c.c, 5 c.c, i c.c, 0.25 c.c C. Some of each lot of eggs in B transferred to too c.c. normal sea- water after 2 hrs. Result: During the first 12 hrs. no changes were mani- fested. At the end of 24 hrs., however, all, including the control, had a few 2 to 4 cell groups. The effect was not at all striking ; it required careful search to find any signs of proliferation. XXIII. The same extract, after having been 24 hrs. longer in contact with the tissue, was again employed. A. Control. B. Extract : 4 c.c, 2 c.c, 0.5 c.c C Eggs from each of B placed in 100 c.c. normal sea-water after i hr., 30 mins. Result : No sign of segmentation. Alcohol extract — Mathews^ has also shown that alcohol affects Arbacia eggs much as ether does. He found that when the ova are placed in sea-water containing 4 to 5 parts of alcohol and are left there for from ten to fifteen minutes, they segment into several cells when they are replaced in sea-water. In these experiments, care was taken, therefore, to determine precisely the influence of the smaller quantities of alcohol employed. A general supply of 10% alcohol in sea- water was kept for the experiments. Quantities not over 25 c.c. of this dilute alcohol, added to 100 c.c. of sea- water, were without mitotic influence. As much as 15 c.c. in 100 c.c of sea-water interfered to no appreciable extent either with normal fertilization or osmotic parthenogenesis, as swimming larvae developed within the usual period in both cases. XXIV. Testes from 12 animals in 60 c.c dilute alcohol solution 48 hrs. A. Controls (2). B. Extract : (a) 25 c.c, (b) 15 c.c, (c) 5 c.c, ^ Mathews: Loc. a't., p. 346. Development of Mature Ova. 69 (d) 2 ex., (e) 0.5 c.c. C. Some of each of Bin 100 c.c. normal sea-water after i hr., 30 niins. Result : No appreciable effect in any during the first 1 2 hrs. At the end of 24 hrs., however, several 2, 3 and 4 cell groups were found in both controls and also in each of those transferred to sea- water. The eggs of (d) which had been put into sea-water had a relatively larger proportion that showed initial division, although the actual number was in reality small — less than 10 in 1,000. XXV. Some of the filtrate used in the preceding series was taken to repeat a part of the experiment just described. A. Control. B. Extract: 2 c.c. C. Eggs from B into 100 c.c. sea- water after i hr., 30 mins. Result: No divisions at any time within 24 hrs. XXVI. Seven sets of testes in 10% alcohol 4 days. A. Control. B. Extract: (^a) 15 c.c, (b) 8 c.c, (c) 2 c.c C. Some of the eggs of each of B in 100 c.c. normal sea-water after 2 hrs. Result : Negative during the first twelve hours. At the end of 24 hrs. there were a very few 2 and 4 cell groups in the control and among those of (a) which had been transferred. No effect in any of the others. Alkaline extract. — Many enzymes show their greatest activity in media which are either acid or alkaline. Fluids of either reaction are also especially efficient in transforming zymogens into enzymes. If the latter cannot be extracted from spermatozoa, as the preceding results may be taken to indicate, might not zymogens be detected .-' Loeb ^ found, in his experiments on Echinoderms and Annelids that the addition of a small quantity of acid or alkali caused the unfertilized eggs to segment much more quickly than when they were left in normal sea-water. NaOH seemed less effective than KOH, but some development occurred in the presence of as little as 2 c.c. xi NaOH in 100 c.c. sea-water. Great care had to be exercised here, therefore. Proportionately smaller amounts were used as a safeguard. A saline solution was made for this series containing 8 c.c. of /q NaOH for every 100 c.c. | n NaCl. This solution was faintly though distinctly alkaline and could hardly be considered destructive to any enzymes in the cells. In control experiments similar to those conducted previously to ascertain the influence of foreign substances it was found that as much as 25 c.c. of this ^ LOEB : This journal, 1901, iv, p. 438 ; also Ibid., 1900, iii, p. 136. 70 William J. Gies. solution when added to eggs in loo c.c. of sea-water caused only a few initial segmentations and that comparatively slight influence was exerted either on osmotic parthenogenesis or spermatic proliferation by the same quantity. XXVII. Twenty sets of testes in loo c.c. alkaline solution 24 hrs. A. Controls (2). B. Extract: 25 c.c, 10 c.c, 5 c.c, i c.c. C. Some of each of B in 100 c.c. normal sea-water after i hr. Result : Not a single divisipn. Extract made in fluid of alternate reaction. — XXVIII. With a view of aiding still further the transformation of any zymogen not affected by previous extractions, twelve sets of testes were macerated in the usual way and allowed to remain in the mortar, covered with a glass plate, for 1 2 hours. The normal alkaline reaction of the fresh tissue became faintly acid to litmus during that interval. 25 c.c. of fresh water was added, the mixture neutralized and then made faintly alkaline with {\^ NaOH and repeatedly shaken up in this mixture for about 6 hours. Finally it was neutralized with very dilute HCl and the filtrate mixed with one-third its volume of 2 n NaCl to bring the concentration of the extract close to that of ordinary sea-water. A. Controls (2). B. Extract: (a) 20 c.c, (b) 10 c.c, (c) 1 c.c C. Samples of Bin 100 c.c. normal sea-water after i hr., 30 mins. Result: No effect during the first twelve hours. At the end of 24 hrs. only an occasional 2-cell division could be found in (c) and among those of (a) which had been transferred. The persistently negative results of the preceding experiments, in which the existence of neither an enzyme nor a zymogen could be indicated, gradually developed the idea that possibly an enzyme is formed from material in the egg, or in the sperm, or in both, on contact of the two living elements. If such were really the case it would seem that extracts of the eggs which had been normally fer- tilized might, under appropriate conditions, possess the power of inducing segmentation in unfertilized ova. Extracts of fertilized eggs. — The general experimental procedure by which this matter was investigated was essentially the same in some respects as for the preceding series. The fresh full ovaries were broken up in sea-water in shallow dishes. Only sufficient ova were kept in each dish to form a single layer at the bottom. The glandular tissue, with such eggs as re- mained entangled in it, was withdrawn. A minute quantity of fresh sper- matic fluid was thrown into 100 c.c. of sea-water and a {t\v drops of this mixture transferred to the dishes containing the eggs. Within a few hours practically all of the eggs were developing and some spermatozoa in excess were in active motion among them. When the eggs were desired for extraction the fluid containing them was Development of Mature Ova. 7 1 thrown into a large funnel, the outlet of which was closed with a stopper. The eggs quickly converged to the neck and soon settled to the bottom of the tube in a thick layer, with a clear supernatant fluid. Practically all of this could be eliminated by decantation, leaving a thick mass of eggs in only a small quantity of fluid. The whole process of collection could be com- pleted in two hours. The segmented eggs were finally thoroughly ground with sand and appropriately extracted. Glycerine extracts — XXIX. Eggs from 15 females, many of which had developed to the i6-cell stage, were ground, in small quantities, with 30 c.c. sea-water and 30 c.c. pure glycerine. They were repeatedly shaken in this mixture. At the end of 24 hours the eggs were considerably swelled and distorted, but were little disintegrated, in spite of the grinding. The latter process was repeated. More of the eggs were broken up, but many were held intact by the fertilization membrane. The extraction process was continued 36 hours longer, by which time at least half of the eggs were still unbroken, though distended. A clear filtrate was obtained. A. Controls (2). B. Extract: (a) 12 c.c, (b) 8 c.c, (c) 4 c.c, (d) I c.c, (e) 0.25 c.c. C. Some eggs in each of B were trans- ferred to 100 c.c. normal sea-water after 2 hours. Result: No segmented cells were found in any except (d). After 12 hours 3 or 4 irregular 2 to 4 cell groups could be found among thousands after diligent search.^ Saline extract. — XXX. Eggs from 20 females. Development was allowed to continue until the more advanced had reached the morula stage, when only a very few remained unsegmented and the majority were at or beyond the 8-cell proliferation. They were ground up in 40 c.c. of fresh water, to which 40 c.c. of ig*' n NaCl was added later. Extraction was continued 36 hours. At the end of that time many groups of cells remained tightly held together in the enclosing membrane ; thorough grinding had not sufficed to disintegrate them as completely as was desired. A. Controls (2). B. Extract : (a) 35 c.c, (b) 20 c.c, (c) 10 c.c, (d) 5 c.c, (e) I c.c. C. Some of the eggs of each of B transferred to 100 c.c. of sea-water after 2 hrs. Result: Negative at first. After 12 hrs. occasional irregular forms in initial cleavage were found among thousands in one of the controls, in (b), (c), (d), and among those of (a), (b), (c), and (e), which had been transferred to normal sea-water — just such forms as are sometimes found among normal unfertilized Arbacia eggs which have been kept undisturbed in sea-water for about 24 hours. Alcoholic extract. — XXXI. Eggs from 18 sets of ovaries, after segmenta- ^ The extracts of the fertilized eggs were no more destructive to the test-eggs than the sperm extracts had been. See page 63. 72 William J. Gies. tion had proceeded in many to the blastula stage, were ground in 20 c.c. of sea-water and extracted in this fluid phis 20 c.c. of 20% alcohol. Extraction was continued for 48 hours. The alcohol favored complete disintegration, for before 24 hours practically all of the cells were reduced to granules. A. Controls (2). B. Extract: (a) 15 c.c, (b) 8 c.c, (c) 5 c.c, (d) I c.c. C. Some of each of B transferred to 100 c.c. normal sea- water after 2 hrs. Result : After 12 hrs. a small number of cells in irregular initial segmentation were found among those of one of the controls, also in (d) and among those of (a) which had been transferred to sea-water. The number was less than 10 in 1,000. Discussion of Results. The chief feature of the results we have obtained is their negative character. Occasionally segmentations were noted, but these were few and rarely went beyond the 2-cell stage. Further, when the test-egcrs segmented those of the controls did also. These few divisions could not have been due to spermatozoa, since not a single group was surrounded with the fertilization or so-called " vitelline " membrane, whose absence, Loeb^ has indicated, practically proves non-spermatic influence. Thousands of eggs in the control and extract series were carefully examined in each experiment and yet only a trifling proportion showed initial segmentation ; excepting very few, none of these went as far as the 8-cell stage; and no morula or swimming larva was ever seen. The conditions of the experiments were made as nearly normal as possible and every precaution was taken to guard against evapo- ration. Special ion parthenogenesis was entirely excluded, therefore. All of the eggs were ascertained to be ripe and susceptible to seg- mentation influences. Sufficient variety of extraction process was employed to guard against failures in withdrawal method and the many experiments excluded accidental sources of error. It seems necessary to conclude, therefore, that the occasional segmentations in initial stages that were observed were only such as have repeatedly been seen in ripe unfertilized Arbacia eggs which have been exposed to sea-water for from twelve to twenty-four hours.^ I have not exhausted the means commonly used for enzyme extrac- tion. The time at my disposal for this work, and the facilities of ^ LoEB : This journal, 1901, iv, p. 454. 2 LoEB : Ibid., 1899, iii, p. 136; 1900, iii, pp. 436 and 437. Development of Mature Ova. 73 this laboratory, have not favored the trial of every known method nor attempts to devise new ones. It may be that sperm enzyme is as intimately connected with the structural elements of the cell, and as resistant to extraction processes, as Fischer has found the invert- ing ferment of Monilia Candida to be. Buchner's experience with zymase has not been overlooked, nor the suggestions it offers ignored. However, unless the hypothetical sperm enzyme were very different from most of the others, the numerous methods employed would have succeeded in bringing it to light, if any enzyme action can be exerted by substance in fluids surrounding the ova. It should be recalled in this connection that Loeb^ has recently made a series of experiments with various foreign enzymes to deter- mine proliferative power on unfertilized Arbacia eggs, but with negative results. He states that " the only enzyme that caused the ^g^, to segment at all was papain," but he could not be certain that this was not due to some accidental constituent of the sample of enzyme used. "The other enzymes were absolutely without effect." Two years ago Mathews, in some unpublished experiments cited by Loeb,- tried the effect of rennin on unfertilized eggs of the sea-urchin. The eggs were placed in sea-water solutions of rennet tablets for a while and then transferred to normal sea-water, when segmentation into a comparatively small number of cells resulted. The effect closely resembled those previously described by Alorgan,-^ and Ma- thews concluded that the results noted had been produced not by the enzyme, but by the salts in the tablets increasing the concentration of the water. Negative results rarely justify sweeping deductions. The outcome of these experiments, negative in detail, rather emphasizes possibili- ties which have not yet been specially considered. It may be that either too much extract was employed in each series for positive results to occur or else possibly not enough was taken. Such pos- sibility led to the wide variations of quantity and condition in these experiments, but as no differences were noted between the effects of the largest as contrasted with the smallest proportions of extract, the results afford no conclusive answer in this connection. Again, since enzymes are indiffusible, or, at most, are only very 1 Loeb: This journal, 1901, iv, p. 456. 2 Loeb: Ibid., 1900, iii, p. 437. ** Morgan: Archiv fiir Entwickelungsmechanik der Organismen, 1S99, viii, p. 448. 74 William J. Gies. slightly diffusible, it is possible that, in experiments of the kind con- ducted by Loeb, Mathews, Fieri, Winkler, and myself, enzyme which may be contained in the extract does not or cannot enter the sub- stance of the ovum. It might be assumed that mere contact with enzyme in such solution would not cause segmentation and that, even if the peripheral portions of the cytoplasm should be directly affected by such immersion, the general effect would be entirely different if contact, or diffusion, occurred within the substance farther toward the nucleus. Further, may not the morphological character of the spermatozoon, specially adapted as it is for great motility and penetration, imply that segmentation by indiffusible enzyme, con- tained in fluid surrounding the ovum, is no more possible in artificial than in normal fecundation. If it be ever found that enzymes, or zymogens, are causative influences in natural fertilization, I venture to predict, in view of the results of these experiments, that their action will also be shown to depend on their direct delivery to points wiiJiin the ovum. The results of this v^ork do not warrant any additions to current speculations on the mechanism of fertilization, but a recent sugges- tion may seem to be connected with these results and therefore should be considered here. Loeb,^ referring to his experiments with Echinoderms and Anne- lids, has expressed the view that " the spermatozoon can no longer be considered the cause or tlie stimulus for the process of develop- ment, but merely an agency which accelerates a process that is able to start ivithout it, only much more slowly." Accordingly it may be assumed that "the spermatozoon carries a catalytic substance into the egg." Loeb considered that enzymes and ions may be among these " catalytic substances." If ions are to be reckoned among the agents of proliferation, why it may be asked, did they not make active the sperm extracts used in these experiments .■* But what is the proportion of dissociated electrolyte in the spermatozoon and in such extracts, it may be in- quired in return .-* The composition of the ash does not furnish an accurate idea of the amount in the spermatozoon of salts pre-existent as salts and dissociable in extracts. Arbacia spermatozoa have not been analyzed in this connection nor the amount of dissociated q\qc- trolytes in these extracts determined. We know little of the relative proportions of the various constituents of spermatozoa and ova. As ^ Loeb: This journal, 1901, iv, p. 456. Development of Mature Ova. 75 we have no knowledge of the absolute or relative quantity of free ions entering or acting within the ovum, we therefore know nothing of the influence or sufficiency in this connection of the methods used in these experiments. Further, the ions which become active in the ovum may be originally a part of the molecules of the proteid com- pounds of the ovum or of the sperm, or of both, until the sperm mingles with the protoplasm of the ovum and forms new and proba- bly simpler combinations. These experiments were neither intended for, nor were their conditions suited to an investigation of these particular problems. The results therefore cannot be interpreted as having any bearing on them. It may not be amiss to state, before concluding, that Vigier's ^ assumptions that unfertilized eggs of Arbacia develop into swimming larvae in normal sea-water were invariably contradicted by my nu- merous experiments. Vigier says he was unable to repeat Loeb's results on artificial parthenogenesis. I have often used Loeb's methods with success in order to determine the responsive character of the eggs used in the extract series.^ Swimming larvae can be produced and reared to the pluteus stage with ease. Summary of Conclusions. The positive experimental results of Fieri should be attributed to the action of spermatozoa which had not been removed from the extracts. Winkler's uncertain results were doubtless the effects of osmotic influences. Extracts of the spermatozoa of Arbacia, which have been made by the ordinary methods for the preparation of enzyme solutions, and used in the proportions and under the conditions of these experi- ments, do not possess any power of causing proliferation of the ripe ovum. No evidence could be furnished of the existence of a zymogen in spermatozoa. Extracts of fertilized eggs in the earlier stages of development seem likewise to be devoid of any segmental activity. The extracts did not produce the typical peripheral "vitelline" membrane always formed immediately in Arbacia eggs, on fusion of the male and female elements. 1 See Loeb's criticism : This journal, 1901, iv, p. 454. '^ See references in this connection on p. 57. 76 William J. Gies. These negative results cannot be put forward as proof that there are no enzymes in spermatozoa which function during the normal process of fertilization. They do not show that enzyme action is impossible after, or at the time of union of the spermatozoon with the ovum within the latter, although the results of Series XXIX-XXXI might be interpreted as suggesting that enzymes are not thus elaborated. In conclusion I wish to thank Professor Loeb not only for the suggestions which led me to undertake these experiments, but also for much kindness and encouragement. CONCERNING THE POISONOUS EFFECT OF PURE SODIUM CHLORIDE SOLUTIONS UPON THE NERVE- MUSCLE PREPARATION.! By HARVEY GUSHING. [From the Physiological Iiistittite of Bern. \ IT was an accidental discovery in the preparation of so-called physiological salt solution that, when sodium chloride was added to tap-water drawn from a certain source of supply, the solu- tion was more efficacious than a corresponding one made from dis- tilled water. Upon investigation the tap-water was found to be rich in calcium and potassium salts. To Sidney Ringer is due the credit of recognizing this and of appreciating the significance of the fact that minute amounts of these salts would antagonize the injurious effects upon animal tissues of the pure sodium salt alone. The results given in his remarkable series of papers in the Journal of Physiology since 1880 have received confirmation on many sides, though the views as originally expressed in explanation of the process have of necessity undergone some alteration. More recently the extraordinary series of papers by Loeb - and his pupils has further laid especial emphasis on the directly injurious or " poisonous " effects of the pure sodium solution when used alone. His study, originally directed toward certain low forms of marine life, has shown that the pure solution of sodium chloride, even though isotonic with the sea-water from which the animals were taken acts as a strong poison whose effects are due to the unantagonized sodium ions. The same is true of an equimolecular solution of any individual salt, as calcium chloride, etc. The presumptive explanation is that the various ions form injurious combinations with the proteids of the tissues and in order to prevent this it is essential to have a medium 1 It is a pleasure here to express my gratitude to Professor Kronecker for giving the incentive which led primarily to the making of these observations, and as well to Dr. Asher for many helpful suggestions during their progress. LoEB : This journal, iii, iv, and v. n 78 Harvey C2isJiing. in which an interchange of ions between the protoplasm and the solution no longer takes place. It is the existence of such a medium which accounts for the superiority of the so-called Ringer's solu- tion. This fact has led Loeb ^ to remark that his experience would seem to necessitate " the introduction of a new conception, namely that of physiologically balanced j-a// solutions " 5 meaning by this " salt solutions which contain such ions and in such proportions as to completely annihilate the poisonous effects which each constituent would have if it were alone in solution." Howell ^ and Greene^ in recent papers have laid stress on the fact that the calcium salt stands in an especial relation to the calling out of cardiac contraction, and some writers go so far as to state that it is the presence of this salt alone in the blood serum which gives to it its superiority as an infusion material. Howell has brought out the interesting fact that the gradual increase in the percentage of calcium chloride in solutions in which a batrachian heart has be- come exhausted will reawaken its spontaneous contractility. The relation of nourishment to the prolongation of cardiac activity was first pointed out in 1875 by Kronecker and Stirling. At the same time these authors laid stress rather upon the general properties of blood than upon its. saline constituents. However this may actually be, experience has shown that a more perfect supporting fluid for an isolated heart or muscle is furnished by the blood serum of certain animals than by any known saline combination. Brunton and Ringer* have made some comparisons between the influence of these various solutions upon cardiac muscle and the ordinary striped muscle of the extremities. Ringer and Locke together subsequently showed that sodium chloride solutions would influence and produce variations in the con- traction curve of a muscle and that certain electrical phenomena would also be called out. Locke,^ furthermore, in a brief provisional note has published the results of some observations upon the effect of placing an isolated sartorius muscle in baths of various salt solu- tions. His observations seem to have been complicated by the ^ LoEB : This journal, 1900, iii, p. 445. 2 Howell: Ibid., 1898, ii, p. 47. ^ Greene: Ibid., 1898, p. 83. * Ringer: Journal of physiology, 1887, viii, p. 22. Ringer and Buxton: Ibid., p. 288. '" Locke: Centralblatt fiir Physiologic, 1894, viii, p. 166. Effect of Pui'e Sodium Chloride Solutions. 79 appearance of muscular fibrillations such as are usually produced by the action of pure solutions of sodium chloride. He also directs attention to the persistence of direct irritability after stimulation from the nerve has ceased to call forth contractions, and also that the injurious effect of the pure solution of sodium chloride could largely be counteracted by the presence in the bath of the proper proportions of calcium and potassium salts. So far as I know, his detailed report has not been published. The observations which the writer will briefly recount are in a measure confirmatory of the results which Locke has summarized. The results, however, may have been more precise, inasmuch as they have been obtained largely by infusion methods in a comparatively intact animal rather than by simply immersing an isolated nerve and muscle in a bath. The particular method of preparation offers possibilities of obtaining more prompt and pronounced reactions than does the bath method, and its employment has led to some interest- ing observations. The various fluids which have been used have been brought into contact with the muscle by perfusion, through the abdominal aorta and its return veins, or by direct injection into the belly of the muscle itself. Preparation. The following method of preparation possesses advantages not only in the ease of its accomplishment but in the fact that it is unas- sociated with mutilation of the animal, so that if desired, as in the second series of these observations, the normal circulation may be continued. The frog is pithed and the brain and cord broken up. A vaso- motor paralysis results which aids the subsequent perfusion, as the vessels remain dilated. The gastrocnemius tendon on each side is exposed, divided, and isolated without injury to the neighboring artery and vein. A median dorsal incision (r/ sketch) is then made, exposing the spine and pelvis from behind. The long Os coccygis is then lifted up by its unattached posterior extremity, carefully cut away from the adjoining muscles and dislocated by the insertion of a knife point at its articulation with the sacral vertebra. This entire performance should be absolutely bloodless. In the bottom of the wound thus made, the distended abdominal aorta with its bifurcation is easily 8o Harvey Cms king. accessible. The lumbar plexus also lies exposed on each side and may be freed slightly and caught by a loose ligature in order to facil- itate the subsequent introduction of the electrode under it. A glass cannula of a particular form (r/". sketch) may then be readily introduced into the aorta, even in small frogs. This is a dif- ficult matter with a collapsed vessel or by the usual method of expos- ure through the abdomen. The cannula is securely fastened to the animal's back and remains firmly in place without risk of tearing the vessel by any subsequent manipulations. One of the iliac arteries may be caught with a delicate clamp, or tied, in order that the infusion shall enter one leg only while the other is preserved in normal con- dition for a control, or for a double preparation to be used in another irrigation. The abdominal and iliac veins are then opened to allow the per- fused fluid to pass out, and so to largely prevent any subsequent oedema. The animal is fastened to a board by sharp pins passing through the joints and pelvis in such a way that the muscle contractions are not interfered with nor the blood- vessels compressed, although the preparation is made immovable. The irrigation should be started as soon as possible lest coagulation occur in the cannula or in the small capillaries and thus prevent a successful circulation. The circulation has usually been controlled during the perfusion by the microscopi- cal examination of the capillaries in the web of the foot. Not only can the freedom of the circulation be thus ascertained, but from the comparative number of corpuscles remaining in the capillaries an idea can be gained of the degree of existing anaemia. The various fluids for washing out the vessels were held in gradu- ated burettes or Mariotte flasks which could be raised or lowered so Sketch illustrating method ot making nerve-muscle preparation. Showing exposure of Aorta abdominalis with the Plexus cruralis on each side, after removal of Os coccygis ; also form of cannula and its introduction, preparation of gastrocnemius, etc. Effect of Piire Sodium Chloride Solutio7is. 8i that the degree of pressure and rapidity of flow might be controlled and the amount of fluid passed through the extremity measured. The muscle was stimulated by opening induction shocks, a battery of two Daniell elements being used. The irritation necessary to call forth the maximal twitch, both for direct and indirect stimuli, was first determined, and by means of a Bowditch clock and Kronecker key ^ (Spiilcontact), a series of maximal opening shocks usually with a four-second interval was then employed. In all cases the muscle was unsupported and carried a weight usually of 20 grams. The direct stimuli were given through needles inserted directly into the muscle belly {cf. sketch). A Pohl's commutator without cross wires was interposed between the coil and the preparation and so arranged that the effect of direct and indirect stimulation might be alternated as desired. A Kronecker^ automatic short circuiting apparatus was inserted with the secondary coil in order to throw out the closure shocks. Observations on the Effects of Artificial Circulation on Muscle Contractions. Pure NaCl solutions. — If a 0.6 per cent solution, presumably iso- tonic with the protoplasm of the frogs' tissues, be allowed to circulate through the extremities there will be a comparatively early disappear- ance of irritability to indirect stimuli, which as a rule takes place long before there can be a complete washing out of the blood-vessels. This may be easily corroborated by the fact that the capillary circu- lation in the foot still contains many corpuscles and the return flow through the abdominal vein is still reddened with them. It will be found, however, that the response to direct stinmli persists. If the irrigation be continued for hours, however, a condition of practical bloodlessness may be reached, as shown not only by the absence of corpuscles in the circulation but by the failure to find them with the microscope on teasing the muscle substance itself. In spite of this fact the closure contractions to direct stimuli, althoicgli usually less high than before, may persist {cf. Fig. i). I have succeeded with a frog thus washed out in obtaining them after a period of seventy-two ^ Kronecker: Zeitschrift fiir Instrumentenkunde, 1889, p. 242. 2 Kronecker : Ueber die Ermiidung und Erholuno; der quergestreiften Muskeln. Ludwig's Arbeiten, vi. Jahrg., p. 177. Auch K. S. Ges. Wiss. Math.-Phys. Bd. xxiii, Leipzig, 1872. 82 Harvey Ciishing. TO ,t_, ^ (!) . , t- ro < ^ ■=? — 0.0 o ,^ • • u "" u ■£ .2 •• "a; 3 5 == r; "o > (« -rt - ti '^ o 33 ^ -3 tj rS < - -° Q, |— ■ ^ CO t* S o -y ^ eL ^ ^. o ^ \0 ■ o CM o -35 • u >■> hours of constant irrigation at room temperature (with but one accidental interruption of three or four hours) and the employment of over three litres of the irrigating fluid. The presence of fibrillary twitchings in the muscle which the sodium chloride solutions often occasion have as a rule in no way complicated our " contrac- tion curves," as may be seen in the figures. Furthermore, this early failure of excitability from the nerve, has been shown to appear at variable times, de- pending upon the percentage of sodium chloride in the solution employed and the success of the perfusion. If a normal fatigue curve be made conse- quent upon a series of indirect maximal stimulations, the usual slow failure of response (in winter frogs) after several thousand contractions will be seen (Fig. 2). If now for comparison the other leg be washed out, with for in- stance a 0.6 per cent NaCl solution the failure of response will occur after a few hundred contractions (Fig. 3.) If a weaker solution be employed, in spite of its more irritating effect on the tissues, failure of response is less rapid. Even with pure water, although tetanic contractions may be produced and " contracture " be a marked fea- ture, the muscle will continue to react when stimulated from the nerve longer than when the saline solutions are used. If we increase the percentage above 0.6 per cent the contractions as a rule fall off very rapidly after the irrigation has begun (cf. Fig. i). Effect of Pure Sodium Chloride Solutions. 83 Inasmuch, however, as under all circumstances responses to maxi- mal direct stimuli are in a great measure preserved, it is evident that the chief injurious effect from the salt occurs somewhere in the course of the nerve itself. Shifting the electrode from the exposed plexus to a section of nerve nearer the muscle, as the sciatic, gives no improvement in the response. The effect in many ways, there- fore, can be seen to be distinctly analogous to that of weak curare solutions and presumably is the result of an injury to the nerve ends themselves. In review, it may be said that generally speaking the number of contractions which may be called forth from a mnscle by indirect stimuli after beginning an irriga- tion with solutions of sodinm chlo- ride diminishes with the percentage of the salt in the fluid. Further- more that inasmuch as direct excitability persists, the process of washing out may be considered not to lead to a more rapid ex- haustion of the muscle, but rather to have an injurious effect upon the nerve-endings. There are, however, many of my observations which would tend to show that the nerve ends them- selves not only are capable of fatigue under ordinary normal circumstances, but that their sus- ceptibility to exhaustion under E 3 jj c •— u ^ - 2i o. •r; C M o 5 .- = ca S! c M^ o ^ (M — — '^ ~ .i; 84 Harvey Gushing. these conditions of washing out may be increased. In the first place it has been found possible by indirect stimuli to reach a period of fatigue of a non-irrigated muscle at which practically no further response from the nerve may be given even when the elec- trode is shifted to a new situation nearer the muscle, and yet when directly stimulated the muscle will contract in degree increasing with the strength of stimulus (Fig. 2)- Inasmuch as nerve itself has been shown to be incapable of fatigue, this would point to a local fatigue of terminal organs apart from that of the muscle itself. Secondly, if, during the process of washing out the muscle with a saline solution, the four-second shock be discontinued for a short period although the irrigation in the interim is kept up, it will be found on restimulating that the contractions called forth will be higher than before although lower contractions would be expected from a poisoning effect alone. It may be shown furthermore that this is not due to rest of the muscle, since, if instead of discontinuing completely the stimuli a series of direct stimuli be thrown in so that the muscle itself shall be irritated in the regular periods and yet the nerve in a measure given a rest, on returning to the indirect stimuli they will be found to have become improved. It is thus seen that the question of increased susceptibility to fatigue, apart from the directly poisonous effects of the sodium ion, bears some relation to the problem, though presumably a relatively small one. After a point has been reached in the irrigation at which no further indirect response is obtainable, upon continuing the perfusion, in spite of rest, no further contractions may be called out by way of the nerve. Thus with all pure sodium-chloride solutions at varying periods the so-called poisonous effect appears, though, as Loeb has shown, inasmuch as the same result follows upon the employment of such a fluid even when isotonic with the tissue fluids, the poisonous action may be considered to be a negative one and due to the absence of other essential salts, rather than to the mere presence of the sodium itself. . Effect of " physiologically balanced solutions." — If after a period has been reached in the process of washing out with a sodium chloride solution when no further contractions to indirect stimuli can be called forth, and then the irrigating fluid is changed for one containing the three ions in combination, contractions will reappear. Often this antidotal effect follows with astonishing rapidity, and before more Effect of Pure Sodium Chloride Solutions. 85 than a few drops of the fiulcl can have reached the muscle, the contractions will once more begin and soon may return to their maximal height {cf. Fig. I). Many factors, of course, influence this readjustment, which may not always be so prompt nor be associ- ated with the return of the maximal contraction. This depends largely upon the percentages o^ salines in the solutions both of the original fluid used to exhaust the muscle response and that employed for its recovery. The exact determination of these rela- tions would take an enormous number of observations. It has been found, however, that a solution slightly richer in calcium ele- ments than a " physiologically balanced fluid " is more efficacious in bringing about this recovery than one contain- ing a smaller percentage, which is sug- gestive of a similarity with Howell's findings in the case of the heart. A solution, for instance, containing 0.06 per cent calcium chloride is often more efficacious than one of 0.03 per cent. In very much the same way, and ordinarily with much more certain results, will a following irrigation of defibrinated blood lead to recovery. Often when the calcium-chloride-hold- ing solution has succeeded in causing only a partial return of the maximal contractions, blood serum will further complete the readjustment. The blood of different animals, however, has a widely different effect, just as in the case of the various artificially " bal- ^ '0 x _0 'ti "^. OJ " .^ ■*-• ^4-4 5 '■J ... 'J ■r. tn cj £ (J t/; a ^ ^ > "rt u rt u '>< r, ■" r- be .S ^ CD o P t5 § *-: rt o o 3 "H ° tj C 3 '*^ -. ;- bX) 3 \o p M 'S ^ rt — Th .n ■5, M c "^3 E^ (U u u N _c OJ .£ " ' >^ % c "rt CJ 01 (LI CJ .^ -d rt ^ ^ e^ '& -c "o ^ 7) en 5 n CTj .2 .2 TJ ^ rj ■^ — IS 3 en en en "3 A _o "rt u c^! .a a _rt rt p ni .b: ^ c^ K Q 3 _o ^ (N ro T^ C < o 90 Harvey Gushing. mortem, than one not so fatigued : indeed the very process of active muscle contraction seemingly bears a close relation to the permanent charge associated with rigor. It is here seen that the addition of lime salts injected into a fatigued muscle may produce intra vitavi, a condition at all events simulating rigor mortis. It is fully realized by the writer that this series of experiments has done little more than evolve a convenient method by which future observations relative to the action of various fluids on striped muscle may be continued. Much has of necessity been left undone relative to the effects of a greater number of solutions carefully graded with respect to their content of ions. There are furthermore many irregu- larities which may be encountered and for the final explanation of which a more extended number of observations will be necessary. It has been noticed, for instance, after a perfusion with a sodium solu- tion until indirect stimuli no longer call out contractions that the direct injection of a " balanced " solution into the muscle substance fails to counteract its effect, as will a perfusion of the same fluid. When the paralysis, however, has been occasioned by the injection method the same method of treatment does lead to recovery. Observations have been made upon fifty frogs (Esculenta and Tem- poraria), mostly during the winter months, at which time as a rule more pronounced reactions were obtained than in the short series made in May on spring animals. The observations permit the following conclusions : — The pure sodium-chloride solutions are injurious to the nerve- muscle preparation. The effect is in a measure related to the percentage of this salt in solution. Inasmuch as the response from indirect stimulation fails, while that from direct stimulation may persist, the result primarily affects the nerve ends. The injurious effect may be promptly counteracted by the blood or serum of certain animals or by the proper " physiologically balanced salt solution." By varying the percentage of the calcium ion in the solution, with certain limitations, proportionately beneficial effects may be produced. An excess of the calcium ion in certain cases of fatigued muscle may lead intra vitam to a permanent contraction of the muscle re- sembling rigor mortis. CEREBRAL PRESSURE FOLLOWING TRAUMA. By W. B.CANNON. (With a Preface by W. N. Bullard.) [From the Physiological Laboratory of the Harvard Medical School}^ CONTENTS. Page Preface, by W. N. Bullard 91 Introduction 92 Pathological conditions in traumatic intracranial lesions 93 Clinical findings in cases of traumatic head injuries 95 Theories to explain the increase of intracranial pressure after trauma 100 Experimental inquiry into the cause of brain pressure 103 Primary effects of cerebral trauma 104 Secondary effects of cerebral trauma 109 Summary . 120 Preface. A SOMEWHAT prolonged clinical study of traumatic cerebral lesions, and more especially those induced by blows upon the head, whether or not accompanied by fractures of the cranium, has shown that operative interference is often necessary. It has been apparent on examination of the condition of the con- tents of the cranium in cases of this character, where an artificial open- ing of the cranium was made, that in most instances the dura was tense and the brain did not pulsate. It was also soon proved that this condition did not necessarily depend upon the increase of intra- cranial pressure following the introduction of an extraneous substance, such as a large hemorrhage or clot, since it occurred in the absence of hemorrhages which could thus increase the intracranial pressure. How far this marked intradural tension was abnormal was at first undecided, but later it became evident that it was frequently greater than the normal average. It has long seemed to me that upon the experimental demonstration of the existence of this increase of intradural tension, and upon the determination of its cause, must logically depend our action in many cases of traumatic cerebral lesion. On account of the great practical 1 The first part of this investigation was made by means of a fund established by Dr. W. N. Bullard. 91 92 IV. B. Cannon. importance of these questions to the neurologist and the surgeon, I induced Dr. Cannon to undertake the following research. Dr. Cannon is entirely responsible for the method used in these ex- periments and for the conclusions drawn. Introduction. The term cerebral pressure has been applied somewhat indiscrim- inately by investigators to two conditions, to the pressure exerted on the brain by foreign objects in the cranial cavity, and to the pressure exerted by the brain, in response to internal forces, against the skull. It is desirable, for the sake of clearness, to separate these two con- ditions by means of distinctive terms. The force acting on the brain from without, for example, from depressed bone or subdural hemor- rhage, results in a compression of the brain, and to this phenomenon should be applied the term cerebral compression. The term cerebral pressure may then be restricted to the force with which the brain itself pushes against the surrounding skull. Since, however, the forces within the rigid cranium are in equilibrium, the intracranial pressure must vary directly with the intensity of either of these varying fac- tors, — the compressing foreign body, or the swelling brain. The result is manifestly the same, therefore, whichever factor increases, and con- sequently either of them may cause the so-called " pressure symptoms " which follow injuries of the head. This research is concerned chiefly with the second of the two forces mentioned above, the pressure ex- erted by the injured brain against the skull. In the following discussion it will be well to consider, first of all, the pathological conditions found in cases of head injury, since thereby the phenomena to be explained may be presented in a sche- matic outline which will be serviceable throughout the subsequent treatment of the subject. The clinical findings which accompany the anatomical changes will next be regarded, and will lead to an exami- nation of the pressure symptoms prominent in the cases. In the third division of the paper the theories put forth to explain the increase in intracranial pressure will be discussed ; and finally I shall give the results of my experiments and point out what seem* to me to be the significance of oedema as an active agent in the cause of death after cerebral traumatism. Cerebral Pressure following Trauma. 93 I. Pathological Conditions in Traumatic Intracranial Lesions. The careful macroscopic and microscopic examination of the patho- logical changes in two hundred and twenty-five cases of traumatic intracranial lesions, reported by Phelps,^ has led to a recognition of the logic and trustworthiness of his classification and terminology of these lesions. Phelps divides the lesions primarily into two general classes : the direct, which include the structural changes following the injury more or less immediately, and the indirect, which in- clude the secondary inflammations. The direct effects of injury are of first interest in this discussion. Of these Phelps names three groups which have clinical significance, — hemorrhages, contusions, and lacerations. It will be unnecessary to present a detailed descrip- tion of the variations of these three pathological states ; only such general appearances will be mentioned as have a bearing on the further development of the present argument. Hemorrhages may be epidural, from the vessels of the diploe, the sinuses or the middle meningeal artery; pial, from rupture of vessels of the pia mater; or cortical, the direct result of injury to the brain substance. In any of these three regions the bleeding may be con- siderable or slight. Two effects of hemorrhage should be noted at this point, an immediate effect from profuse hemorrhage, and a more remote effect resulting from slighter effusions. Duret^ has shown by injecting wax into the cranial cavity that a diminution of the intra- cranial space by five per cent produced somnolence and coma, while a diminution by eight per cent resulted in death. Epidural hemor- rhages are the most liable to be profuse, and if sudden they may produce an immediate loss of consciousness, and result in death by simple mechanical compression. According to Phelps, pial and corti- cal hemorrhages are rarely sufficiently copious to produce marked cerebral compression ; the pial hemorrhage characteristically forms merely a thin sheet over the vertex. "The general symptoms which attend these inconsiderable curtailments of the intracranial space, whatever their nature, must therefore be ascribed to other causes than a general circulatory disturbance occasioned by the contraction ^ Phelps: Traumatic injuries of the brain and its membranes, New York, 1897. -' DuRET : ttude experimentale et clinique sur les traumatismes cerebraux, Paris, 187S, p. 186 et seq. 94 ^' B. Cannon. of cranial capacity." ^ These hemorrhages are, however, rarely uncom- plicated. Almost invariably they are associated with some degree of general or local contusion, and hemorrhage and contusion together result in a vascular disturbance which leads to deficient nutrition of the brain tissues. The consequences of the deficient nutrition of the tissues will be the main subject to be considered in this paper. The phenomena of contusion probably exist in all cases of intra- cranial injury, and may be general or confined to the cerebrum. The visible anatomical changes are as follows,^ " a distention of the parenchymatous vessels, a general formation of minute thrombi, the presence of punctate extravasations, and a more or less distinct oedema. . . . The minute thrombi are the most characteristic of the several morbid conditions which have been enumerated. . . . The oedema, which is variable in amount, sometimes appreciable only after some delay and a close inspection upon section, and at other times so profuse that the fluid can be squeezed from the brain by the hand as from a sponge, is notably frequent. All these abnormal con- ditions, the extravasations, thrombi, and oedema, are simply measures of the general hyperemia which immediately preceded death." Con- tusion, therefore, like hemorrhage, manifestly presents, in the extra- vasations and thrombi, conditions for circulatory disturbance and a consequent deficient nutrition of the tissues; that the oedema is a secondary result of these conditions will be shown later. Laceration of the brain is the most severe effect of external vio- lence. It is characterized by rupture of vessels and hemorrhage so profuse that the brain substance may be broken down in all direc- tions. The changes present in lacerations are, however, of no special interest in so far as they differ from those present in hemorrhages and contusions, and may accordingly be passed without further description. From the foregoing review of the pathological findings in cases of traumatic injuries of the brain, it is evident that extensive lacerations and profuse hemorrhages are in themselves conditions capable of causing immediate death. With slight hemorrhages, and especially with contusions, the primary anatomical alterations are often so in- considerable that a more general secondary process must be invoked to explain the fatal issue. What that secondary process may be remains to be discussed. As primary pathological changes, there are the slight pial and cortical hemorrhages, the minute thrombi and punctate extravasations and the vascular distention manifest in con- 1 Phelps: Loc. cit., p. 51. -' Phelps : Loc ciL, p. 53. Cerebral Pressure following Trauma. 95 tusion. The oedema, which is a constant concomitant of these primary changes, is, I believe, a secondary result from them, and is to be held accountable for the serious conditions which follow from apparently slight pathological alteration. Before considering further the manner in which oedema may act in causing death, it will be well to present the evidence offered by clinical cases. II. Clinical Findings in Cases of Traumatic Head Injuries. The symptoms resulting from injury to the brain vary greatly with the severity of the lesion, the nature of the pathological changes — whether hemorrhage, contusion, or laceration — and with the location of the injury in the brain. To enumerate the many variations in symptoms would not be pertinent to this discussion ; it will be sufB- cient to state the general symptoms. In cases of brain lesion there is usually a history of a blow or a fall on the head, or evidence of the rupture of an intracranial blood vessel. The most notable and most constant of the primary symp- toms is unconsciousness. The initial unconsciousness may persist until death, or the patient may regain his senses. If consciousness returns, it is commonly lost again after gradually increasing dullness passing into stupor. The final appearances, whether consciousness is recovered or not, are deeper and deeper stupor, then coma, from which the patient cannot be aroused, and ultimately death. In this final stage certain typical signs manifest themselves. The temper- ature characteristically rises as the time of death approaches, the pupils do not react to light, there is stertorous breathing and slow heart-beat. Sometimes there is paralysis of the face and limbs, though there may be clonic spasms in various muscles. Passage of urine and faeces is also an occasional occurrence. Between the symptoms preceding death from injury to the brain, and the symptoms produced in animals by increasing intracranial pressure or by otherwise causing cerebral anemia, there is a striking similarity. The pressure symptoms observed in animals have been summarized by Bergmann ^ as follows. First, there is evidence of pain due to tension of the dura mater. This is followed by stupor, sopor, and coma. When the pressure is suddenly applied, clonic spasms and sometimes roll and circus movements are noted ; but 1 Bergmann : Deutsche Chirurgie, Die Lehre von den Kopfverletzungen Stuttgart, 1880, XXX, pp. 341 et seq. 96 W. B. Cannon. these phenomena are not seen when the pressure is applied slowly. A slow heart, slow, deep, snoring respiration, vomiting, and emptying of bladder and rectum are characteristic symptoms. In the experi- ments on animals, it was discovered that in order to produce death, the intracranial pressure must equal the carotid pressure, /'. e., anemia must be produced. The manifest resemblance between the clinical symptoms and the experimental phenomena has led to the applica- tion of the term "pressure symptoms" to the peculiar symptom complex following brain injuries. The evidence of intracranial pres- sure observed in these cases on operation confirms the validity of the term. A further insight into clinical conditions attending intracranial lesions, which will serve to make the problem more distinct, and the nature of the explanation more clear, will be secured by an examination of a series of typical cases. These cases will illustrate three facts : i, that there may be injury with pressure symptoms and recovery ; 2, that pressure symptoms may occur with only slight gross lesion ; and 3, that the oedema, which attends the cerebral lesions, is the result of a process requiring considerable time for its operation. I. Cases of head injury vrith pressure symptoms and recovery. — Two cases reported by Walton ^ will illustrate this type. Case I. A boy of six years was struck by a bicycle at noon one day and rendered dazed, but not unconscious. That evening he had fever and vomited, was restless at night and the next morning vomited again. During the day he became drowsy, then unconscious with unilateral paraly- sis including the face. Defecation and micturition were involuntary. That evening operation was considered, but, in view of slight improvement in the conditions, was postponed. On the fourth day after the accident the pressure symptoms had entirely disappeared, and on the sixth day the child was apparently well. Case II. A child, aged three and a half years, fell from a swing and struck on her head. She was dazed, and later she vomited. The next day the left arm was paralyzed. On the third day after the injury the paralysis began to disappear, and thereafter the recovery was rapid. In each of these cases the gradual onset of pressure symptoms some time after injury, and their gradual subsidence, should be care- fully noted. ^ Walton : American journal of medical sciences, 1898, cxvi, p. 270. Cerebral Pressure following Trauma. 97 2. Cases with pressure symptoms -with only slight gross lesion. — The phenomena presented by this class of cases are especially significant ; evidence of considerable intracranial pressure is demonstrable, but on examination no hemorrhage or depression of bone sufficient to cause the symptoms is to be discovered. Case III. Patient fell down a flight of stairs. Upon admission to hospital, semiconscious, and irritable when aroused. Dilatation of right pupil, which was irresponsive to light; no muscular symptoms. He re- mained in a restless, delirious, or stupid condition until his death on the seventh day. There was loss of urinary control on the fourth day, and coma, with picking at the bedclothes, and subsultus tendinum during the last twenty-four hours. The temperature, which was 98.2° on admission, rose slowly and progressively to 102° on the fifth day, to 104.2° on the sixth day, and to 105.2° one hour before death. No fracture of the skull and no epidural hemorrhage ; no superficial laceration; pia mater and cortical vessels very much congested; some opacity of arachnoid membrane, and moderate subarachnoid effusion ; no pial hemorrhage; limited cortical contusion, area of one inch in diam- eter, at bottom of left fissure of Silvius ; laceration of left optic thalamus in its central portion, one fourth inch in diameter and filled with clot. Sub- cortical lacerations of the left side of the pons, one third inch in diameter, in the transverse fibres ; a few punctate extravasations in different parts of the brain ; general hyperemia and well marked oedema.^ Case IV, reported by Walton and Brooks." A young woman, thrown from her horse, struck on her head and was carried home unconscious. Four hours later was still unconscious, breathing quietly, with a pulse of 100. The pupils were equal, somewhat dilated, and reacted sluggishly. There was partial paralysis of the left side of the face, and complete paraly- sis of the left arm and leg. The patient had vomited once or twice since the accident. Restlessness supervened and consciousness partially re- turned. Respirations were shallow. There was incontinence of urine. The left arm and leg became rigid. On the second day the rigidity of the left arm and leg became less marked, and limited voluntary movements appeared ; but in the evening the rigidity and paralysis of the left side were very well marked, and upon efforts to arouse her no response was made. Operation was performed at this time. A small trephine button was removed about two and a half inches above the right external auditory 1 Phelps: Loc. cif., p. 536. - Walton, G. L. and W. A. Brooks, Jr. : Boston medical and surgical journal, 1897, cxxxvi, p. 301. 98 IV. B. Cannon, meatus. No fracture of either table was discoverable, but the dura was seen to be tense and non-pulsating. Upon its incision an ounce of clear fluid spurted through the opening. The brain appeared somewhat oedeni- atous and prominent, but otherwise normal. Exploration under the dura revealed no sign of hemorrhage. Until the fifth day after the injury there was slight improvement, but from this time on the general condition varied from moderate delirium to som- nolence. I'he patient gradually sank. On the fifteenth day, temperature, pulse, and respiration rose rapidly, and on the sixteenth day the patient died. At autopsy there was no sign of any fracture of the skull, and no evidence of extra- or intra- dural hemorrhage, or hemorrhage of the meninges. On section several hemorrhagic softened areas were found, — two, about the size of beans, in the left frontal lobe, and another near the outer margin of the right optic thalamus. Various minute hemorrhages were scattered over the brain. Such cases as these have led to the opinion that the oedema itself may in some way cause the pressure symptoms. Walton states, in reference to Case IV., "that the oedema played a part in the produc- tion of the hemiplegia can hardly be doubted in view of the disappear- ance of rigidity and improvement in motion following the relief of pressure by operation," and he remarks, furthermore, " the import- ance of remembering that a fatal result may follow concussion with- out tangible gross lesion, unless, indeed, the two small hemorrhages in the frontal lobe, with subsequent softening, are considered ade- quate cause for a fatal issue." ^ Bullard^ had previously called attention to these same phenomena. "The cause of increased intra- cranial tension," he declares, " is not altogether plain. It is not by any means, as is sometimes supposed, always a pressure from intracranial hemorrhage. ... In many cases there is no evidence of any severe hemorrhage, and yet the increased pressure is apparent. Again, the increased pressure in all probability occurs in cases of so-called con- cussion and in other mild cases where unconsciousness exists, but where there can be no question of any profuse hemorrhage. What seems to occur is this : the brain in some way acts as a sponge and pushes so hard against the dura as to inhibit or diminish pulsation. If in these cases the dura is incised, the cerebral pulsation again be- comes visible, and the relief to the patient is instantaneous and ^ Walton and Brooks : Loc cit., p. 304. ^ Bullard: Boston medical and surgical journal, 1895, cxxxii, p. 74. Cerebral Presstire following Trauma. 99 extraordinary." In another paper Bullard ^ again declares that the pathological process resulting in increased intracranial pressure is apparently a swelling of the brain itself, due in part to a filling and dilatation of the blood vessels, in part to oedema of cerebral tissue resulting therefrom, and in part also to the excess of fluid between the pia and the dura. That the subdural fluid is not alone the cause of pressure symptoms is shown by the fact that in many cases the in- crease in fluid is not very great and may not be apparent. Further- more, the tension and protrusion of the brain after cutting the dura indicates that the intradural pressure is not in any great degree due to the fluid which has escaped. The pressure is, therefore, due to a swelling of the brain itself. The third fact to be illustrated, namely, that the oedema attending cerebral lesions results from a process reqiiiring appreciable time for its operation, is made clear by a comparison of the findings in cases like those cited above, in which the time element is noteworthy, with the findings in other cases observed soon after the reception of the injury. The presence of oedema in the former condition is usually in sharp contrast with the absence of oedema when the brain is seen immedi- ately after the accident. A single case will illustrate this negative evidence. Case V. Female, fifty-two years old, fell from an electric car, was ad- mitted to the City Hospital shorijy after. Conscious ; pupils equal ; no paralysis anywhere. There was a depression in the right temporal region. The surgeon trephined in this region, the operation being done under ether. No extradural clot. The dura was tense and scarcely pulsated On cutting the dura the brain bulged. No excess of cerebro-spinal fluid noted : re- covered well and quickly.^ In this instance an almost immediate operation gave no time for the cedematous condition, so characteristic an accompaniment of brain injury, to develop. The bulging pulseless condition of the dura at this early stage will be explained by the experimental data to be presented. A summary of the points thus far made in the discussion will per- haps serve to make clearer the further development of the subject. From a study of the pathological alterations after injuries to the brain it was evident that there might be hemorrhages and lacerations so ^ Bullard : Boston medical and surgical journal, 1898, cxxxviii, p. 272. ^ Bullard : Medical and surgical reports, Boston City Hospital, 1895, p. 64. loo W. B. Cannon. severe as to result in almost immediate death. It was also evident that in many instances the initial lesions — the scattered minute thrombi and punctate extravasations — were so slight as to be in themselves no adequate cause for a fatal issue. That this inference is true is indicated by the recovery from the initial loss of conscious- ness, — a result hardly to be expected if the primary lesions were in- trinsically fatal. In cases of this nature, however, the recovery of consciousness was not infrequently followed by a slow subsidence of conscious life, with a progressive increase of the so-called pressure symptoms until death supervened. Leyden ^ long ago showed that in order to produce death intracranial pressure must equal arterial pressure, and this conclusion has been confirmed by Duret,^ Cybulski,^ Hill,* and others. Apparently what occurs in these puzzling cases is this : the intracranial pressure somehow rises higher and higher until it reaches a degree equal to the blood pressure in the arteries. At this point the circulation in the cerebral blood vessels comes to a standstill, the vital centres in the bulb no longer receive their nor- mal nutrition, they become paralyzed and life ceases. Now the ques- tion presents itself: in what way is the intracranial pressure gradually and progressively increased until the pressure in the blood channels to the brain is overcome and death results? It is clear that the pressure symptoms which slowly manifest themselves are dependent on second- ary processes following the primary lesions of the brain; it is clear also that the oedema follows the primary lesions of the brain. Is there any causal relation between these two phenomena? III. Theories to explain the Increase of Intracranial Pres- sure AFTER Trauma. The most noteworthy theory advanced to explain the secondary increase of cerebral pressure following injuries of the brain is that of Bergmann. This same theory is held by Hill. More recently an explanation has been presented also by Courtney. The position taken by Hill '^ and Bergmann *^ has been clearly stated 1 Leyden: Archiv fiir pathologische Anatomic, 1866, xx.xvii, p. 519. - Duret: Loc. cit., p. 183. 3 Cybulski: Centralblatt fiir Physiologie, 1890, p. 835. *Hill: The physiology and pathology of the cerebral circulation. London, 1896, p. 168. ^ Hill: Loc. cit., p. 188, et seq. ^ Bergmann: Loc. cit., p. 420. Cerebral Pressure following Trauma, loi by Hill as follows. The primal condition is an intracranial hemorrhage which acts as a localized foreign body within the cranium. Since the brain is inclosed in a rigid case, and the brain substance itself is in- compressible, this localized foreign body must occupy the space of a certain vascular area, that is, it must cause an obliteration of the capil- laries and veins in the region it occupies. As a consequence the local cerebral tension will be equal to that of the arteries which normally feed the affected area. In the obliterated area there will be complete stasis of blood. The transmission of the increased tension through the brain substance to the veins and capillaries of the border areas will cause a higher blood pressure and a lessened blood flow in these vessels. In more distant areas the circulation is more normal and the blood-flow may have even a compensatory increase of speed. Accord- ing to Hill, the secondary increase of pressure may now be established in two ways, the first of which alone will be described, as the second does not concern the class of cases under consideration. The high blood pressure in the border areas will lead to increased transudation of fluid, because plasma may pass more easily into the brain substance than blood through the compressed capillaries. The transudation will take place at almost arterial tension, will increase the volume of the foreign body, and so lead to compression of other capil- lary areas. Thus is established a vicious circle of processes, and the cerebral anemia may spread indefinitely. The noteworthy feature of this theory is the assumption that the transudation occurs because of high blood pressure in the capillaries of the border areas. It should be observed, however, that the high pressure in these vessels is due, not to a general increase in blood pressure, but to the fact that external pressure in the brain substance about the vessels increases until it partially overcomes the internal pressure in the capillaries themselves. If " the plasma may pass more easily into the brain substance " (where tension is so high that the blood vessels are being compressed) " than blood through the com- pressed capillaries " (in which flow can still occur) there would be the obviously impossible condition of fluids passing from a region of less to a region of greater pressure. But even the assumption that plasma does pass from the vessels will not remove every difficulty. For if, as the theory states, the secondary increase in intracranial pressure is due to such transudation from the capillaries, this pressure must be dependent upon the pressure of the plasma. The pressure of the plasma is, in turn, dependent on blood I02 W. B. Cannon. pressure, and is as much less than blood pressure as the resistance which the tissue about the capillaries offers to the outflow from them. Hill declares that to produce death intracranial pressure must equal the blood pressure in the carotids. The difficulty arises in attempt- ing to induce a method of compression manifestly less effective than arterial blood pressure to produce such a result ; for in the end by this reasoning the direct pressure of the blood through the free ways of the vessels must be greater than the lessened pressure of the trans- udate, and the flow will persist. This theory that the secondary increase in intracranial tension takes place chiefly in the border areas and is due to pressure of the plasma has, moreover, the great defect of leaving out of account the processes taking place in the portion of the brain substance in which the circu- lation is impaired. It will be shown that in this neglected region swell- ing and pressure occur wholly independently of any blood pressure whatever. Another theory of the cause of the increase of intracranial pressure after trauma was offered by Courtney^ in 1899. According to Court- ney, blows on the head paralyze the cerebral vasomotor nerves. Paraly- sis of these nerves results in dilatation of the vessels which they control, and dilatation of the vessels is accompanied hy acute anemia of the brain substance. Courtney does not make clear the manner in which dila- tation of the arterioles would cause anemia. Howell'^ has shown, by perfusing the cerebral vessels of dogs with defibrinated blood, that, no matter how high the arterial pressure, the venous outflow was always proportional in amount. Evidently the higher pressures would in effect dilate the arterioles in a manner satisfactory to Courtney's theory, and yet Howell's records show no indication of even a tem- porary blocking of the circulation in the brain. The sequence of events when the vascular paralysis is permanent, according to Courtney, is " arterial stasis, with enormous rise of intra- cranial (venous) pressure, thrombus formation, transudation." That dilatation of the arterioles does not produce arterial stasis has already been shown ; it has rather the opposite effect of lessening the resis- tance to the flow and increasing the rate. The reason for assuming an "enormous rise of intracranial venous pressure " is not given. Howell's experiments on the dilatation of the cerebral arteries under high pres- sures certainly do not support any assumption of a consequent venous ^ Courtney: Boston medical and surgical journal, 1899, cxl, p. 347. ' "^ Howell: American journal of physiology, 1S9S, i, p. 69. Cerebral Pressure following Trauma. 103 resistance. That such a result does not occur is explained by Howell by the application of the general rule that the total cross area of the veins in any region is greater than the total cross area of the arteries. Since this is true, the expansion of the cerebral arteries is manifestly always relatively less than the diminution in the size of the veins, and therefore Ihe blood flow is not impeded in arterial dilatation. The transudation which Courtney assumes to be a result of the arterial dilatation cannot be absorbed because of the exceedingly high pressure in the veins. It must therefore remain and compress the capil- laries, thus still further impeding the circulation and taking the first turn in a vicious circle, similar to that noted by Bergmann. Inasmuch as the accumulation of the transudate depends on the high pressure in the veins, and the assumption of a high pressure in the veins has been shown to be unwarrantable, this part of the theory falls to the ground. Moreover in this theory again a mechanical filtration of fluid is called upon to raise a pressure greater than its source, and this feature is, therefore, open to the same objections previously noted in discuss- ing the statements of Hill. The theory Courtney has developed has, furthermore, the defect of failing to regard sufficiently the processes occurring in brain tissue deprived of its proper blood supply. The nature and results of these processes will be presented in the next section of this paper. IV. Experimental Inquiry into the Cause of Brain Pressure. It was noted in the citation of clinical cases of chief interest in this investigation — the cases with least evident cause for the pressure symptoms which they manifested — that in many of them two stages were to be observed. The first stage followed immediately after the injury and was marked by loss of consciousness or by a peculiar dazed condition. From this primary stage the patient might recover, only to pass after a time into a secondary stage characterized by a gradual increase of pressure symptoms and frequently closed by death. The question arises, can a more reasonable explanation of these phe- nomena be given than the theories already cited .'' About two years ago Dr. W. N. Bullard pointed out to me the im- portance of knowing the cause of the increased pressure of the brain after trauma of its substance. The results of the study undertaken upon this suggestion may be divided into two groups, (i) those con- cerned with the immediate effects of injury to the brain, and (2) those concerned with the more remote effects. I04 W. B. Cannon. Primary effects of cerebral trauma. — The method employed to deter- mine the immediate effects of head injuries was a modification of the method used by HilP in his investigation of brain pressure. The apparatus consisted essentially of an inner and an outer brass tube, as shown in Fig. i. The outer tube (O) was bevelled at its lower end and threaded so that it could be screwed tightly into the trephine hole. The inner tube (I) had at its lower end a membrane (M) of thinnest rubber; at its upper end a rubber cork (C) securely fastened by pins passing through both the tube and the cork. The inner cylinder was held rigidly within the outer cylinder by means of a screw (S). Fitting closely. around the inner, and tightly over the outer tube, was a rubber capping (R) which effectually prevented any fluid from passing out of the brain case. A curved glass tube led from the rubber cork (C) and was connected by rubber tubing to a straight glass tube (G). The whole inner ap- paratus was filled with water. A line was marked on the glass tube (G), at which the water level stood when the membrane (M) was even with the base of the inner cylinder. If now, after the dura has been removed, the inner cylinder is inserted with the membrane on a level with the dura and the brain exerts a pres- sure against the membrane, the membrane will rise in the cylinder and cause the water level in G to rise to a corresponding amount. But the glass tube (G), because of rubber connections, maybe raised- and thus the water level within it brought to a greater height above the membrane (M) and the pressure on ]\I thereby increased. As the pressure on M increases, the brain pressure is more and more overcome, the brain sinks toward its normal position, and as it sinks the fluid system falls until the upper level is again at the original mark on the glass tube (G), indicating that the membrane is again ^ Hill: Loc. cit., p. 9. ^ The variations in the level of the water due to raising and lowering the rub- ber tubing were found to be relatively insignificant. Figure 1. — Apparatus for measuring cerebral pressure while preventing the brain from bulging through the opening in the skull. Cerebral Pressure following Trauma. 