anne a zie ee nyt ye aa pean sy - ie Oot Ae | , wt) hy rie ¥)» h Ae beats: ati?r vie anes ey SS eS = ays Hee } eatin tye sit at ate it ean Bee TORY piaie [ee Saeed atont ace ms Naty samy nears Pen era ety ord i A ¢ oe Me fy RAR et ie indo Wa ey: io Poe st vy ah eae any a iy Ow “ON f ry | : “ an ey i < “e THE AMERICAN JOURNAL OF PHYSIOLOGY VOLUME EII Re. mr , , sl? BALTIMORE, MD. 1920 AMT TAA JO VAOLAGE git 0 } c 7, \- CO, AS My Y, fe S We VES ZS oe ee Ye ts M gee {] 4 “Ae y Z//II\ wy ‘@ pox - he he! 1T.§ “Sei iery aM FOL Tae ‘ey CONTENTS No. 1. May 1, 1920 Gastric Response To Foops. X. THe Psycaic SecRETION oF GASTRIC JuicE In Normat Men. Raymond J. Miller, Olaf Bergeim, Martin E. oman ane E hglig ts. TAG als oes ope eden ons Ab. dome = spt Kl f Gastric RESPONSE TO Foops. XI. THE INFLUENCE oF TEA, COFFEE AND Cocoa tron Dicestion. Raymond J. Miller, Olaf Bergeim, Martin E. PEEP UaR AAO BART IN Tes) EDGING «os so las oo Made asa Map Delf YN. I. Rapip BLoop PLasMA PROTEIN DEPLETION AND THE CURVE OF REGEN- ERATION. H.P. Smith, A. E. Belt andG. H. Whipple.......... /i4..\. II. SHock as A MANIFESTATION OF TissteE INJURY FOLLOWING Raprip PuasMA PROTEIN DEPLETION. THE STABILIZING VALUE OF PLASMA Protremns. G. H. Whipple, H. P. Smith and A. E. Belt................ III. Facrors CONCERNED IN THE PERFUSION OF LIVING ORGANS AND TissuEs. ARTIFICIAL SOLUTIONS SUBSTITUTED FOR BLOOD SERUM AND THE RESULTING INJURY TO PARENCHYMA CerLLs. A. E. Belt, H. P. Sep OG Ey CCN Geel 4 Faia 1) 12s eee IR cP eae, tet IO OC CEC eae Ghee Tue SEASONAL VARIATION IN THE GROWTH OF BostTON ScHOOL CHILDREN. Fe OTE psn tc NR: SAM, OR Ae de ED Sh 5-0 SAB INTE Meee ce cag: Certain CHANGES NOTED IN ERGOGRAPHIC RESPONSE AS A RESULT OF ROB ACO-SMOKING. Uheaphile: Hapngel, coco wie coe: was srae.8 44. ofa nies. tele DETERMINATION OF THE CAPILLARY BLOOD PRESSURE IN MAN WITH THE Micro-Capittary TONOMETER. C.S. Danzer and D. R. Hooker....... A PLETHYSMOGRAPHIC STUDY OF SHOCK AND STAMMERING IN A TREPHINED SGANEMERM RY ON UEL: LD ICOBDINSE si 1. ci tees She Oud te Nn cece oo oa ect THE CHEMICAL CONSTITUTION OF ADENINE NUCLEOTIDE AND OF YEAST NDA GSTARITG! PACONTD-» VI CLLLE?® -22eeee 284 Tue ARTIFICIAL PropucTion OF MonsTERS DEMONSTRATING LocALIzED DEFECTS AS THE RESULT oF INJURY FROM X-Rays. W. M. Baldwin.. 296 THE RELATION OF THE EPINEPHRIN OUTPUT OF THE ADRENALS TO CHANGES IX THE RaTe OF THE DENERVATED Heart. G. N. Stewart and J. M. PRONG A) A. cone oie vag sole os one es oe tee eee ie ES 5.5 304 THe Errect orf Acips, ALKALIES AND SALTS ON CaTALASE PRODUCTION. PANE Bap ge eben e 42 Faas = eral weeps EE oe me foie et 364 Own THE PERMEABILITY OF THE PLACENTA FOR ADRENALIN IN THE PREGNANT RABBI AND ALBINO Rav. Yoshitaka Shimidzu........-.-..----+--eses 377 Tur CHARACTER OF THE SYMPATHETIC INNERVATION TO THE RETRACTOR Prexis MUSCLE OF THE Doc. Charles W. Edmunds................+..-- 395 Exercy EXPENDITURE IN HovseHoitp Tasks. C. F. Langworthy and H. G. Barottd he aZek es hte tee tr ee A Ee Po UR Ree 400 No. 3. Jury 1, 1920 AMPLIFICATION OF ACTION CURRENTS WITH THE ELECTRON TUBE IN RECORD- ING WITH THE STRING GALVANOMETER. Alexander Forbes and Cath- arine TRACK? «<<: < ok oc os os os cars ein Da Ee eee 409 VARATIONS IN THE RESPIRATORY Deap Air Space Dur To CHANGES IN THE DreprH oF BREATHING. R. G. Pearce and D. H. Hoover............ 472 Tre Susp-ARACHNOID AND INTRA-ARTERIAL ADMINISTRATION OF SODIUM BICARBONATE AND OTHER EectrotytTes. J. B. Collip............... 483 Tue INFLUENCE OF INTERNAL SECRETIONS ON BLOOD PRESSURE AND THE FORMATION OF BILE. Ardrey W. Downs.....,......-...-.2.20eene eee 498 GASTRIN Stupies. V. CHEMICAL STUDIES ON GasTRIN Boptss. F. C. Koch, Ac Bb. Lbuckhandt and les We Reeon .. fee eee eee Beesicsas ols ESSENTIALS IN MEASURING EPINEPHRIN OUTPUT WITH Pate Onde VATIONS ON Its RELATION TO THE RATE OF THE DENERVATED HEART. G.N. Stewart and JM! Rogofi:.0-22... 