105 even. By this means not only can the brain be kept in its normal relations within the cranium, but by measurement of the height of the upper level of the water above M the pressure which the brain is exerting can at the same time be ascertained. By connecting the tube (G) with a piston recorder a continuous record of the pulsations of the brain and the slight changes in its level can be obtained, and the times noted at which the tube is raised or lowered. The pressure which the brain is exerting, as indicated by the height of the water column, is written above the record, whenever the pressure changes. (See curve a, Fig. 2.) Cats were used for the experiments. In every instance the animal was fully anaesthetized with ether before being operated upon. The cylinders were attached sometimes before, sometimes after the injury. Owing to the elasticity of the bones of the skull and to the force of the con- cussion the rigid outer cylinder would not keep its place, but would be driven out. To obviate this difficulty side projec- tions, level with the scalp, were attached to the cylin- der, and strong rubber bands passed over them to keep the apparatus in position. The concussion was caused by blows of a hammer on the skull. Since the results of injury are to be differ- entiated from the normal conditions, a description of the normal conditions will first be given. The uninjured brain of an etherized cat will bulge through the tre- phine opening after the dura is cut, unless pressure is applied to keep the organ in place. The height of the water column necessary to pre- vent the bulging is a measure of intracranial pressure. In one in- stance this pressure during two hours of observation varied widely within the limits of 4 cm. and 20.4 cm. of water, and showed an average of 12.7 cm. of water. In another instance during two hours of observation the pressure varied between 9.4 cm. and 16.9 cm. water, and had an average of 13 cm. In still another case the average Figure 2. — Records of brain pulsations obtained by attaching a piston recorder to the apparatus shown in Fig. 1. The upper curve {a) was taken after considerable hemorrhage ; the numbers in- dicate the water pressure required to keep the brain from bulging. The lower curve {b) is typi- cal for normal pressure (17 cm. water). io6 W. B. Cannon. pressure during five hours was 13 cm. of water. This observation accords with the observations of Hill on dogs, in which he found the intracranial pressure in normal conditions about 10 to 13 cm. of water.^ The variations from the average were largely due to differ- ent degrees of etherization ; renewing the ether invariably caused an increase of brain pressure, and as the effect of the ether was gradu- ally passing away, the brain pressure would slowly fall. Every factor which increased general blood pressure, such as compression of the abdomen, movement of the limbs, stimulation of a peripheral nerve, had an instant effect in heightening intracranial tension. A hemor- rhage at the time of trephining, if at all severe, caused a marked dimi- nution of brain pressure, which ranged in one case of considerable bleeding from 0.8 cm. to 6.8 cm. of water with an average during two hours and a half of 2.8 cm. Hemorrhage caused also a characteristic change in the nature of the pulsations of the brain ; the excursions through which the surface passed at inspiration and expiration were increased in amplitude, while the excursions due to the heart-beat were diminished. A comparison of curve a (Fig. 2) of the brain pulsations after severe hemorrhage, with curve b, which is typi- cal for normal conditions, will make clear this difference. The average brain tension, then, in a normal uninjured animal is about 13 cm. of water. How does injury to the brain affect this ten- sion .'* The immediate effect of a severe blow on the head is an enormous increase of intracra- nial pressure at the time the injury is received. Figure 3. — Curves ot If the cylinders are introduced into a trephine cerebral pressure [a) hole and held in place by strong rubber bands, and blood pressure (^) , , -i i i -r .1 ^ .. r , ., ^. ^ . . ' as above described, and it the water system 01 at the time of injury _ ' ■'^ ^ (5) and immediately the apparatus is then placed in connection with thereafter. a membrane manometer, so as to avoid extensive displacement of the water by the violence of the traumatism, an indication of the effect of the traumatism on the brain may be secured. In Fig. 3 the upper curve is a record of the movements of the brain thus taken. The extensive movements 1 Hill: Loc. cit., p. 73. Cerebral Presszire following Trauma. 107 of the writing lever at the point marked 5 were caused by five con- cussions. As the head was supported in a uniform position by means of a block, the movements could be due only in a very slight degree to the inertia of the water in the cylinders; they must be almost wholly the consequence of great variations in intracranial pressure. To understand the result of a sudden great increase in intracranial pressure, the general conditions of the skull and its contents must be recalled. The brain case, under ordinary pressures, may be regarded as a fairly rigid box, but when exposed to a violent force it manifests a considerable degree of elasticity. Within this case are the brain substance and cerebro-spinal fluid, both incompressible, and the blood, which is the single variable factor. When a severe injury is inflicted on the head, the bones yield under the blow, there is a sudden violent increase in intracranial pressure, which is distributed by the incompressible brain substance and cerebro-spinal fluid. In so far as this increase in pressure overcomes the pressure of the variable factor, the blood, it must result in a checking of the blood flov^^ into the brain. ^ This result appears in a characteristic diminu- tion of the amplitude of the brain pulsations immediately after the injury, as is shown in the upper curve of Fig. 3. The pulsations soon return, however, to their original extent. Following the injury there are changes in respiration, in general arterial pressure, and in intracranial tension. Severe concussion frequently causes paralysis of the respiratory centre ; respiration entirely ceases, although the heart continues beating for some time. Under these circumstances, if artificial respiration is persisted in, the respiratory centre may wholly recover its function. In this matter, which is of direct practical importance, my observations completely confirm those of Polis,^ Horsley,^ and Kramer.* The typical changes in general arterial pressure w^ere recorded from the femoral artery in order to avoid, so far as possible, any interference with the cerebral circulation. At the time of the injury and a moment thereafter, a slight rise in the blood pressure was usually to be observed. The rise was followed by a considerably more pronounced fall of pressure, after which there was a gradual return to normal. When the injury caused a temporary cessation of respiration, the fall was always ^ Cf. Kramer: Annals of surgery, 1896, xxiii, p. 163. 2 Pons : Revue de Chirurgie, 1894, xiv, p. 730. ^ HoRSLEY : Quarterly medical journal, 1894, ii, p. 309. ^ Kramer : Loc. cii., p. 167. io8 W. B. Cannon. greater than when the breathing continued without interruption. (Fig. 3). The phenomenon of interest, however, is the change in intracranial tension following the injury. The upper curve in Fig. 3 indicates a higher cerebral pressure after the traumatism than before. The ob- jection may be justly raised that in this instance the skull was opened, and the only opposition to the projection of the brain through the opening was in the elastic tension of the manometer. The con- cussion may, therefore, have pushed the brain outward, and thereby produced the appearance of greater intracranial pressure. That there is a real increase of brain tension after injury is proved by causing the concussion before the skull is opened, and afterward applying the apparatus for measuring the tension (Fig. i). A record of the variations in the height of t^e water column needed to keep the brain even with the dura can thus be made precisely as in normal states. After injury, the average height of the water column was always greater than with the uninjured brain. In one case the cat was etherized at 11 a.m., and fifteen minutes later was struck ten times in a region afterward trephined. So soon as the trephine button was removed, the dura was observed to be bulging and pulseless with a little blood beneath. At this time the animal ceased breath- ing; artificial respiration was repeated several times, and then the dura was quickly opened, when regular breathing began and continued throughout the period of observation. The apparatus was applied to the edges of the trephine hole, and the variations in pressure were observed during four hours. In that time the pressure varied be- tween lo.i cm. and 47 cm. of water, with an average pressure of 25 cm. There was no noticeable increase of pressure in the last half over that in the first half of the experiment. Other similar opera- tions gave similar results; in the few hours following the injury, the pressure never rose higher than 50 cm. of water, although in some instances there could be seen at the autopsy pial hemorrhages and contusion of the brain substance. Simultaneous records of the brain and the femoral artery show that there is no increase in general arterial pressure to account for the rise in brain pressure after injury. Whether this rise is to be ascribed to increased resist- ance in the smaller vessels of the brain, or to a relaxation of the vas- cular walls, has not yet been demonstrated. Mosso^ noted a consid- ^ Mosso : Ueber den Kreislauf des Blutes im menschlichen Gehirn, Leipzig, 1881, p. 200. Cerebral Pressure following Trauma. 109 erable increase in the brain pulsations after releasing the carotids from compression, and he maintained that the effect was merely local, and that it was caused by a relaxation of the vessel walls in conse- quence of the interruption of the circulation within them, — a phe- nomenon which can be demonstrated in the arm. It is possible that the concussion and the initial checking of the blood stream through the cerebral vessels cause a similar relaxation of their walls, and a consequent greater pressure of the brain against the skull. A consideration of the phenomena attending the reception of an injury to the head, and following soon thereafter, indicates that in cases of simple concussion {i.e., without severe intracranial hemor- rhage or laceration of the brain) the only pressure which might account for the primary pressure symptoms is that observed at the moment of injury. Men remain conscious through the spasms of strychnine poisoning, although, according to Hill,^ the pressure within the skull must, under these circumstances, rise to 50 mm. of mercury, an equivalent of 68 cm. of water. Evidently the highest pressure after injury secured by the method I have employed, 47 cm. of water, is not sufficient to account for either the primary or the secondary unconsciousness, and the other symptoms of intracranial tension following head injuries. It is clear that the force of a severe blow on the head will overwhelm any blood pressure in the cerebral vessels, and result in circulatory disturbances likely to take a considerable time for their readjustment. The paralysis of the respiratory centre for some moments after the injury is an indication of the lasting primary effects. The readjustment of the circulation in the brain may be further hindered by the general fall in arterial pressure immediately after trauma. In one instance the arterial blood pressure fell from 135 mm. Hg to 95 mm. Hg almost directly following the trauma. Since there is in the brain at this same time a process of recovery from a checked blood flow, the factors effective in producing ordinary syncope are present. The primary unconscious- ness after concussion seems thus to be accounted for by circulatory disturbances alone, although in so sensitive a structure as nervous tissue the possibility of molecular changes should not be overlooked. Secondary effects of cerebral trauma. — In the preliminary study of pressure symptoms it was pointed out that in order to cause them the intracranial pressure must equal the general arterial pressure. 1 Hill : Loc. cit., p. 73. no tV. B. Cannon. It was also made clear that these symptoms frequently appeared after the primary effects of injury had passed away, or as a continu- ation and intensification of the primary effects. In these accidental experiments what may be sought as an explanation of the re- sults? Before entering into a discussion of these results three ante- cedent conditions must be regarded: the anatomical arrangement of the cerebral blood vessels, the anatomical changes caused by in- jury, the effects of these changes on the circulation in the brain. The central nervous system is peculiar in having over its surface a richly anastomosing network of blood vessels, from which branches pass into the substance of the organs. On the surface of the brain the vessels form a vast canaliculate reservoir with very free commu- nications from one part to another. An injection into an afferent artery first fills the network and then goes more easily into other afferent vessels than into the nourishing arteries of the tissue.^ The unity of the surface system is manifest. From the common reser- voir there pass into the brain substance the nourishing arteries, some to the cortex, some to the deeper fibre tracts. Each of these vessels is isolated, independent, terminal, not anastomosing with other vessels from the surface or with branches from the central supply of the brain. The physiological conception of a terminal artery does not signify, however, that there is no communication with other sources of blood supply; ^ it means simply that the com- munication, /. c, by capillaries, does not permit the easy establish- ment of a collateral circulation after the interruption of the arterial supply to a part. The brain substance is, then, nourished by termi- nal arteries preceded by an anastomotic network. The importance of the arrangement of the cerebral blood vessels in understanding the effect of injury is seen when the pathological changes caused by trauma are recalled. Contusion of the brain exists in all cases of intracranial injury and, as has been noted, it is characterized by a diffuse formation of thrombi, the presence of punctate extravasations, and often also by thin patches of hemorrhage in the meninges. Associated with these changes is oedema, which is, however, of secondary development. The effect of the pathological changes on the circulation is appar- ent from the nature of the anatomical arrangement of the cerebral 1 Charpy: Systenie nerveux, Part iii of Traite d'anatomie Inimaine, edited by Poirier, p. 700. -. Baumgarten : American journal of physiology, 1S99, ii, p. 245. Cerebral Pressure following Trauma. 1 1 1 vessels. Because of the free anastomosis in the surface network a compressing hemorrhage in any part must force the blood to take the easier paths and thus partially or completely shut off the blood from the nourishing arteries plunging inward from that part of the net. Moreover, owing to the terminal nature of the arteries the general formation of thrombi and extravasations must result in a general diminution of the blood supply of the injured tissue. Proximal to the interruption there would be stasis, distal to the interruption, anemia, and both conditions must result in impaired nutrition of the brain. In this relation the interference with the normal food supply is not so important as the deprivation of oxygen. As will be made clear, the problem of a secondary in- crease of intracranial pressure is essentially the determination of the action of brain tissue deprived of its normal nutrition and especially of its supply of oxygen. The effect on protoplasm of cutting off its oxygen supply may be seen in such widely different structures as unicellular organisms, the muscular tissue of the frog, and in nerve cells. Budgett ^ noted that when paramoecia were deprived of oxygen they began to absorb water, the contractile vacuole increased in size and new vacuoles appeared, then from the surface there protruded vesicles from which some of the vacuoles usually escaped. Finally the vesicles burst and the cell contents were extruded. Poisons, such as potassium cyanide, caused similar changes to take place more rapidly. When potassium cyanide was applied to a sympathetic nerve cell of the frog the cell soon became larger than normal and continued swelling until there was a marked increase in size. Budgett believed that the swelling was probably due to an extensive splitting of the molecules within the cell by means of the poison, and a consequent rise of intracellular osmotic pressure and absorption of water. Loeb^ has noted in the frog's gastrocnemius similar changes from lack of oxygen. In a large number of frogs he ligatured the leg on one side and after a time removed and weighed the two gastrocnemii. Normally the two muscles are of the same weight; but under the circumstances of the experiment, the muscle deprived of its oxygen supply took up water so that in eighteen hours its weight was from one to three per cent greater than that of the undamaged muscle. After forty-eight hours there was a difference of fifteen per cent and at the end of seven days ^ Budgett : American journal of physiology, 1898, i, p. 211. ■2 LOEB : Archiv fiir die gesammte Physiologic, 1898, Ixxi, p. 470. 112 W. B. Cannon. muscles showed a difference of weight from twenty-five to forty per cent. Loeb concludes that his experiments leave no doubt that the assumption of water by a muscle deprived of its blood supply is due to chemical changes in the muscle increasing the internal osmotic pressure and that these chemical changes are probably due to lack of oxygen. These examples of the effect on various tissues of cutting off the oxygen supply show that the active agent in the production of swell- ing is osmotic pressure. Loeb found that a frog's gastrocnemius, placed in a 4.9 per cent salt solution, at first loses water, but later begins to take it up from this strong solution. The osmotic pressure of this solution is more than thirty atmospheres, the osmotic pres- sure of the normal tissue is that of normal salt solution or about five atmospheres. The original loss of water in this instance must there- fore have caused chemical changes in the muscle which raised its osmotic pressure more than twenty-five atmospheres. Evidently in osmotic pressure there is a force acting which is very much greater than any blood pressure and which is entirely independent of blood pressure in producing a swelling of the tissues. These results may now be applied to the conditions in the brain after injury. The initial changes are hemorrhages and contusion with thrombi and extravasations. The result of the changes is an impaired blood supply to the injured region and a consequent lack of oxygen. Both experiment and clinical observation prove that oedema follows. Dean ^ placed a glass disc between the brain and the dura and sewed the dura in place again. The disc produced anemia of the compressed area. After from three to six days Dean found that the parts about the foreign body were oedematous ; a piece of brain taken from the compressed area contained 3 per cent more water than a similar piece from the opposite hemisphere. In clinical cases Berg- mann says: ^ "One finds about the region of injury merely a small zone, which is scattered with specks of blood and colored yellowish red; farther on, as far as the swelling reaches, the brain substance appears moist, glistening, soft, — that is, highly oedematous." Now the question arises, is this oedema, this result of the primary patho- logical changes, this concomitant of the secondary pressure symp- toms, a passive transudate due to blood pressure, as has been held hitherto ; or is it the effect of chemical changes in the brain sub- ^ Dean : Journal of pathology, 1893, i, p. 39. See also Duret, Loc. cit., p. 194. ■^ Bergmann : Loc. cit.., p. 420. Cerebral Pressure following Tratuna. 113 stance, resulting from diminished blood supply and causing the taking of water into the tissues by increased osmotic pressure? The former process, when required to produce the equivalent of arterial pressure, encounters the difficulties already mentioned in criticising previous theories; the latter process provides a force amply sufficient to overcome any possible blood pressure. Normal salt solution has the same osmotic pressure as the blood and therefore the same osmotic pressure as the tissues. Cohnheim ^ found, after introducing normal salt solution in animals even to 40 per cent of the body weight, not the slightest trace of oedema in the central nervous system. Magnus ^ has gone still farther and infused animals with normal salt solution to iio per cent of the body weight, within one hundred and forty-three minutes and produced no general oedema. Since normal salt solution is isotonic with the tissues, so that in its presence water does not pass into them when they are nourished, the passage of water from the solution into the tissues when they are not nourished indicates an increase of osmotic pressure within them. Upon this statement was based the following series of experiments with brain tissue. The first question to be settled was with regard to the action of the brain when normal salt solution is passed through the vessels at ordinary blood pressure. To this end a cannula was tied into each common carotid artery of the cat and a solution of 0.8 per cent sodium chloride allowed to pass into the cerebral vessels. The solution ran from an elevated source and the pressure of the fluid was recorded by a mercury manometer. The height of the reservoir was such that the manometric pressure was approximately blood pressure, — 120 mm. of mercury. One or two minutes after the solution began passing into the arteries the head of the animal was severed from the body, the vertebraterial canals and the spinal canal plugged with cotton soaked in glycerine, and the skull trephined for the measurement of cerebral tension. For this purpose the two cylinders shown in Fig. i were used, but the inner cylinder was attached by rigid glass connections with a mercury manometer, with mercury conduction between the two. Before the inner cylinder was applied the level of the exposed surface of the mercury was made even with the level of the membrane, and over the mercury surface 1 Cohnheim : Lectures on general pathology, London, 1889, p. 459. - Magnus: Archiv fiir experimentelle Pathologie unci Pharmacologie, 1899, xlii, p. 261. 114 W. B. Cannon. was placed a thin paraffine plug fitting closely into the tubing of the manometer. The head was now placed at such a height that the trephine hole was even with the mercury level and the membrane, and the cylinder (I) was then slipped into its holder (O). In case the brain was bulging beyond its normal level the mercury in the ma- nometer would rise and carry up the paraffine plug. Thereupon mer- cury was dropped into the manometer until the plug was forced downward to its former position, which indicated that the brain was returned to its normal level and at the same time showed the force which the brain was exerting against the membrane. The following figures from an experiment in which was employed the procedure above described will show how the brain pressure changed as time passed. The flow of the salt solution began at 10.45 A.M. Time. Pressure of salt sol. Brain pressure. 11.00 a.m. 11.20 1 11.30 12.15 P. M. 1.15 3.15 2 5.00 11.1 cm. Hg. 11.2 12.5 12.5 12.5 12.5 12.5 1.6 cm, Hg. 2.9 3.5 7.1 9.2 12.1 12.1 1 At this time the tlow from the veins had changed from a running stream to a rapid dropping. ■■^ At this time there was only a very slow dropping from the veins. In this experiment the brain pressure remained at the height of 12. 1 cm. of mercury until i p. m. the following day, when the apparatus was removed. The pressure from the salt solution was removed first, but this did not result in any fall of the mercury level in the manometer; the brain had apparently taken in water until it completely filled the cranial cavity. Care was taken that the turgid tissues in the neck did not prevent the solution from passing into the brain, but it was impossible to keep the brain itself from swelling and shutting off its own supply. This probably accounts for the fact that the brain pressure did not rise above the pressure of the solution ; Cerebral Press2ire follozving Trauma. 115 for in other cases in which the tissue was allowed in the early stages of the swelling to press outward as a hernia from the trephine hole, a pressure of more than 28 cm. of mercury was supported for hours without diminution of the swelling and with a solution pressure of only 0.7 cm. of mercury. Evidently under circumstances of non-nutri- tion the brain will take up water from a solution isotonic with the blood and will thereby exert a pressure sufficient to exclude the blood from the cerebral vessels. The question raised in criticising the previous theories of increase of brain pressure now arises again, — is the increase of pressure due to the formation of a transudate from blood pressure, or is it the result of an active process in the tissues themselves ? That brain tissue deprived of its proper blood supply will take up water from a solution iso- tonic with the blood, entirely independently of any mechani- cal pressure of the solution, can readily be demonstrated by re- moving a brain and placing it in 0.8 per cent sodium chloride solution. Under these circum- stances it will almost immedi- ately begin to increase in weight ; as will be seen upon examining the curve of weight (A, Fig. 4) of such a brain, the increase during the first four hours is rapid, and thenceforth is slow and persistent for days. In one instance after five days the brain had increased 33.2 per cent; since the specific gravity of the brain and the solution is approximately the same this increase of weight means an increase in size of about one-third. As has been previously stated, a diminution of the intra- cranial space by one twelfth results in death. It is evident that impairment of the nutrition of the brain as a whole may cause it to take up water to a degree far greater than that necessary to produce death. Now what is true of the brain as a whole may be assumed to be true also of any parts deprived of their blood supply, and localized regions thus affected would take up the water of the plasma from Figure 4. — Curve A showing the increase in weight when the brain is placed in normal salt solution, 0.8 per cent; B when the brain is placed in 2.0 per cent salt solution. ii6 JV. B. Cannon. neighboring regions and swell in the same manner in which the brain as a whole will swell. That nerve cells increase in size, as does the brain, when deprived of nutrition, has been shown by Scagliosi.^ After repeated slight concussion in rabbits he found definite changes in the cells of the brain, varying according to the duration of life after the injury. The microscopic lesions consisted in varicosities of the dendrites, degeneration and swelling of the cell body with formation of vacuoles within it, and a more and more homogeneous appearance of the nu- cleus till the cell almost completely disappeared. The similarity of these changes with those recorded for infusoria deprived of oxygen is striking, and indeed Scagliosi attributes the alterations to disturb- ances of the circulation preventing normal metabolism. The swelling of nerve cells when deprived of normal nutrition may be observed directly by placing in salt solution under a sealed cover, one of the small sympathetic ganglia of the frog. It will be seen from drawings of such a cell, made with a camera lucida, that the dimensions become greater and greater as the time passes, and that finally the cell begins to disintegrate and lose its definite outline. The last of such a cell is a formless granular mass. It is very pos- sible that this is the series of changes through which the nerve cells pass in the production of areas of softening in the brain. ^ A brain which had been in salt solution for five days was declared by Dr. E. W. Taylor, instructor in neuropathology, Harvard Medical School, to show internally the typical appearance of softening. The observations on brains and nerve cells prove that blood pres- sure is not a necessary factor in causing cerebral swelling and oedema. It may be, as Hill and Bergmann maintain, that there is transudation from the cerebral blood vessels when intracranial pressure is locally increased, but they have not shown that the tran- sudation is a mechanical result of the increased pressure, nor have they shown that, even with a mechanically produced transudate, the pressure resulting therefrom is sufficient to produce death. They have, moreover, not regarded the factors increasing osmotic pressure within the tissues. Hill ^ states, " under normal pressures, the secre- tion and absorption of cerebro-spinal fluid does no doubt follow osmotic laws ; " he overlooks the fact that deprivation of nutrition ^ Scagliosi : Archiv fiir pathologische Anatomic, 1898, clii, p. 522. ^ See Case IV, p. 98. 3 Hill : Loc. cit., p. 23. ' Cei^ebral Pressure following Trauma. 1 1 7 causes within tissues chemical changes which affect a force much more powerful than mechanical blood pressure. A conception of the force involved in the increase of weight, when the brain is placed in normal salt solution, may be obtained by means of a method used by Loeb. Curve B of Fig. 4 shows the changes in the weight of a brain placed in a 2 per cent solution of sodium chloride. For four hours there was scarcely any increase in weight. At first there was even a diminution of weight, because the osmotic pres- sure outside was greater than that within the tissues. Soon there was a sharp rise, however, and after about ten hours the slow, per- sistent increase in weight began. The osmotic pressure of a 2 per cent solution of sodium chloride is about 14.5 atmospheres. The pressure within the tissues, therefore, must have developed to this great height in order that the water should pass into them. Inasmuch as the osmotic pressure of an 0.8 per cent salt solution is only 5 atmospheres, there is evidence of a pressure increase of 9 atmospheres, which is about fifty-seven times ordinary blood pressure. Naturally such a pressure never becomes fully operative in the tissues ; disruption must take place long before that result could occur. The internal osmotic pressure is almost balanced by the external, but because of the greater internal pressure, fluids pass inward till the mechanical pushing of the tissues overcomes the push of the blood, thereby shutting out the source of fluid supply. It is clear, therefore, that in the chemical changes taking place in dying brain tissues, there is a force present abundantly able to overwhelm the blood pressure and cause death. It may be objected that the method used is too unnatural, too remote from actual occurrences, to prove the contention that the increased brain pressure after trauma is due to osmotic changes from malnutrition of the tissues. In this relation, the phenomena attend- ing the formation of a hernia cerebri are instructive. The average pressure required to prevent the brain of an etherized cat from pro- jecting through a trephine hole is about 13 cm. of water. In case a hole over the frontal region is left uncovered, however, the brain begins to rise through it, and as time passes the level becomes higher and higher, until the outer surface of the bulging brain is even with the scalp. At this point a considerable compressing force is needed to drive the bulging brain back into the skull. The surface of the hernia shows no pulsations, the vessels may contain blood, but the circulation is much impaired. How may this phenomenon be 1 1 8 W. B. Cannon. explained ? The brain presses against the edge of the trephine hole. As it does so, a circular area of blood vessels in the pial network may- be regarded as separated by pressure from the remainder of the superficial reservoir. Since the anastomoses in the pial vessels are very free, the blood takes a course of less resistance than that offered at the edge of the trephine hole. The nourishing vessels pass into the substance of the brain at right angles to the surface ; the part of the brain within the hole, therefore, would not be normally nourished. Under these circumstances the osmotic pressure in the tissue rises; the plasma from neighboring areas has the osmotic pressure of the blood ; it is consequently taken into the disturbed tissue, and the process of swelling begins, and continues persistently. In this in- stance nothing is introduced into the cranial cavity to increase blood pressure there; the pressure in neighboring brain areas cannot be much greater than 13 cm. of water, and yet a hernia is produced which drives outward the pial coverings of the brain, and in the end offers a marked resistance to reduction. Examination of the hernia shows an evident oedema. A review of the evidence thus far presented shows that in cases of brain injury, pathological changes are brought about which result in an impairment of the nutrition of the injured region. In conditions of impaired or interrupted nutrition, tissues undergo internal changes leading to increased osmotic pressure, and thereby to increase of water-content and greater size. The swelling takes place, therefore, by means of a force much greater than blood pressure. The condi- tions of the brain are peculiar in that the organ lies in a rigid case. Swelling of a part consequently compresses the only compressible portion of the contents of the cranial case, — the blood vessels. Thereby new areas are shut out from normal blood supply, and changes now take place in these tissues as well, with the result that water passes into them ; thus the swelling spreads until the blood- flow is so greatly excluded from the brain that life is no longer possible. It is by this process that the cases of head injury resulting in death from secondary increase of brain pressure^ may be explained. Cases in which death results soon after the reception of severe injury may also be supposed to be due in part to this same process. The remaining ^ class of cases, those in which the pressure symptoms develop after injury but do not result in death, remain to be consid- ^ See p. 97. '^ See p. 96. Cerebral Pressure following Trauma. 119 ered. In attempting to explain these cases, the fact should be remembered that in osmotic action not only does the water pass inward through the membrane to the stronger solution, but also the dissolved particles pass outward and tend to equalize the osmotic pressure. It is conceivable that, if the immediate result of the blow is slight, the passage of wpter into the tissues will be only slight, before the diffusion of decomposition products from the tissues has taken place to so great an extent as to bring about an equilibrium. Or the progress of the developing necrosis may be slow and diffusion occur gradually, and oedema may not, therefore, result. In this connection it would be interesting to know the careful quantitative urine analysis of the body salts in the cases of secondary pressure symptoms after head injury. The occurrence of accumulations of fluid under the dura or in the ventricles in some of the recorded cases ^ seems to offer evidence against the theory of pressure here propounded ; for why does not the fluid pass into the tissues rather than accumulate if the tissues have the great osmotic pressure attributed to them .'' Before this question can be definitely settled, the osmotic pressure of these fluids must be determined. It is probable that the gatherings of fluid are encapsulated. The diffusion of salts from the injured tissues into even a slight amount of fluid in an encapsulated space would render that fluid of higher osmotic pressure than the blood plasma. The plasma would thereupon pass into the encapsulated space in obedience to osmotic laws, and thus increase the volume of the fluid and its compressing effect. Further change in the injured tissues would lead to greater swelling in them, and to diffusion of more of the dissolved products of decomposition. The diffusion into the encapsulated fluid would still further increase its osmotic pres- sure, and result in still more plasma coming to increase its volume. Thus there would be a passage of salts from tissues to blood in a series of decreasing concentrations, and a passage of fluids to tissues in a series of increasing osmotic pressures. And since water will pass more rapidlythan salts through the membranes, the result is usually a greater and greater pressure till death supervenes. These theoretical considerations merely point the way to further experimental work and more exact clinical observation. The facts of the ground here covered remain ; that injuries to the brain inter- fere with its proper blood supply, that such interference causes an in- ^ See Case IV, p. 97. 1 20 W. B. Cannon. creased osmotic pressure within the tissues, and a consequent taking up of water from the surrounding plasma. The swelling and oedema of the brain after head injuries, therefore, are not wholly due to passive transudation, as Hill and Bergmann have maintained, but are mainly the result of an active process in the tissues themselves, a force many times greater than blood pressure, and amply sufficient to produce all the pressure symptoms, and account for all the signs of intracranial tension which the clinical cases of cerebral trauma often manifest. Summary. At the moment of injury the intracranial pressure rises to a height sufficient to check the blood-fiow into the brain. Immediately after the injury the general blood pressure usually rises for a moment, then falls. Thereafter a gradual recovery of normal blood pressure occurs, with a simultaneous increase in the extent of the pulsations of the brain. The paralysis of the respiratory centre following head injury may be recovered from if artificial respiration is persisted in, and the heart action remains strong. ( Horsley, Polls, Kramer.) The primary loss of consciousness after a blow on the head is apparently due solely to circulatory disturbances, though minute changes in the nerve cells must also be considered. The normal cerebral pressure is about 13 cm. of water; after injury the brain pressure may rise to an average of 25 cm. of water. Since this increase is not Sufficient to account for the symptoms present in clinical cases, there must be other secondary processes causing increased pressure. The secondary increase in pressure is due mainly to three factors: deprivation of normal nutrition in injured parts, passage of water into these parts with consequent swelling, and the rigid inclosure of the brain, causing the swelling in one region to affect markedly neighboring regions. The thromboses, extravasations, and hemorrhages, which accom- pany contusion, impair the blood supply of the injured region, especially since the nutrient arteries of the brain are terminal. Brain tissue deprived of blood undergoes chemical changes result- ing in greater internal osmotic pressure and the passage of water into the tissue. Brains placed in normal salt solution increase contin- uously in weight, even to 33 per cent. The swelling which the tissue Cerebral Pressure following Trauma. 1 2 i undergoes must cause it to compress neighboring regions, and thus further impair the circulation so that new regions are involved in the process. Thus a vicious circle is established. The main force effective in causing swelling is probably osmotic pressure, which, in brain tissue, may attain a degree so much greater than blood pressure, that the blood would really be prevented from entering the cranium. Recovery from head injuries after pressure symptoms, and the accumulation of fluid under the dura and in the ventricles, are possi- bly to be accounted for by diffusion of the products of destruction from the tissues. ON THE ANALOGY BETWEEN THE EFFECTS OF LOSS OF WATER AND LOWERING OF TEMPERATURE. By ARTHUR W. GREELEY. [From the Hull Physiological Laboratory of the University of Chicago.'] TN his experiments on heliotropism, Loeb ^ observed that the larvae ■^ of Polygordius and certain Copepods can be made positively heliotropic either by raising the concentration of the sea-water, or by lowering the temperature, thus indicating an analogy in the effects of a loss of water and a reduction of temperature. Other instances may be cited. Thus among plant lice that exist in two forms, one winged and the other wingless, the growth of wings in the wingless forms can be produced either by lowering the temperature or allow- ing the animals to dry. Loeb ^ has also shown that division of the protoplasm in the fertilized Echinoderm Qgg can be prevented by raising the concentration of the sea-water or lowering the tempera- ture. From these facts Loeb ^ concluded that: — " raising the con- centration of the salt solution, in which an animal or a tissue lives, has qualitatively and quantitatively the same effect as lowering the temperature." It is well known that raising the concentration of the salt solution causes the organism or tissue to give off water, but why a reduction of the temperature has an analogous effect has never been explained. It is hoped that the present experiments will throw some light upon this problem. At first the effect of changing the temperature was tried on the common blue-green Stentor coeruleus. The Stentor were kept in small covered dishes, and the water was renewed each day from the aquarium in which the Stentor had been grown. This con- tained an abundance of food material. The animals were divided into three lots : — Lot I was kept at the room temperature (20° C.) as control material. Lot 2 was surrounded by a mixture of ice and salt. Lot 3 was placed in a thermostat at a temperature of from 25° to 28° C. ^ LoEB : Physiology of the brain, 1900, p. 198. 2 LoEB : Journal of morphology, 1892, vii, p. 253. '^ Ibid. 122 Loss of Watei'' and Lowering of Temperature. 12, When the temperature was suddenly lowered to the freezing point, and ice was allowed to form in the dish containing the Stentor, the animals immediately became quiet. Death took place when the freezing point was reached. The cell outlines became extremely irregular and shrunken, and the blue-green pig- ment, stentorin, was rapidly diffused through the water. These irregular cells remained intact for a short time, but soon disintegrated. A very different result was obtained when the temperature was lowered gradually. The dish containing the Stentor was at first loosely sur- rounded with ice, so that the temperature sank to about lo'^ C. More ice and salt were then added, and the temperature allowed to fall for an hour, until zero was reached, or until a film of ice began to form around the edges of the dish. Under these conditions the Stentor soon became quiet, except for a lively contraction of the buccal and peristomal cilia, and remained so for from one to three hours. Large vacu- oles {v. Fig. i) began to appear in the protoplasm, but otherwise the organism was unchanged. Finally the cell slowly assumed a spherical form, the vacuoles (■?', Fig. 2) increased in size, and the nuclear nodes {11, Fig. 2) separated. The peristome then closed over, the mouth and oesophagus with their cilia disappeared, and the only evidence of continued life was a slow contraction of the small body cilia {b c, Fig. 3) found over the entire cell wall. While these changes were taking place, the Stentor still remained irritable and frequently changed their shape, but soon the peristome and body cilia disappeared, the striations of the ecto- sarc slowly faded out, the protoplasm became densely granular in appearance, and the cell ultimately assumed a more or less regular spherical form. This cell shrunk in size because of the absorption of Figure 1. — Camera lucida drawing of a Stentor exposed to a temperature grad- ually lowered to 2° C. and maintained for thirty minutes. Figure 2. — Stentor ex- posed to a tempera- ture of 2° C. during one hour and twenty- five minutes. Figure 3. — Stentor exposed to a temper- ature of 2° C. during two hours. 124 Arthur W. Greeley. the vacuoles, the ectosarc became separated from the rest of the cell, and some of the nuclear nodes came to the surface, and were given off with the ectosarc. The remaining nodes of the nucleus were em- bedded in the endosarc, and later were found fused together, although the protoplasm became so opaque that it was difficult to follow the process. Finally there was formed a simple resting cell, consisting only of the endosarc, from which all the complex structures of the living animal had disappeared, and which resembled a cyst in every respect, except that it lacked the distinctive tough cell wall of the latter. Such a resting cell is shown in Fig. 4. The ectosarc {ec) is seen adhering to the cell wall. Three of the nuclear nodes (w) are visible. It is generally thought that a reduction of temperature produces only a cessation of the vital activities of an organism, and that these C activities may be resumed when the normal \. temperature is again reached. But this ex- \ periment shows that in the case of Stentor ^^.J..... * a lowering of the temperature brings about ^y certain well defined structural changes, that are not necessarily incidental to a perma- FiGURE 4. — Typical resting . - , • , r • r 1 cell of Stentor, after an "^nt suspension of the Vital functions of the exposure to a temperature cell. The low temperature induces a defi- of 2° C. during two hours nite transformation of the organism which does not involve its death. That this transformation is a vital and not a lethal phenomenon is shown by the fact that when the temperature again becomes normal these resting cells undergo a reverse process, and become active Stentor. This retransformation of the artificially produced resting cells into the normal animals was observed to be complete in three or four cases. In several cases, although the first stages of the process were seen, the retransformation was never completed. In all in- stances, the resting cells lived in the small dishes for from two to three weeks before disintegrating, showing that the Stentor had not been immediately killed by the low temperature. In the few instances in which the complete retransformation was observed, the spherical resting cells were removed from the cold water by means of a pipette, and isolated in small covered dishes at the room tempera- ture. In from three to six hours after removal from the cold water the retransformation was begun by the cell gradually lengthening until it assumed the form of an ordinary Stentor. The nucleus appeared as Loss of Water and Lowering of Temperature. 125 a solid band (;/, Fig. 5) formed by the fusion of the nuclear nodes, but it slowly divided until the common chain-like form was reached. The first indication of the peristomal cilia was a longitudinal ciliated band (/ c, Fig. 5) which extended over the an- terior end of the cell. The lower end formed a slight spiral depression which grew deeper, and finally became the mouth {in, Fig. 5). The upper end of this ciliated band curved slowly about the anterior portion of the animal and at the same time the lower end with the mouth depression was drawn up, until the band formed a circle around the anterior end of the cell with the mouth at one side {jn, Fig. 6). This circle be- came the peristomal circle of cilia (/> c, Fig. 6), inclosing the disk or peristome (/, Fig. 6). The cilia extended spirally down the mouth, forming the cilia of the gullet. While these changes were taking place the ectosarc and its striations appeared, and a typical Stentor resulted. These stages in the for- mation of the peristome, and the division of the nucleus correspond closely with a periodical change which Balbiani ^ ob- served in Stentor. He saw the disappearance of the old mouth and peristome, and their regen- eration, with the attending modification of the nucleus as described above. This experiment recalls also Loeb's^ experiments on the trans- formation of organs among Hydroids, in which the material of the polyps was modified into stolons by a contact stimulus, and the process then reversed. None of these changes was observed in the Stentor that were exposed to a higher tempera- ture (25°-28° C), but another process was noted. As described by Biitschli^ and others an increase of temper- ature greatly stimulated cell division among the Infusoria. The con- ^ Balbiani: Zoologischer Anzeiger, iv, pp. 312 and 323. ■■^•Loeb: This journal, 1900, iv, p. 178. ^ Bronn's Thierreich, i, 3, p. 1243. Figure 5. — First stage in the forma- tion of an active Stentor from tiie resting cell, twelve hours after the tem- perature has become normal. Figure 6. — Second stage in the formation of an active Stentor from the resting cell, sixteen hours after the temperature has become normal. 126 Arthur W. Greeley. trol specimens, kept at the room temperature, showed very few cases of cell division, but in "^almost every instance, the Stentor began to divide immediately upon being put into the thermostat. The divisions followed in quick succession for from five to seven bi-partitions, with- out a marked diminution in the size of the Stentor. After three or four hours, division ceased, and the Stentor soon died. This result indicates an increase of irritability, corresponding with an increase of water in the cell, and brings out very clearly the antagonism in the effect upon these organisms of raising and lowering the temperature, the one stimulating cell division and rapid growth, and the other pro- ducing a contraction of the protoplasm and the formation of a spheri- cal resting cell. This antagonism is still further emphasized by the observation that lowering the temperature will not only inhibit cell division, but will bring about a reverse process, causing a fusion of the partially divided halves. This experiment was performed on Stentor in the process of division, and the fusion of the dividing portions was shown very clearly. The shrunken condition and general behavior of the Stentor which had been exposed to a lowering of the temperature suggested the possibility that the process had been accompanied by a loss of water. To test this hypothesis, the concentration of the water was raised, that it might be ascertained whether the same result could be obtained by this means. The results were as follows: when the Stentor were placed in a - ^ or ™ solution of cane-sugar, the effect was the same as that produced by suddenly lowering the tem- perature to the freezing point. The cells became much shrunken from excessive plasmolysis, and soon disintegrated. If, however, a ";■, solution of cane-sugar was used, the Stentor became quiet in a short time, and formed spherical resting cells, which could hardly be distinguished from those formed by a gradual reduction of the temperature. The same stages were passed through, and the two processes seemed to be identical in every respect, although I was unable to bring about the reverse process in this case. This experi- ment makes it probable that a lowering of the temperature causes the cell to lose water, as is the case when the concentration of the surrounding medium is raised. To test further this hypothesis, some experiments were tried upon Spirogyra in which the plasmo- lysis can be easily studied. 1 m indicates inol., or one gram molecular weight in a litre. Loss of Water and Lowering of Temperature. 127 Several years ago, Klebs ^ succeeded in producing parthenogenetic spores of Spirogyra by placing the filaments in a f solution of cane- sugar. The cell contents shrunk through plasmolysis, and formed oval spores in the centre of the ordinary cells of the filament. I re- peated this experiment and at the same time exposed another lot of Spirogyra filaments to a lowering of the temperature (from 20*^ C. to 1° or 2°). The filaments were kept at this temperature for about three hours, and then removed to that of the room. Examination showed that plasmolysis in these filaments had been as perfect as in the others. Regular oval spore-like bodies were formed in the centre of each cell (see Fig. 7). Upon removal to the room temperature, the cells gradu- ally took up water, the chromatophores expanded, and the cells resumed their normal appearance. Another experi- Figure 7. -A cell of Spirogyra ,, , , 1.1 • 1 • . showing plasmolysis produced ment, well known to plant physiologists, , \ \ ' ^ *^ / & « by an exposure to a tempera- which was brought to my notice by Mr. ture of 2° C. during three B. H. Livingston, shows very strikingly hours, that the Spirogyra cell loses water when the temperature is lowered. If the filaments be carefully dried and placed in olive oil, and then exposed to a low temperature, the water may be seen to escape from the filaments, and collect in small drops in the oil. If the temperature be still further lowered, these drops will freeze, and the ice crystals will be readily seen. The author desires to express his obligation to Dr. Loeb, under whose direction this work was done. Summary. 1. In Stentor a lowering of the temperature does not produce simply a suspension of the vital activities of the cell, but it brings about certain well defined morphological changes. The same changes can be produced by increasing the osmotic pressure of the surround- ing solution. 2. These changes consist in the absorption of the cilia and the gul- let and the throwing off of the ectosarc ; and finally there is formed from the endosarc alone a spherical, cyst-like cell, which may be called a resting stage of the Stentor. 1 Klebs : Die Bedingungen der Fortpflanzung bei einigen Algen und Pilzen, Jena, 1896. 128 Arthur W. Greeley. 3. If the temperature be again raised, a reverse process will take place, and the cyst-like cell will develop into an active organism. 4. In Spirogyra, a typical plasmolysis can be produced by a reduc- tion of the temperature. 5. This fact makes it probable that a reduction of the temperature and a loss of water have similar effects, because the cell loses water when the temperature is lowered, as well as when the concentration of the surrounding medium is raised. NOTES ON REGENERATION AND REGULATION IN PLANARIANS^ {continued^. T By frank R. LILLIE. ^HE experiments herewith recorded have been performed as occasion offered and have consequently extended over a con- siderable period of time. They have been selected from a large body of notes, because the facts are themselves new, and, the experiments being made on genera not hitherto studied from this point of view, the results may serve to control the theoretical deductions made from the study of Planaria, the genus of aquatic triclad Turbellaria hitherto exclusively studied. II. Regeneration of the Head in Dendroccelum Lacteum. In Dendrocoelum, which is closely related to Planaria and simi- lar in structure, the capacity for regeneration is very much less, and as concerns the head is limited to the anterior third of the body, approximately, though a new tail may regenerate at any transverse level. My first experiment was to cut a single specimen in two through the pharynx, July 2^^, 1899. The cut surface healed, and the farther fate of the parts was as follows. The posterior part formed no new tissue, although it lived for twelve days. From the anterior portion, on the other hand, there grew out a pointed piece, which formed a tail. I afterwards repeated the experiment several times with a similar result. Thinking that the failure to regenerate a head might be due to the presence of the pharyngeal pouch, I then cut fourteen specimens transversely immediately behind the pouch. Two days after the operation I noticed an interesting difference in the reactions of the two kinds of parts. While the anterior pieces reacted in all respects like intact individuals, migrating to shaded parts of the dish, the posterior parts showed no definite reaction of this sort, but re- ^ Number i of these notes, " The Source of Material of New Parts and Limits of Size," appeared in volume xxxiv^ of the American Naturalist, March, 1900, PP- 173-177- 129 130 Frank R. Lillie. mained scattered irregularly over the bottom, some with the ventral surface up. In four days five of these pieces died, in five days two more were dead, and in six days all of the posterior pieces were dead without having shown any signs of regeneration ; while all of the anterior pieces, kept in the same dish, were living and regenerating new tails. This experiment was repeated with similar result, and I soon became convinced that, in this form, while a tail might regen- erate at any transverse level from the pharynx back, a new head could not be formed from tissue behind the pharynx. The question now was, could a new head be formed in front of the pharynx at any level .'' Specimens were cut transversely immedi- ately in front of the pharynx. The result of these experiments was that, while a very narrow border of new tissue might be formed at the cut end of the posterior pieces, there was never regeneration of even the semblance of a head. In the next experiment the heads of five individuals were removed by a transverse cut just behind the eyes. The heads, thus removed, did not regenerate, but in five days it was apparent that new heads were forming on the decapitated pieces, and in one of them the eyes could already be seen. In six days new eyes could be seen in all. The capacity for regeneration of a head was thus demonstrated. Two questions now remained : first, is the development of a new head due to the position of the cut, or to the size of the piece "i and second, how far back does the capacity for regeneration of a head extend .? The first question received an answer in a very simple way. The head was first cut off just back of the eyes, and then from the anterior end of the major piece a small transverse part was cut. In six days a rudimentary head with eyes developed on one of three such small transverse parts. (The two parts that did not form heads were probably cut too small.) How far back does the capacity for regeneration of a head ex- tend } We have seen that it cannot be formed from tissue just in front of the pharynx, but that it can be formed just back of the eyes. Twelve specimens were cut transversely about half-way between the anterior end of the pharynx and the tip of the head. The reactions of both kinds of parts were quite normal, though the headless parts reacted much less rapidly than the head-bearing ones. In seven days both parts were rapidly regenerating, and eyes had appeared in the posterior parts ; in each head-bearing part a new pharynx and tail were forming. Both kinds of parts then completed Notes on Regeneration and Regulation in Planarians. 131 the regeneration rapidly, though in nineteen days the normal pro- portions were far from being restored.^ Thus in Dendrocoelum, while tissue may grow out at any transverse level, behind the region immediately back of the eyes, in tiie form of a tail, the capacity for regeneration of a head is limited to the anterior third or fourth of the body. I do not mean, of course, to state that the formation of a head back of this region is completely impos- sible. Some one may, at any time, demonstrate by operating on a sufficiently large number of individuals, that a head may exceptionally re- generate back of this level. But these experi- ments demonstrate very clearly that the power to regenerate a new head is limited in Dendro- coelum, by other conditions than size of the piece or presence of certain parts of the intestine. Dendrocoelum differs in this respect from both Planaria and Phagocata, but strangely enough resembles the earthworm Allolobophora fcetida. In this form, according to Morgan's^ observa- tions, regeneration of a head does not ordinarily occur back of the fifteenth segment and never behind the middle of the body, although a new tail may regenerate at any level back of the tenth segment. Why is it that embryonic tissue will continue to grow and differentiate on the posterior end of ABC (Fig. i), but not on the anterior end of ABD? It is not because the cells of the ecto- derm or of the mesenchyme or of the gut are incapable of growth ' This point deserves emphasis. In Planaria and also in Phagocata regener- ating parts rapidly assume the normal proportions, after the differentiation of the new tissue ; this involves great changes in the form and positions of the organs. The completely regenerated individual thus comes to possess not only all the parts of a normal worm, but also the same proportions. Morgan [Biological Lectures from the Marine Biological Laboratory, Woods Holl, 1899, Boston: Ginn & Co., 1900] has proposed the term " morphallaxis," for this phenomenon of transloca- tion of tissues by which the normal proportions are reassumed. In Dendrocoelum, morphallaxis is very slight, or in some parts entirely lacking. '■^ Morgan, T. H. : Archiv fiir Entwicklungsmechanik der Organismen, 1897, V, pp. 570-586. Figure L — Exact camera drawing of a stained specimen of Dendrocoelum. Note extensive ana- stomosis of gut-di- verticula especially on left side. In other specimens more than one trans- verse anastomosis between posterior gut rami may occur. 132 Frank R. Lillie. and differentiation at the level AB, because the same tissues, ABEF, that grow out into a tail, if forming part of CEF, will not grow at all if forming part of ABD. The explanation of this peculiar fact must lie in some conditions of the piece ABC not found in ABD. I believe that the determining condition is the presence of the brain and anterior part of the cen- tral nervous system in the anterior piece. How may the presence of this part of the central nervous system be thought to influence regeneration? The possibilities would seem to be : — I. That it may exercise a "trophic effect." 2. That it may act by coordinating the activities of the piece, and in conse- quence establishing normal stimulation in the regenerating part. 3. For regeneration of a brain we might suppose that certain cells not found in sufficient number back of a given region are necessary. All of these possibilities presuppose differences, which actually occur, between the central system of Dendrocoelum on the one hand, and of Planaria and Phagocata on the other, cephalization being more advanced in Dendrocoelum than in the other genera. I am unable to determine whether the brain does or does not exercise a " trophic effect," or whether the statement that it does would have any different meaning from what follows. Certain facts concerning the reaction of parts of Dendrocoelum incapable of regen- erating a head lead me to place particular emphasis on the second possibility. Some years ago Loeb made the interesting observa- tion, often since repeated (see Parker),^ that decapitated speci- mens of Planaria react to the usual stimulus of light like normal individuals, but more slowly. Any symmetrical piece of Dendro- coelum capable of regeneration tends to come to rest in the shaded parts of the dish precisely like a normal individual. No doubt the coordination of movements upon which so definite reactions depend, is regulated by the portion of the nervous system within the piece. However in Dendrocoelum, all parts incapable of re- generating a head also become incapable, after a day or two, of performing the usual reaction to light. Thus my notes concern- ing twelve specimens of Dendrocoelum cut in two immediately behind the pharynx read : " Two days after the operation all of the anterior pieces go to the side of the dish and remain in the angle between the bottom and side; but none of the posterior pieces ^ G. H. Parker and T. L. Burnett: This journal, 1900, iv, p. 373. Notes on RegeneratioJi and Regulation in Planarians. 133 shows such reaction ; these remain scattered over the bottom, about half of them with the ventral surface up." I have other similar records in my notes ; but also some in which the reaction seems to have continued normal. However, there seems to me to be a connection between the lack of correlation of movement and the inability to regenerate a head. It is therefore quite possible that the fate of the undifferentiated mass may be. detennined by the stiimilation of the normal movements ; if at the anterior end, the forward extensions and contacts may fitrnish sitch stimuli ; if at the posterior end the stimuli would be of a different nature. In fact, the stimuli would differ for all variants in position of the differentiating mass. But, though I am led by the conditions found in Dendrocoelum to attach particular importance to these normal stimuli, I certainly agree with Bardeen ^ that the internal conditions prior to regenera- tion are of great importance. It seems to me that earlier observers have erred chiefly in not paying attention to the conditions of the internal parts. It is especially the physiological actions of these parts that must be taken into account. Bardeen has insisted on this in the case of the intestine principally. But similar relations obtain also in the muscular system and in the nervous system. The contraction of the muscular coats of the body wall is of the nature of a peristaltic wave. In any piece therefore different conditions, brought about by the direction of the contraction-waves, would obtain at anterior and posterior ends. The direction of the nervous impulses within the portion of the nervous system contained in the regenerating part would be typical. Thus in the gut, in the muscu- lature, and in the nervous system the anatomical and physiological relations clearly distinguish anterior and posterior ends. Possibly similar anatomical and physiological conditions prevail in other parts. Before we have recourse to such dimly conceived hypotheses as "polarity," the possible differential effect of known anatomical and physiological conditions should be considered. It seems to me, therefore, that the factors determining the fate of embryonic tissue at the anterior and posterior ends respectively are probably of this nature: — in the first place, the actual anatomical and physiological relations of parts, in all of which in any transverse part there is some antero-posterior differentiation ; in the second ^ Bardeen, C. R. : This journal, 1901, v, p. i. 134 Frank R. Li Hie. place, the coordination of all organs comprising a separated piece, depending probably on the nervous system, and leading to normal stimulation of any exposed embryonic tissue. III. Regeneration of the Pharynges of Phagocata. The specimens of Phagocata gracilis with which I experimented were found in the same pond with Planaria maculata and Dcndro- cceliun. The manner of life is very similar to that of the other two genera. Apparently it does not reproduce by fission. The power of regeneration of this genus is equal to that of Planaria; for instance, a single specimen may be cut into sixteen parts capable of complete regen- eration. Phagocata, as is well known, possesses a very long pharyngeal pouch (Fig. 2), containing a large number of pharynges ; one of these occupies ;44 the usual position at the ante- rior end of the pouch, while the others are attached laterally and communicate with short branches of the posterior gut rami. ^ a. In most parts, after three Figure 2. — Pha- . , , , gocata gracilis 01* ^0"^ ^ays, several pharynges (slightly dia- are found regenerating simulta- grammatic). neously, though the more ante- The lines 3-3, . , , A A z :l i:. ^ ■ rior ones are more advanced, 4-4, 5-5, 6-6, in- ' dicate approxi- and posteriorly the earliest rudi- mately the nients are found (Fig. 3). source of the The formation of the pharynx Figure 3. — Phagocata geTeratId in and pharyngeal pouch from the gj^f ^^^ Regeneration ° ,1111 ot the part in front of Figs. 3,4,5, mesenchyme has already been the line 3-3 (Fig. 2). ^"■^ ^- noticed by Wood worth ^ for Phagocata and Bardeen for Planaria. The former author gives a detailed account. The first rudiment is an accumulation of cells of embryonic type ; near the periphery of this a semi-lunar cavity appears and forms the rudiment of the pouch ; the tissue projecting into the pouch differentiates into the pharynx. The mouth is a secondary perforation of the floor of the pouch. It may be worth ^ WooDwoRTH : Bulletin of the Museum of Comparative Zoology. Harvard, 1891, xxi, p. I. Notes on Regeneration and Regulation in Planarians. 135 while to notice in passing that this differs from the ontogenetic method of origin, both pharynx and pouch arising in the latter case from an ectodermal invagination. b. The origin of the pharyngeal pouch as a cavity in the mesen- chyma surrounding the free end of the pharyngeal rudiment has the following curi- ous effect: the definitive common pouch arises as a series of separated cavities, which secondarily fuse together. The origi- nal partitions are recognizable for a long time as pointed projections of tissue into the pouch between the pharynges. Figs. 3 and 4. But in some cases of regeneration the fusion is incomplete or entirely absent, so that two or more pouches may be pres- ent, each containing sev- eral pharynges. Fig. 4. A curious abnormality is represented in Fig. 5 : two pharynges have a common termination; evidently two buds arose very near together and partly fused. c. As indicating that there is a special relation between the regeneration of the gut and the pharynges, the regeneration of lateral pieces offers some features of interest. The pharynges form near to the cut edge, as does the single pharynx in Planaria. Fig. 4 shows a case of this sort; there is a long pouch near the cut edge, in this case divided in two parts by a partition, and all of the pharynges are on the side towards the old tissue, with a single exception. Connecting with each pharynx, however rudimentary, is a branch of the intestinal system. The pharynges never begin to develop in such pieces on the outer side of the pouch until a branch from the intestine supplies the neighborhood. It will be seen that a new branch, destined to form the posterior ramus of the regenerat- ing side, has grown along outside the pouch. This branch is very slender, but in connection with it a new pharynx has already differ- FiGURE 4. — Phagocata gra- cilis. Regeneration of the part bounded by the lines 4-4 (Fig. 2). Figure 5. — Phago- cata gracilis. Re- generation of the part behind the line 5-5 (Fig. 2). 136 Frank R. Lillie. entiated ; other pharynges form along this side later. But in many observations made on the regeneration of lateral pieces, I never found the regeneration of pharynges along the cut side begun, until the posterior ramus of the intestine of the same side was formed. (See also Fig. 6.) d. Usually the regeneration of the various parts takes place in a coordinated fashion, the pharynges developing in relation to the intestine, etc., and all parts in such a manner that normal, i. e., adaptive, relations are established. Fig. 7 shows an excep- tion to this rule. The part figured was as nearly as possible a 3^2 P^^^ of a normal individual ; the individual in question was divided into sixteen equal transverse pieces, and each of these was cut through its centre. None of these parts re- generated completely. Their history was as follows. Operation, Aug. ist, 1899, thirty- two pieces; Aug. 2nd, twenty-one pieces liv- ing, the smaller pieces very thick dorso-ven- trally ; Aug. 4th, two pieces living. These were killed and stained, and a drawing of one of these is shown in the figure. The points of interest are that, while the piece has not regenerated a whole, the in- testine has grown out in all directions through the piece, and at each of two places a pharynx has regener- ated. These pharynges are widely separated and turned in opposite directions, a species of heteromorphosis. The side toward which they lay was evidently the injured side. Certain regenerative pro- cesses may go on in such a piece, but there is not full coordination, and the result is non-adaptive. Figure 6. — Phago- cata gracilis. Re- generation of the part bounded by the lines 6-6 (Fig. 2). Figure 7. — Phagocata gracilis. Regeneration of a ^^j. The gut branches were very broad, irregular, and branching in more than one plane ; the superimposed branches are not represented. Notes on Regeneration and Regulation in Planarians. 137 IV. Theoretical and Critical. The phenomena of regeneration offer many problems, some of which not only appear insoluble in the present state of our knowl- edge, but actually offer no point of attack. For instance, if one were to ask why the regenerating rudiment of the head develops into such different forms in Planaria, Dendrocoelum and Phagocata, it could only be said that we do not know ; nor are we able to say why Phagocata has many pharynges and the other genera of planarians only one. No physiological explanation of these phenomena can at present be offered. But if we leave out of account such problems, there still remain certain problems of regeneration that may be attacked with good expec- tation of success. The localization of regenerating organs is one of these. In the case of the regeneration of Hydroids, Loeb^ has treated this question with great success, by showing that the regener- ation of hydranths or rhizoids depends on certain external stimuli, such as, in different cases, light, gravitation, or contact. A recent paper by Bardeen ^ seems to me to make for the plana- rians a decided advance in this direction inasmuch as he pays particular attention to the relations of the regenerating parts to the organs originally present in the piece. I propose, therefore, to devote some attention to Bardeen's theory in the light of experiments described in the second and third of these notes, which deal with dif- ferent genera of planarians from the one (Planaria) used by Bardeen. I shall therefore first state his conclusions at some length. I. " Embryonic tissue is formed in the specimens here studied at two places only, (i) at or near a cut surface, and (2) in the region of the piece just posterior to the point of least intestinal pressure." The author offers the following suggestions as to why the embryonic tissue should be formed at these places. The formation of embryonic tissue at the cut surfaces may be due to one or other of these causes: (i) slight change in the osmotic pressure; (2) exposure of the internal tissues to the water; (3) enzymes set free by the injury.^ " The cause of the formation of embryonic tissue just pos- ' LoEB, Jacques: Biological Lectures Delivered at the Marine Biological Laboratory of Woods HoU, 1893, p. 37, Ginn & Co., Boston, 1894. ^ Loc. cit. ^ LoEB : Tliis journal, 1900, iv, p. 60. 138 Frank R. Lillie. terior to the point of least intestinal pressure is equally dark. The process is much slower and is preceded by a retrograde metamorphosis in the pre-existing adult tissue. VVe might assume that here certain intestinal fluids are set free by pressure." 2. "The differentiation of this embryonic tissue depends on its relations to the intestinal apparatus of the animal, (i) If it lies an- terior to the main axial gut, it becomes converted symmetrically into a head region." (2) " If the embryonic tissue lies lateral to the axial gut or to a line extending directly posterior to this, it becomes converted into a new lateral region." (3) "If the embryonic tissue lies at the posterior end of the axial gut, i. e., behind the point of least intestinal pressure, it becomes converted into a new pharynx and pharyngeal pocket." (4) "If the embryonic tissue lies poste- rior to the pharyngeal region, it becomes converted into a new tail." The author concludes from these facts that the intestinal system has a specific action in determining the nature of the parts developed from the embryonic tissue, rejecting the idea that the development of the intestinal system may be merely a coincident phenomenon. He also develops the outline of a theory of the way in which the intestinal system may be supposed to exert such specific effect. There will be no need, however, to consider this, if it can be shown that the rejected alternative is the correct one. The whole theory is based on the relation of the " axial gut." Now what is the axial gut .-* It is what zoologists call the anterior ramus of the intestine, and it extends from the base of the pharynx to the head. (See Fig. i). That is, it is something that can be defined only by its relation to the pharynx and the head. It cannot, there- fore, be otherwise than as stated, if regeneration is to take place at all. The head must form in front of it, the pharynx behind it, and to its sides the sides of the body. If it happened otherwise there would be no regeneration, but the development of a monstrosity. But the author meant more than a mere truism of this sort ; he meant to assert that, however much embryonic tissue might form, it would remain undifferentiated until the establishment of an axial gut, which then exercises its all-compelling influence. How far this con- clusion is justified by the mere coincidence of the phenomena dealt with we may now pass on to inquire. In Bardeen's opinion, then, the localization of regenerating organs depends entirely on the prior regeneration of the axial gut. I think that it can be shown that this conclusion is incorrect in some respects. We Notes on Regeneration and Regulation in Planarians. 139 have seen what is the criterion of the axial gut in normal individ- uals ; what is the criterion in regenerated parts? Curiously enough, Bardeen does not dwell on this essential point; but I think that one who reads his paper with sufficient care will see that the criterion actually employed is the same here, i. c, relation to the head and pharynx. If this be so, Bardeen has simply argued in a circle. The con- sideration of one class of cases discussed by Bardeen will serve to show that this has actually been done. " In the case of a very oblique cross- piece (in front of the pharynx) a lateral branch of the old axial gut may be transformed into an extettsion of the axial gut. The head then devel- ops symmetrically around the tip of this, and hence somewhat lateral to the axis of the parent worm." ^ Why in this case does Bardeen regard a certain lateral branch of the original axial gut as transformed into an extension of the latter .'* For no other reason than that the head appears lateral to the cut end of the original axial gut. There is, in fact, in such cases, no other criterion of the axial gut. The figures illustrating this very case are conclusive in this respect. Re- generation of head and axial gut are really coincident phenomena. I have moreover a large number of observations that show the form of the regenerating gut to be dependent on the form of the new external parts, and not vice versa. The observations recorded in the second of these notes show that the presence of a certain amount of embryonic tissue in front of an unques- tionable axial gut does not furnish all the conditions necessary for the regeneration of a head. It was shown, that in Dendrocoelum there is a region extending a certain distance in front of the pharynx, in which regeneration of a head will not take place. A piece includ- ing this region after about two days has a certain amount of embry- onic tissue on its cut end ; it possesses also a certain portion of the original axial gut; yet a head does not form. Other conditions are necessary. Concerning the localization of the pharynx in regenerating parts, it is held by Bardeen that wherever gut pressure is least, i. e., at the place in the system where waste matters tend to accumulate, there the new pharynx forms. The criterion is not so indefinite here, because in all parts containing any of the three original main rami of the gut, by "place of least pressure" is meant simply the most ^ Loc. cit., p. 35. The development of the new head lateral to the axis of the parent worm in the case of oblique cuts was noticed first by Morgan and received extensive treatment at his hands. 140 Frank R. Lillie. proximal part in relation to the old pharynx. But I do not see how it is possible to determine the place of least pressure in the regenerat- ing system when no part of the original three main rami are present, until they are again established. But by this time the regeneration of the pharynx is usually begun, so that it is often difficult to say what is the determining factor in phenomena that are so nearly coin- cident. But here again comparison of other forms in which the conditions are only slightly different is very instructive. In Phagocata it is very evident that intestinal pressure has nothing at all to do with the regeneration of the lateral pharynges. Reference to Figs. 3, 4, 5, 6, will show that the lateral pharynges develop at the ends of short intestinal branches that are extremely delicate, and in which the intestinal pressure must be higher than anywhere else. It is at least probable that there is an intimate correlation between regeneration of the intestine and of the pharynx or pharynges. I think that it is also probable that the stimuli inducing pharynx formation proceed from intestinal cells of new formation. But all new portions of the intestine do not induce the formation of new pharynges. There is another determining factor. What is this .-^ It seems to me improbable that Bardeen's answer is in any way correct, because it presupposes a typical form of regeneration of the gut that does not exist. As a matter of fact, the regenerating gut is at first in most cases extremely irregular, and the restoration of its typical form is a matter of secondary regulation, due, I believe, to the form of the body, position of the pharynx and the tendency of the intestinal fluids to flow in the paths of least resistance. In studies 2, 3, and 4 I have attempted to show the following facts : — I. The differentiation of exposed embryonic tissue may be de- pendent on the external stimuli to which such tissue is exposed. The functional correlations of all the parts of a piece capable of regen- eration are the internal factors, and the various stimuli from without thus induced in normal sequence are the external factors which deter- mine the location of organs. The case of Dendrocoelum appears to indicate that functional correlation is dependent on the nervous system. The regeneration of a head lateral to the axis of the parent worm in the case of parts cut off obliquely is thus explained, because stimuli which normally would fall upon the head are received by the most advanced part, which is lateral to the original axis in Notes on Regeneration and Regulation in Planarians. 141 the case of such an anterior cut surface. The regeneration of a tail is similarly explained. 2. The intestinal system regenerates in relation to the new external parts, and not vice versa, as maintained by Bardeen ; from which it follows that the location of new parts cannot be due primarily to the form of the gut. Woods Holl, August, 1901. ARTIFICIAL PARTHENOGENESIS PRODUCED BY MECHANICAL AGITATION. By a. p. MATHEWS. [From the Marine Biological Laboratory, Woods Holl, Mass.'\ THE observations of Loeb ^ and Morgan ^ have established the fact that the unfertilized eggs of many different kinds of ani- mals can be started in development either by increasing the osmotic pressure of the sea-water and thus depriving the eggs of water, or by the action of certain specific ions. The observations in the present paper show that mechanical agitation of eggs of the star-fish (Asterias Forbesii?) without any change whatever either in chemical constitu- tion or osmotic pressure of the sea-water will produce the same result. The mechanical disturbance necessary to start the process is in cer- tain conditions of the eggs so slight as to render this a serious source of error in all experiments in artificial parthenogenesis. Some years ago I discovered that vigorous shaking of the unripe eggs of the star-fish for a few moments would cause the unripe eggs to mature, to extrude their polar globules and to be ready for fertiliza- tion. Morgan'^ confirmed and extended these observations. He showed that shaking ruptured the membrane of the germinal vesicle, that this increased above the normal the speed of development of such eggs after fertilization and that the unripe eggs of the sea- urchin, Arbacia, acted similarly. At the time and subsequently I attempted to induce the eggs to continue their division and to de- velop parthenogenetically but always with negative results owing, as I have since found, to my having shaken the eggs too early. Mor- gan, however, observed that in some of his shaken cultures certain eggs developed to gastrulae even without the addition of sperm, but he attributed these to the accidental introduction of spermatozoa through the sea-water in the pipes of the laboratory. The process which finally led to a successful result is, in brief, as ^ Loeb: This journal, 1901, iv, p. 423. ■^ Morgan: Archiv fiir Entvvickelungsmechanik der Organismen, 1899, viii, p. 448. 3 Morgan : Anatomischer Anzeiger, 1893, ix, p. 141. 142 Artificial Pai-the77 agenesis. 143 follows. The eggs are shed into sterile sea-water and allowed to remain there from two to four hours until both polar globules have been extruded and the female pronucleus has re-formed and reached a considerable size. The eggs are then removed to a test-tube with some sea-water and shaken vigorously five or six times back and forth and poured into a large dish with plenty of sea-water. The eggs then undergo the changes described later and a certain number, vary- ing from less than one to more than fifty per cent, depending on the state of ripeness of the eggs, develop into actively swimming blastulae and gastrulae which live for from twenty-four to forty-eight hours and, under favorable conditions, occasionally reach the bipinnaria state. The amount of agitation necessary to start the process of develop- ment varies in different individuals and depends in part upon the length of time the eggs have been lying in sea-water. After from four to six hours in the water the mere transference of the eggs from one dish to another by a pipette, or the jar caused by setting the dish containing the eggs down sharply on the table is sufficient to start development in a small portion of the eggs, with the result that swimming blastulae appear by morning. Even the most careful transference of the eggs with a large mouth pipette from one dish to another into which they are gently introduced with the mouth of the pipette below the surface of the water will sometimes suffice to yield one or two embryos among several thousand eggs, while the untransferred eggs are wholly undeveloped. In the experiments here cited all dishes and pipettes had been sterilized by washing in fresh water containing hydrochloric acid, followed by thorough washing in fresh water to remove the last traces of the acid. My hands and instruments were similarly steri- lized and all sea-water had been heated to 70° C. before using. The individual star-fish were thoroughly washed with a brush under run- ning tap-water to remove any adherent sperm before opening them. With such precautions I have never seen one fertilized tg^ out of thousands examined. Only in a single instance did the non-trans- ferred controls show any advanced embryos, though here and there the first segmentation stages might rarely be met with. The single exception mentioned was the discovery of one irregular partheno- genetic blastula among many thousand untransferred eggs of a very ripe female. It must not be inferred from the precautions taken that the sea-water in the pipes of the laboratory contains sperm, as has been often stated. For I placed a large number of ripe eggs in a 144 A. P. Mathews. thin layer in the bottom of a large dish and allowed the water from the pipes to flow over them for twelve hours and at the end of that time a careful examination of twenty thousand eggs taken from different parts of the dish did not reveal a single fertilized egg. The precautions described were taken to remove any possible doubt of fertilization. Following are protocols of some of the experiments. Experiment 52. August 5. — Unfertilized Asterias eggs in water four hours. 2.45 P.M., shaken slightly in a test-tube. 3 p.m., the eggs have put out fertilization membranes and look like fertilized eggs. 5 p. M., the eggs begin to divide, some into 2-4 cells very regularly, others irregularly. 9 p. M., clear normal looking morulas are common ; a clear area (nucleus ?) visible in each blastomere. 9 a. m., many eggs are dead and opaque ; many ill-shapen swimming blastulge observed. Some dwarfs, some as large as normal. Several embryos are swimming in the fertilization membranes surrounded by the debris of dead cells. Experiment 54. — 3.30 P. M. Same star-fish shaken hard. 8.45 P. M., many 16-32 cell stages observed. More than half the eggs are dividing irregu- larly with the fertilization membranes out. Next morning almost all dead. Experiment 58. August 7. — Non-mature eggs shed at 10.30 A. m. Shaken at I I.I 5 A.M. 3.40 P.M., all are mature and a large number have fertili- zation membranes. 8.30 a.m., next day: several normal looking non- ciliated blastulse found, but most eggs are dead and disintegrated. Experiment 89. August 12. — Female Asterias. Eggs shaken after thirty- five minutes in sea-water. Aug. 13, 9 a.m., a (qw fairly regular segmen- tation stages. No swimming embryos. Another lot shaken harder was also negative. Experiment 107. — Shaken when one polar body had been put out. 2 P. M., next day : almost all eggs dead. A few blastulge found after long search. Experiments 126, 127, 128, 129, 130, and 131. — 2.30 P. M. Shaken in test- tube with progressive severity after three hours' ripening in sea-water. 5 p. M., all ripe eggs have fertilization membranes and the unripe eggs are maturing. About one tenth of all the eggs are in the 2-8 cell stage. Many appear very regular and look exactly like fertilized eggs. 8.30 p. m., one tenth of the eggs are developing from two cells to normal blastulae. 9 A. M., next morning : every pipette full of eggs has many swimming blastulae except those from eggs shaken hardest. The blastulae are many of them abnormal in appearance, being small and irregular in outline. Most eggs have died after a partial development. Experiment 132. — 2.30 P. M. Same eggs drawn up in pipette and shot into another dish of water. 9 a. m., next day : vast majority have no develop- ment ; but here and there a swimming blastula is found. Artificial Parthenogenesis. 145 Experiments 133 and 134. — 3 p. M. Same eggs as preceding drawn up in a narrow pipette and shot forcibly into a dry dish and then covered with water. 8 P. M., 2-16 cell stages are common. The immature eggs have not matured. 9 a. m., next day : two or three swimming blastulae form in each lot of eggs examined (2000 eggs). Experiment 136. — Same eggs transferred from the control dish to fresh sea- water with great care using a large mouth pipette introduced beneath the surface of the water. 9 a. m., next morning : after long search among many thousand eggs, one embryo was found. Experiment 137. — Control of the preceding experiment. The eggs were left in the dish in which they matured. Long search the next morning failed to reveal a single embyro. The effect of different degrees of mechanical violence is shown in the following experiments : Experiment 147. August 17. — Female Asterias. Eggs left in water from 12 M. to 4.30 P. M. Not all ripened, as too many eggs were in the vessel. Transfers were made from this lot of eggs before the female pronucleus had re-formed. Experiment 147 (a). — A large pipette filled and transferred with great care. 9 A. M., next morning : in 6000 eggs only one undergoing an irregular segmentation (into three cells). All other ripe eggs are dead without development. No fertilization membranes to be seen. Experiment 147 (b). — A few eggs from the same star-fish shed into a large amount of sea-water and left undisturbed. Not an egg shows any sign of development. Experiment 147 (c). — Same eggs as 147 (a) squirted into a dry dish and sea-water added. 9 a. m., no embryos. Experiment 147 (d). — Squirted and drawn up and squirted once more. No segmentation. No embryos. Experiment 147 (e). — Squirted four times into a dry dish. 9 a.m., three irregular non-swimming blastulae in 1500 eggs. Nine other eggs irregularly segmenting. Experiment 147 (f). — Squirted eight times. 9 a. m., some swimming blastulae. Many non-swimming. 46 embryos in 2000 eggs. Experiment 147 (g). — Squirted sixteen times. 9 a.m., one perfect swim- ming blastula, many imperfect non-swimming. 43 embryos ranging from early segmentations to blastulae. Many eggs disintegrated. Experiment 147 (h). — Same eggs. Shaken very lightly in a test-tube. No development. Experiment 147 (i). — Shaken harder. No development. Experiment 147 (j). — Shaken still harder in a test-tube. 9 a.m., many en- tirely normal looking segmenting eggs from 2-4 cells onward. Out of 146 A. p. Mathews. 2000 eggs, 107 observed in various stages of development. No swim- ming embryos seen. Few eggs of this star-fish developed when fertilized. Experiment 170. August 18. — Another star-fish. Eggs in water three hours before transference. Female pronucleus re-formed. 4.30 p. m., carefully transferred under water. 9 a.m., next morning : out of 2500 eggs only one early development of two cells found. Only one third have matured. Experiment 170 (a). — Same as Experiment 170. No embryos. Experiment 170 (b). — Squirted twice. In 2500 eggs a few 3, 8, and 16 cell embryos found. Experiment 170 (c). — Squirted four times. 9 a.m., 15 embryos in 3500 eggs. Irregular blastulte and segmentation stages. Experiment 170 (d). — Squirted eight times. 9 a.m., 24 embryos in 2200 eggs. Only one or two well developed and normal in appearance. Experiment 170 (e). — Squirted sixteen times. 9 a.m., six embryos mostly 2-4 cells. Almost all dead, Experiment 170 (f). — Shaken greatly in a test-tube. 9 a.m., two or three have begun to develop in 2000 eggs. In this star-fish also very few of the eggs could be fertilized and the development of these was mostly abnormal. The following experiment shows that the eggs must be shaken after maturation is complete in order to insure a successful result. Experiment 189. August 19. — Very ripe female. Nearly all eggs mature. 9.30 A.M., eggs of set A were transferred as soon as shed into sea-water without sperm. 1.30 p. M., the second set (B) transferred after the for- mation of the female pronucleus. Experiment 189 a (i). — Transferred carefully. No development of any kind. Experiment 189 a (2). — Eggs not transferred. No development at all. Experiment 189 a (3). — Eggs squirted into water. No development. Experiment 189 a (4). — Eggs squirted into dry dish. No development. Experiment 189 a (5). — Eggs squirted into dry dish four times. No de- velopment. Experiment 189 a (6). — Eggs squirted into dry dish eight times. A few fer- tilization membranes out. No development. Experiment 189 a (7). — Squirted twelve times. No development. Experiment 189 a (8). — Squirted sixteen times. No development. Experiment 189 a (9)* — Shaken in test-tube hard. The female pronucleus re-forms before that of the unshaken eggs and is abnormally large. No development. These eggs, though not developing, are so sensitive that transferring them with a pipette to a small watch glass causes all to put out fertilization membranes and develop irregularly. Artificial Parthenogenesis. 147 Experiment 189 b (i). — 1.30 P.M., transferred carefully as maturation is com- plete. 3.50 p. M., about one in eighty have fertilization membranes and nuclei have disappeared. 8 a. m., next day : one normal swimming early gastrula found in 2000 eggs. Experiment 189 b (2). — Squirted into sea-water. 4 p.m., one half have fer- tilization membranes and nuclei gone. Several have divided into 2-3 cells. 7.50 P.M., about 80 per cent have fertilization membranes and are budding off irregular clear protoplasmic masses. 9.30 a.m., about go per cent of the eggs are dead and disintegrated. No swimming blastulae found in 700 eggs, but many non-swimming. Experiment 189 b (3). — Squirted into dry dish. 4.10 P.M., about one tenth have fertilization membranes out. 7.30 p. m., many 2-4 and 8 cell stages. 9 A. M., most are dead. Two swimming blastulae found in 700 eggs. Experiment 189 b (4). — Squirted twice. 4 P. M., one in seven has fertiliza- tion membranes out. 9 A. m., two swimming blastulae found in 1200 eggs. Experiment 189 b (5). — Squirted four times. 9 A. M., four ciliated embryos in 1500 eggs. Experiment 189 b (6). — Shaken in test tube. 9 A. M., many have started, but only four in 2000 eggs have reached the blastula stage. A similar negative result was obtained in star-fish 217, transferred with and without shaking at 9.30 a. m., before maturation. The same eggs transferred at 1.30 P.M., after maturation was complete, gave some swimming embryos in every culture and in that squirted eight times thirty swimming gastrulae in 4000 eggs were found next morning. Experiment 274. — 11 A.M. Eggs shed into sea-water. No eggs develop in the untransferred control. Lot A was transferred at 1.45 p.m. ; lot B at 3 p. m. ; lot C at 6 P. M. ; and lot D at 8 p. m. Experiment 274 a (i). — 1.45 P.M. Eggs picked up in a pipette and shot into water. 10 a.m., next day: no embryos in 3000 eggs examined. Experiment 274 a (2). — Squirted twice into a dry dish. 10 a.m., no em- bryos swimming. Experiment 274 a (3). — Squirted four times. Seventy-five ciliated blastute and gastrulae in 2500 eggs found next morning. Many eggs had dis- integrated. Experiment 274 a (4). - 1.45 P.M. Shaken very lightly in a test-tube. 1 1 a.m., no swimming blastulae. Many 2, 4, and 8 cell stages up to blastulae with a normal fertilized appearance. All eggs alive. Fertilization membranes out. Experiment 274 a (5). — 1.45 P.M. Shaken harder. 11 a.m., many eggs un- dergoing perfectly regular segmentation. No swimming larvee, but farther along than 274 a (4). About one third are developing. None dead. Experiment 274 a (6). — Shaken still harder. 11 a. m., fully one third are de- veloping and many blastute are just beginning to swim. 148 A. p. Mathews. Experiment 274 b (i). — 3 P.M. Transferred carefully to another dish and then the dish set sharply down on the table and the eggs drawn up in a pipette and shot out. 6 p. m., about one seventh of the eggs segmenting directly into four, eight, or sixteen cells. Three to four clear areas visible in other eggs. 8 p. m., 64 and 128 cell stages are numerous. At 8 a. m. 65 ciliated blastulae found in 500 eggs. Experiment 274 b (2). — 3 P.M. Squirted into water. 2 p.m., next day: many irregular groups of cells and blastulse. Experiment 274 b (3). — 3 P.M. Shaken in a test-tube. 2 P. M., next day: the dishes are full of swimming gastrulse and blastulae and swimming pieces of embryos. About one third of all the eggs are swimming. Some disin- tegrate after partial development. Of the control eggs almost one half not matured. Experiment 274 b (4). — 3 P.M. Shaken harder. 2 p.m., next day: fully one half the eggs are swimming blastulae and gastrulae. Experiment 274 c (i). — 6 P.M. Squirted into water. 8.30 a.m., 65 embryos mostly segmental stages of 100 cells and upward, but some swimming blas- tulse found in 2000 eggs. Experiment 274 c (2). — 6 P. M. Squirted into water. 11 a.m., next day : 157 larvae, many just beginning to swim, were found in 3000 eggs. Experiment 273 d (i). — 8 P.M. Squirted into water. 11 a.m., no swimming blastulae, but about one seventh of the eggs are going as large morulae and irregular heaps of clear cells. Many non-dividing eggs are full of clear areas (asters ?) . Experiment 274 d (2). — 8 P.M. Squirted into water. About one fifth de- velop chiefly into blastulae. A few ciliated and swimming. Experiment 274 d (3). — 8 P.M. Shaken hard. 9 a.m., next day : almost all eggs have put out fertilization membranes, swollen and disintegrated. No swimming embryos. The foregoing experiments show clearly that the eggs of the star- fish can often be made to develop with great regularity into blastulae and gastrulae. As a rule the embryos die in from twenty-four to thirty-six hours and are abnormal in shape and appearance, being gen- erally smaller, more opaque and thicker walled than the normal. The riper the eggs, however, the more normal are the embryos, and in many instances the embryos could not be distinguished from the fer- tilized gastrulae of the same star-fish. I have not as yet succeeded in getting them well into the bipinnaria stage, but it happened that the star-fish at this season of the year were not entirely ripe and only a relatively small number of the eggs even when fertilized would, with- out the aid of pilocarpine, go beyond the gastrula stage. Artijicial Parthenogenesis. 149 1 do not think my results can possibly be attributed to any acci- dental infection with sperm nor to any self-impregnation by any pos- sibly hermaphrodite individuals, which Cuenot^ claims exist. For in every instance but one the eggs not transferred, or transferred before maturation was complete, showed no development whatever. I have repeated these controls at least fifty times with a constant result ex- cept in the one instance mentioned. The great sensitiveness of the eggs after maturation to mechanical shock was very surprising. The majority of all ripe eggs will, if shaken, begin to develop, though, as already stated, only a few of them reach the blastula stage. Merely drawing the eggs into a pipette to transfer them to another dish may bring about development. An in- spection of the experiments shows also that the eggs differ in sensi- tiveness at different periods in their maturation. Immediately after shedding into sea-water shaking causes no development. After two hours larvse begin to appear on shaking. At four hours, hard shaking produces a very large proportion of larvae, while merely transferring gives but one or two. A few hours later, transferring the eggs will cause a large number to begin to develop, though as a rule the de- velopment does not go beyond late segmentation stages. At this time shaking causes all the eggs to begin to develop, but none reach the blastula stage. The question arises whether the star-fish is normally partheno- genetic, as McBride ^ says Asterina gibbosa is. Greef ^ in one instance observed the eggs of Asterias rubens develop without sperm, and Hertwig^has recorded similar observations upon Asterias glacialis and Astropecten. It can hardly be possible, hovs^ever, that the general parthenogenesis of Asterias could be overlooked. Certainly the eggs of the animals I observed were not naturally parthenogenetic, since, if left undisturbed, they never developed. Hence, I am inclined to be- lieve that this star-fish while not normally parthenogenetic is never- theless on the verge and that it may be started in several ways. I feel somewhat doubtful about the extent of parthenogenesis which may occur naturally because in almost all my experiments a certain number of the eggs, and in some cases a large proportion, failed to put out the polar globules when shed, and were hence presumably not ^ CuENOT : Zoologischer Anzeiger, 1898, p. 273. 2 McBride : Quarterly journal of microscopical Science, xxxviii, p. 339. ^ Greef : Cited from Viguier: Annales des sciences naturelles, 1901, p. 117. * Hertwig: Jenaische Zeitschrift fiir Naturvvissenschaft, 1890, xxiv, p. 304. 150 A. p. Mathews. fully mature. It is possible that the examination of the fully mature individuals in June, when, according to Mead, they shed their eggs naturally, may show a different condition. It may be said, however, that in my experience as the eggs approach a condition in which they fertilize readily and develop normally their parthenogenetic develop- ment becomes easier to bring to pass. It may be said positively, however, that the individuals observed during the present work were certainly not parthenogenetic unless the eggs had been disturbed. It may be that Greef s and Hertwig's observations were due to disturb- ing the eggs. I strongly suspect that a large number of eggs of other animals will be found to be in a similar state of unsteadiness, making it necessary to handle them with extreme care in all experiments involving the production of parthenogenesis in other than mechanical ways. Loeb. and Fischer permit me to announce that they have already confirmed my results for Amphitrite, Chaetopterus and Nereis, which are also easily started in this same manner. The possible inference that the parthenogenesis already observed by Loeb in these and other forms may have been due to his transfers of the eggs from one medium to another instead of to the chemical action of the reagents he employs, is, I feel sure, not justified. So far as Arbacia goes, I have repeatedly tried to secure development by agitation without success, and attempts by Loeb and Lewis at Woods HoU this summer have also been nega- tive. In the other forms Loeb employed, with the possible exception of the star-fish, the number of larvae obtained by the methods em- ployed by Loeb, i. e., raising the osmotic pressure, or the action of H, K, or Ca ions, give far more larvae than the controls which are trans- ferred in the same manner. It is, of course, not impossible that in some instances at least the loss of water by the protoplasm, or the action of certain chemicals, may so raise its instability that a very small mechanical shock will suffice to start parthenogenesis. Many of Loeb's results, however, are not open even to this objection, to which, at the most, I attribute no great weight, since the eggs were not retransferred from the altered sea-water to the normal, but left in the former subject to the same conditions as the control. It may be well in passing, however, to mention the fact that if ripe star-fish eggs are transferred, after maturation, with a sterile pipette to a small, clean watch-glass for microscopic examination, and are then retransferred to the original dish, a few embryos are almost certain to appear the next morning in the dish. I have observed this several times. It Artificial Parthenogenesis. 151 may be that the occasional parthenogenesis of various sea-urchins described by Viguier^ may have been produced in this manner. But the difficulty he appears to have experienced in repeating his experi- ments successfully, when strict precautions were taken against sperm infection, lead me to suspect that he was dealing rather with some accidental fertilization than, to use his own phrase, an " accidental parthenogenesis." The microscopical changes in the egg caused by shaking or agi- tation are remarkable. It is surprising that so slight a shock can produce so profound a structural effect. These changes have not been studied in sections, so that I can describe only what may be seen in the living egg, which, however, is large and fairly transparent. The first and most striking change is the development of the fertilization membrane, which appears a few minutes after shaking. At the same time the shape of the egg changes from spherical form to a flattened ellipsoid, the flattening generally appearing in the neighborhood of the polar bodies of the fertilized egg. The nucleus which before shaking is present as a large single nucleus, or as a collection of from three to seven separate spheres, undergoes a change. The nuclear membrane fades away either within a few moments or after several hours and the nucleus quite disappears from view. As a rule, its loss is followed after a time by the appearance of from two to thirty clear areas in the &g^, or by a peculiar budding off of clear portions of the protoplasm about the periphery, although this budding may at times take place without the disappearance of the nucleus. As a result of this budding there is formed a morula-like mass of cells with a larger or smaller undivided mass in the centre. These budded off" portions which look like small cells probably do not contain nuclei since they soon go to pieces. It not infrequently happens that the large undivided piece of the cell subsequently develops into an embryo which acquires cilia and rotates rapidly in the debris of the dying buds. A large number of clear areas may develop in a sin- gle Q^g, and the egg segment at once into a very large number of cells. Blastulae so formed do not, however, so far as observed, develop farther, but die in the course of two hours ; but I have not made a careful study of this point. The manner in which shaking brings about development is uncer- tain. In the case of unripe eggs the maturation is ushered in norm- ^ ViGUiER : Annales des sciences naturelles, 1901, p. 88 ; Comptes rendus, 1901, June 10 and July 15. 152 A. p. Mathezvs. ally by the disappearance or dissolution of the nuclear membrane at one spot where the centrosome appears. If eggs do not naturally mature the nucleus persists unchanged. If now the eggs be shaken, the nuclear membrane either breaks mechanically or is dissolved. This has been observed by Morgan and myself. The ripening of the egg appears to be connected in some way with the dis- appearance of the nuclear membrane and the presumable discharge of nuclear matter into the cytoplasm. Similarly, in the eggs after maturation, shaking causes a change which manifests itself by a dissolution or disappearance of the nucleus. This change of the nucleus, however, is not, in the majority of instances, due to mechanical rupture, for it is frequently inaugurated from many min- utes to from two to four hours after shaking has stopped. These observations, together with the well known fact that the centrosome originates as a rule close to the nucleus suggests that the dissolution of the nuclear membrane is a determining factor in karyokinesis. It is possible that if the agitation cause dissolution in one place only, such as normally occurs, the normal division into two cells results, but that if dissolution occur in the periphery, generally many asters are found leading to division all at once into a large number of cells. In this way the appearance of polyspermy might easily be brought about. On the other hand there is also the possibility that mechanical agitation causes the cell to lose water, as happens in some plant-cells when they are stimulated, for example, in the leaves of a sensitive plant. A condition might result similar to that produced by raising the osmotic pressure of the surrounding liquid. Such a loss of water could most probably be produced by a lowering of the osmotic pressure of the cell and this only by a reduction of the number of molecules in the ^gg. This would mean a combination between pos- sibly the organic and inorganic constituents of the protoplasm, which in turn might lead to dissolution of some of the cell elements. The whole subject is at present in such a state that no definite conclusions may be drawn, but I believe many facts of cell physiology might be explained by such a rhythmical or unstable combination between the organic and inorganic constituents of protoplasm. This is a matter which I hope to consider more carefully later. Too strong shaking of the eggs causes a dissolution of the whole ^g'g. After shaking has ceased and after the process of nuclear dis- appearance already described has taken place the egg begins to swell and ultimately either dissolves altogether or remains a swollen mass Artificial Parthenogenesis. 153 of debris. In such case an increase in osmotic pressure would appear to be the result. The fact brought out by these observations that karyokinesis may be started mechanically is, I think, possibly of general application to protoplasm. Loeb and Fischer's observations of similar processes in the Annelids there indicate this. May it not help us to understand how irritation of the skin may cause the development of callous areas, or irritation of bone the production of bony growths .'' The change, whatever be its nature, which is set up by mechan- ical agitation lasts for several generations of cells and causes a hast- ened development. This Morgan had already seen in fertilized eggs. It is visible also in the unfertilized eggs. In Experiment 274a (4, 5,6) it is seen that those eggs shaken very gently were the next day in the early stages of segmentation up to one hundred and twenty-eight cells or more ; those shaken harder were blastulae just ready to swim; those shaken still harder were swimming blastulae and those shaken hardest of all were already beginning to gastrulate. Perhaps this change may be correlated with the apparent diminution of surface tension which exhibits itself in the change of shape of the egg from a sphere to an ellipsoid and in the budding off of portions and irregularity of the outline of the periphery of the Q.gg. That the cohesion is reduced is shown also by the ease with which such eggs may be shaken to pieces, being strikingly different in this respect from the unripe eggs. This, however, ushers us into an at present unknown field. SUxMMARY. 1. The ripe eggs of Asterias Forbesii may be made to develop to the bipinnarian or late gastrula stages by mechanical agitation or shock. 2. The amount of agitation necessary varies in different individuals from a hard shaking in a test-tube to. transferring the eggs from one dish to another. 3. The speed of development is in narrow limits roughly propor- tional to the amount of shaking the eggs have received. 4. The parthenogenetic developing eggs have, most of them at least, fertilization membranes and many look exactly like fertilized eggs. 5. The eggs become more sensitive the longer they lie in sea-water up to seven hours. The most favorable time to obtain the largest number of swimming embryos appeared to be about three hours after shedding and with relatively hard shaking. 154 A. P. Mathews. 6. The ease with which development may be started in this way makes this a serious source of error in any study of artificial partheno- genesis by other means. 7. The microscopical changes observed in the eggs consist in the development of fertilization membranes, the dissolution of the nuclear wall, the frequent appearance of many clear areas (asters) in the sub- stance of the egg, and the segmentation of the &^^, often directly, into many cells. THE COMPOSITION OF TENDON MUCOID.^ By W. D. cutter and WILLIAM J. GIES. \Froin the Laboratory of Physiological Chemistry, of Coltitiibia University, at the College of Physicians and Surgeons, N^ew Yor/c] CONTENTS. Page I. Percentage content of sulphur and nitrogen 156 Preparation of fractional products 157 Analytic results 160 II. Complete elementary composition 163 Records of analysis 163 Discussion of results 166 III. Relation to other connective tissue glucoproteids 170 Composition 170 Heat of combustion 171 IV. Summary of conclusions 172 TN their paper on the glucoproteid of white fibrous connective tissue ■*- Chittenden and Gies- stated that the average amount of sulphur in three analyzed preparations of tendon mucoid^ was 2.33 per cent. Loebisch,* who previously had been the only one to analyze this substance completely, found in it an average of but 0.81 per cent of sulphur, and ascribed to it the formula Cij.^,H.,..N32SjOg|, with a molecular weight of 3936. Referring to the unexpectedly high results of their sulphur determinations, as compared with those obtained by Loebisch, Chittenden and Gies wrote : " We present these figures ' Some of the results of this work were reported before the American Physio- logical Society. See the Proceedings, Cutter and Gies : This journal, 1900, iii, p. vi. - CnrrTEXDEN and Gies: Journal of experimental medicine, 1896, i, p. 186. 3 Following Cohnheim's suggestion (Chemie der Eiweis-skorper, 1900, p. 259) we use the term "mucoid," instead of the previously accepted "mucin," to desig- nate this substance. We agree with Cohnheim that, for the sake of definiteness, the term "mucin" may be best applied to the glucoproteids elaborated by true secretory cells, and the term "mucoid" to similar substances in the tissues. In the present unsettled state of our chemical knowledge regarding these bodies, such a distinction is at best of only temporary convenience. The original differences have little importance in the light of the results of recent researches. ^ Loebisch: Zeitschrift fur physiologische Chemie, 1886, x, p. 40. 155 156 W. D. Cutter and William J. Gies. with some doubt in our own minds, but, having obtained them as the result of most careful work, we see no possible explanation other than that this amount of sulphur is actually present in the mucin molecule." ^ The divergent results of these two investigations naturally throw some doubt on the question of the elementary composition of tendon mucoid. We have attempted not only to ascertain definitely the amount of sulphur in tendon mucoid, but also to explain the previous discrepancy in experimental data relating to sulphur content. In addition to the results in this particular connection, certain others of significance obtained by us may be appropriately given with them. I. Content of Sulphur and Nitrogen. Historical. — Rollett ^ was the first to show that tendon contains mucin-like material. He described some of the qualities of the sub- stance, but made no elementary analyses of it. Eichwald'^ merely verified Rollett's qualitative results, in this connection. Loebisch used Rollett's method to prepare sufficient quantities of tendon mucoid for analysis. Only three preparations were analyzed by Loebisch. But one sulphur determination was made on each, with the following results : (a) 0.82 per cent ; (b) 0.80 per cent ; (c) 0.82 per cent. Chittenden and Gies, who were the next to ana- lyze this particular glucoproteid material, used improved methods of preparation and purification and, in sulphur analysis, obtained seven concordant results on three purified products, with the following averages: (a) 2.34 per cent; (b) 2.35 per cent; (c) 2.31 per cent. The difference is very striking. With respect to the amount of nitrogen in tendon mucoid, a sim- ilar though not so decided analytic difference was established in these two investigations. Loebisch made onl)'^ four determinations of nitro- gen in his three purified preparations. The average of two closely agreeing results for his first preparation was 11. 80 per cent ; for the second the single result was 11.84 P^r cent and for the third it was 11.59 Ps^ cent. Chittenden and Gies made ten determinations in three preparations with the following averages of results in close 1 Chittenden and Gies: Loc.cit., p. 197. ^ RoLLETT : Untersuchungen zur Naturlehre des Menschen iind der Thiere (Molescliott), 1859, vi, p. i. Also, Ibid., i860, vii, p. 190. 2 EiCHWALD : Annalen der Chemie und Pliarmacie, 1865, cxxxiv, p. 177. The Composition of Tendon Ahicoid. 157 agreement: (a) 11.94 per cent; (b) 11.80 per cent; (c) 11.51 per cent. They found, further, that the nitrogen content of a series of very carefully prepared fractional products varied between 11.51 per cent and 12.26 per cent, data which seem to suggest, though they do not establish, the existence of several related mucoids as components of ordinary tendinous tissue. Preparation of Fractional Products. — At the outset of these experiments we assumed that tendon contains more than one gluco- proteid. This seemed probable for several reasons. Among the latter is the fact that the larger tendons show considerable variation in texture throughout their length. Thus the tendo Achillis of the ox, from which the previously analyzed tendon mucoids were extracted, is comparatively soft and very tough in the main shaft, but toward its connections with the bones becomes more compact, and outwardly somewhat resembles cartilage. The superficial qualities of the thick sheaths enveloping the two large branches of the Achilles tendon in this animal also resemble those of cartilage to a certain extent. These physical modifications within the tendinous tissue naturally suggest chemical differentiation of the constituents. Previous an- alytic variations respecting tendon mucoid may have been dependent on extraction of different mixtures of distinct though closely related bodies. Loebisch does not state which portions of the tendons were employed in his work. Chittenden and Gies used sections of the main shaft, together with portions of the two branches and the sheaths of the latter. In our own experiments these parts were extracted separately. General Method. — In the preparation of mucoid for use in these experiments the Achilles tendon of the ox was employed. Follow- ing the usual method, the tissue, immediately after removal from the animals, was thoroughly freed of extraneous matter and cut into very thin cross sections. These pieces were washed in water and then extracted in half-saturated calcium hydroxide. The mixtures were shaken at regular intervals. The mucoid was precipitated from the filtered extract with dilute hydrochloric acid.^ The precipitated substance was repeatedl}^ washed; first in dilute hydrochloric acid, to ^ We have always found that mucoids may be precipitated from lime-water or sodium carbonate solution much more satisfactorily with dilute HCI than with any- other acid. The substance seems to separate much more quickly and completely in the presence of slight excess of this acid. Chlorides have comparatively slight solvent action on mucoids in the presence of free HCI, unless admixed in excess. 158 Jt\ D. Cutter and William J. Gies. remove all traces of adherent proteid impurity, then in water until it was free of acid. It was next redissolved in dilute alkali and repre- cipitated once with dilute hydrochloric acid. The washing process was repeated. Finally the acid-free substance was dehydrated and purified by long-continued treatment with large quantities of boiling alcohol-ether; then dried /;/ vacuo and weighed. First Experiment. Series A and B. — In this experiment two parallel series of fractional extractions were made and the mucoid in each separated and ana- lyzed. 4600 gms. of the main shaft of the tendon about five inches in length, with from two to three inches of its bifurcations, were employed in Series A. In Series B only the sheaths of the branches, weighing igoo gms., were used. Both lots of finely divided tissue were given identical treatment at each stage of the experiment. All extractions were made with 2 c.c. of half-saturated lime-water per gm. of moist tissue. After the extracts had been strained through cloth, the tendon pieces were thoroughly washed in water to prevent adherent dissolved mucoid from becoming part of the succeeding extract. The first extracts in each series were readily precipitated and brought to the fliocculent condition with very slight excess of 0.2 per cent hydrochloric acid. Subsequent extracts, however, became only turbid with large excess of 0.2 per cent HCl — even with an equal volume. It was necessary, therefore, to add stronger acid (1.5% HCl) to separate the mucoid in flocks.^ In purifying, the substance was redissolved in half-saturated lime-water. Powdered thymol, used in the second experiment also, entirely prevented bacterial action. The summary, Table I, on page 159, gives additional significant facts relating to these fractional preparations. A striking feature of these preparations was the fact that precipita- tion became more and more difificult with each extraction. More acid was required in each successive extract (except the fourth of Series B) to bring the mucoid to the flocculent condition. It will be seen from the data in Tables I and II that this was entirely independent of the proportion of contained mucoid. The alkali could not have effected decomposition, and thereby possible variations, because it was too 1 In each instance the acid was added slowly in small quantities. The mix- tures were thoroughly stirred and allowed to stand for flocks to form. After waiting a suificient time, more acid was added if separation had not taken place. At first 0.2 per cent HCl was used. If after an equal volume of the acid had been stirred in, flocks failed to separate, 1.5 per cent HCl was added little by little. Separation took place instantly upon reaching the proper amount of acid. On reprecipitating, the same procedure was followed. The proportion of acid required was not recorded in the latter case, but great variations were observed. This method was employed in the second experiment also. The Composition of Tendon Mucoid. 159 weak. Further, the volumes of fluid in each series were kept con- stant and the temperature was always about the same, so that the salts formed on acidification of the alkali of the extracts had essen- TABLE I. Extract. Time of extraction; Amount of HCl present to completely precipitate.^ Weight of puri- fied product.- Number. Volume c.c. Hours. Per cent. Grams. Series A. First Second Tliird Fourth 9200 9200 9200 9200 24 24 24 48 0.04 0.18 0.26 0.32 6.52 9.79 3.55 3.13 Series B. First Second Third Fourth 3800 3800 3800 3800 24 24 24 48 0.03 017 0.46 0.37 4.23 1.65 \ 0.93 1 The figures for per cent of HCl necessarily present to precipitate in flocks express approximate values. The precise amount of acid neutralized by the Ca(OH)., was not directly determined. It was the same of course throughout each series. Greater exactness would have emphasized the facts made significant by the above data. 2 These weights are for substance dried in vacuo. The amount of each prepara- tion could not be exactly quantitative, of course, because of slight losses during their purification. The mucoids are very difficult substances to handle and their preparation is decidedly laborious. Every effort was made to reduce inevitable loss to a minimum, however, and, as the loss was relatively the same in each preparation, the weights are entirely reliable for the intended comparisons. tially the same influence throughout. The extracts were strained quickly at practically the same time and were promptly treated with acid, so that no changes could have occurred by reason of delay in final treatment. i6o JV. D. Cutter and William J. Gies. The figures for weights of substance in each extract suggest varia- ble resistance, on the part of the mucoid, to the solvent action of the dilute alkali. None of the extracts were ever saturated and all were distinctly alkaline. The peculiar behavior of these preparations harmonizes with the view that the tissue contains two or more gluco- proteids, and that the products separated by the usual method of mucoid extraction are mixtures of different bodies. ic) Second Experiment. Series C and D. — A second set of preparations was made in essentially the same way as in the first experiment. 6600 gms. of the main shaft of the tendon and its branches, of the same size as heretofore, were extracted in Series C ; 4200 gms. of sheath in Series D. The periods of extraction were shorter at the beginning and longer at the close of this experiment than previously. In purifying, the substance was redissolved in 0.5 per cent sodium carbonate. The summary of results given in Table II, page 161, connected with preparation, is directly comparable with Table I. In this experiment, also, successive increase in proportion of acid was necessary for precipitation, the results harmonizing in detail with those of the first experiment. Variations in the quantities of separ- ated mucoid again pointed to variable resistance to the action of the extractive. Fractions of a single substance would hardly act so differently at successive intervals under essentially the same con- ditions. Analytic results. — Although the differences in the action of our several products indicated the existence of two or more mucoids in tendinous tissue, more direct-evidence than qualitative variation was necessary to justify such a conclusion. We very carefully analyzed these products, with results that confirm the original deduction. The amounts of nitrogen and sulphur in mucoids furnish excellent data for general comparisons of composition. Table III, on page 162, summarizes our results for percentage content of nitrogen and sulphur in the ash-free substance dried at 105-110° C. to constant weight.^ The analyses were made by the customary methods — Kjeldahl for the nitrogen; fusion with NaOH over alcohol flame, and precipitation with BaCl^, for sulphur. ^ The proportion of ash in these preparations was usually much less than i per cent. In only four was it more than that, and in none of these did it exceed 1.78 per cent. It consisted mostly of phosphate and chloride ; only a trace of sulphate was present. The Composition of Tendon Mucoid. i6i These results seem to prove that more than one substance has been extracted — that mixtures have been obtained. The results for every member of each series differ decidedly in one respect or TABLE II. Extract. Time of extraction. Amourit of HCl present to completely precipitate.! Weight of puri- fied product. 1 Number. Volume cc. Hours. Per cent. Grams. Series C. First Second Third Fourth Fifth 2 13200 13200 13200 13200 13200 17 20 26 30 65 0.03 0.15 0.17 0.38 0.45 14.56 24.88 17.26 2.04 4.09 Series D. First Second Third Fourth Fifth 8400 8400 8400 8400 8400 17 20 26 30 65 0.02 0.15 0.45 0.39 0.35 11.85 13.41 3.19 0.29 0.59 1 See notes under Table I. - A sixth extraction lasting 124 hours was made in Series C. A trifle more than a gram of unpurified substance was obtained. The presence of nearly 1 per cent of HCl was necessary in order to bring it to the flocculent condition. This substance was true mucoid — on decomposition it yielded a reducing substance. It is evident from these results that it is verj- difficult to completely extract glucoproteid from tendinous tissue. another from the rest in the group, and this, too, in spite of the fact that the analyses of all were conducted by uniform methods and under conditions as nearly the same as it is possible to attain. The extremes in percentage content are too far apart to be due to un- avoidable analytic errors. l62 W. D. Cutter and Willia7n J. Gies, -1- CO U c CO ro CO (U p CM CM CM ,—1 ^ r^ LO o ■^ 0.-I CO ON t-co CM u tl> 'fl ^'^a^ CM ro •^ CM ^ ^ C cvi (si CM CM c^i CM o tn OS ^ 1) 3 rt C/3 ^ CN 1— 1 '^ £ « ^ VO VO C^ ^ 53 to C^ CM ro 1—1 ON ro 7) o » p o ^ -1- -t- COO CnC.D t^ OOCs p ir-. .,_, d) CM CM (Si CM ro ro -t- ^ 3 s 1— ( .— I 1—1 ' ' 3 O S '^.. OS C D 3 Cv] O] 1-.CO coc« LO -t-"^ 4= u C/2 CM CM 1— i 1— i i-i r-i ^ ^r-( (U tl H Ifl 1) CO O) VO (S> O) y5 5 o cvi •>- o ON .2 Qn \r, O l^o Ti- O ■+ COVO CM LO ro d o 2 ^ LT. 1-7 t^ I- r^ t^ Lo ONO oc> 's 13 ^ 1 a! "53 s O ,, OS ^-. (^ .« m o OS 0.2831 gm. substance gave 0.0351 gm. N = 12.41 per cent N. Total Stilphur. 0.7412 gm. substance gave 0.0982 gm. BaS04 = 1.82 per cent S; 0.6574 gm. substance gave 0.0887 S™- BaS04 = 1.85 per cent S. Sulphur covibined as SO^. 0.6686 gm. substance, after boiling in HCl, gave 0.0653 gm. BaSOi = 1.34 per cent S. Ash. 0.1720 gm. substance gave 0,0018 gm. Ash = 1.04 per cent Ash. Percentage Composition of the Ash-free Substance. c 47.62 47.98 H 6.66 6.53 N S Average. 47.80 6.60 12.74 12.70 12.54 12.66 1.84 1.87 1.85 31.09 No. 4. Mucoid of first extract of Series D. Carbofi atid Hydrogen. 0.0770 gm. substance gave 0.1372 gm. CO.^ and 0.0480 gm. H.2O = 48.60 per cent C and 6.93 per cent H ; 0.0968 gm. substance gave 0.1721 gm. CO.j and 0.0578 gm. H.,0 — 48.48 per cent C and 6.63 per cent H. Aitrogen. 0.3946 gm. substance gave 0.0495 S^^^- ^ — ^^.55 per cent N ; 0.3154 gm. substance gave 0.0396 gm. N = 12.55 per cent N. Sulphur. 0.5967 gm. substance gave 0.1159 gm. BaSOi ^ 2.68 per cent S ; 0.7591 gm. substance gave 0.1603 gm. BaSOi = 2.89 per cent S. Sulphur combined as SO3. 0.8904 gm. substance, after boihng in HCl, gave 0.0886 gm. BaS04 = 1.36 per cent S. As/i. 0.1983 gm. substance gave 0.0015 gm. Ash = 0.75 per cent Ash. Percentage Composition of the Ash-fkee Substance. Average. C 48.97 48.87 H 6.98 6.68 N 12.64 12.64 O 48.92 6.83 12.64 70 2.91 2.80 28.81 No. 5. Mucoid of second extract of Series D. Carbon and Hydrogen. 0.1779 a^- substance gave 0.31 01 gm. CO2 and 0.1028 gm. H2O = 47.54 per cent C and 6.42 per cent H ; 0.0608 gm. substance gave 0.1066 gm. CO., and 0.0365 gm. H.2O = 47.82 per cent C and 6.69 per cent H. 1 66 PV. D. Cutter and William J. Gies. Nitrogen. 0.3046 gm. substance gave 0.0380 gm. N = 12.48 per cent N; 0.2545 gm. substance gave 0.0316 gm. N = 12.45 P^^' ^^"^ ^• Sulphur. 0.7143 gm. substance gave 0.1226 gm. BaS04 = 2.35 per cent S ; 0.9841 gm. substance gave 0.1608 gm. BaS04 = 2.24 per cent S. Sulphur combined as SO3. 0.7130 gm. substance, after boiling in HCl, gave 0.0805 g™- BaS04 =1.55 per cent S. Ash. 0.3477 gm. substance gave 0.0059 gm. Ash = 1.69 per cent Ash; 0.1665 §'^^- substance gave 0.0031 gm. Ash = 1.86 per cent Ash. Pkrcentage Composition of the Ash-free Substance. Average. C 48.40 48.67 48.54 H 6.54 6.81 6.68 N 12.70 12.68 12 69 S 2.39 2.28 2.34 O 29.75 Discussion of results. — The general summary of our results for complete elementary composition, Table IV, may be compared with similar data obtained in the previous investigations. It will be ob- served that although there is some variation within each series — very slight in Loebisch's, quite marked in our own — the three group aver- ages are very nearly the same. This is particularly significant in this connection. It suggests that mixtures of generally uniform composi- tion resulted in each of the previous studies. Leobisch varied his method very little and obtained practically uniform products ; Chitten- den and Gies varied theirs more decidedly, and the result was distinct variation in composition of substance extracted. By the fractional method in our own experiments, still greater differentiation was effected. We do not mean to suggest that our own products are chemical individuals. They are mixtures, just as all the previously described tendon mucoids have doubtless been. Further research, with more elaborate methods, and particularly with reference to inner groupings of the elements, will be necessary for definite differentiation, if such is possible while we remain in our present profound Ignorance of the structure and peculiarities of proteid molecules.^ The amounts of nitrogen in our preparations appear to be slightly greater than those previously determined, although the nitrogen con- 1 Hawk and Gies : Loc. cit., p. 414 et scq. The Composiiio7i of Tendon Mncoid. 167 tent of preparation No. 2 (Second extract, Series C), which was the largest in quantity of all our products, ^ conforms closely with the generally accepted figures for content of this element. The only particularly discordant results in the general averages are those for content of sulphur and oxygen (by difference) obtained by Loebisch. We had hoped that this low figure would be explained by our results, but none of our products contained so little sulphur. Our figures in this connection accord very well with those given by Chit- tenden and Gies. As has already been stated, Loebisch made only a few analyses — only one determination of sulphur in each of his three preparations. He duplicated results in only half of the analyses he reported. In referring to the differences in composition observed amon-g their products, Chittenden and Gies stated: "Our results seemingly jus- tify the assumption that white fibrous connective tissue contains more than one mucin, or else that the mucin obtainable from this tissue is prone to carry with it a certain amount of some other form of proteid matter which the ordinary methods of purification are not wholly adequate to remove. . . . There is at the present time no standard of purity with regard to this body, and it is quite as probable that fibrous connective tissue contains two or more mucins as that there is only one mucin in the tissue, and that any devia- tion from the figures obtained by Loebisch or by us in preparation No. 3 is due to the presence of a larger or smaller amount of proteid impurity."'^ We can no longer believe that proteid impurity is responsible for the observed variations. In the first place the quantity of soluble proteid in tendon, other than mucoid, is very slight. Experiments in progress in this laboratory indicate that it is less than 0.3 per cent. If, however, it were possible for all of this small quantity to combine permanently with the precipitated mucoids, it could not ac- count for the regular rise and fall of nitrogen content observed in each series of our experiments.'^ Although it is conceivable that the mucoid of the first extract could be so affected, such an assumption would not explain the rise of nitrogen in the third and subsequent extracts, particularly in view of the marked fall of the same in the second. Then, too, each product was so thoroughly washed in excess 1 See table on page i6i. - Chittenden and Gies: Loc. cit., p. 194. 2 See the table on page 162. .68 Ctitter an d Willia^^ / G^^^- The Composition of Tendo7t Mucoid 169 of 0.2 per cent hydrochloric acid, that unless very intimate and unusual chemical union resulted, lymph proteids must have been quickly and completely dissolved from the precipitates. We know of no other substance in tendon which would resist the washing treatment and, by mechanical admixture or chemical combination, account for the orderly variations observed in the analytic series.^ It is much more probable, we think, that an answer to these con- siderations will be found in the fact that the mucoids are labile bodies of great variety in the tissues and with more than one function to perform. Their acid radicles doubtless make them prone to enter into numerous ion combinations. The very complexity of these sub- stances makes it natural to assume that exactly the same proportions of the constituent radicles would in metabolic changes be the excep- tion rather than the rule. All of the products separated in these experiments were true gluco- proteids, responding to each of the well known reactions and yielding reducing substance in abundance. We have repeated the experiments of Chittenden and Gies on the osazone substance obtainable with the reducing body and, working with a larger quantity of mixed mucoid products by the same and also improved methods, obtained a crystalline product melting at 182° C.^ In microscopic appearance the crystals are identical with those of glucosazone. We have not yet been able to free the substance en- tirely from the brownish globules that occur with it and which persist 1 Since this paper went to the printer we have seen Nerkixg's recent note on fat proteid compounds, in the Archiv fiir die gesammte Physiologie, 1901, Ixxxv, p. 330. His results indicate that various proteid products, which have been puri- fied by the usual methods, contain fat or fatty acid in close combination ; further, that this fatty radicle may be broken off, and extracted, by Dormeyer's method. No such combination with ovomucoid was shown, but about three per cent of extractive matter was found to be combined with submaxillarv mucin. Nerkixg does not state, however, that the mucin was thoroughly extracted in hot alcohol ether during the preliminary process of purification, in the customary manner. No results are presented for tendon mucoid : but Loebisch, and Chittenden and Gies have already called attention to the fact that tendon mucoid when freshly precipitated is admixed with extractive matter that is removable only after long continued extraction. All our preparations were given careful and extended treat- ment in boiling alcohol-ether, and we do not believe that the variations in com- position noted are due to such fat combination. We hope that studies which have lately been in progress in this laboratory, will soon furnish direct evidence con- cerning this and related questions. ^ The product obtained by Chittenden and Gie.s melted at 160° C. lyo W. D. Cutter and William J. Gies. in spite of all our attempts to purify the crystals. It seems certain that glycuronic acid and glucosamin, or very closely related bodies, are formed together in the decomposition of tendon mucoid with hot dilute mineral acid. Ill, Relation to other Connective Tissue Glucoproteids. Composition. — It appears to be definitely established by the numerous results of these and the preceding experiments that the amount of sulphur in tendon mucoid is relatively high — almost the same as in chondromucoid and osseomucoid- — and that Loebisch's data in this particular connection can no longer be accepted as correct. We have never been able to prepare a tendon mucoid having less than 1.3 per cent of sulphur.^ The sulphur is present in at least two combinations, as in the case of chondromucoid and osseomucoid. After boiling with alkali, lead sulphide may be obtained on addition of lead acetate. The amount combined in the form of SOa is relatively large, varying as the analytic data for each preparation show, between 1.33 and 1.62 percent of the whole molecule. The average amount of SO3 sulphur in chondro- mucoid is 1.76 per cent. In osseomucoid it equals 1.40 per cent. Levene^ has lately separated from tendon mucoid a substance very similar to chondroitin sulphuric acid. The quantity of this substance separable from the mucoid has not been estimated. Two years ago, in our preliminary report, we made the following statement i"^ "Before these experiments were started, the similarity in the percentage composition of Morner's chondromucoid and the tendon mucin analyzed by Chittenden and Gies suggested to us that the two substances are perhaps closely related. This was further emphasized by the fact that the osazone crystals they obtained had the same general appearance as the crystals of glucosazone, and, therefore, might have arisen from glucosamin, one of the decomposition products of chondromucoid." Levene's results and our own increase the prob- ability that the two substances are very much the same. The following summary of average elementary composition shows the general relationship of very nearly identical products: ^ See table, page 162. ■' Levene : Zeitschrift fiir physiologische Chemie, 1901, xxxi. p. 395. 8 Cutter and Gies : Loc. cit. The Composition of Tendon Mucoid. 171 C H N S O Chondromucoid Morner 47.30 6.42 12.58 2.42 31.28 Tendomucoid {a) Chittenden and Gies 48.76 6.53 11.75 2.33 30.63 (/;) Cutter and Gies (1) . 47.47 6.68 12.58 2.20 31.07 Osseomucoid Hawk and Gies . . 47.07 6.69 11.98 2.41 31.85 Average . . 47.65 6.58 12.22 2.34 31.21 Heat of Combustion. — Heat of combustion furnishes important means of estimating chemical relationships, though its indications are not, perhaps, so delicate as those of elementary analysis. The deter- minations in these experiments were made by the method described by Hawk and Gies. In Table V we give the heat of combustion of our five completely analyzed preparations, together with comparative TABLE V. COMKUSTION EQUIV.\LENTS. Preparation. Direct determinations. Averages for ash-free substance. Heat of combustion. Small calories. Percentage content. Heat of combustion. Small calories. Per gram of substance. 1 Car- bon. Oxy- gen. Per gm. of substance. For sub- stance con- taining Igm. of carbon. I II Average. I. Tendomucoid. No. 1 No. 2 No. 3 No. 4 No. 5 4925 4963 4921 4908 5044 4940 4930 4934 4920 5036 4933 4947 4928 4914 5(H0 47.47 47.46 47.80 48.92 48.54 31.07 32.39 31.09 28.81 29.75 4967 4986 4979 4951 5131 10463 10506 10416 10121 10571 Average. 4952 4952 4952 48.04 30.62 5003 10415 II. Osseomucoid. Average of two preparations. 4972 4985 4979 47.16 31.79 4992 10589 III. Chondromucoid. Average of two preparations. 4865 4869 4867 45.87 32.90 4883 10647 172 W. D. Cutter and William J. Gies. data from the summary in a recent paper from this laboratory .^ The figures show only imperfectly the differences among the tendon mucoids. They are valuable chiefly for the indication they furnish that the various glucoproteid products referred to are essentially the same compounds. We still believe " continued investigation will show that the differ- ences among the mucins, mucoids, and chondroproteids are not as great as their varying physical properties and behavior have suggested, but that each is a combination of proteid with a glucosulphonic acid, the qualities of each compound, just as in the case of the nucleoproteids, being dependent largely on the proportions and character of the pro- teid and compound acid radicles." ^ IV. Summary of Conclusions. The more important conclusions to be drawn from the results of this research are : 1. Tendon contains more than one glucoproteid. The average percentage composition of five preparations of mixed mucoid was as follows : C H N s o 48.04 6.67 12.47 2.20 30.62 These figures agree very closely with those published by Chittenden and Gies. 2. The composition of mucoid from the shaft and from the sheath : C H N S Shaft (3) 47.56 6.61 12.34 1.95 31.52 Sheath (2) 48.73 6.7.S 12.66 2.57 29.28 3. Tendon mucoids contain an average amount of sulphur equal to that found by Chittenden and Gies — approximately 2.30 per cent. Not a single product had the very low content of sulphur ascribed to this substance by Loebisch. 4. The average composition of mucoid separated from white fibrous connective tissue by the customary methods is very nearly the same as that of chondromucoid and osseomucoid. 5. Thermochemical studies of the mucoids in tendon, cartilage, and bone emphasize the probability that these bodies are very intimately related. 1 Hawk and Gies : Loc. ciL, p. 422. ■■^ Cutter and Gies : Loc. cit. PHLORHIZIN DIABETES IN CATS. By JULIUS F. ARTEAGA. [From the Physiological Laboratory of the University and Belleviie Hospital Medical College i\ FORMER investigations in Professor Lusk's laboratory have established the fact that when phlorhizin is frequently given subcutaneoLisly to fasting rabbits there is a preliminary sweeping out of the body's sugar through the urine, followed by a continuous elimination of dextrose and nitrogen in a constant ratio of 2.8 grams of dextrose to i gram of nitrogen.^ This ratio was that found by Minkowski 2 in fasting dogs diabetic from the removal of the pan- creas. Later investigations showed that if dogs were treated with phlorhizin according to the method employed above with rabbits, the sugar excretion became much larger in relation to nitrogen eliminated, constantly averaging 3.75 grams of dextrose to 1 gram of nitrogen.^ This represented a higher production of sugar from proteid metabolism than had ever before been obtained. Still later work done by Professor Lusk and the author ^ showed that after treating goats with phlorhizin the urinary ratio became dextrose : nitrogen = 2.8 : i. Since the carnivorous dog showed a ratio of 3.75 : I, and the herbivorous rabbit and goat one of 2.8 : i, it seemed to me important to carry on similar experiments with another variety of carnivora, and to this end the cat was selected. The disadvantage of using the cat lies in the fact that the catheter cannot be passed through the urethra because of its small size. On account of this the urine of one day was always more or less mixed with that of the preceding days. The results show that the collec- tion of the urine must be continued several days before the ratio becomes fixed and constant. The urine was caught in a tray beneath the cage occupied by the cat. ' Graham Lusk: Zeitschrift fiir Biologie. 1898, xxvii, p. 82. ^ Minkowski: Archiv fur experimentelle Pathologie unci Pharmakologie, 1893, xxxi, p. 85. ^ Reilly, Nolan, and Lusk: This journal, 1898, i. p. 395. * Lusk: Festschrift zu Voit, Zeitschrift fiir Biologie, 1901. (Not yet published.) 173 174 Julius F. Arteaga. In the first experiment o.i gram of phlorhizin was dissolved in a few cubic centimetres of a warmed 1.2 percent NaoCOs solution and injected subcutaneously every eight hours. The phlorhizin was given on March 7, after the cat had fasted four days. The result of the urinary analyses may be tabulated as follows : Cat I. 0.1 gm. phlorhizin ev6ry eiglit hours after March 7. Date, 1900. Amount of urine in c.c. Weight in kg. Dextrose. Nitrogen. D. :N. March 3 .... 3.16 7 SO p ? 8 192 5.68 2.89 1.96 9 231 8..S3 3.28 2.60 10 10.58 2.46 4.30 It is apparent that the small dose of phlorhizin (o.i gram) became progressively more and more effective in its ability to eliminate sugar, since the highest ratio was found on the fourth day. As I wished to have the body's excess of sugar removed as quickly as possible I returned to the dose of one gram every eight hours, which had been successfully employed in the case of the rabbit. The first experiment was with Cat No. I, used above. In Cats II and III the diabetic urine was not analyzed during the first two days. The results obtained are expressed in the tables on page 175. All these cases show in the later days a pronounced tendency to approach the urinary ratio D. : N = 2.8 : i, and this is conclusively demonstrated in the last experiment, where, on the third, fourth, and fifth days the ratios were 2.93 : i, 2.80 : i, and 2.93 : i. The slightly higher ratios of 3.19 : i and 3.07 : i on the last days of experiment with the first two animals is to be explained by the admix- ture of the urine of previous days when the body's sugar was being removed; hence appears the disadvantage of not being able to sepa- rate the urine by catheterization and by washing the bladder. About six hours after the last urine of Cat III was passed the animal died, apparently in diabetic coma. This last urine was observed to yield a sediment upon standing. These crystals were examined by Professor John A. Mandel, and to him I am indebted for the report that they Phlorhizin Diabetes in Cats. 175 yielded a reducing substance on cleavage with acids, and showed the same crystalline form and melting point as phlorhizin. This dis- covery of phlorhizin in cats' urine seems especially remarkable, since Cat I. Fasting since April 7. Date, 1900. Amount of urine in c.c. Weight in kg. Dextrose. Nitrogen. D. :N. April 11 59 4.01 1.36 2.95 12 144 17.04 4.03 4.23 13 163 12.19 3.74 3.26 14 104 2.37 9.22 2.89 3.19 Cat II. Fasting since April 15. April 18 p 4.05 p p 19 p p p 20 178 16.10 5.03 3.20 21 144 3.49 13.26 4.32 3.07 Cat III. Fasting since Oct. 2. Oct. 3 ? 2.34 p p 4 ? p P 5 48 339 1.16 2.93 : 1 6 130 4.06 1.45 2.80 : 1 7 186 1.941 0.66 2.93 : 1 1 12 hours' urine. neither Cremer ^ nor Lusk ^ were ever able to detect phlorhizin in rabbits' urine. This inquiry into the nature of phlorhizin diabetes in cats assumes 1 Cremer: Sitzungsberichte der morphologisch-physiologischen Gesellschaft zu Miinchen, 1895, p. 75. "^ Lusk : Zeitschrift fiir Biologie, 1898, xxvii, p. 98. 1 76 Julius F. Arteaga. added interest when considered in connection with the work of Lee and Harrold ^ which has shown the muscular fatigue in cats' muscle after the readily combustible sugars have been removed by phlorhizin. The principal result of my research is that in the fasting cat, just as in the rabbit and the goat, the urinary ratio between dextrose and nitrogen in phlorhizin diabetes is 2.8 : i. This is a striking example of biological uniformity. 