2. . sake : opel THE EFFECT OF SPLENECTOMY UPON GROWTH IN THE YouNG. Samuel Chester Henitts ooo 6 aoe oS aw 8 ee ns eee 562 THE IMPORTANCE OF VAGAL AND SPLANCHNIC AFFERENT IMPULSES ON THE ONSET AND Course oF TETAN(A ParaTHyrRopRiva. Walter Lincoln PALMER. v.58 5555 ok (Rass ws 0 ate vas bos eee . Sol INDEX THE AMERICAN JOURNAL OF PHYSIOLOGY VOL. 52 MAY 1, 1920 No. 1 GASTRIC RESPONSE TO FOODS! X. Tue Psycuic SEcRETION oF GASTRIC JUICE IN NorMAL MEN RAYMOND J. MILLER, OLAF BERGEIM, MARTIN E. REHFUSS anp PHILIP B. HAWK From the Laboratory of Physiological Chemistry, Jefferson Medical College, Philadelphia Received for publication February 24, 1920 As early as 1852 it was shown by Bidder and Schmidt (1) that the mere sight of food called forth the secretion of gastric juice in the dog. It remained, however, for Pavlov (2), some forty years later, to estab- lish more definitely the character of the so-called ‘‘appetite’’ or “nsychic’”’ secretion of the gastric juice. He pointed out that such a secretion was the normal initiator of gastric digestion in dogs, and might be induced by the sight, smell, taste, mastication or thought of food, or even through the stimulation of appetite by the presence of solid matter in the stomach. This was further emphasized by the fact that very appetizing foods, such as meats, induced a greater secretion in ‘‘sham feeding” experiments than milk or bread, which were not so greatly relished by the animals. The results of experiments on man have been by no means so con- clusive. Thus Carlson (3) was able to point out probable sources of error in most of the earlier investigations. This author concluded from the work of Homborg (4) and his own experiments that in spite of apparent evidence in the literature to the contrary, what Pavlov found true for animals was also true for man, namely, that gastric secretion was not induced by the chewing of indifferent substances nor by the taste or smell of chemical substances not arousing the appetite sensation. 1The expenses of this investigation were defrayed by funds furnished by Mrs. M. H. Henderson, The Curtis Publishing Company and Dr. L. M. Halsey. 1 THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 1 2 MILLER, BERGEIM, REHFUSS AND HAWK Accepting the view that the initiation of gastric secretion is de- pendent upon arousing or augmenting the appetite sensation, it becomes of immediate interest to determine whether the thought, sight, smell or taste of food exerts the greatest influence in this direction as well as to estimate the absolute importance of the appetite secretion as a factor in gastric digestion. Carlson found that no secretion could be induced in his gastric fistula subject by the thought of food, and that the secretion produced by seeing or smelling food was relatively slight and inconstant, the signifi- cant appetite secretion being that induced by the tasting and chewing of good food. Luckhardt (5), on the contrary, employing a Rehfuss stomach tube and using as a subject a completely normal man, found under good experimental conditions that the combined sight and smell of food markedly increased the flow of gastric juice. Pavlov believed the appetite secretion to be of very great import- ance in initiating gastric digestion. This is discounted by Carlson, who found in his human subject, as well as in cats and dogs, that the continuous secretion of the stomach served a similar purpose, and that elimination of the appetite secretion did not cause indigestion. That the unpalatable condition of a food need not necessarily influ- ence its ultimate digestion and utilization in the alimentary tracts of normal men was indicated by work carried out in the laboratory of Atwater, who found that nauseating meat, the so-called ‘‘embalmed beef’’ fed to soldiers in the Spanish-American war, was still well utilized. That the economy of man is well served under such circumstances must remain doubtful. In this connection it is also necessary to consider psychic or other influences tending to inhibit the development of the appetite secretion; bearing in mind that emotional excitement may destroy the motor (6) as well as the secretory (7) activities of the stomach. The present paper is a contribution toward determining the relative importance of the various factors involved in the appetite stimulation of gastric secretion, as well as toward estimating the influence of the appetizing or unappetizing character of a meal and of the mental attitude of the subject upon the gastric response and the ultimate digestion of food. Following the appetite stimulation in each case, the stomach was emptied at regular intervals, using a Rehfuss stomach tube, the volume of secretion determined, its free and total acidity, pepsin and amino- acid nitrogen estimated according to procedures previously described (8). PSYCHIC SECRETION OF THE GASTRIC JUICE 3 The sight psychic secretion and the effect of smell. Experiments were carried out to determine whether the sight of food alone, or the sight and odor of food combined gave rise to the production of any