1 Lee and Harrold : Proceedings of the American Physiological Society, This journal, 1900, iv, p. ix. ox THE PRODUCTION OF ARTIFICIAL PARTHENO- GENESIS IN ARBACIA BY THE USE OF SEA-WATER CONCENTRATED BY EVAPORATION. By S. J. HUNTER. \^From the Laboratory of Comparative Zoology a?td Entomology, University of K'ansas.^ IN the parthenogenetic development of the eggs of Avbacia by the Loeb method, the nature of the actions of the solutions used merits consideration. Loeb states that artificial parthenogenesis in Echinoderms is caused by an increase in the osmotic pressure of the solution surrounding the unfertilized eggs.^ Since density can be expressed in terms of specific gravity it is evident that we can ascertain the relative osmotic pressures of the solutions used as well as the degree of pressure required to produce artificial parthenogenesis. The specific gravity^ of sea-water at the Marine Biological Labora- tory, Wood's Holl,^ where this work was carried on was 1.0211. Specific Gravities of Normal Sea-Water and of Sodium Chloride Solutions. Sea-water 1.0211 2| n sodium chloride solution 1.0902 Solution No. 1. — A 10 per cent solution of 2,\ii sodium chloride, com- posed of 25 c.c. 2\ n sodium chloride and 225 c.c. sterilized sea-water 1.0265 Solution No. 2. — A 15 per cent solution of 2\n sodium chloride, com- posed of 37| c.c. 2J « sodium chloride and 212| c.c. sea-water . . 1.0283 All sea-water that came in contact in any way with unfertilized eggs was first sterilized by heating slowly to 65° C, then cooled to ^ Loeb, J. : This journal, 1900, iv, p. 184. - The specific gravity figures given throughout are calculated by Pemberton's tables from readings on Baume's Hydrometer. The temperature of all solutions at the time of reading was 33° C This temperature was used since it was not convenient to reduce to 15.3° C, required by the hydrometer, and since relative densities only were desired. ^ I am indebted to Dr. F. R. Lillie, for the ample facilities afforded and valu- able suggestions given during the progress of this work. 177 178 S. /. Hunter. 22° C, the temperature of sea-water at Wood's Holl, and to replace oxygen which might have been driven off by the heat, the solution was passed through a glass syphon with the longer arm of very small aperture, and allowed to fall in a fine stream through the air for a dis- tance of about four feet into a wide Stender dish. This was done twice with each flask of sterilized sea-water and three times in the case of sea-water concentrated by boiling. It has been shown by Loeb, Wilson, and others that the 10 to 15 per cent 2\n sodium chloride solutions are the best sodium chloride solutions for the production of artificial parthenogenesis in Arbacia. Now it seemed reasonable to suppose that if the density of normal sea-water were increased by evaporation until it stood between the figures given for Solutions i and 2, artificial parthenogenesis should likewise be produced. Accordingly 500 c.c. of normal sea-water were reduced by boiling slowly to 375 c.c, a reduction of 25 per cent in volume. In the same manner another quantity of 500 c.c. of normal sea-water was reduced to 250 c.c, a reduction of 50 per cent, in volume. The specific gravity readings of these two solutions were as follows : — Solution No. '3. — Specific gravity of 500 c.c. sea-water reduced by evaporation to 375 c.c 1.0260 Solution No. 4. — Specific gravity of 500 c.c. sea-water reduced by evaporation to 250 c.c 1 0431 The specific gravity of Solution No. 3 was 0.0005 below that of Solution No. i. This, however, can be accounted for by observa- tional error. Before their use for experimentation, these solutions were cooled to 22° C. and aerated three times. Unfertilized eggs of a single female Arbacia were divided, placed in Solutions 3 and 4, allowed to remain there for two hours, then removed to sea-water. The eggs from Solution 3 segmented. Some of the eggs developed into gastrulae capable of locomotion. No segmentation or perceptible processes of development were observed in eggs taken from Solution 4. These experiments were repeated with two other females and similar results secured. The use of Solu- tion 4 was discontinued. Sea-water that had been reduced 10 per cent in volume,- z. e., 100 c.c. reduced by boiling to 90 c.c, was tried with negative results. The next experiment was with sea-water which had been reduced 30 per cent in volume, i. c, 100 c.c had A rtijicial Parthenogenesis in Arbacia. 179 been boiled down to 70 c.c. In this concentration segmentation and development took place, but not more favorably nor in larger num- bers than in Solution 3. No further experiments with various con- centrations by evaporation were performed. Observations were confined to sea-water reduced in volume 25 per cent. The optimum duration of time for eggs in this solution was found to be two hours and twenty minutes. To ascertain this, eggs were removed to sea- water every thirty minutes until two hours had elapsed, and then every ten minutes for the next thirty minutes. Eggs taken out and placed in sea-water, after being in Solution 3 for one hour, would begin to segment, but would not develop so far as those left in it for two hours and some minutes. The eggs of two females, however, do not always act alike under similar conditions. Hence the opti- mum length of time may vary slightly with the eggs of different females. In order that there may be no question concerning the contami- nation of eggs with spermatozoa the method of operation may be briefly given. All dishes were thoroughly washed with hydrant water, inverted and allowed to dry. All instruments were heated in the flame of an alcohol lamp, then placed in a dish of hydrant water, just before using. Through the kindness of the Curator, Mr. Gray, the females were selected and no males were brought into the laboratory. The female from which the eggs were to be taken was washed for about three minutes under a stream of hydrant water, laid on a sterilized vessel until the hands of the operator had been thoroughly washed with soap and hydrant water, then the sea urchin was washed under a stream of hydrant water for about three minutes longer. If the female had just been brought in from the ocean,^ and was ready to spawn, oviposition was easily brought about by pouring tepid sea-water over the ventral surface. Sometimes oviposition was incited by the hydrant water, which was somewhat warmer than the sea-water. Some of the eggs were kept in steri- lized sea-water to serve as a proof of absence of spermatozoa. Ten experiments with eggs of as many females were performed to observe the relative effects of Solutions 2 and 3. In each experi- ment the eggs of one female were separated into three lots ; one in sea-water for control, one in Solution 2, and the third in Solution 3. With one exception the eggs in Solution 3 developed more rapidly ^ The majority of those selected for experimentation were brought from the ocean the same da}' their eggs were used. i8o ^. J. Hunter. and showed a larger percentage of larvas (locomotive forms) than appeared among the eggs developed in Solution 2. Since my work was concerned primarily with the early cleavage stages of these eggs, I paid little attention to the length of life in those taken from Solution 3. In one case, however, I observed plutei six days old that had been developed in Solution 3. In all, the eggs of fourteen females were placed in Solution 3 and in eleven cases satisfactory results were secured. In some of the experiments as many as 90 per cent, approximately, of the eggs began to segment. Comparatively few, however, reached the freely moving gastrula stage. In one instance I estimated that the num- ber of moving gastrula approached 40 per cent of the eggs in the solution. Were I to -prepare cultures to continue the life of the individuals, I should use fewer eggs and greater amounts of sea- water. As it was, a large number of eggs in early cleavage stages was desired. This same solution of sea-water reduced in volume 25 per cent by evaporation, was used this season by the class in Embryology in the Laboratory and a good percentage of larvae obtained. These experiments with sea-water concentrated by evaporation, then, tend to strengthen Loeb's osmotic theory of artificial partheno- genesis in Arbacia. Sea-water that is condensed until it is isotonic with Loeb's 10 per cent to 15 per cent 2\n sodium chloride solu- tions will cause artificial parthenogenesis.^ Sea-water with osmotic pressure perceptibly less or greater than the 10 to 15 per cent solu- tion of 2% n sodium chloride will not produce artificial partheno- genesis. Furthermore, it is evident that a certain osmotic index or degree of pressure is essential for artificial parthenogenesis. ^ Solution 2, the 15 per cent solution of i\ii sodium chloride, gives more satis- factory results than Solution i, the 10 per cent i\n sodium chloride solution. Solution 2 has greater osmotic pressure than Solution i. while Solution 3, the sea- water reduced in volume 25 per cent by evaporation, which gives even more satis- factory results than Solution 2, is isotonic with Solution i. Possibly the optimum duration of time for eggs in Solution 2 is less tlian the optimum duration of time for eggs in Solution 3. AN ANALYSIS OF THE INFLUENCE OF THE SODIUM, POTASSIUM, AND CALCIUM SALTS OF THE BLOOD ON THE AUTOMATIC CONTRACTIONS OF HEART- MUSCLE. By W. H. HOWELL. l^From the F/iysiological Laboratory of the Johns Hopkhis University. '[ THE relation of the phenomenon of rhythmic contractility to the inorganic constituents of the lymph, blood, or other medium bathing the tissue has attracted much attention in recent years. The relation has been studied most carefully in connection with the typically rhythmic heart-muscle, but investigations have been ex- tended to other tissues exhibiting the same property, among the invertebrates as well as the vertebrates. Historically the pioneer work in this field was accomplished under the influence of Ludwig. Systematic investigations by Merunowicz, Kronecker, Stienon, Gaule, Martins, von Ott, and others^ emphasized in a striking way the importance of the inorganic constituents in all media designed to maintain the rhythmic contractility of the isolated frog's heart. The earlier observers in studying this <|uestion seem to have assumed that the influence of the inorganic constituents upon the isolated heart is secondary. The essential or an essential factor was the presence in the circulating medium of some albuminous material. This point of view was unfortunately over-emphasized by Kronecker and his pupils. Most of the work of this school was directed to prove that a circulating medium cannot sustain rhythmic contractility in heart-muscle unless it contains serum albumin, and that the heart-muscle, as regards its rhythmic contractions, depends upon an ever present supply of this material in the surrounding liquid, having no store of available energy within its own substance. This view of the immediate necessity of serum albumin in the bathing liquid has, I believe, been fully demonstrated to be erroneous by the investigations of the author and his pupils. All students of this problem at present will be forced to admit that the contractions ^ For references see Howell. This journal, 1S99, ii, p. 81. 181 i82 W. H. Howell. of heart-muscle and other rhythmically acting tissues may be maintained for very long periods in solutions containing only inorganic constituents. Merunowicz analyzed very successfully the influence of the various inorganic constituents and stopped short of reaching the present standpoint only because he was unaware of the essential importance of calcium salts. The credit of first recognizing the peculiar importance of calcium is due to Ringer. This author showed so clearly the especial part taken by Ca and K salts in maintaining the rhythmic contractility of heart-muscle that subsequent investigators have of necessity followed the lines of work as laid down by him.' The extraordinary efficacy of the so-called Ringer's mixture of sodium, potassium, and calcium salts in proper proportions in maintaining the contractions of heart-muscle is now a familiar fact. Recent studies have been directed largely to an attempt to discover the nature of their action and the individual influence of each of the elements in question. With regard to the nature of their action only vague hypotheses have been offered. Loeb ^ has imagined the existence of compounds of the metallic ions with the proteids, but, although there is every reason to believe that the proteids may form dissociable compounds with either the kation or the anion or possibly with both (Pauli), the existence of such compounds has not been demonstrated experimentally, and, even granting their existence, no clearer idea is obtained of the connection between the inorganic salts and the periodic katabolisms of rhythmically contracting muscle. If we confine ourselves to experimental facts, it would seem that in this matter of the relation of the inorganic salts to rhythmic contractions we are not justified in going beyond the conservative statement made by Merunowicz — namely, that for the heart contractions both organic and mineral constituents are necessary, the former supplying compounds from which energy is obtained, the latter giving to these compounds such a form that they can be used in the production of contractions. The particular influence exercised by each of the three salts present in a Ringer's mixture, as well as the effect of other inorganic compounds, may, however, be investigated by the methods available at present, and in this respect much has been accomplished in recent years. The present paper is a report of a series of experiments made ^ Loeb: Arcliiv fiir die gesammte Physiologic, 1900, Ixxx, p. 229, and Fick's Festschrift, Braunschweig, 1900. Atitomatic Co7itr actions of Heart- Muscle. 183 with this end in view and is a continuation of previous work of a similar character. The experiments reported were all made upon strips of the ventricular muscle of the terrapin. A longitudinal strip was prepared from the free edge of the ventricle in its lower two thirds, and this strip was then bisected in its long diameter so as to give two strips as nearly similar as possible for comparative experiments. The strips were attached to levers which traced the records of the contractions on slowly revolving drums that made one revolution in from one to twelve hours. The apparatus was so arranged that the strips could be immersed conveniently in baths of liquids of different compositions. It should be borne in mind, as has been shown by work from this laboratory, that strips from this portion of the ventricle do not, as a rule beat spontaneously when freshly taken from the heart and im.mersed in the animal's own serum or in a Ringer's mixture containing Ca and K in the proportions found in blood. Under what might be called normal conditions they are not automatically contractile. By modifying the proportions of the salts in various ways, or by first altering the conditions within the strip itself, a prolonged series of spontaneous contractions may be obtained. As opposed to this behavior of strips from the apex of the ventricle, strips from the venous end of the heart (vena cava) beat with perfect regularity in serum or an equivalent Ringer's mixture. Action of Apical Strips of Ventricle in Solutions of NaCl. NaCl exhaustion. — Ventricular strips placed in an isotonic or ap- proximately isotonic solution of NaCl give a characteristic series of beats that has been described in detail by Greene.^ I wish to empha- size here only certain peculiarities of this series that seem to be of fundamental significance. In the first place, if the strip has been freshly prepared there is always a certain latent period before the contractions begin. The length of this latent period is very variable. In thirty experiments made during the present series in which this factor was observed the period varied from ten to one hundred and eighty minutes, but the average was fifty-six minutes, which corre- sponds fairly well with the interval usually to be expected. In the second place the series of beats begins with contractions which compared with later ones are submaximal and infrequent. The 1 Greene: This journal, 1899, "> P- ^-- 1 84 W.H.Howell. contractions then increase quickly or slowly to a maximum in rate and amplitude showing often a beautiful " staircase." Afterward the contractions decline gradually to zero, or in some cases after reaching a minute size become somewhat infrequent without decreasing in height for a considerable time. The duration of the entire series of beats varied in the present experiments between fifty and two hundred minutes, with an average of about one hundred and twenty minutes. The characteristic feature after the maximum is passed is the steady decline to a minimum or to zero. The contractions are as a rule well coordinated. The so-called exhaustion is due to changes in the strip since in all cases the bathing solution was relatively large in quantity (25 c.c.) and renewal of the solution at the end of the series was not followed by a revival of the contractions. In the third place the strip shows a continuous loss of tone during the latent period and throughout the whole series of beats. This loss of tone in NaCl is one of the most distinct effects of this solution and will be referred to again in connection with an attempt to analyze the effect of sodium salts. With reference to this limited series of contractions in NaCl the questions that naturally arise are : Why do the ventricular strips beat in NaCl and not in serum or an equivalent Ringer's mixture? What is the meaning of the long latent period and why is the series a descending one, ending finally in the state of NaCl exhaustion? A number of facts that have a bearing upon these questions will be described and discussed later. At present, for the sake of clearness it may be said that this series of experiments has led the author to adopt provisionally at least the following hypotheses, i. That the apex of the ventricle in the terrapin beats for a time in pure solutions of NaCl, owing to the gradual removal of the potassium which under these conditions diffuses out from the heart tissue; whereas in nor- mal serum or an equivalent Ringer's mixture the proportion of potas- sium present is sufficient to prevent, or if one prefers the expression, to inhibit, spontaneous contractions. 2. That the latent period is an expression of the time necessary for the removal of the potassium to the point at which its inhibitory influence ceases to be felt. Inhibition from mechanical injury probably influences also the duration of this period in excised strips. 3. That the state of NaCl exhaustion is due mainly to the loss of Ca from the tissue as the result of diffusion. Recovery of contractions after NaCl exJianstion. — After the ven- tricular strips have ceased to beat in isotonic solutions of NaCl they Automatic Contractions of Heart-Mtiscle. 185 are totally insensitive to mechanical or electrical stimulation. They can be made to beat again, however, for a longer or shorter time by altering the composition of the surrounding medium in one of the following ways. 1. Immersion of the strip in a proper Ringer s mixture of K,Na,and Ca salts. — The beats thus produced begin in a few minutes and may last for very long periods, for twenty-four hours or more, without any change of solution. When the potassium and the calcium are present as chlorides in the quantities found in serum, that is, 0.03 and 0.026 per cent respectively, the beats are characterized by their long con- tinuance and by their irregular grouping. A perfectly regular rhythm such as occurs in the normal heart-beat is not obtained. A pause in diastole is usually quite distinct and of variable duration, while at ir- regular intervals the beats fall into groups having a more rapid rhythm. 2 . Immersion of the strip in a proper in ixtiire of Na Clo.J per cent and CaC/^'—ThQ recovery caused by this mixture in which no potassium is present varies in character with the amount of Ca used. If present in amounts equal to that found in blood, the revival of contractions may be of short duration and accompanied by a strong development of tone. If, however, the amount of Ca is smaller (o.ot per cent CaClg) and is subsequently increased, a revival may be obtained for somewhat long periods. It happens that in the present series of ex- periments the action of the mixture of NaCl and CaCU was tested usually when the strip, after exhaustion in NaCl, had been immersed for an hour or more in an isotonic solution of sugar. Under these conditions it may be assumed that the strip had lost most of its dif- fusible Na as well as K and Ca, and when placed in mixtures of NaCl 0.7 per cent and CaCl2 0.0 1 to 0.02 per cent, a series of contractions was obtained lasting for many hours. The characteristic of the revival in NaCl and CaCl2 as contrasted with that occurring in a Ringer's mixture was the perfectly regular rhythm ; systole and diastole fol- lowed with perfect regularity and with no perceptible diastolic pause and no grouping of beats. 3. I mmersion in isotonic or approximately isotonic solutions of cane sugar or dextrose. — If the ventricular strip after coming to rest in a solution of NaCl is transferred soon to an isotonic solution of cane sugar or dextrose, it gives always a more or less definite series of beats lasting for a short time. In fifteen experiments of this charac- ter the average duration of the series was fifty-five minutes, the ex- tremes lying between thirty and ninety minutes. In this series of 1 86 W. H. Howell. beats the contractions are usually small at first and slow, but both rate and amplitude increase somewhat rapidly to a maximum. After passing the maximum the beats again decrease somewhat in size and become slower, the series ending as a rule with several infrequent beats of good amplitude, instead of dying out gradually as in the NaCl series. Soon after the last spontaneous beat the strip becomes entirely unirritable to mechanical or electrical stimuli. When a strip is brought into this condition by exhaustion first in NaCl and then in sugar and is deprived apparently of all its diffusible ions of Ca,K, and Na, it seems to be in the best condition for a long-lasting revival with Ringer's mixture or with an appropriate mixture of NaCl and CaClg. 4. Innnersion in isotonic solutions of LiCl. — The series of beats in this solution is usually shorter than in sugar solutions, but presents the same general characteristics. During the series the tone may increase somewhat, although usually not so much as in the sugar solution. After the cessation of the LiCl series an excellent and long-lasting revival of contractions may be obtained with a Ringer's mixture, while an isotonic solution of NaCl alone is usually entirely ineffective. Some further reference to the action of solutions of LiCl will be made below. Action of Apical Strips of Ventricle in Solutions OF NaCl and KCl. As Greene has previously stated, fresh strips of the ventricle placed in NaCl 0.7 per cent to which KCl has been added in normal amount, 0.03 per cent, usually remain perfectly quiet except for a marked loss of tone. In some cases a series of beats may occur, but they are of smaller amplitude than the usual NaCl series and the whole series is of briefer duration. In such cases, increasing some- what the amount of KCl suffices to prevent spontaneous contractions. It is worth noting that when the mixture contains KCl in amounts not much exceeding 0.03 per cent the effect upon the heart muscle is not injurious. Subsequent immersion in Ringer's mixtures gives an excellent series of beats. Action of Apical Strips of Ventricle in Solutions OF NaCl and CaCl^. In marked contrast to the action of NaCl and KCl, a mixture of NaCl and CaCl.2 may cause a series of vigorous contractions. The Automatic Contractio7is of Heart- Muscle. 187 duration and characteristics of this series depend largely upon the amount of CaCl2 used. The general action of such mixtures has been described by Greene. I have only to add that in the present series of experiments two points were particularly noted. First, contrary to the statement made by Greene, no distinct shortening of the latent period was observed as compared with the action of solutions of NaCl. It should be said, however, that the CaCl2 was never used above the concentration found in the blood, namely, 0.026 per cent; usually, in fact, the solutions were made up to half this strength. Moreover, the latent period is so variable even in companion strips from the same heart that only a long series of observations would be conclusive. Second, if the proportion of CaCl 2 is kept below 0.026 per cent the series of contractions may continue for four or five hours or more, the size of the beats gradually diminishing in amplitude. The series of beats obtained differs from the series in NaCl alone in two partic- ulars ; first, in its greater duration when the proportion of CaClg is not excessive; second, in the fact that the condition of tone in the strip follows a different course. There is first a loss of tone, as in the NaCl alone, but subsequently the tone increases, the rise of the curve being more or less rapid according to the amount of Ca used. If the solution is changed to a standard Ringer's mixture before this rise in tone has become too marked, a typical long-continued series of beats may be obtained. If, however, the rise in tone has been rapid be- cause of the amount of Ca used, or if it has been allowed to continue too long, the subsequent use of Ringer's mixture or solutions of NaCl may be without effect. It would appear that under these conditions the muscle goes into a condition which may be designated as Calcium rigor, from which recovery is impossible. Action of Apical Strips of the Ventricle in Isotonic Solutions of Sugar. Numerous experiments under varying conditions were made upon the influence of isotonic solutions of cane sugar or dextrose. As the results obtained will be used frequently in the subsequent discussion of the effect of the different salts, they may be briefly summarized here. i. Immersion of a fresh strip in an isotonic solution of sugar gives as a rule no contractions whatever. The strip shows only cer- tain changes in tone which follow usually a definite course. The 1 88 W. H.Hoiucll. curve at first falls for an hour or less, owing to loss of tone. This first fall in tone, seen also in solutions of LiCl, may probably be explained by the fact that the handling of the strip during its prep- aration and attachment to the recording lever throws it into a condition of excessive tone from operative violence. From this con- dition the strip slowly relaxes in the more or less neutral solutions in which it is placed. After this initial fall the curve slowly rises again because of a gradually increasing development of tone. A similar curve is obtained with isotonic solutions of LiCl. The rise of tone is due doubtless to the loss of the diffusible ions in the sub- stance of the strip, particularly those of Na. 2. If one adds to the sugar solution salts of K, Ca, and Na, in amounts equal to those found in the blood or present in Ringer's mixtures, either singly or together the efifect is entirely negative so far as spontaneous contractions are concerned. If Ca salts alone are added, especially after the immersion in sugar has continued for some time, the strip shows a tendency to go into excessive tone. If KCl is added alone in small quantities (0.03 per cent) it either exerts no distinct effect, or, like the Ca, causes an increase in tone. If NaCl is added to the sugar solution in amounts equal to 0.7 per cent, giving therefore a hypertonic solution of double the normal osmotic pressure of serum, a loss of tone occurs precisely as when the strip is placed in 0.7 per cent NaCl alone, but no spontaneous contractions result. Such hypertonic solutions are not distinctly injurious to the muscle, since, after an immersion of from three to four hours, transference to a Ringer's mixture is followed by spontaneous beats within a few minutes. If CaClg to the amount of 0.026 per cent is added to the sugar solution in addition to the NaCl the effect is the same in so far as spontaneous contractions are concerned. The fact that fresh strips of the ventricle remain quiet in sugar solution plus NaCl to 0.7 per cent or NaCl and CaCl^, but give a definite series of automatic contractions in NaCl alone or NaCl plus CaCl^ indicates that the sugar in the strengths used (nearly 8 per cent cane sugar) is not entirely indifferent but exerts an inhibitory influence. That the quiescence is not due simply to the high osmotic pressure of the solution is indi- cated by the action of strips in hypertonic solutions of NaCl. 3. The after-effects of sugar solutions. — It is noteworthy that after prolonged immersion in isotonic solutions of sugar transferal to 0.7 per cent solution of NaCl is not followed by the usual series of contractions. A loss of tone results as usual with this solution but Automatic Contractions of Heart-Muscle. 189 no spontaneous contractions are seen. If, however, the strip is transferred from the sugar solution to a Ringer's mixture or a mixture of NaCl and CaClg an excellent series of contractions is obtained. That prolonged immersion in sugar solutions is not injurious to the tissue was well illustrated in one experiment in which the strip had remained for twelve hours in a mixture consisting of sugar 8 percent, 27.5 c.c. and NaCl 0.7 per cent, 2.5 c.c, without giving a single contraction. When transferred to a standard Ringer's mixture a series of beats was obtained lasting over twelve hours. 4. TJie action of isotonic solutions of sugar after exhanstion in NaCl o.y per cent. — After a heart-strip has given its typical series of contractions in NaCl and the stage of so-called exhaustion is reached immersion in isotonic solutions of sugar gives usually a more or less definite series of contractions lasting for about one hour. In some cases the series was short and the contractions small. Generally, however, as previously stated, the beats increase somewhat rapidly to a maximum in amplitude and rate and then fall off a little in ampli- tude as the tone increases, the series ending usually with several separated contractions of nearly maximal amplitude. The precise form of the curve made by the series of beats varied somewhat with each strip. In these experiments we have examples of a series of spontaneous beats occurring in a medium that contains no Na, Ca, or K ions. It must be remembered, however, that these substances, although not present in the outside medium, may be contained in the tissue itself. In Lingle's^ paper the author criticises a statement made by Greene to the effect that fresh heart strips give sometimes a short irregular series of beats in isotonic solutions of dextrose. Lingle explains this result on the gratuitous assumption that Greene's dextrose solutions contained some sodium salt, since otherwise " a rhythm was established in a solution of a non-electrolyte, a fact that directly contradicts Loeb's idea." In Greene's experiments the dextrose was prepared carefully from cane sugar according to the method given in Beilstein, and although it may have contained sodium in traces sufficient to give the flame test it certainly did not contain this substance in quantity. The weakness of Lingle's criticism is evident from the fact that addition of NaCl to a sugar solution, to the extent of 0.7 per cent, does not confer upon it the property of causing a strip 1 LixGLE : This journal, 1901, iv, p. 265. 190 IV. H. Howell. to give spontaneous contractions. And even when the sugar solution is diluted with 0.7 per cent solution of NaCl the addition of as much as \ of its volume of the saline solution does not, as Lingle himself is at pains to demonstrate, bring about spontaneous contractions. The experiments described give clear proof of the possibility of a series of automatic beats in a solution entirely free from electrolytes ; but that the beats thus produced are probably dependent upon the presence of definite electrolytes in the tissue itself is made probable by the following considerations. First, the definite limitation of the series to about one hour, a fact to be explained possibly by the diffusion of electrolytes out of the tissue. Second, if fresh strips are placed first for from one hour and a half to two hours in a mixture of NaCl and sodium oxalate, the latter being present to the strength of 0.3 to 0.4 per cent and the total osmotic pressure being equal to that for 0.7 per cent solution of NaCl, subsequent immersion in a sugar solution fails to give the series of beats obtained after the action of NaCl alone. The result in such cases is either entire absence of contractions or a feeble series of minute beats, and it is explicable upon the hypothesis that ordinarily after immersion in NaCl for from two to four hours some dissociable Ca compound is still present in the tissue, whereas in the case of the solution containing oxalate this Ca is made insoluble. We may suppose that after the usual completion of the NaCl series of beats, which follows upon an immersion in NaCl of three or four hours, some compounds of Ca and possibly K are still left in the strip since the diffusion out of these substances is slow in fresh strips, and that the series of beats in the sugar solutions continues only so long as they are present. After the cessation of the sugar series the tissue may be considered as practically deprived of electrolytes and in this condition cannot be made to give spontaneous contractions by solutions containing NaCl alone, CaCl2 alone, or KCl alone, but will give a capital series of beats in a mixture of the three salts or in a mixture in proper proportions of only the NaCl and the CaCI.,. 5. TJic action of isotonic solutions of sugar plus CaCl^ or KCl after previous exhaustion in NaCl o.y per cent. — After exhaustion in NaCl, immersion in sugar containing Ca in doses of from 0.004 to 0.026 per cent is followed by a series of very vigorous contractions superposed on a rapidly rising tone-curve. The contractions are sharp and vigorous in proportion, roughly speaking, to the amount of Ca in the solution, but the duration of the series may be very brief as the strip Automatic Contractions of Heart-Mnscle. 191 soon goes into Ca rigor. Addition of K salts to any amount will not neutralize this effect of the Ca, indeed the K salts alone under these conditions seem often to lead to an increase in tone. Action of Apical Strips of the Ventricle in Solutions of LiCl. In solutions of LiCl isotonic with 0.7 per cent NaCl the ventricular strips act much as they do in sugar solutions. Distinct tone changes, but no contractions, occur. With regard to the changes in tone there is first the customary relaxation extending over an hour or more and then a slow shortening, approximately to the original base line. Addition of CaCl2 alone to the solution or in combination with KCl in any proportion fails to evoke automatic contractions, but tends to throw the strip into augmented tone. If the strip is allowed to remain in the LiCl for a long time, for three or four hours, — subsequent immersion in 0.7 per cent NaCl causes no contractions. If, however, the strip under these conditions is changed to a Ringer's mixture, recovery may take place and an excellent series of beats be obtained. Effect of solutions of LiCl after exJiaiistion in NaCl. After a strip has ceased to beat in 0.7 per cent NaCl, immersion in LiCl gives a short series of beats the general characteristics of which have already been described. After the cessation of this series, addition of CaCl2 alone or in conjunction with KCl has no effect other than to cause a rapid development of tone. If changed to a 0.7 per cent solution of NaCl the result is a pronounced decrease in tone without any contrac- tions. If, however, the strip is transferred to a Ringer's mixture or to an appropriate mixture of NaCl 0.7 per cent and CaCl2 (o.oi to 0.02 per cent) a good series of beats may be obtained. These reactions are significant from the point of view of the influence of the various salts or their kations. Ca alone added to the LiCl gives no beats but an increased tone, Na alone in 0.7 percent solution of NaCl gives no beats but diminished tone, but Na and Ca together or with K as in a Ringer's mixture cause spontaneous beats and an inter- mediate condition of tone. These facts, like many others, indicate the necessity of both Ca and Na for spontaneous rhythmic contractions. 192 W. H. Howell. Action of Apical Strips of Ventricle in Solutions of Sodium Oxalate. The experiments made with this salt were intended to throw some light on the special significance of Ca for rhythmic contractions. They were made in two ways; first, by adding solid sodium oxalate to a 0.7 per cent solution of NaCl in quantities equal to o. i to 0.3 per cent of the solution, giving therefore an hypertonic solution ; second, by replacing a part of the NaCl in a 0.7 per cent solution by an equivalent amount of sodium oxalate, thus preserving approximately the osmotic pressure of the solution and the total number of sodium ions. The results which will be used subsequently in the theoretical discussion were as follows : 1. Immersion of a fresh strip in a solution of NaCl and Na.2C204 is followed by a relaxation from loss of tone which is apparently more rapid and complete than in NaCl alone, but no contractions occur except possibly a series of very small beats lasting but a short time. Subsequent immersion in NaCl 0.7 percent or in an isotonic solution of sugar fails to give any contractions. That the strip has not been permanently injured by the oxalate is shown, however, by the fact that immersion in a Ringer's mixture gives a good series of beats. It is interesting, moreover, to find that if the Ringer's mixture is fol- lowed by a bath of 0.7 per cent NaCl the usual series of beats from this solution is obtained, although as just stated the NaCl solution without previous immersion for some time in Ringer is not capable of arousing contractions. 2. After a fresh strip has been immersed in a 0.7 per cent solution of NaCl and has gotten well started on its series of beats, addition of sodium oxalate causes a marked diminution in the amplitude of the beats but may not bring them to a complete stop, as would perhaps be expected from the results given under experiment i. After cessa- tion of such a series of beats immersion in a sugar solution fails to give the usual series of beats such as have been described as follow- ing after ordinary NaCl exhaustion. Action of Apical Strips of Ventricle in Solutions deprived of Oxygen. It seems certain that a supply of oxygen is necessary for the maintenance of automatic contractions; but fortunately in experi- Automatic Contractions of Heart- Muscle. 193 ments upon the effect of solutions of the inorganic salts this factor does not have to be considered, since the aqueous solutions used hold sufficient oxygen in solution for the tissues of the cold-blooded ani- mal. A single experiment was performed to ascertain the influence of an inadequate supply of oxygen and the results may be referred to very briefly. A fresh strip was immersed in a solution of NaCl 0.7 per cent which had been boiled thoroughly and then allowed to cool under a layer of oil. This layer was kept upon the solution through- out the experiment. After a latent period of twenty-three minutes the strip gave a series of contractions lasting for sixty-five minutes. The series differed from a normal NaCl series which was obtained from a companion strip from the same heart in unboiled salt solution mainly in the fact that the amplitude of the contractions was small throughout. The deprivation of O2 in this experiment was of course not complete, as some at least was contained in the strip itself when immersed. The most interesting point, however, was that after the completion of the NaCl series subsequent immersion in a mixture of NaCl and CaCl2 that had been deprived of its oxygen by boiling gave almost no effect in the way of contractions, and changing the strip afterward to a similar unboiled solution was equally negative. The absence of oxygen seemingly prevented the strip from recovering its automaticity in a mixture of NaCl and CaCla and permanently altered the tissue. Theoretical Considerations. The importance of sodium salts. — Sodium salts exist in the blood and other liquids of the body in large quantities, 0.5 per cent or more, chiefly as sodium chloride. It is obvious therefore that the osmotic pressure of these liquids is dependent mainly upon this con- stituent. It has been natural to use so-called physiological saline as the basis of the solutions employed in investigating the action of various inorganic salts upon the heart. Experience has shown that solutions of NaCl of from 0.6 to 0.8 per cent are as nearly a normal medium for most living tissues as has yet been obtained by the use of any single substance in solution. The present author therefore in a previous paper ventured the opinion that this salt is essential in the blood and artificial circulating liquids because of its osmotic im- portance, giving abundant proof at the same time that so-called phy- siological salt solution is not sufficient alone to maintain the automatic 194 W.H. Howell. contractility of heart muscle. While the osmotic importance of NaCl in the blood is evident from the relatively large quantity present, Loeb has called attention to the fact that in addition the sodium ion is of importance in a special way in connection with the causation of ^.utomatic contractions. The experiments that have been described in this paper confirm this point of view, although they show equally well that Loeb has been led to attribute an all-importance to the Na ion which is not warranted by his own experiments nor by those of other observers. Loeb's views of the relations of the Na, Ca, and K ions to the rhythmic activity of heart tissue have not, so far as his own work is concerned, been obtained from experiments performed directly upon this tissue, but are based mainly upon experiments on other rhyth- mically contracting tissue such as the bell of the gonionemus and the so-called rhythmic contractions of skeletal muscle in certain artificial media. It is a matter of some difficulty to criticise his views upon the importance of Ca, K, and Na to the rhythmic contractions of the heart, since in the numerous papers published by this author his statements of their relative importance have varied somewhat from time to time. In his earlier papers at least Loeb has assumed that his results obtained upon other tissues apply equally well to the heart, an assumption which is not wholly correct, as may be proved by a comparison of the behavior of heart tissue and skeletal muscle tissue in solutions of NaCl and CaCl2, or by the differences in reaction in the same media shown by heart muscle, skeletal muscle, plain muscle (oesophagus) and cilia. In his most recent paper ^ Loeb's theory seems to be that Ca and K are not directly necessary to the automatic contractility of the heart, but only indirectly in that they neutralize the poisonous action of the Na ions. Stress is laid upon the poisonous action of Na as a fundamental fact that explains the neces- sity of the presence of Ca and K. So far as rhythmic activity is con- cerned only Na ions are directly concerned, or, to quote from Lingle, they are the producers of rhythmic activity. When, however, only so- dium salts are present, when all the Ca and K ions are replaced by Na ions, then the theory assumes that the physical properties of the pro- teid tissues become so altered that rhythmic activity is impossible. It would seem to follow from this last statement that Ca and K are directly necessary to the maintenance of the normal physical properties ^ LoEii : Archiv liir die gesammte Physiologic, 1900, Ixxx. p. 229. Atitomatic Contractions of Heart-Muscle. 195 of the heart proteids, a conclusion that it is difficult to reconcile with the hypothesis that they are not directly necessary to the contractility of the heart muscle. All the arguments adduced to show that Ca is only indirectly of importance in rhythmic contraction in that it serves to neutralize the Na might in truth be reversed and used equally well to prove that the Na ions are not directly necessary to heart activity but only indirectly by antagonizing the poisonous action of an excess of Ca. To lay emphasis upon a poisonous action of sodium salts seems to the present writer to be unfortunate, when one considers that as a matter of fact every other substance soluble in water, when taken alone, is also, so far as known, poisonous to the heart muscle in the same sense. Pure solutions of salts of Ca, K, or other elements, as well as solutions of non-electrolytes such as sugar or urea, are totally unable to support rhythmic activity of the heart muscle. Pure solutions of proteids, fats, or carbohydrates would without doubt show a similar action and would therefore be poisonous to the heart muscle in the same sense as a 0.7 per cent solution of NaCl. To say that pure solutions of NaCl are poisonous to the heart muscle is to say nothing more than that solutions of NaCl alone will not maintain the irrita- bility and automatic contractility of the heart tissue, but the same is true to a more marked extent of every other known constituent of the blood, when taken alone. The really remarkable thing about NaCl is that in pure solution of appropriate concentration it is less injurious to living matter than any other single substance. The uni- versal employment of physiological saline in histological and physio- logical work upon living tissue bears witness to this fact. Loeb's interesting discovery that the young fundulus will live in- definitely in distilled water but dies in a short time when placed in a \n solution of NaCl does not alter this fact at all, so far as the writer can see. The fundulus dies in pure solutions of NaCl, but under such conditions we may suppose that its diffusible Ca, K, etc., are lost, as would be the case if the heart or any other of its tissues were bathed directly in the solution. But, says the author, the fact that the fish lives in pure water shows that neither Ca nor K is directly necessary to the heart's activity or to the act of respiration. How so? The author surely does not mean to assert that the fundulus when placed in pure water loses its store of Ca, K, and Na ions from the liquids and tissues of the body and still continues to live indefin- itely. But if the Ca, K, and Na are not lost from the tissues and 196 JV. H. HoiuelL liquids of the body, how does the experiment given prove anything regarding the direct or indirect necessity of Ca or K to the heart's activity ? It is not necessary to add that heart tissue does not retain its normal properties long in pure water compared with 0.7 per cent solutions of NaCl. If the fundulus can really live indefinitely in pure water we must assume that there is a protective reaction of its super- ficial epithelium toward the water, which prevents the diffusion from the interior of the animal of the soluble constituents of the body- liquids. While the writer is unwilling to admit the conclusion of Loeb and Lingle with regard to the sole direct importance of Na ions in rhythmic activity and believes that the experiments reported by Lingle 1 furnish only a superficial justification for this conclusion, he takes pleasure in acknowledging that the work of these authors has been important in directing attention to a special significance of the Na ions. From the experiments reported in some detail in the pres- ent paper the author has been led to conclude that the presence of both sodium and calcium salts is absolutely necessary for the produc- tion of rhythmic contractions. No experiment that the author has been able to devise, and none that he has seen reported by other observers, has shown in a conclusive way that the ventricular muscle can exhibit spontaneous rhythmic contractions in the absence of either sodium or calcium salts. The long known fact that the ventri- cle will beat, for a time, in solutions of NaCl is of course not opposed to this conclusion. The fact that there is no Ca in the bathing solution does not mean that there is no Ca in the beating strip. Ventricular strips suspended for many hours in solutions of NaCl still give a calcium reaction with ammonium oxalate. If we remove the calcium, that part at least that is in dissociable form, by the only certain method, namely its precipitation as oxalate, then the strip fails to give any contractions at all unless new calcium is supplied in the bathing liquid. If one chooses to make the objection to this experiment that the oxalates prevent the contractions, not by precipitating the Ca, but by some direct action of the oxalic acid ion, one cannot prove the con- trary directly, since by the nature of the reaction the Ca cannot be removed without an excess of the oxalate and vice versa. The indirect evidence, however, shows conclusively that the effect of the oxalate 1 Lingle : Loc. cit. Automatic Contractions of Heart-Muscle. 197 is due to its precipitation of the Ca. When a ventricular strip has been immersed for an adequate time in a mixture of NaCl and NagCgO^ it cannot be made to resume spontaneous beats by immersing it for any length of time in a solution of NaCl, or indeed, so far as the writer's experience goes, by immersing it in any other solution except one containing Ca salts in addition to the sodium salts, the only excep- tion to this statement being the similar though less efifective reaction of the salts of strontium or barium. One of the series of experiments described in the first part of this paper would seem at first sight to indicate the possibility of automatic contractions in the presence of sodium salts alone, but a considera- tion of the conditions makes this conclusion inadmissible. The experiments referred to consist in suspending a fresh strip from the apex of the ventricle in a solution of NaCl 0.7 per cent. After the series of NaCl beats is completed the strip, after being rinsed with water, will give a new series of beats lasting about one hour if im- mersed in an isotonic solution of sugar or LiCl, the contractions usually being larger and of longer duration in the sugar solution. The bathing solution in this case contains neither Ca nor Na, but it might be assumed that since the strip has been removed from a solution of NaCl the beats continue during the time necessary for the diffusion of the Na from the strip into the solution of sugar. This explanation, however, is contradicted by the following facts. I. Ca is still present in such strips, as may be shown by the reaction with ammonium oxalate. 2. In some cases after a very long con- tinued series in NaCl, the subsequent bath in sugar solution gives only a short lasting series of feeble contractions. 3. After immersion in NaCl and Na2C204 for a time equivalent to that required for a NaCl series, subsequent treatment with sugar solutions gives no con- tractions at all. 4. After completion of the series of beats in the sugar solution or in LiCl, addition of NaCl in small or large quantities to the bath causes no revival of the contractions. Transferal again to a 0.7 per cent solution of NaCl gives no new series of beats in the case of LiCl. With the sugar solution, reversing the solutions in this way gives in some cases no contractions, in others a brief series of small beats. If the sugar or LiCl series were due simply to the Na while diffusing out of the strip, one should expect a similar series when opportunity is given for diffusion of the same substance into the strip. 5. After completion of the sugar series, immersion in a proper mixture of NaCl and CaCl2 gives usually a beautiful series of 198 W. H. Howell. contractions, the ventricular strip under these conditions being in fact in a particularly favorable condition for reaction to appropriate solutions. The occurrence of a limited series of beats in a solution of sugar or LiCl after previous exhaustion in NaCl may be referred therefore to the presence of both Ca and Na in a usable form in the strip. The whole matter of the diffusion of the ions out of or into the heart strips is of course an assumption, but a permissible, if not necessary one, to explain the action of the various solutions used. Most of the results obtained are in harmony with the hypothesis that the fresh strip gives up its kations slowly, particularly the Ca, under the influence of diffusion, and that the interaction of Na and Ca in the production of rhythmic contractions is possible only when the quantities present in the tissue do not depart too far from a certain proportion. That the diffusion of the K out of the heart strip when suspended in a bath of NaCl is more rapid than that of the Ca is indicated by an analysis of the experiments reported in the first part of this paper, and is in ac- cord with an interesting experiment related by Ringer.^ According to this author, when a frog's heart is fed with 0.6 per cent NaCl and dialyzed serum the contractions obtained resemble those resulting from a mixture of sodium chloride and calcium salts. Addition of potassium salts to the mixture brings the heart-beats back to their normal character. When potassium is not present in the solutions the proportions of NaCl and CaCJg which seem to give the most regular and forcible contractions with the ventricular strips are NaCl 0.7 per cent and CaCl2 o.oi per cent to 0.012 per cent. Any marked increase of the Ca leads to an exaggerated tone ending in a condition of calcium rigor. On the other hand, if the amount of calcium is greatly reduced contractions become impossible. We may suppose that this last condition exists at the end of the usual NaCl series when the stage of NaCl exhaustion is reached. Subsequent immersion in a solution of sugar or LiCl may give a short series of beats, as we have seen. We have given reasons for believing that this last series is dependent upon the presence of some Ca in the strip in dissociable form which becomes effective as the excess of NaCl is reduced by diffusion. That the Ca plays a part in this series is indicated further by the fact that a fresh strip which is immersed for a longer time than ^ Ringer; Journal of physiology, 1887, viii, p. 288. Automatic Contractions of Heart-Muscle. 199 usual, about four hours, in a solution of sugar or LiCl, remaining quiet during this time except for changes in tone, and which is then immersed in NaCl 0.7 per cent does not contract, or gives at best a few feeble inconspicuous beats. When the strip, however, is trans- ferred to a proper mixture of NaCl and CaCl2 or to a Ringer's mix- ture a long-lasting series of beats is obtained. After completion of its series of beats in NaCl and the subsequent series in solutions of sugar or LiCl, if some CaCl2 or KCl is added in physiological doses to the last solution the only effect is an augmented tone. If on the contrary the strip is transferred to a solution of NaCl 0.7 per cent the only effect is a loss of tone. But transference to a mixture of NaCl and CaClg or to a Ringer's mixture is followed by an excellent series of automatic contractions. The only conclusion that one can draw from such experiments is that both the Na and the Ca are necessary for the production of rhythmic spontaneous contractions. The two substances seem to have a distinctly antagonistic effect upon the heart-tissue, — excess of Na leading always to a loss of tone, excess of Ca to a greatly increased tone, while in mixtures of proper propor- tions the condition of tone is practically unchanged and automatic contractions result. Ringer laid emphasis upon the antagonistic influence of the K and Ca salts in the matter of tone, an excess of the former leading to a loss of tone and of the latter to an increase of tone. In his experi- ments, however, NaCl was always present, so that its influence was a constant factor under both conditions. That potassium salts within certain limits tend to counteract the stimulating influence of calcium salts upon tone cannot be doubted, but that this influence is far less marked than the effect of Na salts is shown conclusively by the experiments described. In solutions of sugar, for instance, the very marked influence of traces of Ca in causing an augmentation of tone is scarcely if at all counteracted by adding K salts, whereas the addition of Na salts, provided the Ca is not present in excess, is promptly followed by a relaxation of tone. The fundamental antago- nism is between the sodium and the calcium, and in this matter of the condition of tone, as in that of automatic contractions, it is noteworthy that for the preservation of normal conditions the sodium salts must be present in relatively large proportions as compared with the calcium salts. The rhythmic play of contraction and relaxation in a series of beats under the combined action of sodium and calcium salts may be 200 W. H. Howell. compared with the opposing influence of the same substances upon the condition of tone. As I have described in a previous paper the changes in tone may themselves take on a rhytlimic character, which indeed is very marked for the tissue at the venous end of the heart. When we compare the rhythmic contractions of various tissues such as the heart, oesophagus,^ stomach, etc., it becomes in fact a matter of difflculty to say when these changes are merely variations in tone and when definite contractions. For the present it would seem necessary to base any distinction made simply upon some arbitrary standard of the rapidity of shortening. The facts so far as they are known at present would seem to show that the contractions and relaxations in rhythmic beats are due to fundamentally the same causes as lead to rhythmic changes in tone. It is natural to suppose that the energy of the contraction or increased tone is referable to a reaction of the proteid tissues of the heart muscle. The influence of the inorganic salts can only be secondary in making this reac- tion possible. This suggested relationship between the proteids and the salts recalls the known facts regarding the dependence of the proteids for their properties upon combinations with inorganic salts, the solubility for instance of the so-called globulin group. Pauli^ has recently called attention to the fact that the globulins are not soluble in solutions of non-electrolytes such as sugar, and he supposes that a globulin in solution is in chemical combination both with the anion and kation of the salts present. A further fact of possible bearing upon this physiological relationship between the proteids and the inorganic salts is the discovery by Starke^ that globulins may be precipitated from their solutions by the action of dilute solutions of CaCl2, a fact which may have some bearing upon the phenomenon of Ca rigor so often referred to in this paper. The special influence of potassium salts. — So far little has been said of the influence of potassium salts on the heart beat, for the reason that the experiments described seem to indicate that the presence of potassium salts is not essential to the production of spontaneous contractions ; under proper conditions an appropriate mixture of sodium and calcium salts suffices to initiate spontaneous contrac- tions and maintain them for a considerable period. However, the 1 Stiles : This journal, 1901, v, p. 338. '^ Pauli : Archiv fiir die gesammte Physiologic, 1899, Ixxviii, p. 315. ^ Starke: Zeitschrift fvir Biologic, igoo. xl, p. 419. Automatic Contractions of Heai't- Muscle. 201 influence of the potassium upon the heart rhythm is most evident, and when used in optimum proportions, as in Ringer's mixture, the spontaneous contractions of the heart strips are maintained for much longer periods than with a mixture containing only sodium and calcium salts. As far as my experiments have gone the potassium ion seems to influence the rhythm chiefly, its influence in general being compa- rable to that of inhibitory agents. In mixtures of NaCl and CaCl2 the strips beat with a rapid and very regular rhythm, although the diastolic phase is gradually shortened as the strip slowly goes into a condition of heightened tone. When KCl is added to the mixture in physiological doses its effect, as soon as it is felt at all, is to lengthen the period of diastole or as one might say to increase the duration of the refractory phase. Records of contractions of heart strips obtained in Ringer's mixture of the composition used in this laboratory (NaCl 0.7 per cent — CaCl2 0.026 per cent — KCl 0.03 per cent) are characterized by a certain irregularity in rhythm ; pauses of longer or shorter duration are intercalated between single beats or between groups of beats. An attempt has been made to determine how completely this influence of KCl on the rate of beat can be controlled experi- mentally. For this purpose ventricular strips were chosen that had been exhausted by an immersion for three or more hours in NaCl 0.7 per cent and subsequent immersion for an hour or more in sugar solutions. Under these conditions the strip must be practically deprived of its diffusible Na, Ca, and K, but, as has been pointed out, is in an excellent condition to give spontaneous contractions in appropriate mixtures. In one such experiment the strip had been transferred to a Ringer's mixture (NaCl 0.7 per cent, CaCl2 0.024 per cent, KCl 0.024 per cent) and had been beating well for about eighteen hours. It was then placed alternately in a solution containing only NaCl 0.7 per cent and CaCl2 0.024 P^r cent, and in the mixture containing KCl in which it had been beating. In the solution without K the rate was very regular, ten to eleven contractions a minute, while in the solution containing K the rate was much slower, the beats for a time occurring at the rate of five per minute, but finally becoming slower, two per minute, and eventually only two in ten minutes. Changing back to the solution without K the beats within one half minute again returned to the previous rapid regular rhythm. In another experiment the 202 W. H. Hozcell. strip after a similar preparation was placed in a Ringer's mixture (NaCI 0.7 per cent, CaClg 0.024 per cent, KCl 0.024 per cent) in which it beat quite regularly at the rate of from twenty to twenty-two beats per minute. KCl was then added to the solu- tion, which consisted of 25 c.c. of the above mixture. Successive additions were made of jV c.c. of a i per cent solution of KCl. The effect upon the rhythm was marked, but it was difficult to hold the strip to a constant rate. That is, the addition of a certain amount of KCl would be followed by a slower beat, but in a little while this effect would pass off and the strip would beat at its former rapid rate. Addition of more KCl would again slow the rhythm, but if too much had been added the strip would gradually pass into a condition of complete inhibition, from which it could quickly be brought back to a rapid beat by changing to the original solution. In this and in other experiments it was found that increasing the amount of K was followed at a certain point by a more or less sudden cessation of the spontaneous beats. While in this state of inhibition the strip could be made to beat again, either by adding more CaCl2 to the solution or by removing some of the KCl, that is, by changing to a solution containing less KCl ; but the latter method always gave a more rapid and regular rhythm than the former. In other words, after sufficient KCl had been added to bring about complete inhibition, increasing the CaCl2 in proportion never completely antagonized the potassium effect. I was not able, therefore, to corrob- orate fully the statement made by Ringer that the effect of a toxic dose of KCl may be completely removed by adding what, if taken alone, would be a toxic dose of CaClg- In my experiments the antag- onism between the Ca and the K was incomplete, excessive doses of the one not being entirely neutralized by increasing the proportion of the other. The experiments, so far as they go, seem to indicate that there is for each ventricular strip a certain amount of KCl which almost or completely inhibits the automatic contractions in the presence of Ca and Na. This dose of potassium varies somewhat with strips from different hearts or with the same strip under differ- ent conditions. If the amount of potassium salt added is not too great this inhibition is not toxic in character, that is, the tissue is not injured. On the contrary, the strip readily beats again when transferred to a mixture containing less potassium, its condition in fact resembling very much Atitomatic Contractions of Heart-Muscle. 203 that of a fresh ventricular strip when immersed in the animal's own serum. In serum, as I have described in a previous paper, the strip usually remains quiet but irritable to artificial stimulation, and it may be made to beat spontaneously by diluting the serum with a 0.7 per cent solution of NaCl, or in some cases by increasing the amount of calcium by the addition of a little of a solution of CaCl^. The latter method is never so effective as the former or the transference directly to a 0.7 per cent solution of NaCl, and in the light of the present series of experiments I should explain the more favorable effect of the latter procedure by the resulting diminution in the percentage of potassium salts. The NaCl series of beats. — The fundamental phenomenon in all the experiments upon the action of the inorganic salts upon the heart rhythm is the series of beats obtained by immersion of fresh strips in 0.7 per cent solutions of NaCl. It will be remembered that this series shows first an increase to a maximum and then a gradual and very regular diminution to zero. The general effect of NaCl solutions was described by Merunowicz, Aubert, and others, and more recently by Greene. In a former paper I suggested that this effect of NaCl solu- tions was explained by the fact that under these conditions both the potassium and the calcium salts diffuse out of the strip into the surrounding bath and that the diffusion of the potassium takes place more rapidly, so that the antagonistic influence between the potassium and calcium is removed by a relative preponderance of the latter, the strip eventually ceasing to beat when the dissociable calcium is completely removed or very much reduced in amount. While still holding to this general explanation, I should be inclined now to give the following theory of the cause and course of this series of beats. When the strip is placed in the solution of 0.7 per cent NaCl, both the potassium and calcium ions begin to diffuse out, and we may suppose that the former pass out more rapidly and com- pletely. After the potassium contents of the strip are sufficiently reduced its inhibitory influence is weakened and under the combined influence of the Na and the Ca still present the strip gives its auto- matic contractions, which, under the conditions assumed, would naturally increase to a maximum and then slowly decrease as the Ca diffused out, the strip, as the foregoing experiments have indicated, being unable to contract spontaneously in the presence of sodium alone. Loeb has given a different explanation of this series of NaCl beats. 204 ^. H. Howell. To quote Lingle, who has applied his views to the ventricular muscle, " the sodium ions act by migrating into the muscle substance and combining with some part of it. And hence when too many sodium ions have combined and taken the place of a number of Ca ions in the muscle, rhythmic beats cease." That is, the series is started by the action of Na ions and subsequently suspended by the toxic action of the same ions when the Ca has been completely, or almost com- pletely, displaced. The cessation of the series of beats, according to Loeb's view, is due directly to the poisonous effect of the NaCl ; according to my view, to the loss of Ca, which forms a necessary factor in the production of the contractions. With regard to the first part of Loeb's theory, that the initiation of the series of beats is due to an excess of Na ions migrating into the strip and replacing the Ca ions, it is to be borne in mind that when the strip is placed in a solution containing both NaCl and CaCl2 it beats as well as, or indeed much better than in a solution of NaCl alone. This fact would seem to contradict the view that the replace- ment of the Ca in the tissue by the Na is the necessary factor in starting the contractions, while it supports my hypothesis that the removal of the excess of potassium by diffusion is what really liberates the tissue and permits it to beat under the combined in- fluence of the Na and Ca. It is quite certain that when the strip is immersed directly in a solution containing the Ca, Na, and K in the proportions present in serum it does not beat, as a rule, and that a proper mixture without the potassium gives rise to a series of beats. As for the view that the gradual falling off of the series of beats in NaCl and their final disappearance is due to a poisonous effect of the Na ions, it would seem, if I interpret their views correctly, that Loeb and Lingle believe that this toxic action is exhibited only in the ab- sence or great reduction in number of the Ca ions, and that only as long as Ca ions are present do contractions occur. Looked at ob- jectively, this way of stating the matter amounts to the same thing as saying that the Ca is necessary to the rhythmic contractions, a view in which I heartily concur. The action of poiassiinn salts and inJiibition, — The especial effect of potassium salts upon the heart has been known since the pioneer work of Bernard, and has been made the subject of investigation by a number of observers. Bottazzi ^ particularly has described the effect ^ Bottazzi: Archives de physiologie normale et pathologique, 1896, p. 882. Automatic Contractions of Heart-Muscle. 205 of potassium salts upon the heart of cold-blooded animals. Like Ranke, Ringer and others, he has been impressed by the inhibitory character of the standstill produced by these salts. So much so, in fact, that he compares their action directly with that produced by stimulation of the inhibitory fibres of the vagus, and suggests that the action of the two are not only superficially alike, but funda- mentally identical. Adopting the hypothesis of Gaskelland Fano that inhibition is an expression of anabolic processes in the heart tissue, he assumes that the potassium salts also bring about cessation of contractions by augmenting the anabolic processes. There are no facts, so far as I know, that give any real support to this hypothesis, but it is somewhat interesting to recall how many physiologists (Aubert, Ranke, Fano, Meltzer, Bottazzi), have been led to suggest in one form or another that the normal diastole of the heart muscle is at bottom a phenomenon of inhibition. To the present writer it seems unnecessary to assume that the standstill produced by a slight excess of K salts is due to an augmented anabolism. If the properties of the proteids in the heart tissue are dependent upon their union with the inorganic salts of the blood and lymph, a preferable hypothesis would be that the compounds with potassium possess too great a stability to undergo that decomposition which we suppose to be the origin of the liberation of energy in a contraction. An interesting point with regard to the potassium inhibition, if the phrase may be allowed, is that the heart tissue is not only incapable of spontaneous contraction, but, according to those who have tested the matter (Ringer-Bottazzi) has lost its irritability toward electrical stimuli. In this respect there would seem to be a difference between the standstill produced by potassium salts and vagus inhibition, but the difference may be one of degree only. While the heart inhibited through the vagus is usually irritable toward electrical stimulation, yet, according to Schiff and Eckhard,^ if the inhibitory stimulus is sufficiently strong this irritability is lacking. On the other hand, although a dose of KCl strong enough to stop the whole heart may remove the irritability of the ventricle toward electrical stimulation it is possible that the dose might be so graduated as to stop spontaneous contractions without loss of irritability to artificial stimuli. If the views advocated in this paper are correct such a condition of affairs exists normally in the living ventricle, the proportion of potassium salts in the blood being sufficient to suspend automatic contractility ^ Quoted from Tigerstedt's Physiologic des Kreislaufes, 1893, p. 254. 2o6 W. H. Howell. in this part of the heart without destroying its irritabiHty toward artificial stimuli. The analogy or relationship between potassium standstill and vagus inhibition is very suggestive and need not be complicated by any assumption as to the cause of inhibition. The very potent influence of the kations Ca, K, and Na upon the liberation of spontaneous con- tractions suggests furthermore the possibility that the opposing in- fluence of inhibitory and augmentor nerve impulses on the heart's contractions may eventually be traced to an influence upon the re- lations of these elements to the living proteids of the heart tissue. Summary. The chief conclusions arrived at in this paper may be stated briefly as follows : 1. Spontaneous contractions of the ventricular muscle of the terra- pin's heart are dependent upon the presence in the tissue of disso- ciable compounds of both calcium and sodium. If either the Ca or the Na is absent automatic contractions are impossible. 2. Sodium salts (NaCl) in the liquid surrounding the heart tissue tend to produce a relaxation from loss of tone. Calcium chloride on the contrary causes a shortening from increase of tone which may pass into a permanent rigor. Potassium chloride exhibits an antag- onistic effect toward the action of calcium chloride, but only to a marked extent when sodium salts are also present in approximately normal proportions. 3. The influence of the potassium ion when present in physiological proportions is shown by a slowing of the rhythm or a lengthening of the refractory phase. Under the combined influence of Ca, Na, and K the automatic contractility is maintained for longer periods than by the action of Na and Ca alone. 4. The fact that ventricular strips do not contract spontaneously as a rule in the animal's own serum or in an equivalent Ringer's mixture is due to the inhibitory influence of the potassium salts. 5. The characteristic series of beats obtained by immersing a fresh strip of ventricle in a 0.7 per cent solution of NaCl is referable to the following changes, a. The latent period and the beginning of the series of spontaneous beats is due to the gradual loss of dissociable potassium from the strip by diffusion, b. The gradual diminution and final disappearance of the beats is due to the loss of the dis- sociable calcium by diffusion. THE ACTION OF PILOCARPINE AND ATROPINE ON THE EMBRYOS OF THE STAR-FISH AND THE SEA- URCHIN. By albert p. MATHEWS. \^From the Marine Biological Laboratory, Woods Hall, Mass.J THE work of Heidenhain and Langley has seemed to show that pilocarpine and atropine do not affect gland-cells, but only the endings of the hypothetical secretory nerves. The chief evidence of this is the observation that atropine stops the secretion which normally occurs in the submaxillary gland of the dog upon stimula- tion of the chorda tympani, but not that which follows stimulation of the cervical sympathetic. This fact indicates either that the secreting cells are not poisoned by atropine or that the sympathetic causes secretion in some other way than by acting on these cells. There is no evidence except in the salivary glands that atropine has not poi- soned the secreting cells, for in no other glands have two secretory nerves been found, one of which is paralyzed by the drug, while the other is not affected. The possibility that the sympathetic produces its secretion by acting in some other way than upon the secreting cells has not re- cently received the attention which it deserves. Langley, in Schaefer's text-book, dismisses it as improbable. The alternative has been so generally accepted that if atropine paralyzes a gland's secretion it is considered strong evidence of the existence of secretory nerves in this organ. In spite of the general acceptance of the view just stated that the sympathetic acts directly on the secreting cells, there is reason for doubting whether this is the case. I have elsewhere put together evidence pointing to the action of the nerve on contractile tissue in the gland. There appears to me to be no difficulty in understanding all the phenomena of the sympathetic secretion on this hypothesis, and several facts exist which are, I think, incompat- ible with any other explanation. The presence in the salivary glands of contractile cells corresponding to the contractile sheath about the sweat glands is now abundantly demonstrated. These, the so-called " basket-cells," closely resemble the similarly situated cells about the salivary glands of the molluscs, and in these animals their contraction 207 2o8 Albert P. Mathews. has been observed under the microscope by Saint-Hilaire. ^ Miss Hyde's^ work, also, on the physiology of the salivary glands of Cephalopods leaves no doubt that these glands contract and thus ex- pel most, if not all, of the secretion observed upon nerve stimulation. Furthermore, suprarenal extract, which causes marked contraction of smooth muscle, causes a secretion from the salivary glands, but not, so far as I have observed, from the pancreas, where contractile cells are absent. These facts, showing a probability that the sympathetic does not cause secretion by acting on gland cells, undermines the main evidence that atropine does not act on the gland cells. The meaning of the fact observed by Bunch ^ that stimulation of the sym- pathetic may cause a slight preliminary increase in volume of the submaxillary gland is not sufficiently clear to offset nearly all other phenomena of sympathetic secretion, which point clearly to a con- traction of the gland upon sympathetic stimulation. The following objections may also be advanced against the as- sumption that atropine does not act on these cells. The same argu- ment applies, for instance, to quinine. Quinine, if injected into the duct, paralyzes the chorda secretion but not the sympathetic. If the sympathetic produces its secretion by action on the gland cells this demonstrates that the chorda is paralyzed at a time when the gland cells are not paralyzed, for the sympathetic is then still active and there is no reason for believing that the gland-cells connected with the sympathetic differ from those connected with the chorda. We are hence driven to the very improbable conclusion that a gen- eral protoplasmic poison, such as quinine, does not act on the cells of the salivary glands, but only on the ends of the nerves, and par- ticularly the ends of the chorda secretory fibres. The inhibitory action of atropine on the spontaneous secretion of saliva is also hard to understand on the nerve-end hypothesis.* There are also well substantiated observations that in cats in which the sciatic nerve has been severed for two weeks, pilocarpine will still cause sweat secretion in the hind foot, thus indicating that the drug acts in part at least on the gland cells. These facts made it desirable to test the action of the drugs on embryos in which there are no nerve cells. I used embryos of the ^ Saint-Hilaire: Anatomischer Anzeiger, xix, 1901, p. 478. 2 Hyde, Ida H.: Zeitschrift fiir Biologic, 35, 1897, p. 459. ^ Bunch: Journal of physiology, 1900, xxvi, p. i. * Mathews: This journal, iv, 1901, p. 483. Action of Pilocarpine and Atropine on Embryos. 209 star-fish, Asterias Forbesii, and the sea-urchin, Arbacia punctulata. The eggs were trans- ferred immediately after fertilization into 100 cubic cen- timetres of sea- water to which had been added from one tenth of a cubic centimetre to two cubic centimetres of a one half per cent solution of pilocarpine hydro- chlorate or atropine sulphate. They re- mained in these solutions during development. The drugs were dis- solved either in sea- water or fresh water, the result being the same in either case. The eggs from each female were kept separate and were invariably com- pared with the control eggs in sea-water. Followingare the protocols of some of the experiments : -a -a ni "3 oi en a; 3 en . 3 < a OS < tn . < ^ s «i . 3 < 30^ No. of experi- ment. 14 Asterias. Fertilized August 3, 3.30 P. M. 2IO Albert P. Mathews. d 6 o o o 1) -a -a ai (u >.; c ii o & i: ci to o u u^ o d 6 6 B 1 astu l£E. No invagi- nation. Opaque and very small. Blas tulas. All at bot- tom. Small and opaque. cs tn a2 P to .3 oi O S =2X o u O ;^ <^ tiS ^ '^ ^ S O ^ a! ,Q 'O CZ3 -a 3 "s m bo «-c X = - • 3. i £ S«cS" u u d in O fl « 2 £ rt - m c« tn -3 "«-E X i: rt .£ -d =5- !£'C tonitn ,■2 oi a> aj CO C-C-v^tM o cj cj 8 o -a -a '5. g o 1> "&, o o cJ O d S5 cj d = bo en oj "" s rt " 0-5 t^' tn S >P o . tT! &| re — : u 6 d u Jin S'C-s S . . g d c— 1 tn -a Mfc, ^11 «J -"-- in (U J 3. (J s- (u rt i; ^ S tn r* [jJc«J3in3CtuJ; to 5 c5 £ bO bfi ^ 1 ?' tuJ S boW c ail > ^ -. J-" l>^ .3-^ X QJ 1- C f^ t^ U r^ '^•S £ « ^g^ hH 3 tn 3 in (>J -< < GO I— 1 bb < 3 »— ' < ^ti <: No. of experi- ment. tn XJ _^ .2 ?^^ ■r. S = 0- Action of Pilocarpine and Atropine on Embryos. 211 d 8 •a •xi (0 u .s ■^ V. "o U u 6 2 cells. None swimming. Few swimming. About \ swimming. Beginning to form plutei. 91.2m. 116.4m- Small gastrulffi, just beginning to change to plutei. "a, . g.2 d 1—1 bb c 11 ■? •= S =" ./ " ^ .^^ .2 (U « 3 1) - fe u S^ £ 'bb 3 . :z^3. a. ii CM tSI 3CN 1^2 S2 ^ a. eg '■ s tiS • s ^-g .S rt g u Sic:; 3 \d c5 i CM 1) bO c a! ■n u bO C _3 'bb a, 13 3 ._; bJO rt_e S c >- -s 06 - ?3 ■33 3 3 3l 8.30 816 8.34 CI ... . 31.73 30.99 31.26 31.52 31.37 The average quantity of chlorine in ligament ash was found by us to be 7.39 per cent. P20g was equal to 28.95 PCi" cent of the liga- ment ash. Fat (ether-soluble matter). — Although the Achilles tendon does not appear to hold as much admixed adipose tissue as ligamentum nuchae, it seems to contain almost as much extractive substance. The following percentage results in this connection, calculated for fresh tissue in each case, were obtained by Dormeyer's method : 1 2 ,3 4 5 6 7 Average Fresh tissue . . 087 1.10 1.21 1.16 0.98 1.05 0.93 1.04 The proteid constituents. — It has been known for a long time that tendon consists mostly of collagen. As we have already indicated the earlier observers considered tendon to be almost pure collagen, Rollett's^ researches on the structure and composition of connective tissues demonstrated the presence in tendon not only of such soluble proteids as might be constituents of contained lymph, but also of mucoid. Numerous histologists have shown the presence also of elastic fibres in tendinous tissue. Coagrilable proteid (albinnin, globiiliii). — Rollett detected only traces of coagulable proteid in aqueous extracts of the Achilles tendon of the horse. Loebisch ^ called attention to the fact that ^ Levene's result does not harmonize with Morner's. The latter's method for the detection of chondroitin sulphuric acid in tendon should have revealed the presence of the acid substance in tendo mucoid identified by Levene. See Hawk and Gies : This journal, 1901, v, pp. 398-399. '^ Rollett : Untersuchungen zur Naturlehre des Menschen und der Thiere (Moleschott), 1859, vi, p. i; also, Ibid.^ i860, vii, p. 190. ^ Loebisch: Zeitschrift fiir physiologische Chemie, 1886, x, p. 43. 2 28 Leo Buerger and William J. Gies. • aqueous extracts of the same tendon of the ox contain slight quan- tities of coagulable proteid — "serum globuHn " and an albumin co- agulating at 78° C. Richards and Gies ^ recently observed that aqueous extracts of this tendon from the ox contain minute propor- tions of two coagulable proteids ; one, a globulin, coagulating at 54°-57° C, the other, an albumin, coagulating at 'j^'^ C. In this work we experienced gredt difficulty in making satisfactory quantitative estimations. The quantity of coagulum for 100-200 grams of tissue was always very slight. Frequently it was impos- sible to obtain the coagulum in a perfectly clear fluid. The results were the same in aqueous and in sodium chloride extracts. One or two indirect methods gave no more satisfactory results. Tendo mucoid is somewhat soluble in the aqueous and saline extracts of the tissue, and possibly the observed interference with perfect coagula- tion of the simple proteids was due to the presence of larger or smaller amounts of this glucoproteid. The following percentage results were obtained in extracts from tissue which had been cut into narrow strips and then very finely divided with scissors : — 12 3 4 5 6 7 Average Fresh tissue . 0.231 0.184 0.191 0.274 0.177 0.219 0.262 0.220 It is possible that not only a small quantity of coagulable proteid was lost in each determination, but also that a small proportion of mucoid was admixed with the coagulum as a result of the addition of the dilute acid ordinarily employed to complete coagulation. We feel satisfied, however, that the above average amount is very nearly that contained in this tissue. Much of it doubtless is a part of con- tained lymph. The average quantity in ligaraentum nuchae is 0.616 per cent. Mjicoid? — The proportion of mucoid in tendon is comparatively large. Halliburton states that the average amount for normal connec- tive tissues is 0.521 per cent.^ The amount in the human tendo Achillis he found varied under normal conditions between 0.298 and 0.770 per cent. Chittenden and Gies* obtained as much as i per cent of chemically pure mucoid from the tendo Achillis of the ox, al- ^ Richards and Gies : Loc. cit. - See Cutter and Gies : Loc. cit., foot-note, p. 155. 3 Halliburton : Loc. cit., p. 477. * Chittenden and Gies: Journal of experimental medicine, 1896, i, p. 186. The Chemical Coiishluenis of Tendinous Tissue. 229 though their experiments were not designed for quantitative deter- minations. The amount in ligamentum nuchae was found by us to average 0.525 per cent. Our percentage results for the Achilles tendon of the ox were the following: 12 3 4 5 6 7 Average Fresh tissue . 1.361 1.420 1.332 1.220 1.043 1.22S 1.3S0 1.283 In these determinations we profited by the experience of Cutter and Gies that repeated treatment with excess of dilute alkali is neces- sary to extract completely mucoid from tendon.^ Halliburton ^ gives a record of determinations of mucoid in human tissues under abnormal conditions. In one case the Achilles tendon contained as much as 1.42 per cent. The tendons of the heart under similar conditions contained 1.65 per cent mucoid. Elastin. — When tendon pieces are boiled in water they rapidly diminish in size and only a small quantity of elastin-like material is left behind. This residual material is not as resistant to the action of dilute acid and alkali as is the elastin of ligamentum nuchae, although it appears to be true elastin.'^ The following results for percentage content were obtained in our quantitative determinations : 1 2 3 ^ 5 Average Fresh tissue .... l.,=;61 2.130 1.634 1.100 1.740 1.633 Munz'^ separated this substance, studied some of its reactions and" decomposition products, and made a few analyses of it. He found its nitrogen content to vary between 14.31 and 14.48 per cent. The accuracy of these analytic results has been doubted, since the nitro- gen content of all elastins has been found to be above 15 per cent. One of our own specially prepared samples of tendon elastin, after it had been extracted with alcohol and ether, gave the following percent- age results on analysis: (a) Nitrogen — by the Kjeldahl method — 15.42, 15.49, 15.45; average, 15.45. (b) Sulphur — by the fusion method over alcohol burner — 0.48, 0.54; average, 0.52. (c) Ash — 1.32, 1.28; average, 1.28. These results agree fairly well with those for aorta elastin obtained by Bergh^: N, 15.20; S, 0.66; Ash, 0.51. ^ Cutter and Gies : Loc. cit., p. r6i. 2 Halliburton: Jahresbericht iiber die Fortschritte der lliier-Chemie, 1888, xviii, p. 324. '* KiJHNE : Lehrbuch der physiologischen Chemie, 1868, p. 356. * MiJNZ : Quoted by Gorup-Besanez, loc. cii., pp. 143 and 645. ^ Bergh : Zeitschrift fiir physiologische Chemie, 1898, xxv, p. 341. 2 30 Leo Buerger and William J. Gies. Collagen. — The great bulk of the solid matter of tendon is col- lagen. We made five quantitative determinations by the indirect method/ with the following percentage results : 1 2 3 4 5 Average 0.63 32.47 30.98 32.27 31.59 31.59 Fiesh ti.s.sue . The proportion of collagen in the fresh tendo Achillis is almost exactly the same as that of elastin in ligamentum nuchae. Recently, in testing his method for the determination of collagen in connective tissue containing little soluble proteid, Schepilewsky ^ COMPOSITION OF TENDO ACHILLIS. Constituents. Fresh tissue. ! Dry tissue. Ash. I Calf. Ox. j Calf. Ox. Ox. Water 67.51 62.870 Solids . . . . 32.49 37.130 Inorganic matter . 0.61 0.470 I 1.88 1.266 SO. i 0.031 1 0.084 6.65 PoOr .... 0.039 0.147 .... 0.106 0.397 8.34 31.37 CI 31.88 36660 98.12 98.734 Fat (ether-soluble matter) . 1.040 2801 Albumin, globulin .... 0.220 0593 Mucoid .... .... 1.283 3.455 Elastin .... .... 1.633 i 31.588 4.398 85.074 Collagen (gelatin) Extractives and mined substance un deter- 0.896 i 2.413 found 80.86 per cent of collagen in dry tendon. The particular ten- don he used is not mentioned. In the dry Achilles tendons of the ox analyzed by us the collagen amounted on an average to 85.074 per cent. 1 See Vandegrift and Gies : Loc. cit., foot-note, p. 295. ■^ SCHEPILEWSKY : Archiv fiir Hygiene, 1899, xxxiv, p. 351. The Chemical Constituents of Tendinous Tissue. 231 Crystalline extractives. — Our results for extractives were only qualitative. Creatin and nuclein bases could readily be detected. The proportion of extractive matter was small. Our results were similar to those previously obtained in this laboratory for ligament. In the table on the opposite page the extractives are included in " Extractives and undetermined substance," the figures for which were obtained by difference. Average Composition. — The data of all our analyses are brought together in the summary on the opposite page, which gives the average percentage composition of fresh tendo Achillis and of the dry solid matter in it, together with the results of partial analysis of the ash. STUDIES ON REACTIONS TO STIMULI IN UNICELLU- LAR ORGANISMS. VIII. — ON THE REACTIONS OF INFUSORIA TO CARBONIC AND OTHER ACIDS, WITH ESPECIAL REFERENCE TO THE CAUSES OF THE GATHERINGS SPONTANEOUSLY FORMED. Bv H. S. JENNINGS and E. M. MOORE. \_Fyoin the Zoological Laboi-atory of the University of Alichii;a?!, Jacob Reighard, Director.] IT is well known that when certain infusoria are left undisturbed they do not remain scattered, but gather in more or less dense groups. Thus, if they are mounted on a slide in a thin layer of water, soon dense aggregations will be formed in certain areas, while the re- mainder of the slide will be nearly deserted. One of the first investi- gators to describe this phenomenon was Pfeffer.^ He observed its occurrence in Glaucoma scintillans, and less markedly in Colpidium col- poda, Stylonychia mytilus, and Paramecium. Pfeffer was inclined to believe that these aggregations were due, partly at least, to a con- tact stimulus, resulting from a striking of the organisms against small solid bodies, and especially against each other. In the first of this series of studies,- this phenomenon in the case of Paramecium was subjected to a thorough examination. It was demonstrated that while the contact stimulus plays a certain part in the production of these aggregations, the chief factor involved is a reaction to carbon dioxide. The Paramecia tend to gather into regions where the water is impregnated with this substance. Since the animals themselves produce carbon dioxide in their respiratory processes, any spot where a few have gathered (owing to the contact stimulus or for any other reason), becomes a centre for the produc- tion and diffusion of this substance. Therefore other Paramecia collect here ; more carbon dioxide is produced ; more Paramecia collect, and in time a dense aggregation is formed. It was farther shown that this 1 Pfeffp:r: Untersuchungen aus dem Botanischen Institut. Tubingen, 1888, ii, p. 618. - Jennings: Journal of physiology, 1897, xxi, pp. 258-322. 2 34 H. S. Jennings mid E, M. Moore. effect of carbon dioxide is due to the fact that it forms in water an acid solution ("carbonic acid") — the Paramecia collecting in the same way in any substance having a weakly acid reaction. As the other infusoria which form similar aggregations of course likewise produce carbon dioxide in their respiratory processes, it seems very probable, as was pointed out in the paper just referred to, that the same factors are at work here as in the case of Parame- cium ; that the 'spontaneous aggregations formed are due to the ten- dency of the organisms to collect in carbonic acid. This probability has been set forth also by later investigators, as for example in the recent paper of Rothert.^ But no one has hitherto undertaken to determine by experiment in how far this may be true. This is the problem which the study here presented has attempted to solve for a certain number of infusoria (sixteen species). The primary question to be answered is therefore as follows : — Are the spontane- ous aggregations formed by certain species of infusoria due to their gathering in carbon dioxide excreted by themselves } The investiga- tion involved a test of the reactions of the organisms studied both to carbon dioxide and to acids in general, and at the same time brought out a number of points as to the method in which the reactions of the organisms are produced ; these secondary matters are likewise set forth briefly in the following paper. Methods. The method of experimentation most used was that described in the first and second ^ papers in this series of studies. The organisms were studied in a thin layer of water, by mounting them on a slide covered with a large cover glass supported near its ends by slender glass rods. Their reactions were tested by introducing with a capil- lary pipette a drop of the substance in question beneath the cover glass, or in some cases by allowing it to diffuse inward from the side of the cover glass. In the case of gases, as carbon dioxide, it was found very convenient to proceed as follows. The gas is introduced into a large rubber bulb, such as is used with syringes or atomizers. To this is attached by the rubber tube a glass tube drawn out to a fine point. By inserting the point beneath the cover glass and pressing the bulb, a bubble of gas is introduced into the preparation. Where ^ RoTHERT : Flora, 1901, Ixxxviii, p. 402. ■^ Jennings : This journal, 1899, ii, pp. 311-341. Reactions of hifusoria to Carbonic and other Acids. -'OD still different methods were used, these are mentioned in the account of results. In attempting to determine the reactions of organisms to carbon dioxide, it is of course absolutely necessary that there should be no considerable quantity of this substance already present in the water. And since the organisms are continually producing carbon dioxide in appreciable quantities, some method of getting rid of the gas is a prac- tical requirement of the highest importance. The simplest method is to aerate the water thoroughly immediately before each test. This may be done as follows. Place a few drops of the fluid containing the organisms, — as much as will be placed on the slide at once, — in a watch glass ; then with a clean pipette inject it repeatedly over the surface of the watch glass, force bubbles into it, and mix it thoroughly with the air. Then place on the slide, cover, and perform the tests at once. Repeat the aeration before every test, as it requires only a very short time for water crowded with organisms to become impregnated with carbon dioxide. Of course it is not reasonable to expect organisms to gather in carbon dioxide when the water in which they are found already contains this substance in the optimum concentration. This precaution is equally necessary in testing other acids, as it is the com- mon factor in all acids to which the effects of the carbon dioxide are due. This or an equally efificacious method of aeration is an absolute necessity, if clear cut and constant results are to be obtained with carbon dioxide or acid solutions in general. This cannot be too much insisted on. Sometimes definite reactions will be obtained without aerating the water, in case it happens not to be already impregnated with carbon dioxide, but a little later the same organ- isms may give negative results. A second precaution worthy of mention is the necessity of having the water containing the organisms relatively free from debris: — fila- mentous bacteria, and the like. Most of the infusoria are markedly thigmotactic, tending to come to rest upon coming in contact with small solid bodies. If a preparation contains a network of fine bacte- rial filaments, frequently the infusoria will not gather in the acid at all, but remain at rest on the filaments, while if the filaments are removed, as by straining through coarse cloth, marked positive re- action is at once obtained. No attempt was made to determine quantitatively the exact strength of solution to which the organisms react. The purpose of the work 236 H. S. Jennings and E. M. Moore. was to determine whether the organisms do or do not give at any concentration a certain reaction to the substances in question. This was accomplished by beginning with a concentration so slight that the organisms did not react to it at all, and gradually increasing the strength till the solution is destructive. Somewhere between these limits will be found the characteristic reaction of the organisms. The value of quantitative determinations of the exact concentrations of acids to which the organisms react is largely illusory, in the majority of cases, as this varies with the amount of carbon dioxide present in the water, — -a. factor not under exact control. It varies also apparently with organisms from different cultures, and with the thickness of the layer of water in which the infusoria are confined. In experimenting with carbon dioxide especially, it is impracticable to attempt the use of solutions of known strengths; the introduction of a bubble of gas into the preparation gives all concentrations, from saturation next to the bubble to zero at some distance from it. In the following account of the work, the organisms will be taken up in the order suggested by the nature of the results obtained. A. Organisms which collect in Solutions of Carbonic and OTHER Acids. Chilomonas Paramecium. — This small flagellate is perhaps the com- monest and most abundant member of the group to which it belongs. It is therefore the most accessible form for experimentation on the Flagellata, and it will probably usually be employed when work on this group is undertaken. It is therefore important that the funda- mental facts as to its reactions should be well established. An exten- sive piece of work has already been done by Garrey^ on the reactions of this organism to chemicals, especially to acids. To our great regret we were compelled to come to results essentially different in some respects from those set forth by Garrey. It is unfortunate that there should be such disagreement, as this is likely to result in leaving the subject doubtful in the minds of other investigators. We believe however that we are able to point out exactly the factor to which the differing results are due, and to show that Garrey's results would probably not have differed from our own if this factor had been taken sufficiently into consideration. This factor is the normal pres- ^ Garrey: This journal, 1900, iii, pp. 291-315. Reactions of Iiiftisoria to Carbonic and other Acids. 237 ence of an acid, — a solution of carbon dioxide excreted by the organ- isms, — in the fluid in which Chilomonas occurs. Reaction to carbon dioxide. — Water containing Chilomonas is aerated in the manner above directed, and any bacterial filaments are removed by straining through coarse cloth. It is then placed on a slide, covered, and a bubble of carbon dioxide introduced. At first there is no gathering of the organisms, but soon they begin to collect about the bubble of gas, and gradually a dense ring is formed. Fig. i gives the general appearance of the progress of the experiment ; it was taken from an actual preparation. This experiment we have repeated many times, always (when the conditions were properly fulfilled) with the same results. The ex- FiGURE \,a. Figure \,b. F"iGURE 1. — Reaction of Chilomonas to a bubble of CO.j. a, Preparation immediately after the introduction of the bubble, before the organisms have collected, h, The same preparation a few minutes later, showing the dense collection of infusoria about the bubble. periments succeed equally well in the apparatus used by Garrey, and figured on page 294 of his paper. With a long capillary pipette a bubble of carbon dioxide can be introduced into the chamber beneath the cover glass. The flagellates at once gather about it in a dense ring, while they do not thus gather about bubbles of air similarly introduced. Chilomonas thus gathers about bubbles of carbon dioxide in dense collections, just as Paramecium does. The conditions above referred to as necessary of fufilment are (i) that the carbon dioxide should be properly removed from the water just before making the experiments ; (2) that the bacterial filaments and other debris in the water should be largely removed. The justification of the first condition is at once seen. It is idle to test the organisms with carbon dioxide when they are already immersed in a solution of that substance. It is undoubtedly to a neglect of this precaution, which is nowhere so much as referred to by Garrey, that the negative results of this investigator are due. 23S H. S. Jennings and E. M. Moore. The necessity for the removal of the debris is evident on examining the behavior of the organisms. Chilomonas is very strongly thig- motactic ; if when swimming through the water it comes in contact with a bacterial filament or bit of debris of any sort, it at once attaches itself by one of its two flagella, and comes to rest. Thus, in a preparation containing such filaments, all the individuals will soon be found quietly attached, which of course prevents their collecting anywhere. Reaction to other acids. — To what factor is the collecting in the solution of carbon dioxide due.^ Is it, as in the case of Paramecium, due to the acid qualities of this solution (to the H ions, according to the dissociation theory) .-' To answer this question, tests were made Figure 2, a. Figure 2, b. Figure 2. — Reaction of Chilomonas to a drop of -ij % HCl. a. Preparation immediately after the introduction of the drop (no organisms either within or gathered about the drop), b. The same preparation a few minutes later. with Other acids, and the organisms were found to collect in weak solutions of these exactly as in the carbon dioxide solution. These results were clear cut and unmistakable ; they were obtained with hydrochloric, nitric, sulphuric, acetic, formic, butyric, propionic, citric and oxalic acids. As a control, the organisms were tested with distilled water; no gathering was formed about or in it at any time, but the organisms remained quite neutral in their behavior toward it. The details of the phenomena vary with the strength of the acid solution used. With a stronger solution (say -^-^ per cent HCl), the animals gather in a dense ring about the margin of the drop, leaving vacant the area within, containing the stronger acid (Fig. 2). With a weaker solution the organisms gather into the interior of the drop, leaving no part of it vacant (Fig. 3). No characteristic difference was to be observed between their behavior to inorganic acids, such as HCl and HNO3, and that toward organic acids, such as acetic and butyric, save that of course Reactions of Infusoria to Carbonic and other Acids. 239 different concentrations were required to produce the same result. In Carrey's work, the results obtained with inorganic acids differed from those obtained with certain organic acids. Garrey studied the organisms in a much thicker layer of water, and introduced the acid through a tube-like opening at one side of the preparation. He observed that a dense gathering was formed about hydrochloric acid, but explained this as follows : When the strong acid reaches the organisms they begin at once to swim violently. This soon takes them outside of the area of acid, leaving it clear. On reaching the outer boundary, they stop, since there is no further cause for move- ment, thus forming a dense aggregation just outside the drop. " That in the zone surrounding the area there is a dense gathering (in other Fkutre 3. Figure 3. — Collection of Chilomonas within a drop of j^q % HCl. words that there is a ring formation ) is in my opinion due to the fact that those individuals which were in the clear area are now gathered in the space immediately surrounding it " {loc. cit., p. 296). The result is thus in a sense accidental, and would occur with any chemi- cal which had the property of setting the organisms in violent move- ment. In Garrey's experiments no gathering was ever formed in the centre of a drop of inorganic acid, no matter how weak it was. In our own work, the gatherings which occurred in drops of inor- ganic acids, were clearly not explicable in this manner, (i) When the drop was first introduced (Fig. 2 a) there were no individuals either within the drop or forming a ring about it. Later a dense ring was slowly formed (b). As there had been no individuals within the drop, of course the ring was not formed by individuals moving out of the drop and then stopping. Moreover, the process of ring formation is clearly evident to observation ; it is due to the swimming of organisms into the ring from outside. (2) If a weak solution of acid is used, it is at first empty, but later becomes com- pletely filled with organisms (Fig. 3). This of course could not possibly take place in the way assumed by Garrey. Collections 240 H. S. Jennings and E. M. Moore. of this sort were observed in the case of all inorganic acids studied. They were not formed when distilled water alone was used. (3) The gatherings about bubbles of carbon dioxide of course could not occur in the manner described by Garrey, as the bubble contained no individuals to move out and form a ring. In the case of the organic acids,, the phenomena observed by us were precisely parallel to those occurring in inorganic acids. It is possible that the gatherings formed last longer in the case of some organic acids, though to us this seemed not usually very marked. The results obtained by Garrey with organic acids varied much in different cases. With oxalic, formic, citric, succinic and valerianic acids the phenomena were the same as with inorganic acids; i.e., a clear area was formed, sometimes with a ring of organisms surround- ing it (if the acid was strong); this ring formation Garrey explained as in the case of inorganic acids. " Malic, tartaric and mandelic acids produce a clear area, often with a ring about it. In the formation of the ring, the phenomena were so inconstant that I was unable to say that it was or was not due to a migration of the organisms from without to it " (^loc. cit., p. 307). Finally, with acetic, butyric and lactic acids, a clear area surrounded by a dense ring was formed, and Garrey was able to assure himself that the ring formation was due to a migration of the organisms from the outside. Thus Garrey's results with acids can be placed in three categories, (i) Some showed a ring formation, which in the author's opinion was due merely to the driving of the organisms out of the area in which the acid was found. (2) Some gave such inconstant results that the author was not certain what he should conclude about them. (3) Some showed a ring formation of such density and clearness that it was evident that the organisms came from outside of the acid area. Now this inconstancy and uncertainty in the results with acids is exactly what is obtained when the carbon dioxide is not removed from the water by thorough aeration before each experiment. The culture water contains varying amounts of carbon dioxide and in some cases a part of it is accidentally driven off in the manipulations pre- paratory to the experiments, in other cases not. The presence of carbon dioxide means also the presence in the water of whatever it is that gives acids their characteristic qualities. Hence in such a fluid the organisms are already in an acid solution and naturally do not Reactions of hifiisoria to Carbonic and other Acids. 241 react with any precision when an acid is introduced, while in cases where the carbon dioxide is partly or entirely driven off, distinct reactions are obtained ; the results thus become inconstant and uncertain. After careful removal of the carbon dioxide before every experiment, the results with all acids are, according to our observations, essentially the same, — i.e., a ring or group is formed by immigration of the organisms from the outside. It was this same neglect to remove the carbon dioxide from the water that led Garrey to deny the results with Paramecium, though these are demon- strable with ease. In view of the contrasted results obtained by Garrey on the one hand and by ourselves on the other, it is much to be desired that some third person should reinvestigate the reactions of Chilomonas, testing the various methods used, observing the precautions set forth, and perhaps taking counsel by correspondence or conversa- tion with both sides, that there may be no omission which might seem to vitiate the work. It remains of course possible that Chilo- monas from different cultures reacts differently, though we have used dozens of different cultures and have observed no such difference. Our own results on Chilomonas may now be summarized. This organism reacts to carbonic and other acids just as Paramecium does, forming dense collections in localized areas, where carbon diox- ide is present. The spontaneous aggregations sometimes formed by Chilomonas may therefore be due to their collection in carbon dioxide excreted by themselves. Cyclidium glaucoma. — This ciliate infusorian likewise gathers in carbon dioxide and in solutions of acids in general. The collections thus formed are dense and lasting. Cyclidium was not observed to form spontaneous collections, though this may occur. Colpidium colpoda. — This is one of the infusoria which was described by Pfeffer as collecting spontaneously into groups. It reacts to solutions of carbon dioxide and other acid solutions, just as Para- mecium and Chilomonas do, gathering in dense aggregations about a bubble of CO.,, or in a drop of weak acid. It is therefore probable that the spontaneous groups are due to carbon dioxide. If Paramecium and Colpidium are mounted together, they will gather spontaneously into groups, each group containing both kinds of infusoria, the boun- dary of the groups being practically the same for each. The cause of the grouping is thus evidently the same in the two cases. 242 H. S. Jennings and E. M. Moore. B. Organisms which form Spontaneous Gatherings, but do NOT COLLECT IN SOLUTIONS OF CARBONIC OR OTHER AciDS. Oxytricha aeruginosa. — This organism, when mounted on a slide, forms spontaneous groups which are similar in every respect to those formed by Paramecium. The method of reaction to a stimulus in Oxytricha is by backing, and turning to the aboral (or right) side, — the side which is not notched. If the organisms are at first scat- tered uniformly throughout the preparation, they will soon be found to be forming groups in one or more regions. If the individuals within the groups are observed, they are found to be swimming hither and thither in all directions. But when one comes to the outer boundary of the group, it at once swims backward a short distance, turns toward the aboral side, and then starts forward again. As this hap- pens every time the boundary of the group is reached, the animal remains within it. Individuals outside, whose course carries them by chance into the areas where a group is forming, do not react at all as they enter the area. But after swimming across, they do react as above described upon coming to the outer boundary of the area. Hence every Oxytricha that enters a group remains within it, and after a time a dense aggregation is formed. The groups thus pro- duced increase in area, spreading out regularly, but maintaining a definite boundary. The phenomena seem thus in every way identical with those ob- served in the case of Paramecium (see the first and second of these studies). It might therefore be reasonably expected that the cause would be found to be the same. But experiment shows that this is not the case ; Oxytricha aeruginosa does not collect in regions where carbon dioxide is. present, nor in other acid solutions. If a bubble of carbon dioxide is introduced into the preparation, the Oxytrichas do not gather about it, but on the contrary give their " motor reaction " when they come into its neighborhood, — reversing the direction of movement, and turning toward the aboral side. They thus leave the space about the carbon dioxide empty. Toward drops of acid solu- tions of all sorts they react in the same manner. If Oxytricha and Paramecium are present in the same culture, or if the two are mixed together and experimented upon in the usual way, the results are as follows. The Paramecia collect about the bubble of carbon dioxide, or in the drop of acid, at once; the Oxy- Reactions of Infusoria to Carbonic and other Acids. 243 trichas do not. Thus a separation of the two kinds of infusoria is soon brought about. If Oxytricha and Paramecium are mounted together and the slide is allowed to stand for a time, both kinds of infusoria will form spon- taneous groups, but the two groups are quite separate. The Paramecia gather in one region, the (3xytrichas in another. Individuals of either kind may pass directly across the groups formed by the other, or swim in and out of the area where the other group occurs. The groups are thus clearly due to different causes in the two cases. Oxytricha therefore forms spontaneous gatherings similar to those of Paramecium, but not due to the same cause. It seems evident that Oxytricha must excrete some other substance, not an acid, which acts upon it in the same way that the excreted carbon dioxide acts upon Paramecium. The nature of this substance remains to be discovered. Loxocephaius granulosus. — In the case of this organism the facts are closely parallel to those described for Oxytricha aeruginosa. It forms spontaneous gatherings, but does not collect about bubbles of carbon dioxide nor in acid solutions in general. Mounted on the same slide with Paramecium, the two organisms form separate groups in different regions of the preparation. Clearly, Loxocephaius, like Oxytricha, excretes some substance which brings about the collec- tions, but this substance is not carbon dioxide. In preparations containing both Paramecium and Loxocephaius, the following may be observed as to the relations of the two organ- isms. Loxocephaius swims in and out of the groups of Paramecia, paying no attention to the limits so strictly observed by the Paramecia. In the same way Paramecium swims indifferently in and out of the groups of Loxocephali, when the latter groups are first forming. But after a group of Loxocephali has become well established and contains very large numbers of individuals, a Paramecium passing accidentally into the group usually remains there. Thus after a time a considerable number of Paramecia may be mingled with the Loxo- cephali. The Paramecia swim about freely within the group, but turn back on coming to an outer limit. It is to be noted that this outer limit is not the same as that which turns back the Loxocephali, but lies a little outside of it, so that the area in which the Paramecia are confined is larger than that which limits the Loxocephali, inclos- ing the latter. These phenomena are probably to be explained as follows. Loxocephaius is not affected by carbon dioxide, therefore does not 244 ^' ^- Jennmgs and E. M. Moore. gather in the groups formed by the Paramecia, but swims in and out of them indifferently. But it does excrete some other substance, not of an acid nature, into which it gathers; hence the spontaneous col- lections formed. To this substance Paramecium is indifferent, hence it swims indifferently in and out of the groups of Loxocephali, at first. But of course Loxocephalus produces carbon dioxide in its respiratory processes, hence after a group of these organisms has been formed for some time, the water becomes impregnated with carbon dioxide, as well as with the other (hypothetical) substance. Paramecia now passing into the group remain, owing to the carbon dioxide. The areas over which the carbon dioxide and the hypothetical substance are effective are not identical, that for the carbon dioxide being a little larger; hence the limit of the excursions of the Paramecia is outside that for the Loxocephali. C. Organisms which do not collect in Carbonic or other Acids, and which were not observed to form Spon- taneous Gatherings. The following organisms were tested with carbon dioxide and with solutions of various acids in the same manner as those hitherto described. Every precaution was taken to remove the carbon dioxide from the water before making the tests, and the experiments were repeated under various conditions, with uniform results. None of these organisms gather about bubbles of carbon dioxide or in solutions of acids. They are as follows: Oxytricha fallax, Euplotes charon, Stylonychia pustulata, Colpoda cucullus, Spirostomum teres, Stentor cseruleus, Enchelys farcimen, Halteria grandinella, Didinium nasutum, Euglena viridis, and Heteromita globosa. Some of these organisms, on coming in contact with a solution of carbon dioxide, at once give their characteristic " motor reaction," backing and turning toward a structurally defined side; thus they turn away from the area in question, leaving it empty. Others do not react at all to carbon dioxide, and to other acids only when very strong. Those that were indifferent were Oxytricha fallax, Stentor caeruleus, Didinium nasutum, Euglena viridis, and Heteromita globosa. Reactions of Infusoria to Carbonic and other Acids. 245 D. The Method by which the Gatherings are Brought About. Throughout the work attention was given to the method by which the infusoria gather together. The point which was especially studied was the question of orientation. Do the organisms collect in the region where a certain chemical is present because they become oriented in the lines of the diffusing ions? Or are the collections brought about in the manner described for Paramecium, in the first and second of these studies? The phenomena were carefully examined in all the infusoria in which collections were observed, — in Colpidium colpoda, Oxytricha aeruginosa, and Loxocephalus granulosus, and additional observations were made on Paramecium and Chilomonas. All the ciliates men- tioned are of sufficient size so that their movements can be exactly observed with the Braus-Driiner stereoscopic binocular, and there can be no doubt as to the method in which the gatherings take place. They collect in essentially the same manner as has been shown in previous studies to be true for Paramecium. The organisms are at first swimming freely hither and thither. When the drop of acid is introduced, or collections are produced in other ways in certain regions, some of the individuals swim into the area in question merely through their usual movements. They do not change their course or react at all as they enter the area. But as they swim across it and reach the opposite side, where they would if unchecked pass out of the area into the surrounding water, each infusorian gives its characteristic " motor reaction." Oxytricha after moving backward turns toward its un- notched side, Loxocephalus to the aboral side, Colpidium toward its convex side, Paramecium toward the aboral side. The animal is thus prevented from leaving the area containing the chemical, but swims in another direction within this area. As it reacts in the same way every time it comes to the outer boundary of the area, it does not leave it at all. Other individuals enter in the same way, through their ran- dom movements, and remain through the same reactions, so that after a time the areas in question swarm with infusoria. If the animals are allowed to come thoroughly to rest before intro- ducing the chemical, usually no collection is formed within it. This shows the essential part played in the reaction by the random move- ments of the organisms. It seems difficult for many minds to believe that the dense gather- 246 H. S. Jennings and E. A/. Moore. ings observed can be produced in this way. That this is the real method by which the collections occur, can be very neatly demonstrated to the eye in the following manner. A number of Paramecia or other infusoria which collect in acids are mounted on a slide. Upon the upper surface of the cover glass a small circle about the size of the drop of acid usually introduced, is made in ink with the pen. By directing the attention to the area within the ring of ink, it will be seen that many infusoria (as many as ten per second or more, in an ordinary mount of Paramecium), cross the area every instant. It is therefore evident that if all of them could be stopped within the area, a dense group would soon be produced. With the capillary pipette a drop of acid is now introduced beneath the ring; the same number of infusoria now enter the area as before, but every one remains and a dense collection soon results.^ In the paper already cited, Garrey maintains that the flagellate Chilomonas collects in certain acids in a manner entirely different from that above set forth. He holds that the collections (in acetic acid, for example), are produced through an orientation of the organ- ism in the lines of the diffusing ions. The reactions to other sub- stances, drops of which are left empty by Chilomonas (as for example a solution of sodium chloride), take place in a way entirely different, according to Garrey. Here there is no orientation ; the chemicals merely cause " swift shooting movements," by which the animal is carried out of the area, or prevented from entering it. The method of reaction exhibited in collecting in acetic acid is denominated by Garrey chemotaxis, while that shown in keeping out of or leaving a drop of sodium chloride he calls chemokinesis. In the sixth of this series of studies, reasons drawn from a study of the movements of the individual Chilomonads have been given for rejecting this distinction in kind between the reaction in collecting and that in avoiding a region containing chemicals. Certainly no such distinction can be made in Paramecium, nor in the other ciliates above mentioned. Leaving out of account the direct observations on the movements of the individuals, there are certain experiments which amount almost to a demonstration that there is no such distinction in kind, — even in Chilomonas. They demonstrate at least that collec- tions exactly similar to those produced through the supposed " chemo- 1 This experiment was demonstrated on the screen by means of the stereopti- con before tlie Society of Western Naturalists at the meeting in Chicago in December, 1900. Reactions of Infusoria to Carbonic and other Acids. 247 taxis " can be produced through the operation of the admitted " chemokinesis." Acetic acid may be taken as a type of the substances toward which, according to Garrey, Chilomonas shows orientation, or " chemotaxis ; " while sodium chloride is an example of the substances which cause no orientation, but merely "chemokinesis." If a drop of acetic acid of a proper concentration is introduced into a preparation of the in- FlGURE 4. Figure 5. Figure 6. Figures 4, 5, and 6. Diagrams showing how the grouping of the organisms depends on the relations of the two fluids to each other in space, a represents the area occupied by the fluid into which the infusoria may pass without giving the " motor reaction ;" h the fluid into which it cannot pass without giving the "motor reaction." When a is water, b is a salt or alkaline solution ; when b is water, a is an acid solution ; in either case the grouping of the organisms (Paramecium or Chilomonas), is that shown in the figures. fusoria, the latter soon collect in the drop ("chemotaxis"), while if a drop of sodium chloride solution is introduced, it remains empty ("chemokinesis"). But suppose we mount our infusoria in a weak solution of sodium chloride, and introduce a drop of water.'' The salt solution is admitted to produce no orientation, but merely "chemokinesis," yet in a short time the drop of water is filled with a dense group of infusoria, — just as was the acetic acid in the former case. Apparently the collection is formed in water just as quickly as in the acid, and the present authors have been able to detect no difference in the method of forma- tion. It is at least demonstrated that collections can as well be formed without orientation as with it, and that if these infusoria 248 H. S. Jennings and E. M. Moore. possess the power of becoming oriented to diffusing ions, this power is a useless luxury. According to our observations, the phenomena are identical in the two cases. The organisms swim about, in the solution in which they are mounted (water, or sodium chloride solution, respectively), and enter the drop (acetic acid, or water, respectively), without reaction. After having entered they give the usual " motor reaction" when they come to the outer boundary of the drop; hence they do not leave it, and the drop after a time swarms with the animals. The following series of experiments is instructive and brings out clearly the facts as they appear to the present authors. Mix a part of the infusoria (Paramecium or Chilomonas) with a weak solu- tion of sodium chloride, not strong enough to injure them, mix others with a weak, non-injurious solution of acetic acid, and leave others in water. Now make mounts with the fluids in various relations to each other : (i). Make a preparation (Fig. 4) in such a way that half the fluid on the slide is water (a) containing infusoria, while the other half is salt solution (h) containing infusoria. After a short time most of the infusoria will be in the half containing water alone. (la). Make a similar preparation (Fig. 4), save that one half {a) is acetic acid containing infusoria, the other half (/;) water containing infusoria. In this case after a time most of the organisms will be found in the acid. (2). Make a preparation (Fig. 5) in sucli a way that the salt solution sur- rounds the drop of water, the water {a) being introduced as a drop into the salt solution (b). After a time the drop i^a) of water contains a dense swarm of the organisms. (2a). Make a preparation as in the last, save that water (Fig. 5, V) sur- rounds the acid {a), which is introduced as a drop into the water. In this case there is likewise a dense aggregation formed in the drop a (of acid). (3). Make a preparation (Fig. 6) such that the water (a) surrounds the salt solution (^), — the latter being introduced as a drop into the water. After a short time the drop b (of salt solution) is empty. (3a). Make a similar preparation, in which the acid solution (Fig. 6, a) surrounds the drop of water {b). Soon the drop b (of water) is left empty. With the same pair of substances we get, therefore, either a dense aggregation (or what has been sometimes called " positive chemo- taxis "), or a certain definite area left vacant ( " negative chemotaxis "), depending upon the relation in space of the two fluids to each other. And this result may be obtained whether we use as our chemical one like acetic acid, to which it has been maintained that the infusoria Reactions of Infiisoj'ia to Carbonic and other Acids. 249 show positive "chemotaxis" proper, or whether we employ a salt to which they are held to react only by "chemokinesis." Thus with either pair of fluids, whether we do or do not get a dense aggregation of infusoria depends " on the configuration of the two fluids" — on the relation of the two fluids to each other in space. General statements embodying these relations may be made as fol- lows. If we distinguish as b that fluid into which the infusorian cannot pass without causing the " motor reaction," as a that into which it can pass without causing the reaction, then If b surrounds a (Fig. 5), a dense aggregation is formed in a (" positive chemotaxis "). If a surrounds b (Fig. 6), the small area b is left empty (" negative chemotaxis"). If a and b occupy equal areas (Fig. 4), after a time most of the organisms will be found in a. (This last case is not so strongly realized in a minute organism like Chilomonas as in a larger creature, such as Paramecium, because the distances to be passed over are so great that a weak swimmer like Chilomonas will not soon reach the area a, and may come to rest in large numbers in b without reaching rt at all. But in any case, a considerable majority will be found in a.) If a is water, /; may be a solution of an alkali or of a great variety of neutral salts; in the case of Paramecium, almost any neutral salt If b is water, a may be any acid. In either case the resulting phe- nomena will be essentially the same. Summary. 1. In order to test the reactions of infusoria to acids, it is necessary to remove with great care from the water containing the organisms the carbon dioxide produced by the organisms in their respiratory processes. 2. Colpidium colpoda, Cyclidium glaucoma, and Chilomonas Para- mecium collect in solutions of carbonic and other acids, just as Para- mecium does. The spontaneous collections formed by these organ- isms may therefore be due to their excretion of carbon dioxide. 3. Loxocephalus granulosus and Oxytricha aeruginosa form spontaneous collections similar to those of Paramecium, but do not gather in carbonic or other acids. The spontaneous collections in these cases must therefore be due to other causes. 4. The following infusoria do not collect in carbonic or other acids, 250 H. S. Jennings and E. M. Moore. nor were they observed to form spontaneous gatherings : Oxytricha fallax, Euplotes charon, Stylonychia pustulata, Colpoda cucullus, Spirostomum teres, Stentor casruleus, Enchelys farcinien, Halteria grandinella, Didinium nasutum, Euglena viridis, Heteromita globosa. 5. The collections, according to our observations, take place in the manner described in previous numbers of this series of studies for Paramecium. In cases where this has been disputed, it is shown that collections essentially similar to those produced by what has been considered " chemotaxis " proper are likewise produced by what is admittedly " chemokinesis." THE MOVEMENTS OF THE INTESTINES STUDIED BY MEANS OF THE RONTGEN RAYS.^ By W. B. cannon. [From the Laboratory of Physiology in the Harvard Medical School^ CONTENTS. Page Introduction 251 The method 254 The movements of the small intestine 256 Rhythmic segmentation of the intestinal contents 256 Peristalsis 260 Rhythmic segmentation and the pendulum movement 261 The course of the food in the small intestine 262 The competence of the ileocaecal valve 264 The movements of the large intestine 264 Antiperistalsis in the colon 265 The changes when food enters the colon 267 The appearance of tonic constrictions . 268 Defecation 269 The question of antiperistalsis 271 The effect of emotions and sleep 275 Introduction. THE investigation of intestinal movements has been beset by the same difficulties that characterized the investigation of the gastric mechanism. Pathological subjects or animals subjected to the disturbing action of drugs and anaesthetics and of serious opera- tions, have been the only sources of our knowledge. A considerable difference of opinion as to the nature of the normal movements in the intestines has resulted from observations made under these neces- sarily abnormal conditions. The slowly-advancing peristaltic wave and the Pendelbezvegung, or swaying movement, described by Ludwig, have been regarded as true physiological processes. Concerning anti- peristalsis and the swiftly-running vermicular contraction, observers are not so nearly in agreement. The. activity of the large intestine ^ The results of this investigation were reported to the Boston Society of Medical Sciences, November 19, 1901. 251 252 W. B. Caiinoit. has been described as simply peristalsis of a slower rate than that seen in the small intestine. The best known of the intestinal movements is the normal peri- staltic wave. This wave is slow, having a rate of about two cen- timetres per minute/ is regular, and by most observers is said to move always in one direction. The progress of the contraction, as suggested by Nothnagel's experiments,^ and, as clearly stated by Mall and by Bayliss and Starling, is dependent upon a local reflex. According to Mall,'^ when an object stimulates the mucosa there occurs above the point of stimulation a constriction which forces the object downward ; and as it moves downward new regions immediately above the mass are by this reflex brought into constriction, and thus the wave and its stimulus advance together. " At the same time," Mall states, " a sucking force, due to active dilatation below the body, may have a tendency to drag it down." In what manner an active dilatation of the intestinal wall may occur so as to produce a " sucking force," he does not make wholly clear. Bayliss and Starling, in describing normal peristalsis in the intestine, state that the contractions above the bolus increase until there is a strong tonic constriction.^ This passes the bolus onward, and as it advances the ring of constriction follows it. While the bolus is passing down, the gut above it is traversed by waves running as far as the con- stricted ring. These observers state the law of intestinal peristalsis thus : " Local stimulation of the gut produces excitation above and inhibition below the excited spot." The pendulum movements are characterized by a gentle swaying motion of the coils, and are accompanied by rhythmical contractions of the intestinal wall. They continue with undiminished force after paralysis of the local nervous mechanism by nicotine or cocaine; they have been called, therefore, myogenic or myodromic contractions. Ob- servers have described them variously as shortenings and narrowings of the gut, rhythmically repeated at nearly the same intestinal circumfer- ence ; " as alternating to-and-fro movements of the long axis without ^ Cash : Proceedings of the Royal Society, 18S6, xli, p. 227. 2 NoTHNAGEL : Aixhiv fiir pathologische Anatomic unci Physiologic, 1882, Ixxxviii, p. 5. ^ Mall: Johns Hopkins Hospital Reports, 1896, i, p. 51. '' Bayliss and Starling : Journal of physiology, 1899, xxiv, p. 106. ^ Ludwig: Lehrbuch der Physiologic des Menschen. Leipzig und Heidelberg, 1861, ii, p. 615. Study of MoveniC7its of Intesti7ies by Routgeu Rays. 253 changes in the lumen ; ^ as local or extensive periodic contractions and relaxations mainly of the circular musculature ; ^ and as waves involving both muscular coats of the intestine, and travelling normally from above downward at a rapid rate (2 to 5 cm. per second).^ They have been seen in the dog,'* and in the rabbit and cat.'' In the cat Bayliss and Starling noticed that when the lumen of the gut was distended by a rubber balloon, there appeared rhythmical contractions, which nearly always were most marked at about the middle of the balloon, /. r., the region of greatest tension. This form of constric- tion, which, as my observation shows, is an indication of the manner in which the rhythmical contraction acts in the cat's intestine, Bayliss and Starling seem to have regarded with slight attention, since it did not accord with the law of peristalsis. The swift vermicular wave may pass the whole length of the intes- tine in about a minute. It is often seen just after death, as well as in pathological states such as intestinal anaemia or hypersemia, and when the bowel contains gases and organic acids from decomposing food.^ Starling is inclined to regard this type of intestinal activity as an ex- aggeration of the rhythmic type; Mall, on the other hand, places it in a class by itself, and declares that its service is to rid the intestine rapidly of irritating substances. Nothnagel, who designates this form of movement with the term Rollbez^'cgung, thinks it is transitional between a physiological and a pathological activity. The existence of antiperistalsis has been so much questioned that it will be considered in a special section of this paper, where my observations may be conveniently introduced. The common understanding of the manner in which food passes through the intestinal canal is that the chyme ejected from the stomach is pressed downward by a peristalsis, which passes slowly over a part or all of the small intestine. The peristaltic waves of the colon are supposed to constitute an independent group, similar to those of the small intestine, but weaker and slower. Movements 1 Raiser: Beitrage zur Kenntniss der Darmbewegung. Dissertation. Giessen, 1895, p. 7. Nothnagel: Die Erkrankungen des Darms und des Peritoneum. Wien, 1898, i, Darmbewegung. p. I. 2 Mall : Loc. cit., p. 48. 3 Bayliss and Starling : Journal of physiology, 1899, xxiv, p. 103. 4 Ibid. ° Bayliss and Starling: Journal of physiology, 1901, xxvi, pp. 127 and 134. ^ Bokai : Archiv fiir experimentelle Pathologie and Pharmakologie, 1887, xxiii, p. 232. 2 54 ^'^- B- Cannon. of the food other than the uninterrupted advance have been men- tioned by some observers. Starling ^ states that the effect of the pen- dulum movement is to mix the contents of the intestine and bring them into intimate contact with the mucous membrane. GriJtzner writes that he has been brought " by strange and peculiar observa- tions " to believe that the fluid contents of the small intestine move irregularly forward, then forward and back, then perhaps remain quiet for some time, only to pass backward for a long distance and finally to move forward steadily to the end. In this manner the food is mixed, and brought into contact with the absorbing walls.- The to- and-fro shiftings of the food Griitzner ascribed to advancing and retrograde contractions of the intestinal musculature, and he argued that even circular constrictions must force the liquid contents away in both directions. To support his contention Griitzner introduced mercury into the intestine and observed it with the Rontgen rays. After noting a backward and forward movement of the mercury he dismissed the method, saying, " Many a flash must come from the Rontgen tube before the normal movement of the intestinal contents is made entirely clear by this method." The following account is a summary of many repeated observations on different animals, and is a contribution to a clear understanding of the normal movements of the intestines and their contents. The Method. The method employed in this investigation is identical with that used in 1897 to observe the movements of the stomach.^ Sub- nitrate of bismuth, one-tenth to one-third the weight of the food, was mixed with what the animal ate. Thus far observations have been made almost exclusively on cats. The subnitrate of bismuth was generally mixed with canned salmon, — a food which cats relish. The animal to be observed was usually not allowed to eat anything during the day previous to the observation, and moreover was com- monly given from four to six teaspoonfuls of castor oil to clear the bowels. A tranquil mood on the part of the animal was found to be quite as necessary for seeing movements of the intestine as it was for ^ Starling : Text-book of physiology, edited by Schafer. Edinburgh and London, 1900, ii, p. 330. 2 Grutzner : Archiv fiir die gesammte Physiologic, 1898, Ixxi, p. 515. 2 Cannon : This journal, 1898, i, p. 362. Study of Movements of hitestines by Rout gen Rays. 255 J?^- securing normal activity of the stomach. For this reason female cats, which submit quietly to the confinement of the holder and the straps, proved to be much' more favorable subjects than the males, which struggle violently when tied into the hammock. Curiously the crackling and rumble of the static machine, which generated the electricity for producing the X-rays, instead of frightening the ani- mal had often a soothing effect. The appearance of the food in the alimentary canal is shown in Fig. I, the reproduction of a radiograph taken five and three-fourths hours after the animal finished eating. The cat lay on her back, and the photographic plate was placed over the front of the abdomen. The intestines move up and down in the body cavity with each respiration ; in order to secure clear outlines, a leaden plate was slipped be- tween the cat and the Crookes tube during inspiration and the beginning of expiration, and then removed till the beginning of in- spiration. Thus the plate was exposed to the rays only during the pause recurrent at the end of each expiration, when the shadows resume approximately their former position. Records were taken, both by means of radiographs and by means of tracings made with a soft pencil on tissue paper laid over the fluorescent surface of the screen. The reliability of the latter method has been proved by comparison with radiographs taken immediately before or afterwards. In Fig. I the spinal column is seen in the middle line with the pelvis below. On the right, above, is the pyloric end of the stomach, and on the left, dimly outlined as a lighter area, because of the pres- ence of gas, is the ascending colon. In the caecum there is a small amount of food present. Loops of small intestine containing food are to be seen in various parts of the abdomen. These loops are often distinct enough to allow movements in them to be seen without any manipulation ; when this is not the case, however, and the loops overlie one another, as on the right side of Fig. i, a slight pressure with the fingers through the abdominal wall will readily separate from neighboring loops the one to be observed. Figure 1. — Appearance of food in the intestines 5f hours after eating. This and other radiographs re- duced two-thirds. 256 W. B. Cannon. The Movements of the Small Intestlxe. When the food has been distributed through the intestine so as to present the appearance shown in Fig. i, a noticeable feature in most or all of the loops is the total absence of movement. If the animal remains quiet, however, only a few moments elapse before peculiar motions appear in one or another of the loops, or perhaps in several, and last for some time. These motions consist in a sudden division of one of the long, narrow masses of food into many little segments of nearly equal size; then these segments are again suddenly divided and the neighboring halves unite to make new segments, and so on, in a manner to be more fully described. I have called this process the rhythmic segmentation of the intestinal contents. Further /"^ C^ rP) r^ r^K ^ observation reveals peristalsis ^^ ^\ J V / \ J \ /^"^ CD here and there, and under cer- (^ C£) CD CZ) CZ) C^ ^ ^^^" circumstances the typical \ I \ ^ yj yj ^ ^ *■ ^ \ / ^ swaying movements may be O C-^ ~^ ^^ ^-^ O Q seen. All these phenomena Figure 2. — Diagram representing the process are noW tO be Considered in of rhythmic segmentation. Lines 1, 2, 3. 4 detail. indicate the sequence of api^earances in the ,-,, ^i • ^ ^. r ^1 „„,,■,. , , r Rhythmic segmentation of the loop. 1 he clotted hnes mark the regions of division. The arrows show the relation of intestinal contents. — This is the particles to the segments they subse- by far the most COmmon and quently form. ^^^ most interesting mechan- ical process to be seen in the small intestine. The nature of the process may best be understood by referring to the diagram in Fig. 2. A string-like mass of food is seen lying quietly in one of the intes- tinal loops, line i, Fig. 2. Suddenly an undefined activity appears in the mass, and a moment later constrictions at regular intervals along its length cut it into little ovoid pieces. The solid string^ is thus quickly transformed, by a simultaneous sectioning, into a series of uniform segments. A moment later each of these segments is divided into two particles, and immediately after the division neigh- boring particles (as a and d, line 2, Fig. 2) rush together, often with the rapidity of flying shuttles, and merge to form new segments (as c, line 3, Fig. 2). The next moment these new segments are divided. 3 A 4 ^ In lieu of any better short e.xpression, "string" of food is used to designate the long slender mass of the contents lying in a loop of the intestine. Study of Movements of Intestines by Rontgen Rays. 257 and neighboring" particles unite to make a third series, and so on. At the time of the second segmentation (Hne 3, Fig. 2) the end par- ticles are left small. Observation shows that these small pieces are not redivided. The end piece at A simply varies in size with each division ; at one moment it is left small, at the next moment it is full size from the addition of a part of the nearest segment, and a moment later is the small bit left after another division. The end piece at B (probably the rear of the mass) shoots away when the end mass is divided, and is swept back at each reunion to make the large end mass again, only to be shot away and swept onward with each recur- rence of the constrictions. Thus the process of repeated segmenta- tion continues, with the little particles flitting toward each other and the larger segments shifting to and fro, commonly for more than half an hour without cessation. P^rom the beginning to the end of a period of segmentation the food is seen to have changed its position in the abdomen to only a slight extent; whether this change is a passing of the food along the loop, or a movement of the loop itself, it is impossible to tell from the shadows on the screen. The change of position, however, is much less conspicuous than the lively divi- sion and redivision which the mass suffers so many times from the busy, shifting constrictions. From this typical form of rhythmic segmentation there are several variations. Sometimes, and especially when the mass of food is thick, the constrictions do not make complete divisions and are so far apart that the intermediate portions are relatively large. Moreover the con- strictions do not take place in the middle of each portion, but near one end ; thus each portion is constricted, not into halves, but into thirds. If a little pointer is placed at the middle of a segment, when the segments are completely divided into halves, in a few seconds the pointer will be in the middle of the clear space between two segments ; but in a few seconds more the first phase will return and the pointer will again be indicating a segment, — two operations intervene be- tween similar phases. When, however, the portions are constricted into thirds, the indicator shows it, since three operations intervene between similar phases. The manner of these changes is made clearer by reference to the diagram in Fig. 3. That each portion is con- stricted into three pieces is proved also by watching the gradual reduc- tion of the portion at the left end of line i through lines 2 and 3 ; and also in the gradual formation of a full-sized portion at the right end of lines 2, 3, and 4. When food undergoing this process is watched, it 258 W. B. Cmino7i. appears to be affected by a series of constrictions, each of which starts at one end of the mass and marches through to the other end, leaving its impress at short intervals along the length. The progression of the dotted lines from right to left in a, b, c, and d, etc., Fig. 3, gives a notion of these advancing constrictions. Another variation of the segmentation is shown in Fig. 4. In this type there are evidently divisions and subdivisions, i. c, one more operation between the appearance and the reappearance of the same phase than is present in the simple division of the small segments in a long string of food (Fig. 2). This form of segmentation is fairly typical for the constrictions seen in food advancing through the in- testine. Sometimes the divisions occur in the middle of a long string of food and leave the ends a '^ wholly unaffected. A remarkable feature in the seg- mentation of the food is the rapidity with which the changes take place. The simplest way of estimating the rate of division is to count, not the number of times the partition Figure 3. — Diagram showing the rela- r ..i r i • ^1 r ^, . , , 01 the food recurs m the same tions of the portions when they are con-stricted into three pieces. Tlie place, but the number of different dotted lines indicate regions of con- sets of segments observed in a striction; the arrows indicate the re- j^^^^ -^^j^ yj^^^^ j^ pj ^j^^ lationship or the pieces to the portions they subsequently form. appearances of Imes I, 2, 3, 4, etc., would be counted, and not merely lines I, 4, etc. Repeated observations on different animals have shown that the most common rate of division in long, thin chains of food varies between 28 and 30 times in a minute, i. c, there is a change from one set of segments to another set every 2 seconds, and a return of the same phase every 4 seconds. In some cases the rate is as low as 23 times per minute. The larger masses seem to be associated with a slower segmentation ; the operations indicated in Fig. 3, for example, occurred from 18 to 21 times in a minute, so that the same phase reappeared only once in 8 or 9 seconds. The segmentation frequently continues for more than half an hour; in one instance it was seen to persist with only three short periods of inactivity for two hours and twenty-two minutes. At the rate of 30 segmentations per minute it is clear that a slender string of food may commonly undergo division into small particles Shidy of Movements of Intestines by Rontgen Rays. 259 more than a thousand times while scarcely changing its position in the intestine. I have seen once, in a cat only lightly etherized, the exterior of an intestine which was dividing the food as above described. An hour and a half after a meal of salmon the anaesthetic was given, the abdomen opened, and the flaps raised so as to form walls. Warm salt solution was then poured into the abdominal cavity, and the floating coils left covered with the transparent omentum. The gastric peri- staltic waves were running regularly; on the intestine there were visible at various places during the period of observation regions of constriction which had the appearance shown in Fig. 3, except that the rings were relatively nearer together. New rings of constriction took place on the same side of all the bulging parts at the margin of the constricted portion (>/. dotted lines, Fig. 3). As new rings occurred the old relaxed, but apparently with tardiness, for the con- tents gurgled as if forced through the narrowed lumen. The con- strictions recurred irregularly and at much longer intervals than in the normal animal. The contracted rings were pale and bloodless. The effect of the process of rhythmic segmentation proves it an ad- mirable mechanism. The food over and over again is brought into closest contact with the intestinal walls by the swift kneading move- ment of the muscles. Thereby not only is the undigested food in- timately mixed with the digestive juices, but the digested food is thoroughly exposed to the organs of absorption. MalP has shown that contraction of the intestinal wall has the effect of pumping the blood from the submucous venous plexus into the radicles of the superior mesenteric vein and thus materially aids the intestinal cir- culation. Moreover, lacteals loaded with fat will in a few moments become empty unless the intestine is slit lengthwise so that the mus- cles cannot exert compression.