gastric secretion, as well as to compare the effects on such secretion of the sight of foods prepared in an unpleasing manner. A breakfast table was set, in a small well-lighted laboratory, with clean linen, bright chinaware, and the following foods were served in a pleasant manner and at one time: ham and eggs, oranges, shredded wheat, bread and butter, hot coffee, cream and sugar. Subjects were brought to this room after any residua had been removed and the level of continuous secretion established. The nose was kept lightly but effectively clamped throughout. In figure 1 are charted the results obtained with a subject who was a laboratory helper, accustomed to eating lunch at the laboratory. Total volumes of secretion removed are charted with total and free acidities, amino-acid nitrogen values and a so-called total actual acidity which indicates the product of the volume by the total acidity and represents the total amount of acid in the gastric samples. It will be noted that this subject showed a marked response to the sight of food as regards the volume of gastric juice secreted and the “total actual’ acidity. The continuous secretion in this case was high and a greater augmentation by psychic stimulation might perhaps be expected. Under similar conditions another subject (fig. 2) showeda lower and less acid continued secretion. Sight augmented the volume but slightly, although the acidity was distinctly increased. Somewhat less accustomed to eating in the laboratory were the subjects of the experiments charted in figures 3, 4 and 5, which were carried out in exactly the same way except that clamps were not used. The distinct but not voluminous secretion in these cases may be attributed mainly to the effect of sight as the odor of the meal was not pronounced and odor was found to have little influence on psychic secretion in these subjects. The subjects were, of course, led to believe they would actually receive the food. To determine whether any elaboration in setting was requisite for the stimulation of secretion by sight, the meal was simplified to a simple half grape fruit served in the usual way. The odor being imperceptible, the noses were not clamped. The results are charted in figures 6 and 7. A stimulation of secretion was brought about by the sight of the food in both cases. It is also probable that some of the psychic secretion may leave the stomach during the intervals of the experiment with the 4 MILLER, BERGEIM, REHFUSS AND HAWK increase in gastric tonus and produce a stimulation of the pancreatic secretion. The effect of allowing subjects to seat themselves at a breakfast table prepared in an unpleasing manner was also tried out on six sub- jects. The same table was used as in the preceding experiments, and the same foods were served. However, the ham and eggs were scorched; the shredded wheat biscuits and bread roughly broken; the coffee and milk weak and diluted; the butter soft; sugar lumpy and dark; the oranges partly squeezed; the dishes generally somewhat greasy and with an appearance of dirtiness induced by the use of charcoal. Newspapers were used in place of linen. The noses of the first two subjects were clamped, of the others free. The results of these experiments are charted in figures 8, 9, 10, 11, 12 and 13. In only one of these cases (see fig. 12) was any secretion induced above the level of the continuous secretion. It is evident, therefore, that food served in an unpleasant manner will not give rise to an appetite secretion under ordinary conditions, although custom and degree of hunger will naturally influence the conception of an appetizing food. Breakfasts served in a pleasant manner and with appetizing foods were set before each of the six subjects just mentioned from 15 to 30 minutes after they had been presented with a view of a breakfast of the opposite and discouraging character and which had evoked no psychic gastric response. The results are plotted in the same charts as the preceding tests (figs. 8, 9, 10, 11, 12 and 18). It will be noted that in the first two cases a marked appetite secretion followed the presentation of the second meal, this being a sight effect as noses of these subjects were clamped. However, the other four sub- jects did not show a psychic secretion under these conditions. The subjects showing a response were accustomed to eating in the laboratory and may have felt that they would not be expected to actually partake of the disagreeable food. The