^ The rhythmic constrictions, therefore, both propel the blood in the portal circulation and act like a heart in promoting the flow of lymph in the lacteals. This single movement with its several results is an excellent example of bodily economy : the repeated constrictions, as already shown, thoroughly churn the food and digestive fluids together, and also plunge the absorbing mucosa into the very midst of the food masses ; but not only are the pro- cesses of digestion and absorption favored by these movements, — they also, by compression of the veins and lacteals of the intestinal 1 Mall: Loc. cif., p. 68. ^ Mall: Loc. cit., p. 47. 26o W. B. Cannon. wall, serve to deport through blood and lymph channels the digested and absorbed material. Peristalsis.^ — The phenomena of peristalsis and segmentation are usually combined in some manner while the food passes through the small intestine. Peristalsis is observed normally in two forms ; as a slow advancing of the food for a short distance in a coil, and as a rapid movement sweeping the food without pause through several turns of the gut. The latter form is frequently seen when the food is carried on from the duodenum ; and it may readily be produced in other parts of the small intestine by giving an enema of soapsuds. When a mass of food has been subjected for some time to the seg- menting activity of the intestine, the separate segments, instead of being again divided, may suddenly begin to move slowly along the loop in which they lie. That this movement is not a swinging of the coil as a whole, but a peristaltic advance of separate rings of its circu- lar musculature, is made probable by the fact that the succeeding seg- ments follow along the same path their predecessors CZ) 1 have taken. The advance of the little pieces may A — \ continue for seven or eight centimetres, when finally — > the front piece stops or meets other food. Then all the succeeding pieces are swept one by one into — ' CZ) 4 the accumulating mass, which at last lies stretched P^igure4 — Dia- ^-long the intestine, a solid string manifesting no gram showing . sign of commotion. combined peri- Another form of slow peristalsis is frequently s a SIS anc seg- observed when the food is pushed forward, not in mentation. ^ small divisions, but as a large lump. The relatively long string of food is first crowded into an ovoid form as the forward movement begins, and as it is collecting thus, it seems at the last to be suddenly formed into a more rounded ball, as if the mass were pulled or pushed together at the two ends. The next moment it is indented in the middle by a circular constriction (as shown in Fig. 4, line 2), which spreads it in both directions along the loop. The trailing portion {a) is next cut in two, and the severed part some- ^ Without the possibility of seeing the relations of a movement to the ends of the intestine it cannot be stated absolutely whether the movement is peristaltic or antiperistaltic. Such relations can be seen on the fluorescent screen only near the stomach and near the ileoca?cal valve. The evidence that advancing peristalsis is the normal movement is so overwhelming that I have assumed that when food is moving in loops not visibly related to fixed points, it is moving forward. Study of Movcmoits of Intestines by Rontgen Rays. 261 times flies back over its course about a centimetre. Now the whole mass is swept together again and slightly forward as shown in line 4, Fig. 4, and the segmenting process is repeated. At stage 3, Fig. 4, a constriction sometimes appears around the middle of the advanced portion {U). Thus with many halts and interruptions the food slowly advances. A slight variation of the movement just described is observed when the amount of food is greater and extends farther along the intestine. Under such circumstances, as the mass moves forward, constrictions appear just in front of the rear end, which separate it from the main body, and cause it to shoot backward sometimes through the distance of a centimetre. The main body meanwhile is not disturbed. No sooner has the rear section been shot away than it is swept forward again into union with the rest of the food, and the whole mass then advances until another interfering constriction repeats the process. Rhythmic segmentation and the pendulum movement. — There is little doubt that the segmentation of the food which I have seen is due to an activity of the intestinal musculature similar to that which causes the so-called pendulum movement. This activity, as already noted, is rhythmic, and, although accounts differ,^ analytical methods prove that it involves both the longitudinal and the circular layers of muscle. Observations of the effect of the rhythmic con- tractions upon the food show that the action of the circular fibres is most prominent. It is probable, however, that the longitudinal fibres also play an important part in the process of segmentation. Exam- ination of Fig. 2 makes clear that in line 2 the regions of constric- tion appear between the regions of constriction in line 3 ; before c can be formed, therefore, the constriction between a and b must re- lax. Contraction of the longitudinal fibres between two segments would help to enlarge the constricted lumen of the gut. It seems probable that, as the constrictions on either side of c occur, the lon- gitudinal fibres between them contract; almost simultaneously the constriction between a and b relaxes, and the two particles are thus brought swiftly together. A similar process naturally would take place for each of the shifting segments. Thus the function of the longitudinal muscles would be to contract between new rings of con- striction and thereby aid in relaxing the former ring between them. During my one observation of the segmenting process as seen on '' See page 252. 262 IV. B. Cannon. the surface of the intestine, I could not be sure that the distance be- tween neighboring segments was shortened as the constriction relaxed ; that activity of the longitudinal fibres is present, however, is indicated by observations of Raiser ^ on the intestines of the rabbit and the cat. Raiser observed the outer surface of the coils, and describes the normal movement as an alternate contraction and relaxation of single divisions of the longitudinal fibres ; he notes that these short divi- sions shift. But whether they shift in alternation with the shifting circular constrictions, as seems probable, is an interesting point not yet determined. Bayliss and Starling state that the swaying pendulum movements are essentially due to peristaltic waves recurring in the same place and running rapidly downward. This form of the m.ovements I have seen only once. At this time about 90 c.c. of soapy water had been injected. This procedure has the effect of exaggerating in g every particular the movements of the small intes- tine. In this instance a broad constriction appeared Figure 5.— Trac- about the middle of a long string of food and per- ing showing gisted there while it spread down the gut. As the segmentation . , , i 1 1 ^ r f h me in Contraction spread, the gut swayed slowly to and fro the duodenum, before it. Then there was a relaxation, followed by This and other ^ recurrence of the constriction in the same place, tracings re- spreading of the contraction, and a swinging of the duced two- ^ . ^ . . & & ti^ij-ds. loop just as before. This phenomenon was repeated again and again till finally the string was divided and the forward piece pushed through a tortuous course to the colon. The course of the food in the small intestine. — Chyme is not forced from the stomach by every wave that passes over the antrum, but only at intervals.^ When the pylorus relaxes, the food, moved towards the pylorus under considerable pressure, is squirted along the duodenum for two centimetres or more. Careful watching of this food shows that usually it lies for some time in the curve of the duodenum until additions have been made to it from the stomach, and a long, thin string of food is formed. While it is resting in this place it is ex- posed to the outpouring of the bile and pancreatic juices. All at once the string becomes segmented (see Fig. 5), and the process of rhyth- mic segmentation continues several minutes, thoroughly mixing the ^ Raiser: Loc. cit., p. 7. - Cannon: Loc. a'i., p. 369. Study of Movements of Intestines by Rontgen Rays. 263 intestinal digestive juices with the chyme. In this region the alternate positions of the segments are sometimes far apart, and the to-and- fro movement of the particles may be a relatively extensive and very energetic swinging. Finally the little segments unite into a single mass, or form in groups, and begin to move forward. The peristalsis here, as already mentioned, is much more rapid than the normal peri- stalsis elsewhere in the small intestine. The masses once started go fiying along, turning curves, whisking hither and thither in the loops, moving swiftly and continuously forward. After passing on in this rapid manner for some distance the food is collected in thicker and longer strings resembling the strings seen characteristically in the other loops. Towards the end of digestion the small masses shot out from the stomach, after a few segmentations, may move on in the rapid course without being accumulated in a larger mass until the swift movement ceases. During the first stages of digestion in the cat's small intestine the food usually lies chiefly on the right side of the abdomen ; during the last stages the loops on the left side contain the greater amount of food. In these loops the food remains sometimes for an hour or more with no sign of movement. All at once a mass begins to show irreg- ular depressions and elevations along its length, and then suddenly it is divided, at first partially, later completely, into many little equal parts, and these repeatedly undergo division and reunion, division and reunion, over and over again, in the manner described above as rhyth- mic segmentation. After a varying length of time the activity wanes, and the little segments are carried forward individually and later brought together, or join and move on as a single body, or they may reunite and lie quietly for some time without further change. Thus by a combined process of kneading and peristaltic advance the food is brought to the ileocaecal valve to enter the large intestine. Rec- ords froVn ten different animals show that salmon does not appear in the small intestine until an hour or an hour and a half after the food is eaten. Inasmuch as five or six hours elapse after eating before this food begins to be seen in the colon, it is evident that the chyme takes four to five hours to pass the length of the small intestine. It is in- teresting to note that the operations are considerably shortened if the meal has consisted of bread and milk. 264 W. B. Cannon. The Competence of the lLEOCii:cAL Valve. The ileocaecal valve in the cat is situated three or four centimetres from the blind end of the caecum. Its position is usually marked in shadows of the food in the colon by a slight indentation, towards which masses about to enter the colon are ordinarily directed from a point somewhat distant in the small intestine (see Fig. 6). Regarding the competence of the ileocaecal valve many observa- tions have been made. Griitzner has reviewed the evidence bearing on the question ^ and concludes that the valve is not competent, least of all for liquids. He declares that as soon as liquids or thin fluid masses appear in the upper part of the colon, they pass in many in- stances into the small intestine the moment that the pressure on the colon side rises slightly. If the colon contains a solid or a thick mushy mass, the passage towards the small intestine is scarcely pos- sible, because every increase of pressure in the large intestine must force the two lips of the valve together and close it. The importance of the competence of the ileocaecal valve under normal conditions cannot be appreciated until the function of the first part of the colon is considered. In order that this part of the intes- tinal mechanism may perform its service, the competence of the valve for the food which enters the colon from the ileum should be perfect. As a matter of fact such is the case. Not only does the activity of the colon prove this statement, but the failure of every attempt to drive the food in the colon back through the valve into the ileum confirms the proof. Again and again I have tried, by manipulation through the abdominal wall, to press the normal contents of the colon downward with sufficient force to cause them to return to the small intestine, but without success. The valve held perfectly. The Movements of the Large Intestine. When the large intestine is full, palpation through the abdominal wall demonstrates that the material in the lower descending colon '^ and in the sigmoid flexure is usually composed of hard, incompres- sible lumps, while that in the ascending and transverse colon and ^ Grutzner : Archiv fiir die gesammte Physiologic, 1S98, Ixxi, p. 495. 2 The large intestine in the cat has no fixed ascending, transverse, and descend- ing portions ; in this paper these terms are used to designate the parts of the curve of the colon which occupy these relative positions. Study of Movements of Intestines by Roiitgen Rays. 265 the caecum is soft, permitting the walls of the gut to be easily pushed together. The condition of the contents in these two regions seems to indicate a rough division of the large intestine into two parts, and the mechanical activities of these two parts verify the differentiation. In the descending colon the material is very slowly advanced by rings of tonic constrictions (see Fig. 7) ; in the ascend- ing and transverse colon and in the caecum by far the most common movement is an antiperistalsis. Autiperistaisis in the colon. — The colon of cats which have been without food for a day usually contains enough gas to make the posi- tion of the gut distinguishable with the fluorescent screen (see Fig. i). The first food to enter the colon from the small intestine is carried by antiperistaltic waves into the caecum (Fig. i), and all new food as it enters is also affected by these waves. Thus the contents of the colon, instead of being driven immediately toward the rectum by slow peristalsis, as is the general opinion, are first repeatedly pushed toward the caecum by an antiperistaltic action. These antiperistaltic waves follow one after another like the peri- staltic waves of the stomach (see Figs. 5, 6, and 10). They begin either on the more advanced portion of the food in the colon (when only a small amount is present), or at the nearest tonic con- striction, which is usually at the turn between the transverse and descending colon (Figs. 7 and 8). The waves rarely run continu- ously for a long time. When the colon is full, it is usually quiet. The first sign of activity is an irregular undulation of the walls, then very faint constrictions passing along the gut toward the caecum. These constrictions may first appear only on the ascending colon. As they continue coursing over the intestine they become deeper and deeper until there is a marked bulging between successive constric- tions. When the waves have thus become more prominent, they are seen to start near the end of the transverse colon and pass with- out interruption to the end of the caecum. After these deepest waves have been running for a few minutes the indentations grow gradually less marked until at last they are so faint as to be hardly discernible. The final waves are sometimes to be observed only at the end of the transverse colon. Such a period of antiperistalsis lasts from two to eight minutes, with an average duration of four or five minutes. The periods recur at varying lengths of time; in one instance a period began at 1.38 p. M. and was repeated at 2.06, 2.34, 2.55, 3.15, and at 3.36, 266 W. B. Cannon. when the observation ceased ; in another instance a period began at 2.43 P. M. and was repeated at 2.57 and at intervals of from ten to fifteen minutes thereafter while the animal was being watched. The waves have nearly the same rate of recurrence as those in the stomach; about five and a half waves pass a given point in a minute, /. I?., eleven waves in two minutes. This rate has proved fairly con- stant in different cats and at different stages in the process of diges- tion ; in one case, however, the waves passed at the rate of nine in two minutes. The stimulating effect of rectal injections on the movements of the small intestine has already been noted. Enemata have also pronounced stimulating action on the antiperistalsis of the colon. Usually the almost immediate result of a rectal injection of warm water is the appearance of deep antiperistaltic waves, which often continue running for a long period. In one case, after an injection of 50 c.c. of warm water, the waves followed one another with monoto- nous regularity during an observation lasting an hour and twenty minutes. The manner in which this antiperistaltic mechanism affects nutrient enemata introduced into the bowel will be discussed in the section devoted to the question of antiperistalsis. These constrictions passing backward over the colon do not force the normal contents back through the valve into the small intestine again. I have seen hundreds of such constrictions, and only twice have there been exceptions to this rule, — once under normal con- ditions, when a small mass slipped back into the ileum, and at another time when a large amount of water had been introduced into the colon. The importance of the competence of the ileocaecal valve is now apparent; indeed, antiperistalsis in the colon gives new mean- ing and value to the location of a valve at the opening of the ileum. For, inasmuch as the valve is normally competent, the constrictions repeatedly coursing toward it force the food before them into a blind sac. The effect on the food must be the same as the effect seen in the stomach when the pylorus remains closed before the advancing waves. The food is pressed forward by the approach of each con- striction ; but since it cannot go onward in the blind sac, and is, moreover, subjected to increasing pressure as the constriction comes nearer, it is forced into the only way of escape, i. c, away from the caecum through the advancing constricted ring. About twenty-five waves affect every particle of food in the colon in this manner during each normal period of antiperistalsis. The result must be again a Study of Movements of Intestines by Rontgen Rays. 267 thorough mixing of the contents and a bringing of these contents into close contact with the absorbing wall — a process which has already been variously repeated many times in the stomach and in the small intestine. Two other movements have been observed in the ascending colon, but they are rare appearances. The first of these was a serial sec- tioning of the contents noticed in an animal given castor oil with the food. A constriction separated a small segment in the cxcum ; another constriction then cut off a segment just above the first, and with the disappearance of the first constriction the two separated segments united. A third segmentation took place above the second, and the changes occurred again. Thus the whole mass was sectioned from one end to the other; and no sooner was that finished than the process began again and was repeated several times. A slight modification of this movement was observed in a colon containing very little food. The mass was pressed and partially segmented in the manner characteristic of the small intestine, and was thus again and again spread along the ascending colon and each time swept back into a rounded form by antiperistalsis. The second of the two movements mentioned above consisted in a gentle kneading of the contents. This was caused by broad constrictions appearing, relaxing, appearing, relaxing, over and over again in the same place. When several of these regions were active at the same time they gave the food in the colon the appearance of a restless undulatory mass. Once a constriction occurred and remained permanently in one place while the bulging parts on either side of it pulsated alternately, at the rate of about eighteen times in a minute, with the regularity of the heart-beat. Although these phenomena are some- what striking, they are not usual, and are in no way so important as the antiperistalsis. The changes -when food enters the colon. — The passage of food through the ileocsecal valve seems to stimulate the colon to activity. As food is nearing the ileocaecal valve the large intestine is usually quiet and relaxed (Fig. 6, 4.00), though occasionally indefinite move- ments are to be observed ; and sometimes just before the food reaches the end of the ileum the circular fibres of the colon in the region of the valve contract strongly, so that a deep indentation is present there. The indentation may persist several minutes; it disappears as the muscles relax just previous to the entrance of the food. The food is moved slowly along the ileum and is pushed through the 268 W. B. Cannon. valve into the colon. The moment it has entered, a strong contrac- tion takes place all along the coecuni and the beginning of the ascending" colon, pressing some of the food onward, and a moment later deep antiperistaltic waves (Fig. 6, 4.03) sweep down from the transverse colon and continue running until the caecum is again normally full, /. e., for two or three minutes. The appearance of tonic constrictions. — It has already been noted that as the food accumulates in the ascending colon it is at first confined to this region by antiperistaltic waves. With further acces- sions, however, the contents naturally must be pressed more and more into the transverse and descending colon. In the early stages of this accumulation, while the food lies chiefly in the ascending colon, the only activity of the muscular walls is the anti- peristalsis. As the contents extend along the intestine a deep constriction appears near the advancing end and nearly sepa- rates a globular mass from the main body of the food (Fig. 6). The contents of the the colon and also the first large intestine progress farther and farther tonic constriction. 4.00, the {^^^^ ^j^e CECcum ; meanwhile new tonic constrictions appear which separate the contents into a series of globular masses. And as the number of these divisions increases they take a position farther from the caecum so that they are present chiefly in the descending colon (Fig. 7). Raiser^ has recorded a similar appearance in the terminal portion of the rabbit's colon, in which deep circular constrictions separate the scybalous masses. He maintains that these masses are pushed onward by the con- strictions. Comparing tracings made at rather long intervals (forty- five minutes), I found that the rings disappear from the transverse colon and then are present with the waste material in the descend- ing colon. Thus in the cat also these rings which seem with short observation to be remaining in one position are in reality moving slowly away from the caecum, pushing the hardening contents before them. The contents at this stage are no longer fluid, and conse- quently they must offer considerable resistance to a force pushing Figure 6. — Tracings showing changes wlien food enters colon relaxed as food ap- proaches in the ileum. 4.03, the colon contracted and traversed l^y antiperistaltic waves after the food has entered. 1 Raiser: Loc. cit.^ p. 12. Study of Movements of Intestines by Rout gen Rays. 269 them through the colon. It is an advantage to have this pultaceous substance propelled in divisions rather than in a uniformly cylindrical mass, since the fibres along the length of the mass are thereby rendered effective. Such are the functions of the persistent rings, — the\' form the waste matter into globular masses at the end of the transverse colon and slowly push these masses onward. In the transverse colon, which is free from the slowly-moving rings, the antipei istaltic waves have full sway. In the region of the tonic rings an infrequent or even a slowly periodic relaxa- tion and contraction are often to be observed. These changes seem to take place in all the rings at about the same time. Once I saw antiperistaltic waves running over the upper- most of four segments, but, since the rings on Figure 7.— Radiograph either side of the segment held tightly, the waves had merely the effect of churning the material of the segment and did not move it onward. Inasmuch as the material in these segments at first is soft, so that the segments are easily compressible, while the fecal masses which are the final result are relatively hard and dry, it follows that even within the confines of these persistent rings some absorption is taking place. Defecation. The process of clearing the colon is a process of repeated reduction of the amount of material present. Figure 8 (3-ii) is a radiograph showing the food in the colon at 3.1 1 p. :m. About 3-25, with a slow sweeping movement, the gut swung around so that the ascending colon was lying in the position of the last half of the transverse colon, and the transverse colon had taken the position of the descend- ing part (Fig. 8, 3.25). At the same time the tonic constrictions disappeared and were replaced by a strong, broad contraction of the circular muscle, tapering the contents off on either side in two cones. The region of strongest contraction was apparently drawn downward with the rest of the gut by a shortening of the descending colon. As the intestine swung around, more material was forced into the rectum, showing the region of tonic constrictions (descending colon), and the region of an- tiperistalsis (trans- verse and ascending colon). 270 IV. B. Cannofi. and when the swinging of the intestine stopped, the constriction which divided the lumen passed slowly downward and with the aid of the muscles surrounding the abdominal cavity pushed the separated mass out of the canal. ^ After the terminal mass had thus been pushed out, the colon with the reniainder of its contents returned to nearly its former position (Fig. 8, 3.46). .About two hours afterward this remnant had been spread throughout the lengtli of the large intestine by means of the slowly-moving rings. Fig. 7 is a radiograph of the same colon pictured in Fig. 8 ; the radiograph was taken at 1 1.50 a. m., and at 12.15 ^- ^- the material in the lower descending colon was forced out in the manner above described. Within three hours the remaining portion had been spread into the evacuated region, as shown in Fig. 8, 3.1 1. The manner in which the material is spread from the region of the antiperistaltic waves into the region of the slowly ad- vancing rings presents a problem. During normal living new food constantly arriving in the colon must force the old contents forward 3.11 3.25 3.46 Figure S. — Two radiographs and a tracing showing the changes taking place in defe- cation. 3.11, material in the colon. 3.25, colon carried downward and terminal mass separated. 3.46, after defecation, when the colon returns to former position. Defe- cation occurred at 3.27. just as the later parts of a meal force forward the earlier parts ; there is no doubt, however, that most of the contents of the caecum and the ascending colon may be passed onward even during starvation. The ^ In thi.s case the feeces were soft. Shidy of Movements of Intestines by Rontgen Rays. 271 emptying of these regions, according to my observations, is never complete ; for, after considerable time has elapsed and the large intestine is cleared and dilated with gas, some substance is still to be detected in the caecum and clinging to the walls of the ascending colon. The only activities manifested here are the antiperistaltic waves, and the strong tonic contraction of the whole circular muscu- lature shown in Fig. 6. It is clear that the latter activity would serve to press into the transverse colon a considerable portion of the con- tents of the ascending colon, and the remnant seen clinging to the walls would be the part not thus pressed forward. Twice I have seen appearances which might account for the empty- ing of the first portion of the large intestine in a more thorough manner than that above described. At one time, without apparent stimulation, strong tonic contraction occurred along the entire length of the ascending colon, which forced the contents almost wholly into the transverse portion. This action seemed merely an exaggerated form of that observable after food passes the ileocaecal valve (see Fig. 6). At another time, after a mass of food had passed through the ileocascal valve, after the ascending colon had contracted gener- ally and the antiperistaltic waves had coursed over it in the usual manner, a deep constriction appeared at the valve and ran upward without relaxation nearly the length of the ascending colon, pushing the contents before it. For an instant the wave paused ; then the constriction relaxed, and the food returned toward the caecum. These observations indicate that either a general contraction of the wall of the large intestine or a true peristalsis may be effective in pressing waste matter from the region where antiperistalsis is the usual activity into the region where the slowly-advancing rings may carry it on to evacuation (see Fig. 7). The Question of Antiperistalsis. In 1894, Grlitzner^ published an observation and made an assump- tion, about which there has since been much controversy. He main- tained that when normal salt solution, holding in suspension hair, powdered charcoal, or starch grains, is injected into the rectum, it is carried upward into the small intestine and may even enter the stomach. These experiments have been repeated by several ob- servers. Some have Confirmed Griitzner's results; others have failed, 1 Grutzner : Deutsche medicinische Wochenschrift, 1894, xx, p. 897. 272 JV. B. Cannon. after using most careful methods, to find any evidence of the passage of the injected material back to the stomach, and they have declared that the apparent success was due to carelessly allowing the food of the animal to become contaminated with the test materials, so that these were introduced into the stomach by way of the mouth. That antiperistalsis does not occur in the small intestine seems to be proved by Mall's experiment^ of reversing a portion, sewing it in place, and then finding that the food does not pass the reversed region, but col- lects at the upper end. Sabbatani and Fasola^ reversed stretches of small intestine of varying length, and found that the reversed portions allowed fluids to pass, but that the persistence of the physiological direction of movement caused an accumulation of undigested food in the region of the upper suture. However a portion of the intestine lay in relation to the rest, it always manifested the normal peristalsis. Many other observers'' working directly on the intestine confirm this testimony and state that the progress of the constriction-rings is always downward and that antiperistalsis is not physiological. In 1898, however, Griitzner* took his stand again in favor of a backward movement in the intestines, and in a somewhat metaphysical manner argued that peristalsis and antiperistalsis belong to each other just as relaxation of muscle is related to contraction. He assumed that as the contents are advanced by slow peristalsis, so are they returned by a similar movement in the opposite direction, and he mentions several pathological cases (fistula of intestine) to substantiate the assumption. By means of the X-rays it is possible to see just what takes place when a fluid is injected into the rectum. For the purpose of deter- mining how nutrient enemata are received and acted upon in the intestines, I have introduced thin fluid masses in large and small amounts, and thick, mushy masses in large and small amounts, in different animals. The enemata consisted of 100 c.c. of milk, one egg, ten to fifteen grams of bismuth subnitrate, and two grams of starch to hold the bismuth powder in suspension. To make the thick enema all these were stirred together and boiled to a soft mush ; to make the thin enema all the parts were boiled together except ^ iMall : Johns Hopkins Hospital Reports, 1S96, i, p. 93. ''■ Sabbatani and P'asola: Archives italiennes de biologie, 1900, xxxiv, p. 195. ^ See FuBiNi and Luzzati : Moleschott's Untersuchungen, 1888, xiii, p. 386, for an array of observers who declare that antiperistalsis is not a normal movement. * Grutzner ; Archiv fiir die gesammte Physiologic, 1898, Ixxi, p. 513. Study of Movements of Intestines by Rontgen Rays. 273 the ^gg, which was added after the boiled portion was cooled. The small amount injected was 25 c.c. ; the large amount almost 90 c.c, about the capacity of the large intestine when removed from the body. The animals were given first a cleansing injection, and after this was effective the nutrient material was introduced. In order to make sure of the observation, a control radiograph was first taken to show no bismuth food present, and other radiographs taken at Varying intervals after the injection to record the course the food was following. 1.50 2.15 3.00 Figure 9. — Radiographs showing that after a large nutrient enema (about 90 c.c.) has been given, the food is forced more and more from the large into the small intestine. The enema was introduced at about 1.40 p. m. At 3.00 segmentation was occurring in many loops. These experiments show that when small amounts of nutrient fluid are introduced they lie first in the descending colon. In every instance antiperistaltic waves are set going by the injection and the material is thereby carried to the caecum. When large amounts are injected they stop for a moment in the region between the transverse and descending colon as if a constriction existed there. Then a con- siderable amount of the fluid passes the point and antiperistaltic waves carry it to the caecum. In any case the repeated passing of the waves seems to have the effect of promoting absorption, for in the region where these waves continue running, the shadows become gradually more dim and finally the bismuth appears to be only on the intestinal walls ; in other regions, c. g., in the descending colon, the shadows retain their original intensity. Small injections have never in my 2 74 W. B. Cannon. ^~ experience been forced even in part into the small intestine; but with the larger amounts, whether fluid or mushy, the radiographs show many coils of the small intestine containing the bismuth food. The passage of the injected material beyond the ileocaecal valve is ^obably due entirely to antiperistalsis in the colon, — a factor un- known to both Griitzner and his opponents. The valve, which is thoroughly competent for food coming normally from the small in- testine into the large, is curiously incompetent for a substance, even of the consistency of thick cream, introduced in large amount by rectum. When the valve first permits the food to enter the ileum, the fluid pours through and appears suddenly as a winding mass occupying several loops of the intestine (Fig. 9, 1.50, about ten minutes after the injection). The mass is continuous from the valve to the other end; antiperistalsis is therefore not visible in the small intestine under the circumstances of this experiment. The antiperistaltic waves of the colon, however, continue running ; the transverse and ascending colon are thus almost emptied, and the small intestine more and more filled with food (Fig. 9, 2.15 and 3.00). After a short time the typical segmenting movements can be seen in the loops, busily separating the food into small masses, and over and over again dividing and redividing them. I have never seen food material pass back from the colon so far as the stomach ; but once, about ten minutes after an injection of 100 c.c. of warm water, the cat retched and vomited a clear fluid resembling mixed water and mucus. In the fluid were two intestinal worms still alive. The importance of the mechanism by which nutrient enemata are passed backward in the intestine is evident. In the colon the nutrient material is worked over by the antiperistaltic waves, inti- mately mixed with whatever digestive juices may be present, and exposed to the organs of absorption in that region. If the enemata are large, the digestive and absorptive processes are by no means confined to the colon, but may take place along extensive surfaces of the small intestine. I have repeatedly seen rhythmic segmentation active throughout many loops of the small intestine, thus exposing the injected food to the same mixing and absorbing processes as affect the nutriment which has come through the stomach in a normal manner. Study of Movements of Intestines by Rontgen Rays. 275 The Effect of Emotions and Sleep. Observations on the stomach of the cat showed that the peristalsis is inhibited whenever the animal manifests signs of anxiety, rage, or distress.^ Since the extrinsic innervation of a large part of the intes- tinal tract is the same as that of the stomach, it is of interest to note the effect of emotional states on the movements of the intestines. Esselmont,^ in a study of the dog's intestine, noted constantly after signs of emotion a marked increase of activity lasting for only a few moments. Fubini'" also observed that fear occasioned more rapid peristalsis. There is no doubt that many emotional states are a strong stimulus to peristalsis, but it is equally true that other emo- tional states inhibit peristalsis. In the cat the same conditions which stop the mo\'ements of the stomach stop also the movements of the intestines. The female cats used in these observations ordinarily lie quietly on the holder and make no demonstration. Sometimes, however, with only a little premonitory restlessness the cat "" "" suddenly flies into a rage, lashing her tail from Figurk 10. — Tracings side to side, pulling and jerking with every showing the effect of ,. , 1 1 • • 1 • excitement on antiperi- hmb, and biting at everything near her head. stalsis in the colon. During such excitement, and for some moments after the animal becomes pacified again the movements, both of the large and small intestine, entirely cease. Such violence of excite- ment is not necessary to cause the movements to stop; a cat which was restless and continually whining while confined to the holder, showed no signs of intestinal movements during any period of obser- vation (one period lasted more than an hour), although the changes in the distribution of the food observable from one period to the next proved that movements were going on during the quiet intermissions. In another cat, uneasy and fretful for fifty minutes, no activity was seen ; then she became quiet for several minutes, and peristalsis of the small intestine appeared. When the segmentation process in the small intestine is stopped by 1 Caxxon : Lot. cit.. p. 3S0. - EsSELMONT : Report of the British Association for the Advancement of Science, 1899, p. 899. •^ Fl'BIXi : Moleschott's Untersuchungen. 1892, xiv, p. 528. 276 W. B. Cannon. excitement the segments unite and the series of parts returns to the form of a solid string. The change occurring in the large intestine when the antiperistalsis is inhibited by excitement is shown in Fig. 10. The tonic constrictions in the descending colon are apparently not affected by emotional states, for they do not seem to relax in the excitement which causes the movements to cease. By holding the mouth and nostrils closed, or by pressing between the rami of the jaw, the breathing may be stopped. As soon as the cat shows distress from lack of breath every form of intestinal movement stops. The statement is sometimes made in text-books of physiology that the gastric and intestinal mechanisms cease to act during sleep. It is worthy of note that nearly all the animals curled up and slept during the time between observations ; nevertheless, the progress of the food through the intestines continued. The statement is also made that at night, even without sleep, the intestines are almost entirely at rest; that this is their normal time for repose.^ I have seen both large and small intestines actively at work, however, from half past nine until half past ten o'clock at night. Summary. 1. Bismuth subnitrate, 10 to 33 per cent, mixed with the food renders the movement of the intestinal contents and thereby the movements of the intestinal walls visible on the fluorescent screen. 2. The activity most commonly seen in the small intestine is the simultaneous division of the food in a coil into small segments, and a rhythmic repetition of the segmentation each time applied to the new segments formed from parts of those just divided. In the cat this rhythmic segmentation may proceed at the rate of 30 divisions per minute. The effects of the constrictions causing the segmentation are, the mixing of the food and the digestive juices, the bringing of the digested food into contact with the absorbing mechanisms, and the emptying of the venous and lymphatic radicles of their contents by compression of the intestinal wall. 3. Peristalsis is usually combined with segmentation. As the food is advancing, interfering constrictions often separate the rear end of the mass from the main body. The separation is momentary, how- 1 Hess: Deutsches Archiv fiir klinische Medicin, 1S87. xl, p. 104. LuDWiG ; Loc. cit., p. 617. Study of Movements of Intestines by Rout gen Rays. 277 ever; the rear end is swept into union with the main body again, and the whole mass is pushed onward until another constriction repeats the changes. 4. The ileoca;cal valve is thoroughly competent for food entering the colon from the ileum. 5. The usual movement of the transverse and ascending colon and the caecum is an antiperistalsis. This recurs in periods about every fifteen minutes, and each period lasts commonly about five minutes ; the waves recur during a period at the rate usually of eleven waves in two minutes. This antiperistalsis gives new significance to the ileocascal valve ; for the food, now in a closed sac, is thoroughly churned and mixed by the constrictions running towards the caecum, and again exposed to absorbing walls without any interference with the processes in the small intestine. 6. As soon as new food enters the large intestine a strong general contraction takes place along the caecum and ascending colon, forcing some of the food onward ; a moment later antiperistaltic waves begin to pass. 7. With the accumulation of material in the transverse colon, deep tonic constrictions appear one after another and carry the material into the descending colon, leaving the transverse and ascending por- tions free for the antiperistaltic waves. 8. In emptying the large intestine the material in the lower descend- ing colon is first carried out by combined peristalsis and pressure of abdominal muscles; the remainder of the material is then spread into the evacuated region, and this region is again cleared ; the second remainder may be similarly affected. In normal life the new food arriving in the colon must force forward the old contents of the ascending and transverse colon. 9. The observations have revealed no evidence of antiperistalsis in the small intestine, but, since the ileocjecal valve will allow nutrient material under pressure to pass backward, the antiperistalsis of the large intestine may force into the small intestine a considerable por- tion of a large nutrient enema. Segmentation in the small intestine aftects such an enema precisely as it affects food which has passed normally through the stomach. 10. Signs of emotion, such as fear, distress, or rage, are accompanied by a total cessation of the movements of both large and small intes- tines. The movements continue in the cat both during sleep and at nig-ht. THE REFLEXES CONNECTED WITH AUTOTOMY IN THE HERMIT-CRAB. T. H. MORGAN. \^From the Laboratory of Biology at Bryii Mmur College.'] THE throwing ofif of the legs of crabs after injury has been long known, and the physiological processes involved in the act, have been examined in some detail by Fredericq.^ There are, how- ever, certain " instincts " (reflexes) connected with the autotomy that have as yet received very little attention. If a leg is quickly cut off distal to the " breaking-joint," the stump of the leg that remains may not be thrown ofif at once, although if held for a moment it will in- variably drop ofif. If the stump of the leg is not held and the crab is returned to the water, the stump is generally drawn up against the side of the body or pressed down against the bottom of the aquarium, and in consequence of the resistance that it meets with it is often cast ofif. If the leg is not gotten rid of in any of these ways the other walking legs of the same side are brought against or across the stump and may assist in the autotomy. Very often the leg cannot be thrown off in any of these ways, and this is especially true for hermit- crabs {Eupag-ur/is), in which case the stump is caught hold of by one or by both claws of the first pair of legs, or chelae and these, pulling at the stump, furnish the resistance necessary for the autotomy to take place. It is this instinct of the hermit-crab that has espe- cially interested me. The act appears to be one carried out as though it were a deliberate attempt on the part of the crab to get rid of the useless stump. The advantage of the result is obvious, not only because a better surface for regeneration is present at the breaking-joint, but also because relatively little blood is lost when the leg is thrown ofif at this place. In fact crabs often die when the leg is not thrown off at the base. The catching hold of the stump by one or by both of the claws is an act of such a kind that if we could imagine the crab endowed with intelligence, we should not hesitate to praise the action as one calculated to preserve the individual after ^ Fredericq : Archives de zoologie experimentale et generale, 1883. 278 Reflexes connected with Atttotoniy in the Hermit-Crab. 279 injury. On the other hand, Fredericq has shown that the throwing off of the leg is a simple reflex act that involves the ventral nerve- cord. The leg may be thrown off even when the brain has been de- stroyed. Furthermore, I have found in the hermit-crab, that if the whole anterior end of the head (with its appendages) is quickly cut off, and then if one of the walking legs is cut off distal to the break- ing-joint and the crab returned to water, the claws of the first pair of legs almost invariably bend over and catching hold of the stump of the leg, pull at it until it comes off. I have seen this act repeated by the same individual until a number of legs have been successively cast off. A similar act takes place in the decapitated crab if one of the first pair of legs, or chelae, is cut off distal to the breaking-joint. The claw of the opposite side bends over and, catching hold of the stump, supplies the resistance necessary for the throwing off of the leg. Shall we call this reflex action that leads to a beneficial result an in- stinct .'' If so, it is an instinct that involves only the nerve cord, and not the brain. I refer, of course, not simply to the reflex, described by Fredericq, by means of which autotomy takes place at the base of the leg, but especially to that part of the process involving the catch- ing hold of the stump by the claws of another pair of legs. The fol- lowing observations put the matter in even a more paradoxical light. The fourth and the fifth pairs of legs of the hermit-crab are small in comparison to the more anterior legs. They serve to brace the animal against the shell and also to clean out the branchial chamber. They do not have a breaking-joint at the base, and cannot be thrown off after injury. If they are cut off at any level no effort is made by the crab to catch hold of them with its claws. Obviously such an act would be useless, since the leg would not come off, as can be shown by holding the legs with a pair of forceps. The fact that no attempt is made by the crab to catch hold of these legs, if they are injured, shows that the presence of a cut end does not in itself lead to catch- ing hold of the injured part. This was further shown in another way. If the last pair of abdominal appendages are partly cut off, the piece that is left is not caught by the claws of the first pair of legs. A still more satisfactory demonstration of the same thing is given by cut- ting off one of the walking legs inside of the breaking-joint. Al- though profuse bleeding may follow, there is no effort made by the crab to catch hold of the injured region. This is an unexpected re- sult, for if the reflex action takes place through the nerve that goes 2 So T. H. Morgan. to the ventral cord, it would seem that when the same nerve is cut off proximally to the breaking-joint, a reaction would be set up in the ventral cord that would lead to the first pair of legs being brought to the injured region. Since this is not the case we must suppose that there is another factor in the process in addition to the exposure of the cut end of the nerve. This other factor involves the presence of the stump of the leg distal to the breaking-joint. Its presence can- not be made known to the animal, however, through its cephalic sense organs, since the reaction takes place when these have been cut off. Fredericq has shown that the throwing-off of the leg at the breaking- joint is brought about by certain muscles present in that region. If the muscles are injured the leg cannot be thrown off. This relation suggested the following experiment. With a pair of fine oculist's scis- sors it is easily possible to cut, either on the ventral or on the dorsal side, some of the muscles that bring about the autotomy. The leg can no longer be properly used for walking after the operation, but the crab makes no effort to catch hold of it. If the crab really acted intelli- gently we should expect it to attempt to get rid of the leg by catching hold of it with the claws. Or shall we suppose the crab acts even more intelligently and prefers to allow the breaking muscles to regen- erate rather than regenerate a new leg? It would be gratuitous on our part to assume that either alternate has anything to do with the result. We can only infer that the autotomy reflex is not started, hence there is no action. Suppose, however, a leg that has had its muscles cut in this way is itself cut off outside of the breaking-joint. The crab will at once make violent and oft-repeated attempts to get rid of the stump by grasping it with the claws of the first pair of legs. If the muscles have been sufficiently cut apart, the stump of the leg cannot be thrown off, and may, subsequently, regenerate the distal part of the leg from the cut end. If some of the muscles at the breaking-joint are cut, as in the last operation, and if at the same time the nerve of the leg is also cut proximally to the breaking-joint, no attempt is made by the crab to catch hold of the leg. If the leg is then cut off distal to the breaking- joint, still no attempt is made to catch hold of the stump of the leg. This latter result is what we might have anticipated from the results of the experiment in which the nerve was cut in two proximally to the breaking-joint, the leg itself being left attached. How shall we formu- late these results? Shall we conclude that the entire series of reac- tions are purely machine-like, and that nothing takes place which Reflexes conuecteei ivitJi Aiitotomy in tJie Hej^mit-Crab. 281 involves processes akin to intelligence on the part of the crab? There seems scarcely any room for doubt on this point, for we must suppose the crab to be entirely ignorant of the beneficial results of its action. The reactions are simple, reflex ones, and they only appear intelligent to us, since we can see that, on the whole, the result is useful to the animal. But why so often do just those reflexes take place that we should ourselves as intelligent agents recommend because they are useful to the future welfare of the individual? Why does the crab catch hold of its leg and bring about its breaking off when the leg is cut off" out- side of the breaking-joint, and why does not a similar action take place when the leg is cut off inside of the same region ? Why does no simi- lar reaction take place when other appendages without breaking-joints are injured .? Does the problem for the physiologist go no further than an examination of the reflex-action itself? We meet here with only a striking case of the familiar phenomenon of machine-like reflex ac- tions taking place in the organism as though they had been specially devised as ends to means. It is at present customary to assume that the physiological problem involves only a study of the mechanism by which a process takes place, and that it is beyond the scope of phy- siology, and by inference beyond the range of scientific inquiry, to investigate the origin of useful reflexes, or rather to investigate how it is that so many responses are just those which are useful to the organism. In other words, is there a higher law that will account for those physiological changes being present which are useful to the organism? Pfliiger, in 1877, drew attenion to this problem in his " Law of Teleology." It is unquestionably the easiest solution of the difficulty to deny that this is a proper field for scientific investigation. Moreover, it is undoubtedly true that reactions sometimes take place that are injuri- ous to the organism, but this does not do away with the fact that the great majority of responses are useful. Many zoologists believe that they have found a plausible interpretation of this relation in the Darwinian theory, for if only useful responses survive, their prepon- derance in organisms appears to be accounted for, but I doubt if all useful responses can be explained in this way, because many of them are not of vital importance to the organism, and without a death-struggle the action of natural selection, sensii slrictu, cannot take place. But even if it be admitted that a selective process takes place, the physiological problem remains exactly the same, for we 2^2 T. H. Morgan. have to account in the first place for the appearance of the first use- ful variation, and in the second place find an explanation of the accu- mulation of variations by inbreeding. It would not be profitable to enter here into a general discussion of the origin of useful responses, but I cannot but believe that it is by no means certain that the prob- lem is beyond the possibility of scientific investigation. A PHYSIOLOGICAL STUDY OF THE PULMONARY CIRCULATION. By HORATIO C WOOD, Jr. \^From the Laboratory of Pharmacodviiaviics of tlie University of Pennsylvania^ ^ I "^HE earliest investigation concerning the vaso-motor supply of -■- the lungs which I have been able to find is that of Brown- Sequard on the changes in the appearance of the lungs after various injuries to the pons varolii or the crura. Brown-Sequard observed, as the result of injuries to these parts, sometimes ecchymosis and effusion of blood, sometimes pallor or oedema of the lungs, which changes he attributed to vascular alterations in these organs. The untrustvvorthiness of direct observation of the calibre of blood-vessels, as well as the impossibility of determining whether these changes were the primary or secondary effects of the injuries to the brain, makes this evidence of little value. In my studies I have relied upon what is after all the most generally useful means of determining the calibre of the vessels, namely, the alteration of the blood-pressure. The method which I have employed is as follows. The experiments were made upon dogs of moderate size (about lo kilos) ; all the dogs but one were curarized and were rendered entirely insensible by morphine. In the greater number of the experiments the carotid pressure was taken simultaneously with that in the pulmonary vessels, although in one or two the pulmonic pressure only was observed. After placing a cannula in the carotid artery, and connecting the trachea with the artificial respiration apparatus, an incision was made the whole length of the chest, the chest-walls drawn apart, and after careful dissection a small branch, sometimes of the right, sometimes of the left, pulmo- nary artery was exposed, and a cannula placed in it. Usually I chose a branch supplying one of the superior lobes of the lung, being care- ful not to injure the blood-vessels or lung tissue of the neighboring lobes. Sufficient air was forced into the lungs by the artificial respira- tion apparatus to neutralize the diminished area in the respiratory surface. Absorbent cotton wrung out in hot water was placed lightly over the opening in the chest in order to prevent drying or chilling. The exposure of the lung is associated necessarily with a 2S3 284 Horatio C. Hood, Jr. considerable amount of shock to the animal, and I have nearly always found the carotid pressure low, and presumably the pulmonic pressure is correspondingly low. Under the above conditions I have found that the pressure in the pulmonary artery of the dog, at the beginning of the experiment, varies between ten and twenty-five millimetres of mercury, and its relation to the carotid pressure to be usually between 1-4 and r-io. The relations between the two arterial systems may be greatly distorted by various influences. TABLE OF PULMONARY PRESSURE. Mean of Average ratio of Author. ]5ressure in pulmonic to pulmonary- aortic pressure. Velich 22.7 1 : 6 Dog Knoll 12.2 1 : 7 Rabbit Bradford Dog and Dean 18 1 : .s Wood 16 1 : 4.3 Dog That the pressure in the pulmonary artery may be distinctly altered by appropriate measures is so generally admitted that I need only refer to the appended tables of my experiments. In the inter- pretation of these variations there is, however, considerable diversity of opinion, and a close analysis of these changes is necessary. Asphyxia, when practised in animals thoroughly curarized, is a very efficient and at the same time comparatively uncomplicated vasomotor stimulant. In every case of nine experiments in which asphyxia was employed, there was a distinct elevation of the blood- pressure in the pulmonary, as in the carotid artery. Knoll, in a long series of experiments made upon rabbits, obtained an " outspoken " elevation of the pulmonary pressure in only a single case; but as the results of every other experimenter with whose work I am familiar are in accord vi^ith my own in finding that asphyxia elevates the pressure in the lesser circulatory system, as well as in the larger, it would seem that Knoll's results were due to some fault in technique, or that rabbits react very differently from dogs. It may be, however, that this investigator expected too considerable changes in the pul- monary pressure. In my experiments the rise absolutely expressed A Physiological Study of the P7ilino?iary Circulation. 285 was never so great in the pulmonary as in the aortic pressure, but in proportion to the original heights of the pressure the elevation of the pulmonic might be relatively even greater than the elevation in the carotid artery. In other words, the difference between the minimum TABLE I. Time in min. and sec. Pulmonary pressure, mm. Hg. P''^-'^^"''^- per min. mm. Hg. ' min. sec. 20 47 162 5 Hegin asphyxia. 17 20 52 162 20 22 5 5 162 25 27 71 162 40 End asphyxia. 43 40 137 162 3 25 68 180 3 2 to 3 9 Stimulate intact right vagus. 3 5 22 60 84 3 15 28 97 192 3 25 ■■ Divide right vagus. 3 30 22 57 210 3 45 20 50 120 Stimulation of peripheral vagus. Weigh t of dog, 11.3 kilos. light pulmonary. and maximum pressure in the carotid artery, when expressed in millimetres, bore much the same relation to the difference between the high and low pressure in the pulmonary artery as the original pressure in the carotid bore to that in the pulmonary, although by certain influences the proportion could be absolutely destroyed. I have found, also, that stimulation of the central end of a divided pneumogastric nerve causes simultaneous rise of the pulmonary and carotid pressures. The rise in the pulmonary pressure did not seem to be proportionately as high as that in the carotid. Frangois-Franck 286 Horatio C. Wood, Jr has found that stimulation of other sensory nerves, as the intercostal or crural, likewise elevates the pressure in both circulatory systems. Three explanations of the elevation of pressure which takes place in the pulmonary arteries have been offered ; first, that it is due to a damming back of the blood : second, that it results from a greater flow to the right heart: third, that it is due to direct contraction of the arteries of the pulmonary circulation. Let us consider these in order. TABLE II. Time in miii. and sec. Pulmonary pressure. mm. Hg. Carotid pressure, mm. Hg. Pulse rate per min. min. Bee. 13 82 138 5 Begin asphyxia. 40 13 93 144 1 16 150 168 1 15 20 139 192 End asphy.xia. 15 16 124 168 Begin asphyxia. 16 21 162 16 20 23 1.56 198 16 25 24 131 198 End asphyxia. 16 30 25 157 20 20 119 20 1 to 20 11 Compress inferior vena cava. 20 11 IS 51 20 12 24 51 20 19 20 105 ^Yeight of c og, 6.8 kilot According to the first theory it is supposed that the increase of resistance offered by the contracted vessels in the general circulation holds back the blood in the left ventricle, thereby increasing corre- spondingly the pressure in the left heart, which in turn dams back the blood in the pulmonary vein, and in this way increases the resistance A Physiological Sliidy of the PiiLmoiiary Circulation. 287 offered to the flow of blood out of the right heart. VeUch has found that the rise of arterial pressure caused by the intravenous injection of the extract of the suprarenal capsules was accompanied by an elevation of the pressure in the left ventricle from 32 to 100 mm., and in the left auricle from 6 to 30 mm. These figures represent different ex- periments, apparently only one of each having been made. It is very difficult to interpret properly these single experiments, but they would seem somewhat to favor the stagnation theory. It has also been argued that the pressure in the aorta is elevated sooner than the pressure in the pulmonary artery, and the conclusion has been reached by a sort of" post hoc : propter hoc" logic that the lateness of the rise in TABLE III. Time in Pulmonaiv Pulse rate per min. mill, and sec. pressure, mm. Hg. min sec. 12 240 5 Begin asphyxia. 30 19 228 45 Ih 228 End asphy.\ia. 2 20 Begin asphy.xia. 2 .5 30 10 20 2S 222 3 40 24 240 End asphy.xia. 3 50 18 240 Weight jf dog, 20.5 kilos. the lesser circulation is due to the necessity of producing a sufficient resistance to dam back the blood through the heart. I have found that while it is true that the aortic pressure begins to rise before the pulmonic pressure (see Tables I and 11), yet this elevation need only amount to a few millimetres (7 mm. in Table I) before the pulmonary pressure begins to go up ; and it is hardly plausible that seven or eight millimetres' elevation in the pressure of the larger arteries is sufficient to overcome the reserve power of the heart. If the rise of pressure in the pulmonary vessels is a passive one, 288 Horatio C. Wood, Jr. any condition which affects the aortic pressure must of necessity affect also the pulmonic pressure; but such is not the case. The simplest method of increasing the pressure in the aorta is by occlusion of its thoracic portion. In compressing the thoracic aorta I have found that while the carotid pressure might be more than doubled there was almost no rise at all in the pressure in the pulmonary artery (Table IV), Again, under the influence of digitalis the systemic pressure reached a point three times its previous height, while the pulmonary pressure rose one millimetre (Table IVj. Brad- ford and Dean have obtained similar results by stimulating the peripheral end of the splanchnic nerve. Knoll suggests that the rea- son that these manipulations fail to elevate the pressure in the lesser circulation is that the normal heart is able to overcome the excessive resistance of high pressure and does not permit the blood to be dammed back ; but that after the heart has been weakened by the accumulation of carbon dioxide in the system it permits of regurgita- tion. But this modification renders the theory none the more tenable. In the first place, I need only call attention again to the short interval which elapses between the beginning of the elevation of the carotid pressure and the beginning of the elevation of the pulmonic pressure; in the second place, certain influences such as the injection of suprarenal extract, irritation of a sensory nerve (see Table IV), which can hardly be accused of weakening the heart, cause a simultaneous rise in the two pressures. Moreover, as I shall show later, it is possible to produce an elevation of the pulmonary pressure without any rise whatsoever in the aortic pressure. It therefore seems proven that the rise of the pressure in the pulmonic artery cannot properly be attributed to the damming back of the blood. The second explanation of the elevation of the pressure in the pul- monary arteries is, strangely enough, more or less the direct opposite of the foregoing. It supposes that as the blood flows more rapidly through the contracted blood-vessels of the general circulation, a greater supply of blood flows to the right heart. This supposition, I must confess, seems to me rather opposed to the laws of physics; it is hard to understand how more blood flows through a vessel of nar- row calibre than would flow through a vessel of wide calibre, the pumping force being constant. This view has been most strenuously urged by Openchowski, who has brought forward a certain amount of evidence to support his belief. His most important evidence is the following experiment. A Physiological Study of the Pulmonary Circulation. 289 TABLE IV. Time in min. and sec. Pulmonary pressure, mm. Hg. Carotid pressure, mm. Hg. Pulse rate per min. min. sec. 17 44 174 Right vagus has been cut. 5 Stimulate central end of vagus. 17 19 62 174 20 19 64 180 1 19 52 184 ' 1 .Mo 1 3.=; Stimulate central end of vagus. 1 30 22 70 192 1 35 21 69 192 2 19 46 192 2 5 to 2 16 Digital compression of thoracic aorta. 2 15 20 91 192 3 18 48 192 3 5 to 3 38 Compression of thoracic aorta. 3 20 19 87 3 38 20 100 204 3 38 J 16 42 3 41 26 69 204 5 18 52 194 5 5 to 5 25 Inject 1 gram sodium nitrite. 5 ?>h 20 38 194 6 15 19 30 7 to 7 18 Inject 12 c. c. tinct. digitalis. 8 15 20 90 264 • 9 25 21 112 194 11 16 54 11 15 14 31 6 Weigh t of dog, 1 1.5 kilos. Le ft pulmonary. 290 Horatio C. Wood, Jr. Openchowski severed the spinal cord in the thoracic region (exact location not given), stimulated the cervical cord and obtained no elevation of the pulmonary pressure. He concluded from this that the reason the pulmonary pressure failed to rise was because the section of the spinal cord prevented the systemic vessels from receiv- ing a constricting stimulus, and that there was consequently no increase of flow of blood to the risfht heart. T AISLE V. Time in min. and sec. Pulmonary pres.sure. mm. Hg. Carotid pressure, mm. Hg. Pulse rate per min. min. sec. 12 121 162 5 to 20 Inject 008 gram nitroglycerin. 23 16 70 180 50 11 67 190 1 5 11 88 200 1 10 to 1 20 Inject 0.008 gram nitroglycerin. 1 25 13 50 190 1 40 Begin injection of 0.7% NaCl solution. 2 45 17 114 186 10 2,S 149 186 Has received 950 c.c. salt solu- tion. 11 30 38 161 180 Begin asphyxia. 13 41 205 126 Traube waves in the carotid hut not in the pulmonary. 14 30 31 115 16 18 64 81 Weight Df dog, 12 5° H 1-2 5 5° 2 2 5 5° 3 3-4 5 5° 5 3-4 5 5° 7 3-4 5 5° 9 None 5 15° 2 " 5 15° 3 5 15° 5 " 5 15° 7 " 5 15° 9 " 14 5° 9 " 1 One blast ula in the entire lot of 3000 eggs. But eggs after an exposure to 4° C. for forty-five minutes formed one, and after an exposure of one hour formed five swimming gastrulae in five hundred eggs. After an exposure to 4° C. for two and three, four and a half, and seven and nine hours, there were formed respectively ten, JO 2 Arthur W. Greeley. forty-two, and one hundred gastrulte in five hundred eggs. After an eleven hours' exposure very rarely an egg reached the blastula stage, and after an exposure of eighteen hours there was no development at all. The control series contained a few irregular segmentations, a small pro- portion of which reached the morula stage. An exceptionally large proportion of these eggs were mature. Ninety per cent maturated. EXPERIMENT X. September 2, 1901. Temperature of the room, 23° C. Time between removal of eggs from fe- male and ex- Temperature. C. Duration of exposure to low tempera- ture. Number of gas- trula; formed in periment. 500 eggs. hours. hours. ^\ 4° 1 4 None 4i 4° 4 ■" 4| 4° 4 1 4* 40 1 5 4i 4° 2 10 4i 40 3 10 4^ 4° 4| 42 4i 4° 7 100 4i 40 9 100 4| 4° 11 1 5 40.50 18i None From the foregoing experiment.s it will be seen that an exposure to a temperature of from 4° C. to 7° C. during one to nine hours will cause fully maturated eggs of Asterias to develop into swimming larvje. The optimum temperature is 4 or 5° C, and the optimum period of exposure about six or seven hours, although this varies greatly in different lots of eggs. Unfortunately in August, when the experiments were performed, great difficulty was experienced in ob- taining mature eggs, viz: — those that would maturate on standing in sea-water. The proportion of mature eggs in those experimented on varied considerably, and, for this reason, there is a decided dissimi- larity in the quantitative results of the different experiments, although Artificial Parthenogenesis in Starfis/i. 363 in no case did an experiment fail when maturated eggs were used. In some of the experiments, the control series contained a very few irregular blastulas, which, however, did not develop into motile larvae. Their formation was probably due to a slight agitation the eggs re- ceived in being transferred from dish to dish, although every precau- tion was taken to prevent it. Their number is too small to affect the result. The development of the eggs exposed to low temperatures re- sembled closely that of normally fertilized eggs. Segmentation began in from one and a half to two hours after the eggs were trans- ferred to the temperature of the room, and was preceded by the for- mation of a fertilization membrane. The eggs divided regularly into two, four, eight cells, and at any stage of development, the parthen- ogenetic eggs could hardly be distinguished from fertilized eggs- The parthenogenetic blastulae, when removed to fresh sea-water, lived five or six days, and formed bipinnaria. A comparison of the time occupied in the development of parthenogenetic and fertilized eggs is given, the development of the parthenogenetic eggs being measured from the time of their removal to the temperature of the room. Beginning of segmentation . . Formation of swimming blastulre Formation of swimming gastrulas Formation of mouth invagination . Delage,^ in experiments on the unfertilized eggs of Asterias, ob- serves that immediately after the eggs have maturated, while they are in the so-called critical stage, they may be made to develop in several ways, viz: — by an increase in the concentration of the sea- water, by the chemical effect of certain ions, and by an increase in the temperature of the sea-water (to 30 or 35° C). His conclu- sion that the eggs may be made to develop by warming the sea- water, is open to criticism because he was not aware of the possibility of development through mechanical agitation, and apparently used no precautions to prevent it. In several experiments, I subjected eggs of the same lot as those used in the low temperature experiments to an increase of temperature ranging from 31° to 37° C. during periods of exposure of from one to seven hours. When great care was exer- cised in handling the eggs, not a single segmentation was produced. 1 Delage : Comptes rendus, 1901, cxxxiii, p. 346. Fertilized Parthenogenetic eggs. eggs. 1\ hours 1| hours 12 " lS-20 " 24 " 36 40 " 60 " 304 ArthiLr W. Greeley. Summary. 1. After maturation has been completed, the unfertilized eggs of Asterias Forbesii cat) be made to develop regularly into bipinnaria by an exposure to a temperature of 4" to 7° C. for from one to nine hours. 2. Segmentation of the Asterias ^g^ cannot be produced by rais- ing the temperature of the sea-water. ON THE PROLONGATION OF THE LIFE OF THE UNFERTILIZED EGGS OF SEA-URCHINS BY POTASSIUM CYANIDE. Bv JACQUES LOEB and WARREN H. LEWIS. \Froin the Hull JViysiological Laboratory of t/ie Univt-rsity of Chicagoi\ Introduction. IN a former paper Loeb pointed out that the experiments on arti- ficial parthenogenesis have a bearing upon the problem of the prolongation of life.^ The unfertilized mature eggs of a sea-urchin die comparatively soon when deposited in sea-water. The same eggs, however, live a longer time when caused to develop either artificially, by extracting a certain quantity of water from them or naturally, by allowing a spermatozoon to enter. Loeb concluded from this that there are two kinds of processes going on in the ^gg: one which leads to the death and disintegration of the ^gg — a mortal process: and a second which leads to cell divisions and further development. The latter process inhibits or modifies the mortal process. With this assumption the problem of the prolongation of the life of a cell was given a concrete form. According to this idea death and disintegration are due to specific processes which take place in the egg, and possibly in other or all living matter. These processes must be checked in order to render life possible. If this theory was of any value it was certain to lead to the discovery of artificial means by which the life of unfertilized eggs might be prolonged. The specific life phenomena are, as far as their chemical side is concerned, chiefly, if not altogether, catalytic phenomena. Hence it was to be expected that a checking of the specific mortal processes should be brought about by agencies which inhibit catalytic phenom- ena without permanently altering the constitution of living matter. Among all the agencies which act in this way, potassium cyanide seemed to meet this condition most perfectly. It weakens or inhibits a number of enzymatic processes in living matter without necessarily ^ Loeb : This journal, iv, 1901, p. 455. 305 3o6 Jacques Loeb and Warren H. Lewis. altering the constitution of the latter. When the potassium cyanide is permitted to evaporate, the original condition of the system may be restored. A series of experiments on the effects of KCN on the unfertilized eggs of sea-urchins confirmed our expectations and proved that by adding a small quantity of KCN to sea-water the unfertilized eggs of the sea-urchin can be kept alive a comparatively long time at a tem- perature of 20° C. or above. Method. The experiments were made at summer temperature, which prob- ably never went below 20"^ C. during our experiments, but was con- siderably higher at times. As a rule the ovaries of several females were used for each experiment, and the precautions described in former papers were applied to guard against contamination of the eggs by spermatozoa. A large quantity of eggs were kept in normal sea-water to serve as control material. The rest were distributed into various finger-bowls or flasks which contained sea-water to which various quantities of KCN had been added. After certain intervals some of the eggs were taken out of these solutions and put into nor- mal sea-water. Half an hour after the transfer had been made sperm was added to the eggs to determine whether they could be fertilized. The development of these eggs was carefully watched. In a number of experiments tests were also made to see whether or not the eggs were still able to develop parthenogenetically by extracting water from them. Eggs kept in Normal Sea-water. It was found that unfertilized eggs, when kept from i to 23 hours in normal sea-water, could not only be fertilized but also reach the pluteus stage (at a temperature of about 20" C). After that time they began to weaken. They could either not be fertilized at all, or their development stopped at an early stage. Kggs that had been in sea-water for from 24 to 32^ hours reached at the best only the gastrula stage, and only a small percentage of eggs developed. As a rule the eggs could no longer be fertilized after 32 hours. In the few cases where they were fertilized after that time they showed only the beginning of a segmentation. The latest date at which eggs that had been kept in normal sea-water could be fertilized was 48 hours. In this case, however, the effect of the fertilization consisted Prolonging Life of U^ifertilized Eggs of Sea-Ur chins. 307 in a few eggs showing the first segmentation : no membrane was formed. We were not able to repeat this observation. As a rule the eggs after from 24 to 32 hours became a sticky mass and assumed a dirty brownish color and this was the beginning of the complete dis- integration and putrid decay. Eggs kept in KCN Solution. Our stock solution of KCN was yV Of this solution we added small amounts to sea-water. A solution of 99 c. c. sea-water and i c. c. of the y^ KCN will be called an iq-Vo^ KCN solution. When we say that the eggs were kept in an g^f q- KCN solution, it means that they were kept in 98 c. c. of sea-water to which had been added 2 c. c. of the y\ KCN solution. All the KCN solutions in which the eggs were kept were made of sea-water to which a small amount of {\ KCN solution had been added. This way of proceeding was neces- sary, as the eggs of sea-urchins are very sensitive to changes of osmotic pressure as well as to KCN solutions that are stronger than gf 0- In the first series of experiments the solutions, in which the eggs were kept, were put into finger-bowls which were covered by glass plates. These glass plates did not fit air-tight, and hence a slight but continuous evaporation of KCN occurred. This evaporation was in- creased once or twice a day when we removed the cover in order to get out some of the eggs. It is thus evident that in these experi- ments the concentration of the KCN was a maximum at the begin- ning and decreased continuously during the experiment. We shall see later that this fact is not without influence upon the result. We found that the duration of life was longest when the eggs were kept in KCN solutions that varied between g^^'-Q and x-l^is- Most experiments were made with y o'Vo KCN solutions (99 sea-water -|- I y'o KCN). While the control eggs (that had been left in normal sea-water) could no longer reach the pluteus stage when fertilized 24 hours after they had been put into the sea-water, the eggs that had been kept in a yo\o" KCN solution in every case reached the pluteus stage when fertilized 72 hours after they had been put into the solution. When taken out of the KCN solution and put back into normal sea-water they developed into normal plutei, when sperm was added. In a few experiments we even got plutei from eggs that had been in the poisoned sea-water from 90 to 100 hours. These eggs not only reached the pluteus stage, but kept on devel- 3o8 Jacques Locb and Warren H. Lezvis. oping, and remained alive as long as eggs that had been fertilized immediately after being taken out of the ovary. Some of the plutei which developed from eggs that had been fertilized after a stay of 72 hours in poisoned sea-water lived five or six days and longer, which is a considerable time if we remember that the temperature of the water was never lower and often higher than 20° C. The longer the eggs remained in the poisoned sea-water the smaller the percentage became of those that yielded to fertilization and the earlier their development stopped. Eggs that had been in a loosely covered io\o KCN solution for five days never developed beyond the blastula stage. Eggs that had been in the poisoned sea-water over 72 or 80 hours as a rule no longer formed a membrane upon impregnation. More- over, the first segmentations were liable to be irregular. The Optimal Concentration of the KCN Solution. We have said that we used in most experiments a y^'Vo" KCN solu- tion. The weaker as well as the stronger solutions were less effective. Concentration of KCN. Result of fertilization after a 75 hours" stay in the solution. Pure sea-water No egg segments. 6T50 KCN No egg segments. 3"2^oo KCN No egg segments. re-ffoo KCN No egg segments. WW KCN Very few eggs show a beginning of segmentation. WW KCN Very few eggs show a beginning of segmentation. , 2 Wo KCN Few eggs go through the early stages of segmentation. iWo KCN Many eggs segment and develop into swimming larvae. tW KCN Many eggs segment and develop into swimming larvje. Wo KCN A few eggs develop into swimming larvas. Wo KCN No egg segments. 2^5 KCN No egg segments. Wo KCN No egg segments. Wo KCN No egg segments. Prolonging Life of Unfertilized Eggs of Sea-Urchins. 309 When the sea-water contained too little KCN it had little or no effect upon the prolongation of life. When we added too much it evidently altered the constitution of the egg and the latter could no longer be fertilized when taken out of the solution. The results of experiments with various concentrations of KCN can be seen from the preceding table. The first vertical column at the left indicates the concentration of the KCN in the sea-water. Eggs of the same females were distrib- uted into these mixtures, and after 75 hours eggs were taken out, put into pure sea-water (which was changed repeatedly) ; half an hour later sperm was added. The second column indicates the develop- ment of these various groups of eggs. It should be added that the last two solutions were more harmful than pure sea-water, inasmuch as they completely annihilated the power of development of the eggs in less than 24 hours. The Role of the Evaporation of KCN. In all the experiments mentioned thus far the KCN solutions which contained the eggs were kept in finger-bowls covered loosely with glass plates. It goes without saying that the KCN evaporated slowly but steadily from these solutions and that therefore the experi- ments were actually experiments with KCN solutions whose con- centration diminished steadily. This could be ascertained without titration by the fact that the odor of the solutions grew steadily weaker. It was necessary to determine whether or not this diminu- tion in the concentration of the KCN solution had any effect upon the result of these experiments. We put the KCN solution and the unfertilized eggs into the same loosely covered finger-bowls, but renewed the solution every 24 hours. When we used y^Vo" KCN solutions the prolongation of life which we had noticed before was diminished. Eggs that were put into a Y^Vo" KCN solution which was renewed every 24 hours lost their power of development almost completely in less than 75 hours. After that time only a small percentage began to develop when fertilized, and those that segmented never reached the swimming stage. Of those, however, that were kept in y-g^Vo KCN solutions which were not renewed, not only practically all segmented but all reached the pluteus stage. We then tried the effects of ^ ^\ ,, KCN solutions in small corked flasks, in which the evaporation of the KCN 3IO Jacques Loeb aitd Warren H. Lewis. was prevented. As was to be expected, the results were decidedly poorer than with jo'^q KCN solutions, which evaporated slowly. After ^6 hours a few of the eggs were able to segment when ferti- lized, but they did not reach the swimming stage. But even in these experiments the unfertilized eggs that were put into the xf^ KCN solution lived and preserved their capability of being fertilized consid- erably longer than the eggs kept in normal sea-water. These experi- ments were repeated many times with the same striking result. Maximal Prolongation of Life. From the previous experiments it appeared as if there were two conditions to be considered in the attempt to prolong the life of the unfertilized eggs of the sea-urchins. First, a comparatively high initial concentration of the KCN solution (about yoTu) seemed neces- sary (perhaps to stop suddenly certain injurious processes in the ^&&)- Second, if this high initial concentration was maintained it injured the constitution of the egg. We thought that by decreasing the concentration of the KCN solution more carefully, still better results might be obtained. This was indeed the case. We will describe the most striking experiment of this kind. The unfertilized eggs were put into a y-^ KCN solution. After 24 hours they were transferred into a j^oiy KCN solution, after 48 hours into a 2 o'o •> after 72 hours into a o^Vo' after 96 hours into a g^'oo , and after 120 hours into a go'oQ, ^" which they remained for the rest of the time. The solutions were kept in closed flasks. After certain intervals a portion of the eggs was transferred into normal sea-water and brought together with fresh sperm to test their power of develop- ment. The first portion of eggs was fertilized after having been in the solution 66 hours. About 80 per cent of the eggs seg- mented regularly, formed membranes and reached the pluteus stage. The second portion was taken out of the solution and fertilized after 90 hours. About 30 per cent of the eggs segmented regularly, but formed no membranes. They also reached the pluteus stage. The third lot was taken out and fertilized after 99] hours. About 20 per cent of the eggs segmented, some regularly, some irregularly. None had formed a membrane. They also reached the pluteus stage. The same occurred with the next lot of eggs, which were taken out after 112 hours, with the difference only that fewer eggs segmented. The eggs taken out after having been 120 hours in Prolonging Life of Unfertilized Eggs of Sea-Urchins. 311 the poisoned sea-water, upon fertilization, yielded a number of swim- ming gastrula; but no plutei. After 139 hours a new lot of eggs was taken out and fertilized. But few eggs segmented, some regu- larly, and none formed a membrane. They reached the blastula stage and swam about. After 144 hours the results were still similar, with the exception that the blastulae did not swim. The eggs that were taken out after 161 and 168 hours were still alive, and upon the addition of sperm reached the eight-cell stage, but did not develop further. They formed no membranes and few seg- mented. We discontinued the experiment at this point. We performed, with the same material, several other series of experiments. One series of experiments was performed with a yf KCN solution in finger-bowls covered with glass plates, which allowed some evaporation to occur. The solutions were changed every 24 hours simultaneously and in exactly the same way as in the chief experiment mentioned above. The slight evaporation seemed to be less advantageous, as the eggs ceased earlier to reach the pluteus stage, and after 144 hours were no longer even able to segment. It might be said that on the whole their duration of life was a little over 24 hours shorter than that of the eggs of the first series of experiments. Two further control experiments accompanied this series. One lot of eggs was left in the ^f q- KCN solution into which the eggs had originally been put. They were kept in finger-bowls covered with glass plates. These eggs lived but a little over five days ; that means, when fertilized, after having been in the poisoned sea-water 120 hours only a few eggs began to segment and these did not develop beyond the four-cell stage. The eggs that had been 139 hours in this solution did not segment when put back into normal sea-water and exposed to sperm. Eggs taken from this solution after 66 hours reached the pluteus stage when fertilized, but those taken out after 90 hours reached the gastrula stage only when fertilized. Finally one lot of these eggs was left in normal sea-water. It need hardly be said that after 66 hours not a single Q.g^ could be fertilized or was alive. Everything was disintegrated. We have varied these experiments in several ways, but it would be tedious to enumerate all these detailed experiments. It seems possible that by a further improvement of the methods the unfertilized eggs of the sea-urchin may be kept alive even longer than seven days. 312 Jacques Loeb and IVarren H. Lezvis. Artificial Parthenogenp:sis. In all the experiments mentioned thus far the power of develop- ment of the unfertilized eggs had been tested by adding sperm to the eggs. We tried whether the KCN preserved also the power of the Qgg to develop parthenogenetically. It was also our inten- tion to find out whether or not the power of the eggs to develop parthenogenetically disappeared sooner than their power of develop- ing sexually. In a number of experiments we removed unfertilized eggs at various intervals from the KCN solution, put them into lOO c.c. of sea-water, to which was added about 15 c.c. of a 2.7 ;/ KCl solution. After from li to 2 hours they were put back into normal sea-water. Loeb has shown that the unfertilized eggs of Arbacia reach the pluteus stage when treated in this way. In order to be able to express ourselves briefly we shall call this method osmotic fertilization. In one experiment we took eggs that had been in a Yo^Vo KCN solution (in loosely covered dishes) for 42 hours, and fertilized them osmotically. About 75 per cent of the eggs developed and many reached the pluteus stage. A second lot of unfertilized eggs were exposed to osmotic fertilization after they had been in the lo'Vo^ KCN solution for 66 hours. About 12 per cent of the eggs developed and some of them reached the pluteus stage. At the same time eggs of the control material of the same culture which had been kept in normal sea-water were also fertilized osmotically. From the eggs that had been in the normal sea-water for 23 hours we were able to produce plutei by osmotic fertilization. But after 42 or 66 hours not an egg even segmented when fertilized osmoti- cally. In another series of experiments we got almost identical results. In this series we obtained swimming blastulae from the control material by osmotic fertilization after the eggs had been in normal sea-water for 31 hours. In a third series we obtained parthenogenetic gastrulae from eggs that had been in a (loosely covered) 5^f KCN solution for 74 hours. In a fourth series we obtained a beginning of a parthenogenetic development up to the 32-cell stage from eggs that had been in a yf ^ KCN solution for 113 hours. Finally we tried how long the eggs would yield to osmotic fertilization when kept in a KCN solution whose strength was diminished a certain amount every 24 hours. We began with a sf Tj KCN solution and dropped gradually down to a o o'Vo solution. In this case we obtained parthenogenetic gastrukc from eggs that Prolonging Life of Unfertilized Eggs of Sea-Urchins. 313 had been in the poisoned sea-water for 99 hours and blastulae from eggs that had been kept in the KCN solution 113 hours. Even after 140 hours some eggs reached the 4 to i6-cell stage when fertilized osmotically. In each of these experiments the power of developing upon osmotic fertilization lasted almost, but not quite, as long as the power of developing upon natural fertilization. On the whole we might say that the power of parthenogenetic development ceased from about 12 to 24 hours earlier. This difference is probably due to the fact that the exposure o<^ the egg<=; to sea-water of a higher concentration in the act of osmotic fertilization injures the eggs slightly. Effects of Lack of Oxygk:-. The poisonous effects of KCN upon higher animals are chiefly if not wholly due to the inhibition of the oxidaiiv'e process in the tissues. The observations of Geppert, Spitzer, and Bredig point in this direc- tion. We had, therefore, to consider the possibility that the pro- cesses that lead to the death of the unfertilized ^.g^, and which are stopped by the KCN, are oxidations. We tried whether or not the life of the unfertilized eggs of the sea-urchin can be prolonged by depriving them of oxygen. We used two different methods of depriv- ing the eggs of oxygen. First, we kept them in gas chambers in which the air was driven out by a powerful current of carefully cleaned hydrogen. The second method consisted in putting the eggs with a few c.c. of sea-water into minute open flasks which were put into test-tubes containing a fresh mixture of 180 gms. KOH in 120 c.c. of water to which a solution of 5 gm. pyrogallol in 15 c.c. of H^O was added. The test-tubes were then sealed. At various intervals a tube was broken and the o.^^ exposed to fresh sperm in normal sea-water. In the latter series of experiments, the eggs died a little earlier than the control eggs kept in normal sea-water. After 29 hours the former were dead. Those taken out and fertilized after 22 hours, reached only the blastula stage. It is possible that the slight rise in temperature while the pyrogallol was put in and the tubes sealed, accounts for this difference. The eggs that were kept in gas cham- bers through which a constant current of hydrogen was sent, lived longer. At 4, [4, 27, and 38 hours after the beginning of the experi- ment, the unfertilized eggs could not only be fertilized, but developed into swimming larvae. Even after 64 hours, a few of the eggs were 314 Jacques Loeb and Warren H. Lczvis. still able to segment into two cells when taken out and brought into contact with sperm. But this was the ultimate limit. While this result is slightly better than those obtained with eggs kept in normal sea-water, it is so far inferior to the results obtained with KCN solu- tions, that we may say that the prolongation of life of the unfertilized egg by KCN is, if at all, only to a slight degree due to the prevention of the oxidative processes by KCN. The Effects of a Low Temperature. A priori one should e.xpect that the best way to prolong the life of the unfertilized eggs of the sea-urchin would be to keep them at a low temperature. The lowering of the temperature ought to stop the action of the enzymes. O. Schultze has shown that the fertilized eggs of frogs can be kept alive on ice for two weeks. We do not know whether any experiments of that kind have been made on the unfer- tilized eggs of frogs. It would not be correct to conclude that fer- tilized and unfertilized eggs behave necessarily alike in this respect. At the suggestion of one of us, Dr. Lyon investigated the effects of a KCN solution upon the fertilized eggs of the sea-urchin, and his results show that there is a characteristic difference between the two cases. As Dr. Lyon's paper will appear shortly, we do not need to discuss this difference here. We put one lot of a culture of unfer- tilized eggs on ice in normal sea-water (Lot I), a second lot (Lot II) on ice, in an y5^ KCN solution. A third lot of the same eggs was put into an Ytffny KCN solution, and kept at room temperature (20° C. or above) (Lot III), and a fourth lot was kept in normal sea- water at room temperature. After 50 hours 75 per cent of the eggs of Lot I, when they were taken from the ice, and sperm was added, developed, and many of them reached the gastrula stage, while 90 per cent of those of Lot II de- veloped. Lot III were equally good, but all of Lot IV were dead. After 63 hours, another portion of the eggs were put back into normal water at 20° C, and exposed to fresh sperm. Only 35 per cent of the eggs of Lot I developed, while 90 per cent of Lot II developed. After ']6\ hours, again a portion of eggs was taken out of these solutions and fertilized. Those of Lot I and II only reached the early segmen- tation stages, while those of Lot III developed into swimming larvae. This experiment shows clearly that the lowering of temperature (provided it does not go below the freezing point) produces changes Prolonging Life of Unfertilized Eggs of Sea-Urchins. 315 in the egg which weaken its chances for development, although one series of experiments should not carry too much weight. But while it seems as if in certain forms (/. r., the frog) the lowering of the temperature had only the effect of diminishing the velocity of chemi- cal changes, and of getting the animal into that condition which Claude Bernard calls latent life, in other forms it is decidedly different. A lowering of the temperature causes, for example, the growth of wings in plant lice. Hence a decided process of cell multiplication and growth is called forth by a lowering of temperature. In several former papers Loeb has pointed out the similarity of the effects of a lowering of the temperature and of a loss of water in cells. In certain Copepods and larvae of Polygordius, he found that loss of water on the part of the animal, as well as lowering of tem- perature, transformed them from negatively to positively heliotropic forms. At his suggestion, Mr. A. W. Greeley studied the effect of a reduction of temperature, and showed that by a lowering of the tem- perature (above the freezing point of water), Stentor coeruleus under- goes a definite series of morphological changes, which can also be produced by an abstraction of water from the animal. He made it probable that a reduction of temperature causes the cells of certain organisms to lose water. In unfertilized starfish eggs Mr. Greeley succeeded in producing the development of larvae, by keeping the unfertilized eggs on ice at a certain stage and for a certain length of time. The unfertilized eggs of sea-urchins could be caused in the same way to at last reach the 32-cell stage. This, then, shows that the eggs undergo a change when their temperature is kept for some time at a little above 0° C. Hence it appears intelligible that keep- ing the eggs of sea-urchins on ice does not prolong their life as much as if we keep them in an y^oIi I'^CN solution. Experiments on the Eggs of Starfish. The eggs of Arbacia are not transparent enough to permit us to convince ourselves whether or not the internal processes in the unfertilized o^gg are brought to a standstill through KCN. If this were not the case, we should have to consider the possibility that KCN preserves the unfertilized eggs longer on account of a bacte- ricidal effect. But observations on the more transparent eggs of the starfish (Asterias Forbesii) left no doubt that the preservation or lack of disintegration of the unfertilized egg is due to the interrup- v3 1 6 Jacques Locb and Warren H. Lewis. tion of certain progressive changes which in normal sea-water are going on in the Qg%. When the &g^ of the starfish is laid it is still immature. Its nucleus is large, and during the first two hours the reduction of the nucleus and the throwing out of the polar bodies occurs. If the eggs are not shaken, the nuclei become invisible, and the eggs, after 24 hours, look dirty, indicating the beginning of disintegration. We put the eggs of starfish, immediately after they were taken out of the ovary, into an y-q'q-o KCN solution. None of the changes characteristic of the process of ripening occurred in these eggs, and after 48 hours even, the nucleus was as large and dis- tinct as at the time they were put into the poisoned sea-water. But the eggs which had been put into normal sea-water had undergone the above-mentioned steady series of internal changes to complete disintegration. Nothing could be more striking than to compare the progressive series of changes in the eggs kept in normal sea-water with the unaltered appearance of the eggs put into the poisoned sea-water. Conclusions. 1. We have shown that the unfertilized eggs of the sea-urchin, kept in normal sea-water at a temperature of about 20° C, gradually lose their power of development. Their power to reach the pluteus state disappears, as a rule, after they have been in sea-water for about 23 hours. In the majority of cases the power of segmentation is lost when the eggs have been in sea-water 48 hours or even less. At that time, as a rule, the eggs form a sticky and discolored mass. 2. We have shown that the life of the unfertilized eggs of the sea- urchin can be prolonged materially by adding KCN to the sea-water. The best concentration of the KCN for this purpose is a mixture of about 100 parts of sea-water and one part of an jo KCN solution. But while this concentration is necessary at the beginning, life lasts longer when the concentration of KCN is gradually diminished during the experiment. By allowing part of the KCN to evaporate gradually, or by transferring the eggs into weaker and weaker solutions of KCN in sea-water, we could obtain plutei from eggs that had been in the KCN solutions for 112 hours, and we could get the beginning of a development of eggs that had been 168 hours in such poisoned sea- water. The eggs had, during all that time, been kept at a tempera- ture of 20° C, or above. Prolonging Life of Unfertilized Eggs of Sea-Urchins. 317 3. Not only the power of sexual, but also that of parthenogenetic development is prolonged through the potassium cyanide. 4. These experiments are another proof of the fact that, while weak solutions of KCN are able to stop certain processes in the cell, the old conditions of the system may be re-established when the solution of KCN is allowed to evaporate. If the KCN solution is of a higher concentration than -f^^ or gfo, the eggs may be injured permanently. 4. Lack of oxygen does not prolong, or prolongs but little, the life of the unfertilized eggs. 5. The lowering of the temperature seems to be far less effective for the prolongation of life in the unfertilized eggs of sea-urchins than the addition of KCN to sea-water. 6. As long as we consider death as something merely negative (namely, the cessation of certain processes), it must appear extremely paradoxical that the life of the unfertilized ^gg of the sea-urchin should be prolonged by applying one of the most deadly poisons. But the paradoxical element disappears, when we start from the assumption which led us to these experiments, namely, that in the unfertilized eggs specific mortal processes are going on, which are checked or modified by the process of sexual or osmotic fertilization. These specific mortal processes are also checked by potassium cyan- ide, which substitutes for the destructive action of these processes a condition of suspension of life ("vie latente" of Bernard). 7. We may next consider the question, What is the nature of the mortal processes in the unfertilized egg, and how can fertilization check them? No definite answer is possible at present. The mortal processes may consist in self-digestion, or in other enzymatic pro- cesses, or they may not be catalytic at all. How the act of fertiliza- tion can modify or check such processes, is not beyond analogy. We know that a full supply of oxygen decreases the fermentative action of zymase. Spitzer has made it probable that the cell nucleus con- tains an oxydizing agent, namely, the nucleoproteids. The process of fertilization results, in the egg of the sea-urchin, in a rapid series of successive cell divisions, in each of which the contents of the nucleus are scattered throughout the cell. It is easily conceivable that this periodic spreading or mixing of the contents of the nucleus and the cells may modify the chemical processes in the egg and check the mortal processes. CONTRIBUTIONS TO THF: PHYSIOLOGY OF THE CALI- FORNIA HAGFISH, POUSTOTREMA STOUTI. — II. THE ABSENCE OF REGULATIVE NERVES FOR THE SYSTEMIC HEART. By CHARLES WILSON GREENE. \Fi om the Pliysiological Laboratory of the University of A/issoiiri.] ^ THE Cyclostomes, or hagfishes, have become classical in mor- phological literature. Their primitive vertebrate characters and their position in the vertebrate series make them in much demand for embryological and morphological research. Very little physio- logical work has been done with this group, yet their physiological interest promises to be great. The hagfishes are relatively large and tenacious of life, two qualities very favorable to physiological research. The representative of the group on our western coast, especially abundant in Monterey Bay, reaches a length of forty to fifty centimetres. It is readily taken by means of the trawl or in traps, and can be kept alive in the aquarium with ease. Its tissues live under experimental conditions for hours or even days. In this brief paper I shall present the results of a series of experi- ments made to establish the relation of the nervous system to the activity of the heart in Polistotrema stouti. The slight cartilaginous skeleton of the animal permits the preparation of organs with great facility. On the other hand, the slightest muscular movement of any part of the body is sufficient to displace the heart and to render the recording of its movements a matter of peculiar difficulty. The successful records finally made were secured by pinning the body muscles firmly to a holder and independently supporting the heart, and by curarizing the animal. Stimulation of the vagus nerve. — Johannes Miiller described the 1 The experiments upon which this paper is based were performed at the Hop- kins Seaside Laboratory, Cahfornia. I take this opportunity to express my obligations to the Directors, Dr. C. H. Gilbert and Dr. O. P. Jenkins, for the facilities of the Laboratory. 318 Regitlative Nerves for Systemic Heart of HagfisJi. 319 vagus nerve of the hagfish as sending branches to the gill sacs, to the heart, and to the stomach-intestine. A small filament, doubtless Miiller's cardiac branch, runs toward the heart in Polistotrema stouti but by macroscopic methods I have never been able to trace it nearer than the tissue just dorsal to the pericardial wall. The motor fibres to the gill sacs furnish a ready means of deter- mining whether or not a given stimulus is effective. The muscles of the gill sacs are always thrown into contraction by even a weak stim- ulus applied to the vagus, so also are the well developed constrictor cardiae at the entrance of the stomach. The stren2:th of current 1 1 n I I I 1 1 I I 1 1 1 1 ni l i imi mi ll I I II III I I I mi II m i l I nii iiiii i I I 1 1 i ni I I 1 1 mi l l n il I I m i 1 1 mi l l Figure 1. — Experiment, Decevtber 2S, IS'JO. Record of contractions of the ventricle upon simultaneous stimulation of the vagus nerves in the hagfish, Polistotrema stouti. Two thirds the original size. Strength of the interrupted current, 500 units. Du Bois Reyniond coil (Petzold manufacture) fed by one Edison-Lalande cell. Time in seconds. necessary to produce contraction of these gill sacs is from twenty to thirty units by the proportionate scale of the Petzold inductorium used when fed by one Edison-Lalande cell. The vagus was stimulated at different points along its course from the point of origin within the cranium to a point just anterior to the heart itself. It was stimulated with the nerve intact and with the nerve cut. Each nerve was stimulated by itself and also both together. The strength of the current was varied from one to one thousand units and the rate of interruption was also varied, The results of a series of such tests are presented in the table on page 320. The heart rates given in this table are computed from counts made for equal periods of time immediately preceding, during, and following vagus stimulation. The error of measurement reaches a maximum of four to five tenths of a beat per minute. Slight variations due to causes which produce a general increase or decrease of rate may, of course, fall within the period under consideration. Of these outside factors among the most important are those influences affecting the return of blood to the heart. Taking into consideration these factors, it seems to me that in no case has a change of heart rate of sufficient magnitude occurred to justify the assumption of a direct ;20 Charles Wilson Greene. vagus influence. The accompanying figure gives one of a series of experiments in which the strength of current was varied from ten to one thousand units. The failure to discover inhibitory nerves in the vagus of the hagfish was a great surprise to me, hence I immediately began experiments to determine by what other path the heart received its supply of regula- tive nerves. The spinal nerves are too delicate to isolate, hence experiments were necessarily confined to an exploration of the brain and spinal cord. TABLE SHOWING HEART RATE WITH VAGUS STIMULATION. Date 1S99. Nerve stimulated. Strength of stimulus. Duration of stimu- lation. Rate be- fore stim- ulation. Rate dur- ing stim- ulation. Rate after stimula- tion. units. sec. Vagus 200 18 26.6 26.4 26.5 Vagus 400 18.4 25.2 25.2 25.2 Dec. 28 Right vagus 500 10 25.2 24.8 24.8 a u Right and left 500 30 24.5 24.6 24.5 .« « vagus Right and left 10 13 24.0 23.8 24.0 .< ■. vagus Right and left 100 17.6 24.0 24.1 24.0 ., <. vagus Right and left 1000 18 23.3 23.3 23.3 vagus Cranial stimulation. — The brain was stimulated with platinum elec- trodes, the bipolar and unipolar methods both being used. The stimulation of different parts of the brain, especially of the medulla and of the roots of the cranial nerves, gave entirely negative results in so far as any influence affecting the heart contractions is concerned. Stimulation of the spinal cord. — It was hoped that by Stimulating different sections of the spinal cord any regulative nerves that might reach the heart by this path would be discovered. My results here also were wholly negative, although numerous experiments were per- formed. It is true that in this series of experiments the heart rate was sometimes slightly increased. The stimulation of the cord was always followed by vigorous contractions of the great lateral muscles, and any one who has ever handled a live hagfish will realize what a pronounced effect the muscular action will have on the tension of Regulative N^erves for Systemic Heart of Hag fish. 321 the heart. These efifects cannot be absohitely eliminated except in the curarized animal and the changes in tension and pressure are great enough to produce a change of one or two contractions per minute. I have reached the conclusion, therefore, that regulative nerves for the heart do not pass out through the spinal nerves in the hagfish. At the suggestion of Dr. Howell, to whom I am greatly indebted for inspiration and advice. I have repeated the experiments along two lines, namely, stimulation of the vagus with the heart not exposed to the air, as it was in the experiments quoted above, and stimulation by applying the electrode directly to the venous sinus. These experiments were kindlv performed for me by Mr. W. F. Allen during the summer of 1900, and I have since repeated the experiments myself. Stimulation of the vagus in the neck had no effect on the undisturbed heart. Upon stimulating the venous sinus there was an increase of two or three beats per minute following the first stim- ulation, an increase that remained permanent for that series. In the experiment quoted the rates were counted in the order given. Hagfish No. X. Natural heart rate per minute 28, 27, 28, 28, 28. Vagus stimulated, heart not exposed to air 30, 30, 31, 30, 30. Sinus venosus stimulated 30, 31, 30, 31, 31. If the experiments occurred in the order : i normal rate, 2 vagus stimulation, 3 sinus stimulation, 4 normal, 5 vagus stimulation, etc., it was found that an increase in rate of one or two beats per minute followed the first stimulation but that the rate did not return after- ward to the former normal. My interpretation of this slight change is that it is a direct effect of the general muscular contractions of the animal. When the heart is left in the pericardium it is covered by a sheet of ventral muscle and the pericardial epithelium. Any body movements occurring greatly facilitate the flow of the blood and at the same time produce a decided change in the pressure on the heart as suggested above. These two factors act to increase the heart rate. Curara acts in the usual way when injected into the hagfish. How- ever, it takes a much larger dose to produce paralysis, and it acts more slowly than on a frog or a mammal. Curara eliminates the influence of the motor nerves of the cord on the body muscles, and also the vagus influence on the muscles of the gill sacs. Stimulation 32 2 Charles Wilson Greene. of the vagus in the curarized animal does not produce the usual con- tractions of the gill muscles nor of the constrictor cardiac, yet direct stimulation of these muscles is followed by contractions. In the curarized hagfish, stimulation of the vagus produces no visible cardiac effects. Records taken under these conditions continue with unin- terrupted rhythm, force, and sequence. The experiments outlined in this paper were performed first during the summer of 1899, but the results were so exceptional that no report was published until further verification could be had. The experiments were repeated at intervals during the late summer, the early autumn, the following Christmas holidays; in March and Sep- tember, 1900; and in July, 1901. The animals were taken directly from the aquarium, and were in good condition. They gave unquestioned evidence of the physio- logical activity of the muscular and of the nervous tissues for hours after being experimented upon. The experiments were performed during the four seasons of the year. The results are uniform throughout the entire series of experiments, and lead to the conclu- sion, which may now be announced with confidence, that tJie Califor)iia hagfish, Polistotreiiia stoiiti, docs not possess regulative nerves for the heart. Since the demonstration ot the inhibitory nerves for the heart by the Weber brothers, this class of nerves has been shown to be present in numerous species throughout the vertebrate series. I can find no exception whatever in the literature available. It never occurred to me that an exception was possible until I attempted to use the hag- fish for demonstrating the vagus influence upon the heart to my class in physiology at the seashore. Harrington ^ found a certain amount of cardiac resistance to vagus inhibition in guinea pigs during October and January. At this sea- son, stimulation of the vagus nerves produced only a certain amount of cardiac slowing with fall of blood-pressure, but never heart stand- still, no matter how strong the stimulus. From February to April, Harrington secured the usual complete standstill of the heart with a sudden and pronounced fall of blood-pressure. He suggested that the difference noted in his experiments was due to a variation in vitality associated with the season, possibly due to a diminution of fresh air and light during the winter months. But in Harrington's 1 Harrington : This journal, 1898, i, pp. 383-394. Regulative Nerves for Systemic Heart of Hag fish. 323 experiments there was not a total absence of vagus influence on the heart. A large number of bony fishes and several sharks and rays have been shown to possess depressor nerves for the heart. Fishes are mentioned in the original list of animals given by the Weber brothers in which cardiac inhibition was produced by vagus stimulation. I have myself demonstrated the presence of such nerves in eleven different species of bony fishes and one shark, common in Mon- terey Bay. Among the invertebrates investigated, some have been shown to possess complex cardiac regulative nervous mechanisms. Perhaps the best known of the invertebrate cardiac nervous systems, is that of the cephalopod molluscs. In the octopus and in the squid, both inhibitory and accelerator nerves are distributed to the heart. ^ The cardiac fibres pass to the heart by way of the visceral nerve. The fibres to the branchial heart are exclusively motor, i. e., accelerator. Conant and Clark,^ in an excellent research, demonstrated that the American edible crab, Calinectes hastatus, possesses a cardiac ner- vous mechanism. From the anterior part of the thoracic ganglion, are given off two pairs of accelerator, and one pair of inhibitory nerves, which run to a pericardial plexus. Illustrative tracings are presented, showing cardiac acceleration and inhibition following the stimulation of these nerves. The land snail, Helix, possesses cardiac inhibitory nerves, as proven by the papers of Young, Foster, Biedermann, and Ransom.'^ On the other hand, the sea snail, Aplysia, possesses a cardiac accelerator nerve, but no inhibitory nerve. Ransom stimulated the " visceral nerve " and always obtained cardiac acceleration. More recently, Bottazzi and Enrique have demonstrated accelerator nerves for Aplysia depilans and A. limacina. They also failed to find in- hibitory nerves. The presence of cardiac regulative nerves in so many invertebrates has tended to strengthen the assumption that such nerves were present in all of the lower vertebrates, an assumption which has not before been questioned. 1 Ransom : Journal of physiology, 1884, v, pp. 261-347. Bottazzi et Enrique: Archives itahennes de l^iologie, 1901, xxxiv, pp. 111-143. - Conant and Clark : On tlie accelerator and inhibitory nerves to the crab's heart, Journal of experimental medicine, 1896, i, pp. 1-7. ^ Ransom : Loc. a't., p. 261. Other literature is cited in Ransom's article. 324 Charles Wilson Greene. The present experiments show that in the hagfish, the lowest of the craniata, there are no cardiac regulative nerves. In this fact we have a striking illustration of the automaticity of cardiac muscular tissue. The hagfish heart is comparable to the heart of an embryo before nerves have entered. Any regulation of the heart's action must depend upon the conditions which affect the muscle directly, i. e., tension, nutrition, etc. The volume and pressure of the blood coming to the heart and the changes in the pressure upon the inter- nal organs produced by the ever-varying movements of the plastic body of the animal are the factors that must have a decided influence on the hagfish heart. Questions as to the influence of nutrition I wish to discuss at another time. Among zoologists there is some discussion as to whether or not the apparent primitive structure of the hagfish may be, in reality, a retrograde or degenerate condition. This question may be raised concerning the physiology of the heart. It is, indeed, a question that cannot be arbitrarily settled, and one that must be taken into con- sideration in any discussion of the acquirement of a cardiac-regulative nervous mechanism in the vertebrate series. THE PHYSIOLOGICAL ACTION OF FORMALDEHYDE. By WALDEMAR KOCH. [F?-oi/i the Hull Phvs'ologkal Laboratory of the University of C/iicaqv, III.'] THE disinfectant action of formaldehyde was first pointed out by Loew.i A number of years later F. Blum '^ introduced for- mol or formalin (a concentrated solution of formaldehyde) into his- tological technique as a hardening and preserving agent. Since then, with cheapened methods of manufacture, formaldehyde has found more general application in the arts and trades. Its use as a preservative for food materials, naturally led to the investigation of its physiological properties. F. Blum ^ had already studied its action on proteids, and Benedicenti,^ extending this in- vestigation, called attention to the relation of these proteid compounds to aldehyde ammonia combinations, on account of the ease with which formaldehyde could be again split off by means of steam distillation. He was also the first one to point out the action of formaldehyde on the blood, changing haemoglobin to hasniatin. The more recent experiments of M. Fischer^ on animals leave no doubt that formal- dehyde is an intense protoplasmic poison. The observations of F. W. Tunnicliffe and O. Rosenheim*' on children directly contradict Fischer's results, but can be explained by considering the extremely dilute solutions used. The results obtained by various investigators with reference to the action on enzymes indicate that proteolytic enzymes are more or less hindered in their action, while starch-, sugar-, or fat-splitting enzymes are very little affected. The first attempt to offer an explanation of the intense bactericidal action of formaldehyde was made by Van't Hoff. In his " Vorlesungen iiber theoretische und physikalische Chemie " Vol. Ill, p. ii8, he states: "Oxide of 1 LoEW : Miinchener medicinische Wochenschrift, 1888, p. 412. '^ F.Blum : Anatomischer Anzeiger, 1895, xi, p. 718. ^ F. Blum : Zeitschrift fiir physiologische Chemie, 1896, xx, p. 127. ^ Benedicenti : Archiv fiir Physiologie, 1897, p. 210. ° M. Fischer : Journal of the Boston Society of Medical Sciences, 1900, V, p. 18. ^ F. W. Tunnicliffe and O, Rosenheim : Journal of hygiene, 1901, i, p. 319. 325 326 Waldemar Koch. methylen (formaldehyde or formalin) oxydizes itself even on ex- posure to the air. The active oxygen formed as a result of this reaction may be made to account for the antiseptic properties of formaldehyde." ^ The present investigation was begun with a view to testing the above suggestion. If formaldehyde must form active oxygen in order to develop its antiseptic action, it should not be poisonous to anaerobic forms of life. As yeast can easily be made to grow anas- robically, it was used in this research. Two tubes half filled with a fermenting solution were inoculated with a little compressed yeast, and into one was placed a small thin glass bulb, filled with a one per cent formaldehyde solution, freed from oxygen by boiling and sealing the bulb while steam was escaping. The air was then expelled from the tubes by a stream of hydrogen and after fermentation was well under way the formaldehyde bulb was broken and both tubes were closed. After an hour the tube containing the formaldehyde had developed little or no gas, while the control tube gave considerable pressure. A microscopic examination of the yeast cells exposed to the formalin showed them to be irregular in outline, with the protoplasm of a granular appearance, reminding one of starved or drying yeast cells in which katabolic reactions have come to a stop. In order to study more carefully the immediate effects of formaldehyde, fermentation tubes filled with a fermenting solution were sealed by filling the bend between upright tube and bulb with mercury and inoculated with about ten milligrams (dry weight) of compressed yeast. The amount of gas collected in such a tube can be considered a measure of the growth taking place. The fermenting solution was made up as follows : Glucose 100 grams. Sodium potassium tartrate .... 4 grams. Ammonium nitrate 2 grams. Sodium carbonate 2 grams. Water 1000 c.c. Sodium indigo sulphonate sufficient to give a blue color. The indigo in this solution soon decolorizes, forming with glucose an unstable compound, which breaks up in the presence of air or oxygen, thus forming a valuable indicator for small quantities of ^ " Methylenoxyd (Formaldehyde oder Formalin) o.xydiert sich schon an der Luft, womit wobl Sauerstoffactivierung und dadurch antiseptische Wirkung zusammenhanst." The Physiological Action of Formaldehyde. 327 oxygen. The mixture has no power of absorbing oxygen like pyro- gallol, but serves merely as an indicator. Formaldehyde does not give such a colorless compound with indigo. Concentra- tion of formal- dehyde added. Before adding \ formaldehyde. After adding formaldehyde. Time in hours. Gas collected. c.c. Time in hours. Gas collected. c.c. I. Control 1 2.5 3.75 9.0 9.0 II. 0.1 per cent 2.5 ' 2.S 21.0 12.0 III. 1.0 " " 2.0 4.7 21.0 3.6 IV. 1.0 " " 0.0 0.0 20.0 00 V. 5.0 " " 2.5 3.0 14.0 0.6 VI. 10.0 " " 2.5 5.1 18.5 1.5 1 A Capac dded one-half c.( ty of tubes 12-15 c :. of water instead c c. f formaldehyde al 'ter yeast had grown two and one half hours. The amount of formaldehyde necessary to prevent the growth of yeast will be seen from the above table. In each case one-half cubic centimetre of formaldehyde of the strength indicated was added, after the yeast had grown for several hours, except in IV, where it was added immediately. As usually about ten cubic centimetres of fermenting solution remained in the tube, the concentration of the formalin added must be divided by twenty to give the strength of solution to which the yeast was exposed. We see thus that 0.05 per cent (i : 2000) kills growing yeast, while 0.005 per cent (i : 20,000) does not. Another interesting fact is obvious on comparing III and IV, namely, that the immediate addition of the formalin prevents all fermentation, while adding the formalin to fermenting yeast permits varying and always considerable amounts of gas to collect. Two factors must be considered in explaining this result. First the reten- tion of carbon dioxide in the meshes of the growing yeast from which the gas is gradually liberated, and secondly fermentation due to zymase after the cell is dead. The fact that zymase is not sensitive 328 Waldemar Koch. to formaldehyde was first indicated by the work of A. Macfayden, G. Morris, and S. Rowland,^ and is also very strikingly shown in the followino- table. For a number of hours after the addition of the Control. Potassium arsenite. Formaldehyde. 15 per cent solution. 1 per cent. Time in hours. Gas Gas Gas collected. Rate. collected. Rate. collected. Rate. 1.0 c.c. c.c. c.c. 1.2 0.4 0.9 0.4 0.9 0.4 1.5 1.6 0.7 1.3 0.5 1.3 0.7 2.0 2.3 O.S 1.8 0.6 2.0 0.7 2.5 3.1 O.S 2.4 0.6 2.7 0.7 3.0 3.9 0.7 3.0 0.6 3.4 0.5 3.5 4.6 0.5 3.6 0.6 3.9 0.9 4.0 51 0.9 4.2 0.7 48 0.9 4.5 6.0 09 4.51 0.3 5.72 0.4 5.0 6.9 1.2 4.8 03 6.1 0.3 5.5 8.1 1.0 5.1 01 64 0.2 6.0' 9.1 10 5.2 0.2 6.6 01 6.5 10.1 1 09 5.4 0.3 6.7 0.2 7.0 11.0 1.0 5.7 0.3 69 0.3 SO 12.0 1.4 6.0 0.1 7.2 0.2 9.0 13.4 0.9 6.1 0.2 7.4 04 10.0 14.3 07 6.3 0.3 7.8 0.3 11.0 15.0 6.6 0.3 81 0.6 23.0 Tube empty 6.9 8.7 1 Added on e half c.c. K.^HAsOg sol. 2 Added fo maldehyde one half ( ;.c. arsenite as well as the formalin a slow but steady evolution of gas continues, gradually coming to a stop. The complete parallelism of 1 A. Macfayden, G. Morris, and S. Rowland : Berichte der deutschen chemischen Gesellschaft, 1900, xxxiii, p. 2764. The Physiological Action of Formaldehyde. 329 the arsenite and the formalin is apparent from the above table, and seems to point pretty clearly to the fact that the zymase continues its action in the presence of formaldehyde as well as in the presence of potassium arsenite, though a much stronger solution of arsenite is required. We have seen then, that the action of formaldehyde does not depend on active oxygen, neither does it affect zymase. Another possibility to be considered is its interference with the proteolytic enzyme, which has been found to be present in yeast by Kutscher' and others. The fact that formalin interferes with tryptic digestion has been observed by a number of investigators. Whether or not the trypsin is itself affected by formalin has never been clearly shown. The following experiments were therefore performed to settle this point. Experivient I. — To 10 c.c. trypsin solution (covered with toluol) added 2 c.c. 10 per cent peroxide of hydrogen. Fibrin digested. Experimerit II. — To 10 c.c. trypsin solution (covered with toluol) added i c.c. 1 per cent formaldehyde and immediately afterwards 2 c.c. 10 per cent peroxide of hydrogen to destroy formalin. Ftbrifi digested. Experiment III. — To 10 c.c. trypsin solution (covered with toluol) added i c.c. I per cent formaldehyde. Fibrin 7iot digested. After 12 hours added 2 c.c. 10 per cent peroxide of hydrogen to destroy formalin. A fresh piece of fibrin was digested. Experiment IV. — To 10 c.c. of a fresh trypsin solution (covered witli toluol) added the fibrin which had not been digested in Exp. III. // was shnvly digested. A fresh piece of fibrin treated with a stronger solution of formaldehyde was not so easily digested. These experiments show that the enzyme, like a starch or fat split- ting enzyme, is not directly affected by formaldehyde. The fibrin, on the other hand, is rendered more or less indigestible, depending on the strength of the formaldehyde and the time of exposure. We may conclude from this that formaldehyde does not act by forming active oxygen, nor by destroying the zymase, but brings about the death of the cell indirectly by rendering its proteid food supply useless and by preventing the digestion of proteids always going on within the protoplasm of the cell, a reaction intimately connected with the life of the organism. ^ KuTSCHF.R : Zeitschrift fiir physiologische Chemie, 1901, xxxii, p. 476. ON THE RELATION OF LIPASE TO FAT METABOLISM — LTPOGENESIS. By a. S. LOEVENHART. [From the Laboratory of Physiological Chet?iistry of the Johns Hopkins University^ SINCE it has been shown that lipase is reversible in its action ^ the author has felt that this fact should throw light on the history of fat in the organism, its absorption, its storing up and its utilisation. Inasmuch as all that follows depends on the reversible action of lipase, it may be well to give a brief account of the experi- ments which demonstrate this fact. It was found that lipase hydrolyses ethyl butyrate with great readiness and in the work referred to this ester was used because it offers many experimental advantages over a true fat. The evidence that lipase is reversi- ble in its action may be briefly stated as follows: — first, the hydrolysis of ethyl butyrate by lipase is incomplete; second, the hydrolysis is inhibited by the products of the reaction, and third, we were able to synthesise ethyl butyrate by the action of lipase on butyric acid and ethyl alcohol. When a fresh aqueous extract of the pancreas is treated with a mixture of dilute butyric acid (f^ to 2%) and ethyl alcohol (sufficient in quantity to bring the whole to i^ per cent) the very characteristic odor of ethyl butyrate soon develops even at ordinary temperature and in the presence of anti- septics, whereas if the pancreatic extract is first boiled the mixture never develops the odor of ethyl butyrate. When this experiment is performed on a large scale a light oil can be distilled from the mixture which possesses the odor and general properties of ethyl butyrate. Further, it can be hydrolysed by lipase, and butyric acid can easily be proven to be one of the products of hydrolysis. The experiment is readily performed and with an active pancreas the odor of butyric ether can be detected in an hour. These experiments leave no doubt that lipase is reversible in 1 Kastle and Loevenhart : American chemical journal, 1900, xxiv, p. 491. 331 332 A. S. Loevenhart. its action. Since the above work appeared, Hanriot,^ unaware of our work has found that lipase can synthesise monobutyrin from butyric acid and glycerine. In the physiological application of the reversible action of lipase the inference is made that lipase is also reversible in its action on the higher fats. The chemical analogy between the fats and other ethereal salts puts this inference almost beyond question. Both are hydrolysed by the same agencies and all the methods of synthesising the one are applicable to the other. Thus hydrochloric acid will i accelerate the hydrolysis of both fats and ethyl butyrate. Under proper conditions it can also effect their synthesis. Similarly we know lipase can hydrolyse both the fats and ethyl butyrate, and since we have proven that it is capable of inducing the synthesis of ethyl butyrate it appears highly probable that it will also induce the synthesis of fats. The light which these experiments throw on the physiological transformation of fat will be considered under two heads, viz. i. The absorption of fat, 2. The storing up and utilisa- tion of fat, — Lipogenesis. The Application of the Reversible Action of Lipase to the Theory of Fat Absorption. Serious objections have been urged against the various theories which have been advanced from time to time to explain the manner in which fat passes from the lumen of the intestine into the central lacteal. Without going too deeply into a discussion of the previous work on fat absorption, it may be said that the old view first advanced by BriJcke, that the fat is taken up by the intestinal mucosa in par- ticulate form has been gradually abandoned in the light of more recent work. Recently, however, Hofbauer^ has attempted to show that fat is absorbed without being previously hydrolysed. He fed fat stained with alcanna red and other pigments and showed that although these pigments are insoluble in water the fat appears in the thoracic duct colored red. Pfiiiger ^ has shown, however, that the coloring matter is soluble in bile, soaps, etc., thus explaining its absorption. Hence 1 Hanriot: Comptes rendus de la society de biologic, 1901, p. 70. ~ '^ HoFBAUER : Archiv fiir die gesammte Pliysiologie, 1900, Ixxxi, p. 263. 3 Pfluger: Ibid., p. 375. Relation of Lipase to Fat Metabolism — Lipogenesis. 333 Hofbauer's work seems to throw no light on the subject. As Rach- ford 1 has shown that the pancreatic juice is capable of hydrolysing all the fat of a fatty meal in the period of pancreatic digestion there is no reason to believe that much of it escapes hydrolysis. As to the absorption of the products of cleavage there is a diver- gence of opinion, Munk maintaining that it is absorbed as free fatty acid, while others believe that it is absorbed as soap. However the fatty acid is absorbed, there is certainly a synthesis of fat in the epithelium, as these cells can be seen to contain fat granules during absorption. Ewald ^ found indication that the dry powdered mucous mem- brane of the small intestine is capable of synthesising fat from a proper mixture of glycerine and soap. Hamburger ^ has recently repeated Ewald's experiments with some modification using the mucosa of large intestine and has reached the same results. In both cases the only test for fat formation was an increase in material soluble in ether. The question as to how the fat which is synthesised in the epithelial cells reaches the central lacteal, has also been a matter of controversy. These granules cannot be egested as such from the attached border of these cells. The view that the fat granules leave the cells as such from the attached border is perhaps less acceptable than that they are taken up in this form from the lumen of the intestine. It seems even more certain that the fat leaves the cells in solution than that it enters them in this condition. It is believed that the demonstration of the reversibility of the action of lipase throws new light on the question of fat absorption and its passage to the central lacteal. In considering fat absorption in the light of this view two facts must be borne in mind, first, the hydrolysis of fats by lipase is incomplete unless the products of cleavage are removed from the field of activity, and second, lipase is capable of synthesising fat from fatty acid and glycerine, and here too the action of the lipase is incomplete. A mixture of fatty acid and glycerine reaches equili- brium in the presence of lipase only when a certain amount has been synthesised into fat, and it may be said that lipase merely has ■ 1 Rachford : Journal of physiology, 1891, xii, p. 72. 2 Ewald: Archiv fiir Physiologic, supplement, 1883, p. 302. ^ Hamburger: Ibid., 1900, p. 433. 334 ^' '^- Loevenhart. the power of hastening the production of equilbrium in such a mixture. The absorption of the products of digestion makes the splitting of fats in the intestine more or less complete. In the epithelial cells there is undoubtedly a synthesis of fat from the absorbed fatty acid and glycerine, and it is believed that this synthesis is occasioned by an enzyrne contained in these cells, which is capable of effecting fat synthesis or fat decomposition according to conditions. The occurrence of a lipolytic enzyme in these cells is proven by the following experiments : Intestinal mucosa. — The small intestine of the pig was removed very soon after death and thoroughly washed with running water, this being continued for some time after the washings became per- fectly clear. The intestine was then opened and the mucosa scraped from the upper part. Five grams of this were ground with white sand and extracted with 50 c.c. water. To test the activity of the extract, tubes were prepared containing 4 c.c. water, i c.c. of the extract, 0.26 c.c. ethyl butyrate and o.i c.c. toluene. After acting at 40° for fifteen minutes the tubes required 0.2 c.c. KOH to neutralise the butyric acid formed, this corresponding to a hydrolysis of 0.5 per cent. On standing ten hours at room temperature (21°) they required 0.95 c.c. ^ KOH, corresponding to 2.38 per cent hydrolysis. Blanks in which the boiled extract was used remained neutral. The lipolytic activity of the intestinal mucosa shows great indi- vidual variation. Occasionally much greater activity than that given above has been found and some specimens show very slight activity. The part played by the lipase in these cells during fat absorption would seem to be as follows : when the fatty acid and glycerine are taken up by the cells, the lipase in them begins to establish equilibrium between these products, which requires that a certain amount combine to form fat. In the epithelial cell at this stage, fat, fatty acids, and glycerine are simultaneously present. Inas- much as fatty acid and glycerine are constantly diffusing out of the cell through its attached border, the fat contained in it would soon disappear if these products were not being simultaneously taken up by the cell from the lumen of the intestine. But while the cell is absorbing the products of fat digestion, fat will remain in it and the quantity found, together with the length of time it remains there, will depend on the relative rates Relation of Lipase to Fat Metabolism — Lipogenesis. 335 of diffusion of the fatty acid and glycerine through the free and attached borders of the cell, viz., the relative rates of entrance and exit. Hence during rapid absorption the epithelial cells are seen to contain more fat than when absorption is slower. According to the view I have just advanced, the free fatty acid and glycerine pass into the central lacteal and the occurrence of fat in the intesti- nal mucosa is not an important factor in fat absorption, since it must again be hydrolysed. In the thoracic duct, however, we find that the split products have been largely but not entirely recom- bined to form fat. Munk ^ found that the lymph from the thoracic duct contains from 5 to 10 per cent of its fatty acid in the free state, thus showing that the synthesis is not complete. Some have supposed that this synthesis takes place in the epithelial cells, but this leaves no expla- nation as to how the fat leaves these cells. Munk thought that the synthesis occurs in the lymphatic glands. With a view of throwing some light on this question a study of the lipolytic activity of lymph and lymphatic glands was undertaken. Lymphatic glands. — The lymphatic glands of a large bitch were removed immediately after death and 20 grams were extracted with with 100 c.c. water. Tubes containing 4 c.c. water, i c.c. extract, 0.26 c.c. ethyl buty- rate, and o. i c.c. toluene and kept at 40° showed a hydrolysis of 0.5 per cent in fifteen minutes. After forty hours at room temperature (21°) they showed 2.76 per cent hydrolysis. Lymph. — The lymph was collected from the thoracic duct of dogs. The following may be taken as representing the average activity of the lymph. To 2.5 c.c. of lymph serum was added 0.2 c.c. litmus and 1.3 c.c. f^ butyric acid, which brought the litmus to the neutral tint; then i c.c. water was added, making the total volume 5 c.c. To this 0.26 c.c. ethyl butyrate and o.i c.c. toluene were added. After thirty minutes at 40° the active tubes had become bright red. After acting for two and a half hours 0.5 c.c. l^ KOH was re- quired to neutralise the butyric acid, therefore 1.25 per cent of the butyrate had been hydrolysed. Blanks using the boiled serum sim- ilarly treated remained perfectly neutral. When the lymph was not neutralised previous to adding ethyl butyrate it was found to be- come speedily acid from the hydrolysis. The activity of lymph was 1 Munk : Virchow's Archiv fiir pathologische Anatomic, 1884, xcv, p. 407. 336 A. S. Loevenhart. tested in still another way. To 11 c.c. of lymph serum, 0.3 c.c. lit- mus and 2.4 c.c. 2^ butyric acid were added to bring it to a neutral tint. 5 c.c. of this mixture was placed in each of two test-tubes and to No. I there was added 0.26 c.c. ethyl butyrate, while No. 2 was kept as a check. After acting seven hours at 40° and sixteen hours at room temperature, No. i required 1.7 c.c. !l^ KOH to bring it to the color of No 2, this corresponding to a hydrolysis of 4.27 per cent. It will be thus seen that both lymph and lymphatic glands possess marked lipolytic activity, which, however, is not great compared with the liver and pancreas. It must be borne in mind in this connection that the results reached with ethyl butyrate may be applied qualitatively but not quantitively to fats and other ethereal salts. The equilibrium in every case will depend on the nature and strength of the combining acid and alcohol. There is reason to believe, however, apart from physiologi- cal considerations that in the case of the higher fats equilibrium would be established when most of the fatty acid and glycerine are com- bined as fat. Euler,^ working on the equilibrium between ethereal salts, their constituent acid and alcohol, and water, finds that the hydrolysis of the ester is greater, the stronger its constituent acid. He also finds that the state of equilibrium varies greatly for different alcohols, and in an entirely irregular way for alcohols of the same series, hence the equilibrium in the case of each ethereal salt will have to be investigated separately. The theory of fat absorption ad- vanced above and supported by the experimental evidence appears to me to be in accordance with all the facts in the case and offers a clear and satisfactory explanation of the process. The Storing up and Utilisation of Fat. — Lipogenesis. In connection with the proof of the reversibility of the action of lipase it occurred to the author to ascertain to what extent the fat synthesis known to occur in the body could be induced by the lipo- lytic enzyme. In spite of the advance made in the study of syntheses occurring in the organism we have as yet but little or no conception as to the vital agencies which induce these syntheses. In speaking of the great progress in synthetic organic chemistry 1 EuLER : Zeitsclirift fiir physikalische Chemie, 1901, xxxvi, p. 405. Relation of Lipase to Fat Metabolism — Lipogenesis. 337 Bunge^ remarks: "Nevertheless the processes employed in no way represent the synthetic processes in the living cell, for all artificial syntheses can only be achieved by the application of forces and agents which can never play a part in vital processes, such as extreme pres- sure, high temperature, concentrated mineral acids, and free chlorine, agents which are immediately fatal to any living cell." It would certainly seem strange if there should be operating in the living organism special synthetic agencies, and yet that no investigator had found the slightest evidence of their existence. It was with a hope of throwing some light on these vital synthetic agents that a study of the reversible action of lipase on ethereal salts was undertaken by Kastle and the author.^ If lipase be the agent which occasions both fat synthesis and fat splitting, it will follow that the enzyme occurs wherever these processes are taking place in the body, and as fat is found in all tissues to a greater or less extent we may expect to find lipase in varying quantities in all the tissues. This has been borne out by experiment. In searching for the greatest occurrence of lipase, Kastle and I tested the lipolytic activity of several organs and tissues. In our work extracts of 10 grams of the tissue in 100 c.c. of water were used, i c.c. of this stained extract was diluted with 4 c.c. of water, and after heating five minutes at 40°, 0.26 c.c. ethyl butyrate and 0.1 c.c. toluene were added. After forty minutes they were titrated with 2^ KOH, using litmus as the indicator. The initial acidity of the extracts, though quite small, was always taken into account. The following results are quoted : c.c. 2'^ KOH Per cent of required. Hydrolysis. Pancreas 1.4 3.52 Kidney 0.7 1.76 Liver 4.1 10.29 Submaxillary gland . . 05 1-26 We proved also that intestinal and gastric mucosa possess lipolytic activity. In a similar manner I have studied the lipolytic activity of the following tissues: liver, pancreas, mammary gland (active and resting), blood and lymph, intestinal mucosa, lymphatic glands, brain, spleen, somatic muscle, and heart muscle. In most cases these 1 Buxge: Lehrbuch, 1894, p. 288; Schafer's Text-book of physiology, 1898, i, p. S93. 2 Kastle and Loevenhart : American chemical journal, 1900, xxiv, p. 491. 338 A. S. Loevenhart. tissues were taken from the pig and dog. Usually 10 per cent ex- tracts were employed and these were allowed to act on the ethyl buty- rate for fifteen minutes, but for tissues of weak activity 20 per cent extracts were employed and allowed to act for a longer time.^ Mammary gland. — The enormous fat synthesis which occurs in this gland when active suggested the occurrence of a large amount of lipase in it and the facts agree perfectly with the theory. A 10 per cent extract of the active mammary of a bitch was tested as described above. After thirty minutes at 40° it required 1.55 c.c. g'o KOH, corresponding to 3.79 per cent hydrolysis. On standing at room tem- perature for forty hours the tubes required 9.1 c.c. 3"^ KOH, corre- sponding to 23.3 per cent hydrolysis. This is even slightly greater than the activity which has been found for the pancreas of the dog. The resting gland was found to possess less than one-tenth the lipolytic power of the active gland. The above is the only active mammary gland that has been tested. In this connection it is interesting to note that Bartlet,^ Henriques and Hansen,^ and others have found that increasing the fat in the diet of cows increases the fat in milk ; moreover, small quantities of the foreign fat taken with the food were found in the milk. On con- tinuing the fatty diet for some time, however, the fat content of the milk fell to normal. It should be stated that in the following account of the lipolytic activity of the tissues the figures given are the averages of the sev- eral concordant determinations. In every case blanks in which the boiled extract was used were carried through and in no case did they show the slightest hydrolysis of the ester. In many cases blanks in which the unboiled extracts without the ethyl butyrate were used were carried through in order to detect any increase in the acidity of the extract itself. The results were always negligible. Liver. — In the dog, pig, and man this has been found to have the greatest lipolytic activity of any tissue tested. When 10 per cent ex- 1 In speaking of the occurrence of lipase in the tissues, Hanriot : Comptes rendus de la Soci^t^ de biologic, 1896, p. 925, states that the liver, blood, and pancreas possess large quantities of it, while the thyroid gland, spleen, testicle, adrenal, urine and lymph contain only small amounts. He gives no data from which the relative activities of these tissues can be estimated. 2 Bartlet : 14th Annual Report. Maine Agricultural Experiment Station, 1898, p. 114. 3 Henriques and Hansen: Extract in the Journal of the Chemical Society, 1900, p. 668. Relation of Lipase to Fat Metabolism — Lipogeuesis. 339 tracts of these livers were used, the results, on testing in the usual way were as follows : Duration of experiment 15 minutes; temperature, 40°. c.c. 5'^ KOH Pel - cent of required. Hydrolysis. Man 1 . 0.9 2.26 Dog . 1.5 3.76 Pig . . 3.4 8.54 Pancreas. — Dog . 1.0 2.51 Pig . . 1.4 3.51 Kidney. — Conditions same. Dog . 0.82 2.06 Pig . . 0.70 1-76 Lung. — Conditions same. Dog . 0.6 1.51 Brain. — Conditions same. Dog . 0.15 0.38 Adrenal. — Conditions same. Duration of experiment 60 min. Dog . . 0.12 0.30 Spleen. — This organ possesses slight activity. A 20 per cent ex- tract when tested as usual showed no change in one hour. After forty hours at room temperature it required 1.4 c.c. ^ KOH for neutralisation, hence, 3.51 per cent hydrolysis. Heart muscle. — A 20 per cent extract of the human heart from the same subject as the liver tested was used. The tube became bright red in one hour and in two hours it required 0.3 c.c. ^ KOH for neutralisation, hence 0.75 per cent of the ester was hydrolysed. Dog and pig hearts show the same activity as the human heart. In these tests the myocardium was carefully cleared of any fatty tissue. Somatic muscle. — This has about the same lipolytic activity as heart muscle. A 20 per cent extract acting for fifteen minutes at 40° and then for sixteen hours at ordinary temperature required 0.5 c.c. 2V KOH 1.25 per cent hydrolysis. Blood. — Hanriot,^ in trying to determine the mechanism by which ^ The individual died of intestinal obstruction. The autopsy was held eight hours after death and all of the organs were found normal. The tissue was tested immediately. 2 Hanriot: Comptes rendus de la Societe de biologie, 1896, p. 753. Archives de physiologie, 1898, p. 797. 340 A. S. Loevenhart. reserv^e fat is utilised by the organism, in 1896 discovered that the blood possesses great lipolytic power. He made an interesting com- parison of the lipolytic activity of the blood of a series of animals, using monobutyrin as the reagent in testing the activity. In my work serum was obtained by centrifugalising cat's blood. It was tested as follows: i c.c. serum, 4 c.c. water, o.i c.c. toluene and 0.26 c.c. ethyl butyrate were kept at 40° for thirty minutes when the mixture required 0.25 c.c. 2^0 KOH for neutralisation, corresponding to 0.63 per cent hydrolysis. On returning to the bath for one hour and a half it contained a coagulum and required 0.7 c.c. -^^^ KOH for neutralisation, the hydrolysis therefore being 1.76 per cent. From this we see that blood has great lipolytic activity. In this calculation the amount of acid used in overcoming the initial alka- linity of the blood was not taken into consideration. Hence, to form a more accurate idea of the activity of blood, the following experi- ment was tried: 5 c.c. serum was neutralised with 1.7 c.c. ^'o butyric acid, using litmus as the indicator. This was divided into two equal parts, and to each was added o.i c.c. toluene. They were heated five minutes at 40°, and to A was added 0.26 c.c. ethyl butyrate, none of the ether being added to B. After one hour, A required 0.9 c.c. ^ KOH to return it to the color of B, hence 2.26 per cent of the ether was hydrolysed. Bile. — Notwithstanding the great activity of the pig's liver the bile possesses but the merest trace of lipolytic activity, as is shown by the following experiments: Tube No. i, i c.c. fresh pig's bile, 4 c.c. water, 0.1 c.c. toluene, and 0.26 c.c. ethyl butyrate. After acting for five hours at 40° and nineteen hours at ordinary temperature, it re- quired O.I c.c. f^ KOH 025 per cent hydrolysis. Blanks with the boiled bile, toluene, litmus, and ethyl butyrate, and with fresh bile to which ethyl butyrate had not been added, remained unchanged, and only required one drop y| KOH to turn them from a dirty yellow to a bluish green. Bruno ^ stated that the bile contains a lipolytic enzyme, and assists the pancreatic juice in its fat-splitting function. The results obtained above are in accordance with the old view, that the bile aids in fat absorption only through its solvent action and not by any power to split fats. The lipolytic activity of intes- tinal mucosa, lymph, and lymphatic glands, was discussed in connec- tion with fat absorption. ^ Bruno : Archives des sciences biologiques, St. Petersburg, 1899, vii, p. 87 ; Chemisclies Centralblatt, 1900, [ii] p. 916. Relation of Lipase to Fat Metabolism — Lipogenesis. 341 Adipose tissue. — The theory of the part played by lipase in fat synthesis indicated the occurrence of lipase in adipose tissue, and the findings are entirely in accordance with the theory. The follow- ing experiment illustrates the degree of activity found. Approxi- mately a 40 per cent extract of the subcutaneous fat of the pig was prepared and tested in the usual way. After standing for thirty minutes at 40°, the tubes required 0.5 c.c. .^^ KOH for neutralisa- tion, corresponding to 1.25 per cent hydrolysis. After forty hours at room temperature, the tubes required 2.55 c.c. f^ KOH, correspond- ing to 6.40 per cent hydrolysis. Inasmuch as it is impossible to extract this tissue with any degree of thoroughness, its lipolytic ac- tivity, in all likelihood, exceeds that found in the above-quoted experiments. In this connection it was decided to test the power of the extract of subcutaneous fat to synthesise ethyl butyrate in a way similar to that used to test the synthetic action of pancreatic extract. When a 25 per cent extract of subcutaneous fat is mixed with butyric acid and ethyl alcohol in the proportions and under the condi- tions described in connection with the pancreas experiments and in the presence of thymol the odor of ethyl butyrate was very apparent after twenty-four hours, whereas the blank in which the boiled extract was used remained free from the odor of ethyl butyrate. This proves the synthetic action of the lipase of this locality. The lipolytic power of fatty tissue from other parts of the pig was also tested. Pericardial and perinephric fat were both found to be active, but markedly less so than subcutaneous fat. This is in accordance with the fact that, during inanition, the fat in these localities is the last to disappear. The lipase which is accountable for the formation of the fat here seems to have subsequently largely disappeared, and hence the diffi- culty of absorbing it during inanition. It is a well known fact that during inanition, or a state of malnu- trition, an animal is capable of absorbing its own fat as food. What is the mechanism by which this is brought about? In answer to this question we may suppose that blood and lymph, as the result of con- tinual oxidation, become poor in fatty acid and glycerine. When the lymph bathing the fat cell becomes poor in these products, how- ever, the lipase restores equilibrium by effecting an hydrolysis of the fat. If the passage of the fat from the lumen of the intestine into the lymph and blood and thence to the subcutaneous tissue, be accom- 342 A. S. Loevenhart. plished by a series of hydrolyses and hydrosyntheses, it may be asked why an animal does not ordinarily lay on the fat taken as food. This would seem to be very well accounted for by the old explanation, that since the fat laid on differs only quantitatively from the fat of the food, it may be due to the greater destruction of one constituent of the fat than the others, whereby the whole is brought to the normal fat of the particular animal. That the fat of the food may be laid on di- rectly has been conclusively proven by Lebedeff,^ Munk,^ and others. The same result has recently been reached by Rosenfeld ^ also, who found that by feeding one dog with cocoa butter and another with mut- ton fat, the fats deposited in each case corresponded with the fat of the food. Further, he was able to produce a deposit of mutton tallow in goldfish and carp by feeding it to them. He concludes, therefore, that the peculiarities in the fat of different animals is accounted for by the fat in the food. As food the fats and carbohydrates occupy co-ordinate positions, both serving for the production of body heat, and the ultimate fate of both being oxidation. Starting with Bernard's discovery of glycogen and the sugar-producing power of the liver, much has been said both for and against his theory of " glycogenesis ;" but it stands to-day as the best explanation of the facts regarding the storing and transpor- tation of carbohydrates in the body. The liver is the primary store for glycogen, secondary deposits being found in all the tissues of the body. In the case of fats we have somewhat analogous condi- tions, and hence we may speak of the storing and utilisation of fats as " lipogenesis." In the case of fats the areolar tissue is the great primary store, secondary deposits being found in all the tissues. In some animals even this difference in the storing of fats and carbohy- drates is not to be noted. In many fish, notably the cod, the liver, at certain seasons of the year, becomes the great depository for fat. The liver. we have found to possess powerful lipolytic activity, and hence, under proper conditions, it should be capable of storing fat. More- over, this is in accordance with the experiments of Noel Paton,^ who 1 Lebedeff: Centralblatt fiir die medicinische Wissenschaften, 1882, p. 129. Zeitschrift fiir physiologische Chemie, 1882, vi, p. 149. Archiv fiir die gesammte Physiologic, 1883, xxxi, p. 11. 2 MUNK : Archiv fiir Physiologie, 1S83, p. 273. ViRCHOW's Archiv fiir pathol- ogische Anatomic, 1884, xcv, p. 407. ^ Rosenfeld: Verhandlungen des XVII. Congresses fiir innere Medicin, 1899, p. 503. * Paton : Journal of physiology, 1896, xix, p. 167. Relation of Lipase to Fat Metabolism — Lipogenesis. 343 showed that the fat contained in the liver of frogs is increased after a fatty meal. It is believed that both phases of lipogenesis are induced by lipase, a fat-splitting and fat-forming enzyme. Against this view- it may be urged, however, that the amount of lipase found in the tis- sue is no index of the amount of fat they contain. Thus the liver contains more lipase than any other tissue, and yet it contains nor- mally comparatively little fat. It seems that besides the presence of lipase, there must exist in the tissues certain conditions which favor the storing up of fat. For instance, in the liver the great lymph for- mation and the enormous and varied activities of this organ may not be favorable for fat accumulation. Indeed, it seems that the conditions for this are more favorable in the more sluggish connective tissue. Under pathological conditions, however, where the function of the liver is interfered with, we note the ease and rapidity with which fat may be stored up. Further, it will be seen below, that the hydrolysis by hepatic lipase is much more complete than with pancreatic lipase. Whether, on the other hand, its synthetic power is equally great or correspondingly less, has not yet been determined. The observa- tion of Hanriot that foetal blood does not contain lipase up to the sixth month, is not opposed to the theory of the part played by lipase in lipogenesis. The absence of lipase from the blood should offer no obstacle to the laying on of fat if the tissues contain lipase. Fate of Lipase in the Body. In this connection it has been found that the urine possesses but a trace of hpolytic activity, and that at times none at all can be noted. Extracts of faeces, on the other hand, show some activity. A 20 per cent extract of fresh faeces was prepared and tested in the presence of toluene to prevent the action of bacteria. After two and a half hours, it required 0.2 c.c. -1^^ KOH, corresponding to 0.5 per cent hydrolysis. The blank containing the boiled extract remained neu- tral. Whether this activity is derived from the pancreatic juice or from lipase contained in the bacteria I am unable to say. We have found that pancreas kept until putrefaction was well ad- vanced, still showed lipolytic activity in the presence of toluene- though it was much diminished. 344 ^- '^- Loevenhart. Distribution of Lipase in the Tissues in Phosphorus Poisoning. In the light of the recent work on phosphorus poisoning^ it was decided to poison dogs slowly by the subcutaneous injection of sub- lethal doses of a one per cent solution of phosphorus in olive oil- Three dogs were experimented with, and during the poisoning they were fed largely on fatty food. They were killed after two weeks. The organs showed fatty degeneration, and on testing the lipolytic activity of the heart, liver, and kidney it was found that in no case did it vary far enough from the normal to indicate that the fatty changes were due to changes in the distribution or amount of lipase in the tissue. On the Non-Occurrence of Soaps in the Body. After having synthesised ethyl butyrate from butyric acid and ethyl alcohol, the question presented itself as to whether it would be pos- sible to synthesise the ether from sodium butyrate and ethyl alcohol by means of lipase. This was soon found to be impossible. This fact indicates that perhaps soaps do not exist in the blood and lymph, and that instead, the free fatty acid occurs there in simple solution.'"^ On looking into the matter, it was found that there had already been accumulated a number of facts which would lead us to believe that no soaps exist in the blood. The toxicity of the soaps and the non-toxic nature of the constituent free fatty acid, offer us strong evidence against the existence of soaps in the blood. Rassmann^ first showed the poisonous action of sodium oleate on injection into the blood current. Ten years later, Munk,* unac- quainted with Rassmann's work, studied very thoroughly the pois- onous action of soaps on rabbits by intravenous injection. He found that 0.13 gm. of sodium oleate per kilo caused death in spite of artificial respiration. On injection he found that the pupils became widely dilated, a fall of about f in blood pressure occurred, the gas ^ Taylor : Journal experimental medicine, 1899, iv, p. 399. Athanasiu : Archiv fiir die gesammte Physiologic, 1899, Ixxiv, p. 511. ^ The solubility of free fatty acid in the blood and lymph has not as yet been determined. ^ Rassmann : Ueber Fettharn, Inaugural Dissertation, Dorpat, 1880. ■* MuNK : Archiv fiir Physiologic, Supplement, 1890, p. 117. Relation of Lipase to Fat Metabolism — Lipogenesis. 345 exchange decreased \ to \, and the heart stopped in wide diastole. The blood also lost its coagulability. On injecting soaps into animals which had been given morphine, they fell into a coma resembling that caused by albumoses and pep- tones. Injection of smaller quantities than 0.13 gm. per kilo, caused more or less grave symptoms in proportion to the dose. The sodium soaps of palmitic and stearic acids are even more toxic than sodium oleate. Munk found that injection of free oleic acid has no effect. This alone, it seems to me, strongly indicates that the blood can only form soaps from the free fatty acids with extreme slowness if at all, or otherwise he would have gotten some symptoms of intoxication upon injecting the free oleic acid. Rachford,^ in working on the prop- erties of the pancreatic juice, performed an experiment which well illustrates the inability of the body juices to neutralise the higher fatty acids. He found that if pancreatic juice was v^rell shaken with neutral olive oil, and then allowed to stand twenty-four hours, the oil separated at the top and was strongly acid, while the juice at the bottom was strongly alkaline, and still in twenty-four hours no soap had been formed and no emulsion produced. It seems that there might be some union between the alkali of the blood and the proteids which prevent this neutralisation. Pfluger- has recently shown, however, the slowness and incompleteness with which one per cent sodium carbonate saponifies the higher fatty acids at 37°. Hoppe- Seyler^ prepared soaps from the blood and lymph, and from this he maintained that they normally exist there. In his method of extract- ing the soaps, however, heat, and large amounts of alcohol were em- ployed, and this method would surely break up such combinations between the alkali and the proteid, if they existed, and would be most favorable for the formation of soap from the free fatty acid. No conclusions can be drawn from the work of Hoppe-Seyler. Friedenthal^ believes that the poisonous action of the soaps is due to their power to precipitate calcium. Whatever may be the manner of their action, our inability to synthesise the ethereal salts from their soaps by means of lipase, is in harmony with the facts above mentioned, all of which point to their non-existence. 1 Rachford : Journal of physiology, 1891, xii, p. 72. ^ Pfluger: Archiv fiir die gesammte Physiologic, 1901, Ixxxvi, p. i. ^ Hoppe-Seyler: Zeitschrift fiir physiologische Chemie, 1883-4, viii, p. 503. * Friedenthal: Archiv fiir Physiologic, 1901, p. 145. 346 A. S. Loevenhm^t. On the Limit of the Action of Lipase on Ethyl Butyrate. With reference to the mechanics of the action of lipase it is important to determine the limit of its action in order to see if lipase of different origins reaches the same limit, and also to determine to what extent this limit is affected by the amount of ethyl butyrate and enzyme present. In order to see whether lipase from the liver and pancreas reaches the same limit, extracts of these organs were prepared and diluted until they were about equal in their power to hydrolyse ethyl buty- rate ; the pancreatic extract was about 20 per cent, while the hepatic was approximately a 7 per cent extract. Tubes containing i c.c. of the extract, 4 c.c. of water, o. i c.c. toluene, and o.i c.c. litmus were heated at 40° for five minutes and then 0.26 c.c. ethyl butyrate was added. After acting for fifteen minutes they were titrated. c.c. ^"0 KOH Per cent of required. Hydrolysis. Pancreatic . . 2.2 5.52 Hepatic . . . 1.95 4.S9 It was intentional here to have the pancreas extract slightly stronger than that of the liver. Mixtures were now made up as follows : Pancreas. Liver. c.c. Water 37.2 Litmus 1.0 Pancreas extract . . 10.0 ^KOHi .... l.S the total volume being in each case 50 c.c. To each was added 0.65 c.c. ethyl butyrate and 0.5 c.c. toluene. These were placed in bottles and kept at 40°. They were frequently vigorously shaken. At certain intervals 5 c.c. of the mixture were drawn off and titrated, with the results shown in Table, page 347. From the series in the accompanying Table we see that solutions of pancreatic and hepatic lipase of the same strength reach very different limits. Although the pancreatic and hepatic extracts used in these series were of equal strength when acting for fifteen minutes, yet when each was allowed to act until the limit was reached the liver extract had decomposed one and three-fourths times as much ethyl butyrate as the pancreas extract. This would naturally lead us to 1 To neutralise the initial acidity. c.c. Water . . . . 38.5 Litmus . . . . 1.0 Liver extract . . 10.0 ^^KOHi . . . 0.5 Relation of Lipase to Fat Metabolism — Lipogenesis. 347 doubt the identity of the lipase of the liver and pancreas. Hanriot ^ states that the lipase of the blood and the pancreas are not identical. Time in hours. Pancreas. Liver. fo KOH required. c.c Per cent of Hydrolysis. fo KOH required. c.c. Per cent of Hydrolysis. 27 46 72 144 4.05 4.45 4.50 4.90 40.5 44.5 45.0 49.0 7.37 7.62 7.85 8.5 73.7 76.2 78.5 85.0 He bases this statement on two facts : First, although extracts of the pancreas and serum have the same activity in the presence of sodium carbonate, the serum shows almost twice the activity of the pancreas extract when the carbonate is neutralised. Second, if serum and extracts of the pancreas be prepared having at 15° the same activity, at 42° the serum becomes twice as active as the pancreas extract. Kastle and I, in a comparison of pancreatic and hepatic lipase, reached a result quite similar to Hanriot's second experiment. Hence it would seem that the lipase of the blood and liver may be identical. As further pointing to the non-identity of liver and pancreatic lipase we found that strychnine sulphate and also phenol acting at a dilution of one part in 5000 lessen the activity of pan- creatic extract 30 per cent while they are without effect on hepatic extract. On the other hand osmic and salicylic acids are much more harmful in their effects on liver than on pancreas extract. Hence we see that pancreatic and hepatic lipase seem to differ in three ways : 1. The velocity of their action is affected differently by changes of temperature. 2. They are affected differently by certain substances. These points were previously brought out by Kastle and myself. 3. When solutions possessing the power of hydrolysing equal amounts of ethyl butyrate in fifteen minutes are allowed to act until the limit is reached in each case, the liver extract hydrolyses 1 Hanriot : Comptes rendus de la societe de biologic, 1897, p. 778. 348 A. S. Loevenhart. nearly twice as much of the ether as does the pancreas. It seems that the lipase of the blood, liver, and kidney presents the same characteristics, while that occurring in other localities resembles the pancreatic. This has not yet been definitely proven. These facts would lead one to believe that these enzymes are chemically different. Yet they both manifest their presence in the same way, viz., by the hydrolysis of ethereal salts. There are other cases where substances of totally different chemi- cal natures show the same catalytic activity. The catalysis of hydro- gen peroxide is effected alike by colloidal metal solutions (Bredig), by many metallic peroxides and by all plant and animal extracts (Schonbein, Loew). Among these substances there is no apparent chemical relation. Other similar examples are to be found among the proteolytic and diastatic enzymes. The effect of the amount of lipase on the limit of its action on ethyl butyrate is brought out in the following series in which a 10 per cent liver extract was used. Into tightly stoppered tubes the following mixtures were placed. A Extract, c.c. . . . 5.0 Water used, c.c. . . 0.0 To each of these tubes were added o. i c.c. toluene, 0.2 c.c. litmus, and 0.26 c.c. ethyl butyrate. They were placed in the thermostat at 38°. After seventy-two hours they were titrated : B C D E F G 4.0 3.0 2.0 1.0 0.5 0.1 1.0 2.0 3.0 4.0 4.5 4.9 A B C D E F G 35^ KOH required, c.c. . . 23.52 19.20 14.85 10.80 7.00 4.25 1.90 Hydrolysis, per cent . 59.04 48.19 37.27 27.11 17.57 10.67 4.77 C D E F G 14.75 10.90 6.90 4.80 2.10 37.02 27.35 17.32 12.05 5.29 After one hundred and seventeen hours tubes similar to the first and carried through simultaneously were titrated : B 3^ KOH required, c.c 18.70 Hydrolysis, per cent 46.93 Thus the limit had been reached after seventy-two hours. From this series it follows that when a large amount of enzyme is acting the quantity of ether hydrolysed when the limit is reached is pro- portional to the amount of enzyme present. For small quantities of the enzyme, however, the hydrolysis is somewhat greater pro- portionally than when a larger amount is used. Relation of Lipase to Fat Metabolism — Lipogenesis. 349 In interpreting the above results, it must be remembered that Hpase is readily destroyed by acid and that the butyric acid produced by its action in the above experiment undoubtedly had a very detri- mental action on it. Kastle and I found that the velocity of the hydrolysis is proportional to the amount of enzyme acting and hence in the limit experiments in which large amounts of enzyme were used there was acting a greater amount of acid for its destruc- tion. On the other hand, in these cases the mixture contained much more proteid and this may possibly have exercised a protective influ- ence over the enzyme. In order to determine the relation of the quantity of ethyl buty- rate to the limit of the action, mixtures were prepared containing liver, or pancreas extracts, and varying quantities of ethyl butyrate, as follows : Pancreas. — No. 1. 18.75 c.c. water. 1.25 c.c. ^KOHi. 0.30 c.c. toluene. 5.00 c.c. 20 per cent pancreatic extract. 1.30 c.c. ethyl butyrate. No. 2. Mixture was made up in the same way, except that it contained 0.65 c.c. ethyl butyrate. The mixtures were kept in stoppered bottles at 38°. After two hundred and fifty-nine hours they had reached the limit, and titration showed the following: Quantity titrated. c.c. i'^ KOH Ester hydrolysed, c.c. required. mgr. No.l 5.0 10.5 68 No. 2 5.0 9.32 61 Liver. — No. 1 19.25 c.c. water. 0.75 c.c. 2% KOn 1. 0.30 c.c. toluene. 5. 00 c.c. 10 per cent liver extract. 1.30 c.c. ethyl butyrate. No. 2. Mixture was made up in the same way, except with 0.65 c.c. ethyl butyrate. Temperature, 38°. After two hundred and fifty- nine hours equilibrium had been reached. Quantity titrated. c.c. 2"^ KOH Quantity ether hydrolysed. c.c. required. mgr. No. 1 5.0 13.6 87 No. 2 5.0 12.3 80 1 To neutralise initial acidity. 350 A. S. Loevenhart. From these series it follows that the amount of ethyl butyrate hydrolysed by lipase when the reaction is allowed to proceed to the limit, is largely independent of the excess of ethyl butyrate present. Summary of Results. The conclusions reached in this paper may be briefly summed up as follows : 1. The reversible action of lipase furnishes us with a clear expla- nation of fat absorption. 2. Lipase has been found to occur in all the tissues of the body that have been tested, most notably in the liver, active mammary gland, blood, lymph, and intestinal mucosa. As pointed out in the above, special interest attaches to the fact that lipase has been found in considerable quantities wherever fat synthesis is known to take place as in active mammary gland and subcutaneous fat. 3. The close analogy between the storing up of fat and carbohy- drates in the body is pointed out, and as we conveniently call the storing and translocation of carbohydrate " glycogenesis/' simi- larly we may call the corresponding process in the case of fats " lipogenesis." It seems, further, that both phases of lipogenesis may be brought about by the enzyme lipase, which is either fat- forming or fat-splitting, according to conditions. 4. The inability of lipase to synthesise ethyl butyrate from sodium butyrate and alcohol, together with much collateral evidence, has led to the belief that the free fatty acids rather than the soaps exist in the blood and lymph. 5. The fatty changes occurring in phosphorus poisoning are not due to changes in the distribution or amount of lipase in the tissues, as no disturbances of this character were noted. 6. Lastly a study of the limit of the action of lipase on ethyl butyrate has revealed the following: 1. The limit is nearly proportional to the amount of enzyme acting. 2. It is nearly independent of an excess of ethyl butyrate. In conclusion I desire to express my thanks to Professors Abel and Howell, in whose laboratories this work was done. THE PHYSIOLOGICAL ZERO AND THE INDEX OF DEVELOPMENT FOR THE EGG OF THE DOMESTIC FOWL, GALLUS DOMESTICUS, A CONTRIBUTION TO THE SUBJECT OF THE INFLUENCE OF TEM- PERATURE ON GROWTH. By CHARLES LINCOLN EDWARDS. \_From the Biological Laboratory of the Department of A\itural History, Trinity College.^ CONTENTS. Page I. Introductory 351 II. Apparatus and methods 352 III. Incubations of Series A 358 IV. Incubations of Series B 366 V. Incubation of Series C 373 VI. Incubations of Series D . . . . 374 VII. Blastoderms of Series E 382 VIII. The physiological zero 387 IX. The inde.x of development 388 X. Normal size and variation of the blastoderm and of the area pellucida . . . 391 XI. Growth of the blastoderm independently of the appearance of the primitive streak 392 XII. Normal size and variation in the volume of the egg and the relation of diameter of blastoderm to volume of egg 395 XIII. Summary 396 I. Introductory. QINCE the time of the Egyptians it has been recorded that warmth ^ is thechief factor in producing development in the eggs of birds. Those workers who have incubated eggs of the domestic fowl, from Von Baer and Harvey down to the present have established the op- timum temperature at 38° C, with a range from 35° C. to 39° C. within which the development is usually normal. It has been shown also that cold below 35° C. when not too intense or too prolonged to destroy the embryo retards, or even suspends, the processes of growth during its application. Prevost and Dumas, Dareste, and others give 28° to 30 ° C. as the physiological zero for the Q.^g of the domestic fowl. Rauber places this minimum at 25° C. Below this temperature it was presumed 351 352 Charles Lincoln Edwards. that no development takes place. The disagreement as to the exact degree of the physiological zero as shown by the above opinions, founded as they must have been largely upon guesswork, together with the lack of precise data, led me to undertake the following in- vestigation of the subject. Besides establishing the physiological zero at 20.oo°-2i.oo° C, I have determined more exactly the normal size and variation of the blastoderm and of the area pellucida, the index of development from the physiological zero to 30.75°, the normal volume of the ^%g, its variation and relation to the diameter of the blastoderm, the growth of the blastoderm independently of the appearance of the primitive streak, and the dependence of ontogenetic differentiation upon rise in temperature.^ The description of the variations (monsters) produced during these experiments is reserved for later publication. II. Apparatus and Methods. In Series A, a Cyphers incubator, provided with its own meta- thermostat, a normal thermometer, calibrated and divided to one- fifths of a degree, and an oil lamp were used. In Series B, the same outfit was employed with the substitution of gas, flowing from a gas-pressure regulator, and a calibrated thermo- meter divided to one-tenths of a degree. For this series it became possible to place the apparatus in a basement room in the Hall of Natural History, Trinity College. Having four brick walls, the range of temperature in this room for the period of incubation was gener- ally within one degree. In Series C, the basement room itself became the incubator. The room was cooled by a ton of ice buried in sawdust in order to reduce the summer temperature to that required. The clutch of eggs was kept at the desired temperature by regulating the distance of the eggs from the ice. In this and Series D, thermometers calibrated and certificated by the Groszherzogliche Sachsische Priifungsanstalt fiir Glasinstrumente, were used. In Series D, through the courtesy of Mr. W. C. Wade, one of the large rooms of his cold storage building was used for incubations 16, 17, 18, and 19. Here the air is kept constantly at about 7.5° C. by means of circu- 1 The results embodied in Series A were read June 28, 1900, before Section F, American Association for the Advancement of Science, at the New York meeting. Edwards, C. L. : Science, 1900, xii, pp. 310-3x1. Physiological Zero for the Egg of the Domestic Fowl. 353 lating brine cooled by ammonia. Incubations 20, 21, 22, and 23 were carried on in the room used for Series B and C. A Bausch and Lomb copper incubator, 45 cm. x 35 cm. x 70 cm., provided with Roux's bimetallic thermostat and a gas pressure regu- lator, was used. Besides the main gas flow in which the thermostat was placed, a separate tube was connected with the pressure tank for a minimum safety flame. For both tubes micro-burners were employed. All temperature readings are from Centigrade scales. The read- ings were recorded a number of times each day during the period of incubation, and their average taken for the index of development. The extreme range of temperature fluctuation was generally within one degree and sometimes within about half of one degree. For Series A^ eggs were procured from a countryman. For all other incubations and the data concerning fresh eggs I had my own poultry yard. In order to eliminate any question of special races a mixed flock of brown leghorns, barred plymouth rocks and white wyandottes was used. The eggs were gathered every half hour and during summer weather were placed in a refrigerator at 16° C. In order to test the possibility of injuring the embryos by subject- ing the eggs to a temperature of 16° in a refrigerator the following data were secured. Date when freshly ^^*f ^'^f ^§2 ' Length of , ., 1 J was transferred ^. °- laid egg was placed ' ^ , ■ time in re- . ,, °°f . ^ ^ ^1 from retrigera- r ■ in the refrigerator at ^ ^ • ^, tngerator ^fP lOm tortomcuba- J? ,.o Date of taking out embryo. July 13, 3.30 p.m. i July 14, 7 p. m. 27.5 hrs. j July 18, 10 a. M July 13, 2.30 P. M. 1 July 14, 7 p. m. j July 15, 9 A.M. I July 17, 9 a.m. JulyLS, 9 a. M. I July 17, 9 A.M. July IS, 10 a. M 28.5. hrs. 48.0 hrs. 48.0 hrs. I July 22, 9 a. m. July 19, 9.40 p. M. Age of embryo. Condition of embryo. 3 days, 15 hrs. 3 days, 15 hrs. 2 days, 12f hrs. 5 days. Normal Normal Normal Normal It was shown by Colasanti,^ in eggs kept at a temperature of —7° to —10° C. for from one to two hours that the germ was absolutely 1 CoLASANTi, G. : Reichert's und Dubois-Raymond's Archiv fiir Anatomic, Physiologic, und wisscnschaftliche Mcdicin, 1875. 354 Charles Lincoln Edwards. uninjured, Pictet ^ demonstrated that after a longer period at from —2° to —3° the embryo is killed, but from —1° the ^gg would survive, and Rabaud ^ kept eggs at —18° for half an hour without fatal results. In Series A, B, and C, the eggs were given an average incubation of six and one-half days in order that the effect of each temperature should be complete. It is apparent, however, that the exact length of the period is not important, provided it extends over a number of days. In Series D, each incubation lasted five days. The blasto- derms and embryos were fixed in 10 per cent nitric acid, placed in Kleinenberg's picro-sulphuric acid for about twenty-four hours, grad- ually dehydrated, and then stained in Mayer's cochineal. Such un- fertilized eggs as occasionally appeared were thrown out of the series. In general it was found best to take the measurements from the blastoderm as it rested upon the yolk after fixation, as sometimes there is a slight change in the dimensions, due to the reagents, or some distortion caused by separating the early blastoderms from the vitelline membrane, when the blastoderm is removed. In order to get some standard for the age of embryos included within the first forty-eight hours, the following table was made based upon the data given by Duval.'^ This author has drawn the 5, 10, 15, and 16 hour primitive streaks. Considering the magnification as given I find from Duval's figures that the actual length of the primi- tive streak at 5 hours is 0.625 mm.; at 10 hours, 1.25 mm.; at 15 hours, 1.6 mm.; at 16 hours 1.93 mm. The total growth in length between the 5 and 10 hour stages is 0.625 mm., or 0.125 ^''"^- P^^ hour. It is possible that during the first hour or two there is no detectable trace of the primitive streak but in the absence of data and in order to include those cases in the following incubations where the primitive streak measures 0.5 mm., 0.4 mm., 0.3 mm., and 0.2 mm., I have ex- tended the table over the first four hours by employing the above hourly increment of 0.125 mm. The hourly increment of growth be- tween the 10 and 15 hour stages is 0.07 mm., and between the 15 and 16 hours, 0.33 mm. Fere ^ did not find his chicks developing to the Duval ages. After 48 hours Fere embryos equal 28-33 hours of Duval. After 72 hours Fere embryos equal 46-52 hours of Duval. 1 PiCTET, Raoul : Extrait des Archives des sciences physiques et naturelles. October 1893, Geneve, 1895. 2 Rabaud : Comptes Rendus, t. 128, 1899, pp. 1183-5. ^ Duval, Mathias : Alias d'embryologie, Paris, 1889. * Fere, Ch. : Journal de Tanatomie et de la physiologie, 1894, xxx, No. 4. Physiological Zero for the Egg of the Domestic Fowl. 355 DUVAL. 3 Prim- Noto- Neural Neural r*^ « Secon- dary optic vesi- tJ3 J5 itive streak. Lengtii in mm. |S| ciiord. folds. groove. Auditory E .s § § Length Length Length I. a, pit. 1) ^ OJ !fl in mm. in mm. in mm. Ph > cles. < '^ 1 0.125 2 0.250 3 0.375 4 0.500 36 5 6 7 8 9 0.625 0.750 0.875 1.000 1.125 47 10 11 12 13 14 1.250 1.320 1.390 1.460 1.530 64 15 1.600 65 16 1.930 67 19 1.930 0.463 68 20 1.580 Low 0.790 70 21 1.930 1 1.501 Present. 71 22 1.666 1 Approach ant. " 72 23 1.600 3 Lie together ant. Post, f open 76 24 1.540 5-6 Lie together ant. Post. 1.540 open 77 25 1.150 6 Lie together ant. Post, f open 81 26 1.070 8 Lie together ant. Post. \ open 86 27 0.710 8-9 Partly united ant. Only last open Begin- ning 89 29 0.666 10-11 3 cerebral ves. Mostly closed Present 92 32 Present Rudiment 93 33 0.500 15 Mostly closed Present Deep, wide open 98 38 trace 17 Mostly closed Present Deep, wide open 100 41 " 18 102 43 " 19 Rudi- ment. Begin to close 107 46 " 27 Present Begin to close 109 48 27 Nearly closed Nearly closed 356 Charles Lincoln Edwards. The development of the primitive streak is shown in the following curve plotted from the above table. ~b fO 15 20 25 30 ^5 40^ DEVELOPMENT OF PRIMITIVE STREAK When considered in direct relation to the length of the incubation period there is a considerable individual variation in the growth of the embryo. This well known fact can be shown in an interesting manner by some instances taken from the data given by Keibel and Abraham,^ but arranged by hours rather than in accord with the serial order of normal stages as given by these authors. Before the 20 hour stage, where the Normentafel begins, four fig- ures are drawn. The two for 9 hours of incubation have primitive streaks, 1.2 mm. and 1.6 mm. long, thus being of the 9.5 and 15 hour stages according to the scale I have constructed from Duval. The two for 12 hours have primitive streaks 1.5 mm. and 1.6 mm. long 1 Keibel, F. : Normentafeln zur Entwicklungsgeschichte der Wirbelthiere. Zweites Heft. Keibel, F., and Abraham, K. : Normentafel zur Entwicklungs- geschichte des Huhnes (Gallus Domesticus), Jena, 1900. Physiological Zero for Ihe Egg of the Domestic Fowl. 357 and so represent the Duval 13.5 and 15 hour stages. An embryo incubated for 20 hours has 4 mesodermic somites, making it equal to the 23-24 hour stage of Duval. Of the seven 24 hour chicks one has i-(2) mesodermic somites, a second 3, and a third, 7-8, while the length of the primitive streak in the four remaining is 1.20, 1.30, 1.35, and 1.40 mm. For this period then there is presented a range of stages from 9.5 to 26 hours of Duval. One 32 hour embryo has 6 mesodermic somites, while a second has 9. A 31 hour chick has 7 somites, while one incubated 40.5 hours has 5 (-6), and one 48 hours only 2 ! Having 18-20 somites are chicks of 42, 43, 43.5, and 48 hours' incubation. The eleven chicks of 48 hours' incubation have respectively, 2, 16, 17-18, 19-20, 20, 22-23, 23, 23, 23-24, 25-26, and 28 mesodermic somites. The average is 20.04 somites, instead of 27 found in the figure by Duval of the 48 hour stage. In spite of this variation, I believe that the scale of development should be measured by hours. It is desirable that ultimately this scale should be established upon averages found by a statistical study of a very large number of embryos. However, for the present neces- sity, we are not, in all probability, very far wrong in following the Atlas of Duval as a standard. In the following series I have given the age according to the above table founded upon Duval, and, wherever possible, the serial number of Keibel and Abraham. 358 Charles Lincoln Edwards. III. SERIES A.i — Incub.\tion 1. Twelve eggs, incubated at 30.75° from April 2, 2 P. M. to April 10, 11 A. M., 7 days, 19 hours. No. of egg- Age in hours. 1.1 108 1.2 108 1.3 % 1.4 % 1.5 % 1.6 96 1.7 % 1.8 72 1.9 72 1.10 72 1.11 72 1.12 48 Serial number. Keibel and Abraham. Age in hours. Keibel and Abraham. Hydropic vesicles.- General remarks. 73 73 58 58 58 58. 58 52 52 52 52 39 c 104 104 78 78 78 78 78 67 67 67 67 51 Present Two embryos on one blastoderm.^ Temperature range from 30.25° to 31.25°, • 54.83 per cent of normal development. 1.00°. Average age, 85.92 hours or 1 Including incubations of the spring of 1900 at the University of Cincinnati, O. 2 In a large proportion of the following cases hydropic vesicles occur. They are of various sizes and sometimes so numerous that they form a foamwork. Sec- tions show that these vesicles are formed by the expansion of spaces in the primitive lower layer or later in the mesoderm. Sometimes the vesicles result from the folding in opposite directions of ectoderm and lower layer cells. 3 For cases showing from two to seven distinct embryos on the same blasto- derm cf. Banchi, a. : Monitore Zoologico Italiano, 1895, vi ; also cases by G. F. WoLK, Von Baer, Allen Thomson, Reichert, Donitz, Ahlfeld, Rauber, Dareste, Gerlach, Moriggia, Bourchardt, E. Hoffmann, P. Mitrophanow, a. Ptizin, Klaussner, and Hancock, given by Banchi. Physiological Zero for the Egg of the Domestic Fowl. 359 SERIES A. — lNCUB.vno.\ 2. Six eggs, incubated at 29.25° from April 10, 5 p.m. to April 16, 11 a. m., 5 days, 18 hours. No. of egg- Length of primitive streak in mm. Number of meso- dermic somites. Age in hours. Serial number. Keibel and Abraham. Age in hours. Keibel and Abraham. General remarks. 2.1 9 27 18 32.0 Embryo abnormal. 2.2 8 26 16 24.0 Neural folds contorted. 2.3 6 25 14a 32.0 Neural folds contorted. 2.4 5 24 12 33.0 Cephalic end of neural groove trifid. Three ad- ditional primitive streak- like branches posteriorly. 2.5 3 22 8a 24.5 Neural folds enlarged; con- torted. Primitive streak bifid. 2.6 2.00 16 3 24.0 Temperature range from ^ >9.00° to 29.50°, — 0.50°. Av erage age, 23.33 hours or 14.91 per cent of normal developi nent. Coi isiderable variation is shown in this clutch, being especially n larked i \ 2.4 and 2 .5 with ch; iracters as given above. Mitrophanow,^ describes a number of anomalies of the primitive streak similar to that occurring in egg No. 2.4, which were pro- duced by lowering the temperature to 32°-34°. In one case of a posteriorly bifid primitive streak the structure assumes a crescentic shape, and Mitrophanow suggests that here is an in- stance of atavism to the phylogenetic blastopore, the trace of which, as Duval believes, is apart of normal development. Mitropha- now believes that irregularities in the form of the area pellucida are found with these lateral branchings of the primitive groove, one part directly following the other because of the interdependence during the early stages of all parts of the formative area of the blastoderm. As Dareste^ demonstrates, by varying the particular point of appli- cation of the heat, the growth of the vascular area may be changed from the ordinary circular form to an ellipse, the direction of whose long axis directly depends upon the source of the incubating heat. ^ Mitrophanow, P.: Archiv fiir Entwickelungsmechanik der Organismen, i805, i, pp. 370 to 376. '■^ Loc. cit., pp. 288 to 293. 360 Charles Lincoln Edwards. Kaestner,^ following in lines known since Panum,^ secured many interesting deviations from the normal development by suddenly lowering the temperature a number of degrees, and thus causing a suspension of development for a greater or less time. This cooJing must come within the first two days of incubation, must last for a considerable length of time, and during the interruption, the z^g must be placed horizontally. The duration of the interruption period is inversely proportional to the ontogenetic stage. If, for instance, the &gg is cooled to about 7° below 28° (assumed as the physiological zero), the development can be suspended for three weeks, if at the beginning of the first day of incubation ; if at the end of the first day, for six days ; at the sixth day for seventy-two hours ; at the ninth day for forty-eight hours, and finally, in the second half of in- cubation up to the time of hatching, for twenty-four hours. As Warynski has observed, the yolk when cooled rises and presses the blastoderm against the vitelline membrane, to which it sticks. If this happens during the first two days, while the embryo is unpro- tected by the amnion, the pressure causes an arrest of development and consequent malformations. The exact character of these, and the region of the embryo in which they occur, cannot be predicted, since it is a matter of chance as to the part of the blastoderm which will adhere to the vitelline membrane. As Kaestner pointed out, the rising of the blastoderm probably depends upon a change in specific gravity consequent upon a change in volume due to cooling. Harvey^ found that in eggs opened after three days' incubation, the heart of the embryo beat slower and slower until it ceased, when, after a period of suspended animation, the warmth of tepid water, or even of the finger, caused the pulsations to return. Dareste * opened the shell, leaving the vitelline membrane intact, examined the embryo with a lens, and thus was sure of the cessation of the heart- beat. Upon the application of warm water the contraction recom- menced, and normal chicks were hatched after cooling at 20°, 15°, 8°-io°, and even i°-2°. The development was delayed, so that the eggs cooled for two days hatched in 23 instead of 21 days. Oxygen also has a direct influence upon the growth of the various 1 Kaestner: Anatomischer Anzeiger, 1896, pp. 136-45. - Panum: Virchow's Archiv fiir pathologische Anatomic, 1878, Ixxii, pp. 69-91, 165-197, 289-324. * Harvey, William : Exercitationesde generatione animalium. London, 1651. * Loc. cit.y p. 131. Physiological Zero for the Egg of the Domestic Fowl. 361 parts of the blastoderm and of the embryo. And so other environ- mental conditions. It is not yet possible to give an exact analysis of the part taken by these conditions in stimulating the cells to pro- duce local malformations, which, of course, are the direct manifesta- tions of increased or suppressed growth. SERIES A. — Incubation 3. Twelve eggs, incubated at 28.25° from April 16, 11 a. m. to April 23, 11 a. m., 7 days. Length No. of meso- dermic Serial Age in No. of egg- of prim- itive streak Age in hours. number. Keibel and hours. Keibel and Hydropic vesicles. General remarks. in mm. somites. Abraham . Abraham. 3.1 9 27 18 32 One posterior lat- eral branch of prim- itive groove. 3.2 7 25 15 31 3.3 5 23 11 11 3.4 4 22 9 20 Neural folds only developed anteriorly and posteriorly. 3.5 4 22 9 20 3.6 3 22 7 1 24 ! Few 1 present 3.7 3 22 7 1 24 3.8 3 22 7 24 3.9 3 22 7 24 3.10 3.01 16 Not given •• Lateral branches of the primitive streak. 3.11 1.5 •• 13.6 Fig. 3 12 3.12 00.0 Temperature range from 28.00° to 28.50°, — 0.50°. Average age,2 19.72 hours or 12.60 per cent of norm al development. Variations i ire shown in 3.1 and 3.10. 1 The Duval table give 3 1.93 mm. as the longest pr imitive streak which before the development of notochord and mesodermic somites indicates the 16 hour stage. Embryo 3.10 shows continued growth, then, at the 16 hour stage. ^ This is taken by dividing by the full number of eggs even though in some cases no trace of the embryo is present. ;62 Charles Lincoln Edwards. Ten eggs, SERIES A. — Incubation 4. incubated at 27.00° from April 24, 11 a m. to April 30, 11 a. m., 6 days. 6 li boi; in 3 O _c General remarks. 4.1 7.0 2.0 Not Not Present Primitive groove in form of a 4.2 1.8 15.5 8.0 3.0 given given None central blastopore-like pit. 4.3 1.8 15.5 7.0 2.8 " " Present 4.4 1.5 13.6 8.0 2.5 X 3.0 3 12 4.5 1.5 13.6 8.0 (2.75) 3.0 3 12 '■ 4.6 1.5 13.6 8.0 1.9 3 12 " 4.7 1.0 8.0 7.5 2.0 Not Not " Primitive groove widened. 4.8 1.0 8.0 5.5 2.0 given given " 4.9 1.0 8.0 4.5 1.8 « " " 4.10 0.8 6.4 0.8 2.5 (( " " T emperature ■ange from 26.75° to 27.25°, — 0.50°. Average length of primitive streak,! 1.32 mm. Average age, 10.22 hours, or 6.53 per cent of normal develop- ment. Ave age diameter of blastoderm, 7.15 mm. Average diameter of area 1 pellucida, 2. This and the 37 mr J aver: n. ige diameter of blastoderm and are a pellucida, are taken from the number of eggs presenting these features clearly defined. Physiological Zero for the Egg of the Domestic Fowl. 363 SERIES A. — Incubation 5. Nine eggs, incubated at 26.00"^ from April 30, 4.45 p. m., to May 8, 11.45 A. M., 7 days, 19 hours. Length of Diameter Diameter No. of primitive Age in of blasto- of area Hydropic General remarks. egg- streak in mm. hours. derm in mm. pellucida in mm. vesicles. 5.1 1.3 11 4.5 1.8 Present 5.2 1.3 11 7.0 2.0 " Primitive groove slightly widened. 5.3 1.2 9 6.0 2.0 " Several lateral branches of the primitive groove. 5.4 1.2 9 7.0 2.0 *' 5.5 1.0 8 6.5 2.0 " 5.6 1.0 8 6.0 3.0 " 5.7 1.0 8 6.0 3.0 None 5.8 4.0 1.0 " No trace of primitive streak. Degenerated. 5.9 ' • 5.0 2.0 Present No trace of primitive streak. Temperature ran; ;e from 25.80° to 26.20°, — 0.4 0°. Average length of primitive streak, 1.14 mm . Aver age age, 7.11 hours, or 4.54 per cent of normal develop- ment. Averag ; diamet er of blastoderm, 5.78 mm. Average diameter of area pellucida, 1.98 i nm. ;64 Charles Lincoln Edwards. SERIES A. — Incubation 6. Nine eggs, incubated at 25.5° from May 15, 11 A. m. to May 22, 1 P. M., 7 days, 2 hours. Length of primi- tive Diameter Diameter No. of Age in of blas- of area Hydropic General remarks. egg- streak in mm. hours. toderm in mm. pellucida in mm. vesicles. 6.1 2.0 16.3 11 3.0 X 4.5 (3.75) Present 6.2 2.0 16.3 6.0 3.0 6.3 1.5 13.6 8.0 3.0 X 2.5 (2.75) Primitive groove widely open 7 mm. long. 6.4 9.0 2.0 X 3.0 (2.5) Mere trace of primitive streak. 6.5 _ 9.0 2.0 X 4.0 (3.0) Central embryonic shield. ^ 6.6 4.5 2.0 Same as 6.5. 6.7 4.0 2.0 X 1.5 (1.75) " " " , degenerated. 6.8 6.0 2.0 Embryonic shield. Sev- eral primitive streak-like structures of indefinite form. 6.9 7.5 3.0 Same as 6.8. Temperature range from 24.80° to 26.20°, — 1.40°. Average length of primitive streak, 1.83 mm. Average age, 5.13 hours, or 3.28 per cent of normal develop- ment. Average diameter of blastoderm, 7.2 mm. Average diameter of area pellucida, 2.31 mm. In this clutch were found one 3-day chick and one with 22 mesodermic somites (40 hours), which probably show misplaced confidence in the farmer who gath- ered the eggs, and so they have not been tabulated. 1 In this and the follow ing groups of blastoderms wher e the primitive streak has not yet appeared, the first stage of ontogeny known as the " embryonic shield " is often present. This stage is due to the rather indefinite multiplication of cells of the primitive lower layer. The cell mass is quite as frequently central and circular as posterior and shield-shaped but I shall speak of it by the con- ventional name. Physiological Zero for the Egg of the Dotnestic Fowl. 365 SERIES A. — Incubation 7. Nine eggs, incubated at 25.0° from May 23, 3.30 p. m., to May 29, 5 P. M., 6 days, 1.5 hours. Length of primi- tive Diameter Diameter No. of Age in of blas- of area Hydropic General remarks. egg- streak in mm. hours. toderm in mm. pellucida in mm. vesicles. 7.1 1.6 15 4.0 2.0 None Primitive streak and groove bifid almost the whole length. 7.2 1.5 13.6 6.0 2.0 X 2.3 (2.15) Present Primitive streak sickle- shaped. 7.3 1.3 11.0 7.0 3.0 X 2.5 (2.75) None 7.4 0.5 4.0 5.0 2.0 X 2.5 Present (2.25) 1 1 7.5 6.0 2.0 Central embryonic shield. 7.6 5.5 2.5 " Same as 7.5. 7.7 10.0 3.0 X 2.0 (2.5) " " " 7.8 •• 7.0 3.0 X 2.5 (2.75) " 7.9 6.0 2.0 Temperature r ange fror 11 24.75° to 25.25°,— 0.50°. Average length of primitive streak, 1.23 mm. Av erage age, 4.84 hours, or 3.09 per cent of normal develop- ment. Ave -age dian leter of blastoderm, 6 28 mm. Average diameter of area pellucida, 2., 51 mm. ;66 Charles Lincoln Edwards. IV. SERIES B.i — Incubation 8. Eleven eggs, incubated at 23.13" from April 18, 330 P. m., to April 25, 4 P. M., 7 days, 5 hours. No. of Length of primi- tive Age in hours. Diameter of blasto- derm in mm. Diameter of area pellucida in mm. Form of area Hy- dropic vesicles. General remarks. egg- streak pellu- cida. in mm. 8.1 2.4 17.5 10.0 4.0 X 3.2 (3 6) Oval Present 8.2 1.8 15.5 124 X 14.0 (13.2) 3.5 X 2.5 (3.0) 8.3 1.7 15.3 80 2.35 X 2.0 (2.18) Wings of meso- derm prominent posteriorly. 8.4 1.5 13.6 10.0 X 8.6 (9.3) 3.2 X 2.7 (3.05) " 8.5 1.25 10.0 7.0 2.6 Round 8.6 1.2 9.6 8.0 2.7 X 2.5 (26) Oval Primitive groove indistinct. 8.7 1.15 9.2 7.5 X 6.0 (6.75) 2.55 X 2.3 (2.43) Wings of meso- derm prominent anteriorly. 8.8 11 8.8 8.0 X 8.75 (8.38) 3.0 X 2.9 (2.95) " 8.9 1.05 8.4 10.0 2.85 Round 8.10 0.95 7.6 9.0 X 10.0 (9.5) 3.1 X 2.6 (2.85) Oval Very few present 8.11 0.9 7.2 10.0 3.2 X 2.7 (2.95) Present Temperature range fr om 22.39° to 23.79°,— L.40°. Average length of primitive streak, 1.36 mm. A verage age 11.15 hours, or 7.12 per cent of normal develop- ment. Average dia meter of blastoderm, 9. 10 mm. Average diameter of area pellucida, 2.82 mm. "•■ This and the follow ing Series of incubatio IS were c arried on during 1901, at Trinity College. Physiological Zero for the Egg of the Domestic Fowl. 367 SERIES B. — Incubation 9. Ten eggs, incubated at 22.44° from April 27, 4.30 P. M. to May 4, 4 P. M., 6 days, 23 hours. Length of primi- Diameter Diameter Form Hy- dropic vesicles. No. of Age in of blas- of area of area General egg- tive streak hours. toderm pellucida pellu- cida. Remarks. in mm. in mm. in mm. 9.1 1.7 15.3 5.2 X 5.6 2.7 X 2.9 Nearly None Merest trace of (5.4) (2.8) round primitive groove. 9.2 1.5 13.6 7.8 2.8 X 2.9 (2.85) Oval Present 9.3 1.5 13.6 6.6 2.8 X 3.6 (3.2) Oblong None 9.4 1.4 12.0 7.6 2.65 X 2.85 (2.75) Slightly oval Present 9.5 1.2 9.6 7.6 2.6 X 3.1 (2.85) Oval None 9.6 1.2 9.6 6.8 2.5 X 2.8 (2.65) Nearly round Present 9.7 1.1 S.S 8.2 2.8 X 3.0 (2.3) Round 9.8 1.1 8.8 7.8 3 Round '* 9.9 1.1 8.8 7.0 X 7.8 (7.4) 2.3 X 2.85 (2.58) Oval " 9.10 0.4 3.2 6.4 2.2 Nearly round Only a trace of primitive streak. Temperature ra nge froii 1 22.17° to 22.67°,— 0.5 0°. Average length of primitive streak, 1.22 ir im. Ave rage age, 1( ').iZ hours, or 6.60 per cent of normal develop- ment. Avers ige diam :ter of bias toderm, 7.16 mm. Average diameter of area pellucida, 2.7 8 mm. 368 Charles Lincoln Edwards. SERIES B. — Incubation 10. Eleven eggs, incubated at 21.87° from May 6, 7 p. m. to May 13, 11 A. M., 6 days, 16 hours. No.of egg- Length of prim- itive streak in mm. Age in hours. Diameter of blas- toderm in mm. Diameter of area pellucida in mm. Form of area pel- lucida. Hydropic vesicles. General remarks. 10.1 1.2 9.6 5.4 3.2 X 3.5 (3.35) Slightly oblong Present Primitive groove only very slightly devel- oped. 10.2 1.1 8.8 5.8 X 6.0 (5.9) 2.5 Nearly round No primitive groove. 10.3 0.9 7.2 5.6 2.6 X 2.8 (2.7) Very few present Primitive streak 0.7 mm. broad. No primitive groove. Transverse diameter 12 mm. the longer. 10.4 0.8 6.4 4.8 2.3 X 3.0 (2.65) Oval Present No primitive groove. Primitive streak only faintly marked off from rest of blastoderm. 10.5 0.7 5.6 5.2 2.5 X 3.0 (2.75) Oblong ** Same as 10.4. 10.6 0.3 2.4 4.2 2.65X2 85 (2.75) Round " Same as 10.4. 10.7 6.0 2.8 Indefinite central embry- onic shield. 10.8 5.8 2.7 X 3.5 (3.1) Oblong 10.9 •• 4.6 2.4 Round None 10.10 •• 4.2 1.8 X 1.9 (1.85) Nearly round Present Same as 10.7. 10.11 3.6 X 3.8 (3.7) Not well defined None Same as 10.7. Ter nperature range f rom 20.77° to 22.12°, — 1.35°. A verage len ^th of primitive streak. 83 mm. Average age, 3.64 \ lours, or 2.33 per cent of normal development. Average d ameter of blastoc erm, 5.03 n im. Average diamete r of area p( iUucida, 2.44 mm. Five b lastoderm s show 1 lo trace of the embryo. Physiological Zero for the Egg of the Domestic Fowl 369 SERIES B. — Incubation 11. Eleven eggs, incubated at 21.38° from May 14, 7 P. M. to May 20, 2 P. M., 5 days, 19 hours. Length Diameter Diameter Form No. of egg- of prim- itive streak in mm. Age in hours. of blas- toderm in mm. of area pellucida in mm. of area pellu- cida. dropic vesicles. General remarks. 11.1 1.0 8.0 6.4 2.0 X 2.3 (2.15) Oval Few present No primitive groove. 11.2 0.9 7.2 6.0 3.15 X 3.5 Nearly round Present No primitive groove. Primitive streak not clearly defined. 11.3 0.8 6.4 7.0 3.4 X 3.7 (3.55) No primitive groove. 11.4 5.8 X 6.4 (6.1) Not defined Few present Embryo undeveloped. 11.5 5.6 " Present «< 7^ Charles Lincoln Edwards. SERIES D. — Incubation 18. Twelve eggs, incubated at 22.51° from September 6, 2 p. m., to September 11, 2 P. M., 5 days. No. of egg- Length of prim- itive streak in mm. Age in hours. Diameter of blas- toderm in mm. Diameter of area pellucida in mm. Form of area pellu- cida. Hy- dropic vesicles. General re- marks. Vol. of egg in CO. 18.1 0.7 5.6 5.2 2.6 Round None 47.5 18.2 18.3 18.4 18.5 0. 2 1 6 5.4 X'5.9 (5.65) 5.0X5.2 (5.1) 4.4 X 4.9 (4.65) 4.6 2.7 X 2.8 (2.75) 2.4 2.2 X 2.6 (2.4) 2.4 " Few present None Of crescentic form Embryonic shield. 52.0 51.0 48.0 50.0 18.6 4.6 2.2 " " 52.5 18.7 4.5 2.2 " " Embryonic shield. 48.0 18.8 4.4 2.0 " Present Degenerated. 60.0 18.9 18.10 4.2 4.2 2.0 X 2.4 (2.2) 2.2 Oblong Round None Embryonic shield. 50.0 47.0 18.11 4.1 2.2 " Few 45.0 18.12 4.1 2.1 " present None Embryonic shield. 55.0 Tei nperature range from 22.' 0° to 22.90°, — 0.60°. Average length of prin- itive s treak, 0.45 mm. Average age, 0.6 hours, 0.5 per cent of normal developr nent. / Average diameter of blastoc erm,4.61 mm. Average diameter of area pellu cida, 2 .30 mm. All show more or less degeneration. Physiological Zero for the Egg of the Domestic Fowl 377 SERIES D. — lNCUB.\TioN 19. Nine eggs, incubated at 22.89° from August 26, 5 P. m., to August 31, 5 p. m., 5 days. Length Diameter Diameter Form No. of egg. of prim- itive streak in mm. Age in hours. of blasto- derm in mm. of area pellucida in mm. of area pellu- cida. Hy- dropic vesicles. General remarks. Volume of egg in c.c. 19.1 1.0 8 5.2 2.6 Round None Primitive streak not well defined. 47.5 19.2 0.9 7.2 4.4 2.1 " '• Like 19.1. 50.0 19.3 4.7 X 5.1 (4.9) 2.2 " " Embryonic shield. 62.5 19.4 4.8 X 5.0 (4.9) 2.2 52.5 19.5 4.6 X 5.0 (4.8) 2.4 Present Embryonic shield. 45.0 19.6 4.8 2.8 Few present Small indefinite patches of lower layer cells. 45.0 19.7 4.6 X 4.8 2.2 " None Enibryonic shield. 53.0 (4.7) 19.8 4.5 2.4 " ** " " 51.0 19.9 •• 4.15 X 4.5 (4.33) 2.1 X 2.3 (2.2) Oblong '• •• 52.5 Temperatu re range from 22.5° to 23.2°, — 0.7°. Average length of primitive s treak, 0.95 mm Average age, 1.67 hours, or 1.39 per cent of normal develo] jment. Average diameter of blastoderm, 4.73 mm. Average diameter of area pel ucida. 2.34 mm. 378 Charles Lincoln Edwards. SERIES D. — Incubation 20. Ten eggs, incubated at 23.40° from September 17, 9.30 A. M., to September 22, 9.30 A. M., 5 days. No. of egg- Length of prim- itive streak in mm. Age in hours. Diameter of blasto- derm in mm. Diameter of area pellucida in mm. Form of area pellu- cida. dropic vesicles. General remarks. Volume of egg in c.c. 20.1 0.7 5.6 4.5 2.4 Round Present 54.0 20.2 0.4 3.2 5.7 2.6 " 47.0 20.3 5.0 2.4 None 50 20.4 4.8 2.2 Present Degenerated. 45.0 20.5 4.7 2.7 None Embryonic shield. 49.0 20.6 20.7 4.2 X 5.2 (4.7) 4.5 2.0 2.5 Present .< 51.0 49.0 20.8 4.5 2.15 " u 53.0 20.9 20.10 4.2 X 4.7 (4.45) 4.4 2.1 X 2.4 (2 25) 2.0 Oblong Round " (( If 50.0 56 Temperatur( ; rangf : from 23 2° to 24.7°, — 1.5°. Average length of primitive streak, 0.5 5 mm. Average age, 0.88 hours, or 0.73 per cent of normal develop- ment. A\ erage diameter o f blastoderm, 4.73 mm. Average diameter of area pellucida, 2.32 mn n. Physiological Zero for the Egg of the Domestic Fowl 379 SERIES D. — Incubation 21. Twelve eggs, incubated at 24.24° from September 12, 2 P. M., to September 16, 2 P. M., 4 days. No. of egg- Length of prim- itive streak Age in hours. Diameter of blas- toderm Diameter of area pellucida Form of area pellu- cida. Hy. dropic vesicles. General remarks. Vol. of egg in c.c in mm. in mm. in mm. 21.1 1.40 12.0 5.8X6.0 2.7 Round Present 62.0 (5.9) 21.2 1.20 9.5 4.9 X 5.1 2.6 (5.0) 52.5 21.3 1.20 9.5 4.8 1.9 " " 50.0 21.4 0.80 6.5 5.6 X 5.4 (5.5) 3.1 " •• 61.0 21.5 0.35 2.5 4.9 2.8 " None 48.0 21.6 •• 5.4 2.2X2.7 (2.45) Oblong « 50 21.7 5.0 X 5.1 (5.05) 2.8 Round Present Embryonic shield. 45.0 21.8 •• 5.0 2.6 X 2.8 (2.7) Oblong None 55.0 21.9 •• 4.9 2.2 X 2.6 (24) " Present Embryonic shield. 45.0 21.10 1 .. 4.6 X 5.1 2 6 X 2.8 " None 50.0 (4.85) (2.7) 21.11 .. 4.1 X 4.6 (4.35) 2.15 Round Present . " " 54.0 21.12 .. 4.0 X 4.4 (4.2) 2.0 55.0 Temperature range from 203° to 25.8°, — 5.5°. Average length of primitive streak, 0.9* ? mm. Average age, 3.33 hours, or 3.47 per cent of normal develop- ment. Av erage c iameter of blastoderm, 4.99 mm. Average diameter of area pellucida, 2.53 mn [1. The large temperature range was due to the stopping of the gas for a few hours, during which time the incubator cooled to 20.3, but since the a verages fit into the general curves this was not a matter of enough importance ; to ex elude the data. 38o Charles Lincoln Edwards. SERIES D. — Incubation 22. Tiiirteen eggs, incubated at 25.22° from September 23, 3 p. m., to September 28, 3 P. M., 5 days. No. of egg- Length of prim- itive streak in mm. 22.1 1.70 22.2 1.50 22.3 1.30 22.4 1.20 22.5 1.20 22.6 1.15 22.7 1.00 22.8 0.90 22.9 045 22.10 22.11 22.12 22.13 Age in hours. 15.3 13.6 10.7 9.6 9.6 9.2 8.0 7.2 3.6 Diameter of blasto- derm in mm. 6.0 X 6.6 (6 3) 5.0 X 5.4 (5.2) 6.2 X 7.0 (6 6) 6.2 5.8 6.3 5.4 X61 (5.75) 6.0 X 6.9 (645) 5.4 5.6 5.4X5.7 (5.55) 5.0X5.15 (5.08) 4.7 X 5.0 (4.85) Diameter of area pelliicida in mm. 2.5 X 3.1 (2.8) 2.4 3.2 2.8 2.2 3.0 3.1 2.9 2.0 2.8 2.9 X 3.2 (3.05) 2.0 2.4 X 2.6 (2.5) Form of area pellu- cida. Round Oval Round Hy- dropic vesicles. General remarks. Present None Present None' Present Volume of egg in c.c. Embryonic shield 47.0 58.0 55.0 52.0 55.0 47 53.0 50.0 52.0 49.0 54.0 52.0 50.0 Temperature range from 24.6° to 256°, — 1.00°. Average length of primitive streak, 1.17 mm. Average age, 6.68 hours, or 5.57 per cent of normal develop- ment. Average diameter of blastoderm, 5.78 mm. Average diameter of area pellucida, 2.67 mm. Physiological Zero for the Egg of the Domestic Fowl. 381 SERIES D. — Incubation 28. Six eggs, incubated at 28.92° from October 1, 2 P. M. to October 5, 2 P.M., 5 days. Length No. of egg. of prim- itive streak and groove in mm. Age in hours. Diameter of vitelline area in mm. Diameter of area pellucida in mm. Form of area pellu- cida. Hydropic vesicles. Diameter of area vasculosa in mm. Vol. of egg in c c. 23.1 2.9 16 20.0X25.0 3.0X5.8 Pear Present 8.0 X 10.0 56.0 23.2 2.9 16 (22.5) 21.0 X 22.0 (4.4) 3.2 X 5.3 u (9.0) 7.4 X 8.4 47.0 23.3 2.9 16 (21.5) 15.6X17 2 (4.25) 3.5 X 5.15 « (7.9) 11.6 X 12.0 46.5 23.4 2.7 16 (16 4) 15.8 X 18.3 (4 32) 3.8 X 5.0 u (11.8) 9.8 X 10.4 48.5 23.5 2.0 16 (17.05) 20.5 X 22.3 (4.4) 3.4 X 5.0 « (10.1) 7.1 X8.8 48.0 23.6 2.0 16 (21.4) 13.7 X 14.8 (14.25) (4.2) 2.7 X 3.8 (3.25) " (7.95) 6.6 X 8.1 (7.35) 50.0 Temperature range i "rom 28.6° to 29.2°, —0 6°. Average length of primiti ve streak, 2..S7 mm. 1 \verage age, 16 hours, ot 13.33 per cent of normal develo P- ment. Ave rage dis imeter of blastoderm, 18. 85 mm. Average diameter of ar ea pellucida, ' -.14 mm . 382 Charles Lincoln Edwards, VII. SERIES E. — Group a. Nine eggs, kept at 17.4° from May 15, 9 a. m. to May 21, 9 A. M.,— 6 days. Diameter Diameter No. of of blasto- of area Form of area egg- derm pellucida pellucida. in mm. in mm. a.l 5.0 2.9 X 3.3 (3.1) Oblong a. 2 4.6 X 5.2 2.8 X 2.85 Nearly round (4.9) (282) a. 3 4.8 2 25 X 2.4 (2.32) Oblong a. 4 4.4 X 5.0 2.5 X 2.6 Nearly round (4.7) (2.55) a. 5 4.5 2.3 X 2.4 (2.35) a. 6 4.4 21 X 2.8 (2.45) Oblong a. 7 4.3 2.1 Round a. 8 4.0 2.0 X 2.2 (2.1) Nearly round a. 9 3.5 Not defined Temperature range from 16.9° to 17.9°, — 1.0°. Average diameter of blastoderm, 4.46 mm. Average diameter of area pel- lucida, 2.47 mm. Physiological Zero for the Egg of the Domestic Fowl. 383 fe rt' S rt 3 E 1) tfl £ £ be O O Pi t: 3 £ c t« O 3 ^ (C u t> O O (u XrX X, .CO LO x: LO ,_, eg ::f^ X2X x: ^^^ , to . CO t^ t— I O fO LO f-l 1— ( t--. CO S S . . s s 10 ir-j 1—1 < O ^ O O O "^ "~. ^ to 10 10 t'^ vO to 10 to to o •< o < <^ o <^ o •< o S S o "^ (M c>a t^l^l:^l>.t^t^t^t>.t^ f— ICOl~-COC^ 384 Charles Lincoln Edwards. Ill this group it was of interest to see if tliere would be any growth of the blastoderm at 17"-! 8° in eggs kept for periods of time varying from 1-17 days. The largest and smallest of the blasto- derms (b. I and b.9) were kept about the same length of time (between 8 and 9 days), the one (b.5) kept for the longest time (nearly 18 days) had a diameter (4.35 mm ), slightly under the nor- mal (4.41 mm.) while the measurements for the group show about the same variation as in the other groups of Series E; therefore it is evident that the given temperature does not influence growth either during short or long periods and that the physiological zero is above this temperature. SERIES E. — Group c. Twelve eggs, kept at 19.11° from July 12, 1.45 p. m. to July 19, 8.30 a. m., 6 days, 19 hours. No. of egg. Diameter of blastoderm in mm. Diameter of area pellucida in mm. Form of area pellucida. Hydropic vesicles. c.l 5.3 2.3 Round Present C.2 4.8 X 5 2, (5.05) 2.6 " " C.3 4.8 Not defined None C.4 4.7 .. " Present C.5 4.6 2.3 X 2.6 (2 45) Nearly round " C.6 4.4 3.0 X 3.1 (3.05) None C.7 4.2 X 4.6 (4.4) Not defined Present C.8 4.0 X 4.4 (4.2) None C.9 3.8 X 4.4 (4.1) 2.1 Round Present c.lO 4.0 2.45 " " c.ll 3.6 Not defined " C.12 3.2 X 4.0 (3.6) " " " Tetnperat are range from 18 50° to 20.10°, — 1 80°. Average dia meter of blas- toderm, 4.39 mm. Avera: yt diameter of art 3a pellucida, 2.49 mn 1. Physiological Zero for the Egg of the Domestic Fowl. 385 SERIES E. — Group d. Eggs 6, 8, 12, and 13, laid July 30; 1, 2, 3, 4, 5. 7, 9, 10, and 11, laid July 31. All kept in a refrigerator at 16° and fixed August 2. No. of ■ egg. Diameter of blastoderm in mm. Diameter of area pellucida in mm. Form of area pellucida. d.l 4.8 X 5.6 (5 2) 2.9 Round d.2 4.6 X 5 (4 8) 2.2 « d.3 4.4 X 5.0 2.9 X 3.2 Oblong (4.7) (3 05) d.4 4.4 2.6 Round d.5 4.3 X 4.5 (4.4) 2.0 " d.6 4.3 2.4 X 2.8 (26) Oval d.7 3.9 X 4.6 2.6 X 3.1 Oblong (4.25) (2.85) d.8 4.2 2.5 Round d.9 4.1 2.6 " d.lO 4.1 2.3 " d.U 3.6 X 4.0 2.3 X 2.8 Oblong (3.8) (2.55) d.l2 3.6 X 4.0 (3.8) 23 Round d.l3 3.8 2.0 ' Average diameter of blastoderm, 4.30 mm Average diame ter of area pellucida, 2.50 mm. 386 Charles Lincoln Edwards. SERIES E. — Group e. Eggs 2, 3, 4, 6, 9, 11, laid August 8; eggs 1, 5, 7, 8, 10, 12, 13, 14, and 15, laid August 9. All kept in refrigerator at 16*^ and fixed August 13. No. of egg- e.l e.2 e.3 e.4 e.5 e.6 e.7 e.8 e.9 e.lO e.U e.l2 e.l3 e.l4 e.l5 Diameter of blastoderm in mm. 5.5 4.7 X 5.1 (4.9) 4.9 4.6 X 5.0 (4.8) 4.6 X 4.9 (4.75) 4.4 X 5.0 (4.7) 4.5 X 4.7 (4.6) 4.1 X 4.7 (4.4) 4.4 4.4 4.4 4.1 X 4.6 (4.35) 3.7 X 4.3 (4.0) 4.0 3.7 X 4.2 (3.95) Diameter of area pellii- cida in mm. 2.5 2.8 2.2 2.6 2.5 X 2.6 (2.55) 2.6 2.5 X 3.0 (2.75) 2.5 2.3 2.0 2.5 2.4 1.9 1.9 X 2.2 (2.05) Form of area pellucida. Round Nearly round Round Not clearly defined Oblong Round Oblong Volume of egg in c.c. 65.0 57.0 60.0 53.0 60.0 48.0 60.0 50.0 45.0 47.0 53.0 50.0 50.0 49.0 50.0 Average diameter of blastoderm, 4.54 mm. Average diameter of area pellucida, 2.40 mm. Physiological Zero for the Egg of the Domestic Fowl. 387 VIII. The Physiological Zero. Prevost and Dumas, ^ using an ordinary incubator without thermo- stat, stated that development begins at from 28° to 30°. Dareste,^ as we have seen, places the physiological zero at 28°. He says emphati- cally, " Every observation which I have mentioned shows that no development takes place below 28°, and that development already commenced is fatally arrested at a certain point between 30° and 34°." Kaestner^ also adopted 28° as the physiological zero. Rauber^ gave the physiological zero at 25°, without, however, pre- senting any evidence for his statement. From my incubations, of those with average temperatures between 20°-2i°, two, D.16 at 20.13° and B. 13 at 20.92°, give no trace of the embryo, while two, C.15 at 20.05° ^"d B.14 at 20.72°, give respec- tively 0.07 and 0.54 per cent of normal development. The four in- cubations between 2i°-22° — D .17 at 21.08°, B .12 at 21.11°, B. 11 at 21.38° and B.lOat 21.87° — give respectively 0.56, 1.07, 1.25, and 2.33 per cent of normal development. In the series of incubations, eggs were kept at 17.4° (E.a) and 19.11° (E.c), but did not show any development of the embryo. Since there is in no case any trace of the embryo below 20°, and with four incubations between 2i°-22°, each showing some percent- age of embryonic development, I would place the physiological zero at the degree between 20°-2i°. The two groups of blastoderms showing traces of the primitive streak at this range of temperature may be considered as due to the normal variation in the constitution of protoplasm. Regarding the influence of temperature on protoplasm in general it has been shown by many authors that for some forms (plants, pro- tozoay amphibia, reptiles and others) the activities of metabolism and movement do not entirely cease until within a few degrees above zero. Early in the last century Macaire ^ showed that in the metabolic pro- ^ Dumas: Article " CEuf," in Dictionnaire classique d'histoire naturelle, xii, p. 121. 2 Dareste, C. : Recherches sur la production artificielle des monstruosites, ou Essais de teratogenic experimentale. 2 ed. p. 129, Paris, 1891. 2 Loc. cit. * Rauber: Sitzungsberichte der Naturforschenden Gesellschaft zu Leipzig, 1884. ^ Cf. Davenport, C B. : Experimental morphology, part i, p. 222, New York, 1897. Charles Lincoln Edwards. cesses resulting in phosphorescence in fireflies light appears just above 20°, while a few years later Artaud ^ demonstrated the same thing for marine organisms. For homoiothermic animals Martin and Applegarth^ showed that the isolated cat's heart may be cooled to 16.5° and will revive if soon warmed again, but that usually it dies at about 17° or 18°. So it is seen that the physiological zero for the egg of the common fowl, 20°- 21°, is also near the lethal temperature for the mammalian heart, and for the production of phosphorescence in fireflies and other organisms. IX. The Index of Development. Reaumur,^ and later Bonnet,* state that at a lower temperature than the optimum, development is retarded, while at a higher temperature it is accelerated. Dareste^ says that from 40°-42° an embryo of 24- 30 hours is equal to a 3 day chick of normal incubation, while on the other hand from 30°-33° ^ an embryo of 7 or 8 days is only equal to a normal 3-day chick. He gives the maximum for development at 43°,'^ 44° being fatal, and declares that there is very little development at 41°, 42°, and 43°.^ Rauber^ and Kaestner^ agree with Dareste in all essential points. Davenport^'' gives the following index of development for the em- bryo of the fowl, founded upon a paper by Fere.^^ Temperature .... Index of development . 34° 0.65 35° 0.80 36° 0.72 37° 3S° (l.CO) 39° 1.06 40° 1.25 41° 1.51 1 " The stage at 37° is taken from too few observations to be trustworthy. The stages at 35° and 36° are irregular, doubtless because of too few observations. As we go beyond 41° the ratio must decline again with great suddenness to 0°." ^ See note 5 on p. 387. ^ Martin, H. Newell, and Applegarth, E. C. : Studies from the Biological Laboratory of the Johns Hopkins University, 1890, iv, p. 275 ; also Martin : Physiological Papers, p. 103, Baltimore, 1895. ^ Reaumur : L'art de faireeclorre et elever en toute saison des oiseaux domes- tiques, foit par le moyen de la chaleur du fumier. Paris, 1749. * Bonnet : Experiences pour servir k Thistoire de la g^n^ration des animaux et des plantes. Trad, de Senebier, p. 188. ^ Loc. cit., p. 121. ^ Loc. cit. ^ Loc. cit., p. 129. I*' Loc. cit. Part II, p. 459. " Loc. cit., p. 1(8. 1^ Loc. cit. * Loc. cit., p. 128. Physiological Zero for Ike Egg of the Domestic Fowl. 389 After working over Fere's data, I get the results tabulated below. Temperature .... 34° 35° 36° 37° 38° 39° 40° 41° Index of development . 0.65 0.77 0.71 0.61 (1.00) 1.11 0.75 1.50 Number of embryos ( with age given . . \ 29 37 53 25 37 9 21 The differences in my table are mainly due to employing the full number of cases given by Fere and the addition of the index for 37°. If the objection given above in Davenport's foot-note should hold for 37° it could be urged with even greater force for 40° and 41° and almost as much for 34°. In Series A, B, C, and D I have carried the index of development from 30.75° to the physiological zero, 20°-2i° and present the results there found in the following graphic form. PERCENTAGE OF NORMAL DEVELOPMENT 390 Charles Lincoln Edwards. The break between Curves A and B represents a complete change in apparatus and environment from Cincinnati to Hartford. How- ever, this gap is not of great importance since it involves merely the number of blastoderms with embryos of the same general development. PERGEiNTAQE OrMOR/nAL DEVELOPnE/ST lOr I I I I I — I rr" I I I TrH""* I I i-i L PERGE/STAQE Or/SORnAL DEVELOPnEMT 20 22 SERIES B ANDC 24 Temperaturc ^u , , ■ ' ■ '■'■ ■ - Tip ~~~^~. 1 ^_ _ T 1 ■ " --hr --1- "1 1 1 ^« j.. 1^ I ^ '^ _ZL. ii. ^^^1 i: j Q __ __ _ ___ ^^^ [ ^ ^ ^x ;^*' :: ■"^.. .. _,,." ■^ ^ ^■^ _ _j_ JT. ■ jp It _^. |x ^ ^ =i±=^-::==s-¥^^^ --=== . p - L_T 20 22 24 2^ 2 " Temperatube SERIES D The last incubation of Curve A is 3 per cent less than the first of Curve B and C, but in both, the ontogenetic stage, that of the fully developed simple primitive streak, is practically the same. In Curve D, representing the incubations of Series D, I have given a contin- uous index of development from the minimum to 28.92°, which shows a general agreement with curves A and B-C and at the same time bridges over the gap between the first two. From the data shown in these curves two phases are distinctly represented. First between the physiological zero and 27°-28^, where only the primitive streak is found, the percentage of normal development not rising above 14 per cent. Second from 27°-28° to 30.75°, wherein the ontogenetic differ- entiation as shown by the appearance of mesodermic somites, nervous system and other features of the embryo, is marked by the abrupt rise from 14.91 to 54.83 per cent. Physiological Zero for the Egg of the Domestic Fowl. 391 X. Normal Size and Variation of the Blastoderm and of THE Area Pellucida. V. Baer^ noticed the great amount of variation in the early stages of the chick, but concluded that by the later stages each abnormal deviation had developed as nearly as possible back into the normal condition. Kupffer and Benecke,^ as well as Keibel and Abraham/^ have also called especial attention to this fact. The diameter of the blastoderm of the unincubated egg of the domestic fowl is given by Foster and Balfour'* at " about 4 mm.," by Dareste ^ from 3 to 5 mm., by Assheton ^ at 4.3 mm. ; and from figures 33 and 35 of DuvaP at 2.5 mm. and 3.09 mm. respectively. The last diameters, which are manifestly too small, are obtained from the magnifications as given. I get 4.41 mm. as an average value from the exact measurement of the blastoderms of the fifty-nine cases included in Series E. Until a larger series has been measured this may be taken as the normal average diameter of the blastoderm. The standard deviation from this average is 0.4792 mm. and the coefficient of variability, 0.1087. In regard to the form of the area pellucida there is considerable variation. Among the 136 blastoderms in which embryos have not developed, taken from the above series, there is a frequency of 59|f per cent round, \2\ per cent nearly round, 23^^y per cent oblong and 4j7y per cent oval. From the measurement of fifty unincubated blastoderms I find that the average diameter of the area pellucida is 2.51 mm., with a standard deviation of 0.3382 mm. and a coefficient of variability of 0.1347, This may be taken as the normal until a larger series is obtained. The average from twenty measurements by Dursy ^ is ^ Baer, C. E. V. : Ueberdie Entwicklungsgeschichte der Thiere. Beobaclitung und Reflexion. Konigsberg, 1828 und 1837. * Kupffer, C. und Benecke, B., Photogramme zur Ontogenie der Vdgel. Nova Acta Acad. Leop. Carol.. Bd. xli, 1879. 3 Loc. cit., p. 7. * Foster, M. and Balfour, F. M.: The Elements of Embryology, p. 4, London, 1896. ^ Loc. cit. : p. 287. ^ Assheton, R. : Proceedings of the Royal Society, 1896, Ix, p. 354. "^ Loc. cit. ^ DURSY, E. : Zeitschrift fiir rationelle Medicin, 1867, xxix, pp. 227. 229. 392 Charles Lincohi Edwards. 2.9r mm., while Moleschott^ gives the diameter at "something over 3 mm." XI. Growth of the Blastoderm Independently of the Appearance of the Primitive Streak. Panum ^ observed blastoderms which had developed without an embryo. In eggs not incubated until twenty-seven days after being brought into the laboratory, Broca^ found a transformation of the blastoderm with an absence of the embryo. Dareste* found that such cases could be produced by either the lowest or highest tem- peratures within the range determining development. This author also observed the continued growth of the blastoderm with a disorganiza- tion and disappearance of the embryo. Rabaud ^ demonstrated that in eggs kept at — i8° for one-half of an hour and afterwards incu- bated at 38° the large majority of blastoderms had extended to some distance over the yolk but without any trace of embryonic differentia- tion beyond that involved in cell proliferation. Data from the above incubations are arranged according to tem- perature in the table on page 393. Even before a trace of the prim- itive streak appears there is usually a multiplication of lower layer cells toward the centre, or in the posterior part of the area pellu- cida, forming the embryonic shield. This stage, which occurs gen- erally, I have noted as particularly prominent in A. 6, A. 7, B.IO, B.12, B.14, D.17, D.18, D.19, D.20, D.21, and D.22. Now, taking 4.41 mm. as the normal average diameter of the blas- toderm of the unincubated egg, I find in A. 4 at 27.00"", the highest temperature showing a blastoderm without a trace of embryo beyond that of the indefinite proliferation of cells known as the embryonic shield, that there is an average diameter of 7.00 mm., or an average growth of 2.59 mm., the greatest shown. Toward the lower limits there is not the same close relation. However, the least diameter (4.44 mm. in D.18) is still above the normal, but the difference (0.02) is not large enough to be of importance. The relation of temperature ^ MOLESCHOTT : Untersuchungen, X, Zur Embryologie des Hiihnchens, I. {cf. DuRSY, Loc. cii.). ^ Panum: Untersuchungen iiber die Entsteliung der Missbildungen, zunachst in den Eiern der Vogel. Kiel, i860. ^ Broca : Annales des sciences naturelles, i