other subjects having no knowledge of the character of the test might well be more strongly repulsed by the first meal, this effect being carried over for the period of 15 to 30 minutes until the palatable meal was set before them. A secondary effect might also be their suspicion that since they were not permitted to partake of the first meal they might not have a chance at the second, although the contrary view was impressed upon them. The psychic secretion and the odor of food. The influence of the odor of food alone on psychic secretion was tried out on seven subjects (see PSYCHIC SECRETION OF THE GASTRIC JUICE 5 figs. 14 to 20). The odor of frying beefsteak was used as a stimulus, the odor being pleasant, strong and unmistakably that of an appetizing food of common consumption. Subjects were blindfolded and the ears were muffled in order to exclude the influence of the sight of the steak and of hearing it fried. Subjects inhaled liberally the fumes arising from the frying steak. As in our other tests, the subjects had had no food for fourteen hours. Three of the subjects showed no increase in the volume of secretion under the influence of these odors. The four others showed some increase, but in no case was the secretion voluminous. It would appear from these tests that odor was considerably less important than sight in inducing the appetite secretion, at least in man. These results are supported by the experiments previously mentioned, in which com- bined sight and odor of food brought forth no greater secretion than sight alone. Odor may have an influence on the motor activity of the stomach and may be of importance in animals with a more highly developed sense of smell. There may very well also be considerable differences in individuals of the human species. The psychic secretion and the tasting and chewing of food. The sub- jects of the preceding tests on the influence of odor were permitted to rest for half an hour to reéstablish the level of continued secretion which had in most cases been little affected. They were then per- mitted to chew for five minutes portions of tenderloin steak with strict caution to swallow none of the pieces, this possibility being checked by careful examinations of the gastric contents. In all cases the sub- jects remained blindfolded. In the first four cases the noses of the subjects were also clamped so that none of the vapors could be inhaled by that channel. The results are charted in figures 14, 15, 16 and 17, and show no distinct influence of tasting and chewing meat under these conditions upon the secretion of gastric juice. In one case the volume of gastric contents was somewhat increased, but the low acidity shows that very little acid gastric juice could have been secreted. Ap- parently the taste and chewing of food in the absence of sight or odor produced no marked psychic secretion. In three other cases the same procedure of chewing and tasting beef- steak was carried out, but the noses of the subjects were unclamped, sight, however, being excluded (see figs. 18, 19 and 20). As illustrated in these cases the influence of the combined tasting, chewing and smelling of food on the secretion of appetite gastric Juice was very pronounced and was much greater than that of smell alone. 6 MILLER, BERGEIM, REHFUSS AND HAWK The influence of the sound or thought of food. Subjects were blind- folded and had their noses clamped to exclude the sight and smell of food. They were seated before a frying pan in which a steak was being broiled with plainly audible sputtering and sizzling. They were told what a fine, juicy steak was being prepared for them and a general attempt made to keep their attention on the subject of appetizing meats. The results are charted in figures 21 and 22. In one case the result was negative; in the other case a distinct stimulation of secretion resulted. The variation must be ascribed to individual differences. After one-half to three-quarters of an hour rest, the nose clips were removed and the subjects permitted to smell as well as hear the sputtering of frying steak. Results are plotted on the same charts and show that in one case a very slight rise in secretion took place. In the other case a definite stimulation occurred, although the earlier level for hearing and thought of food was not surpassed. In one case the subject was permitted to smell feces of a repulsive odor fifteen minutes after smelling steak. Any psychic secretion appears to have been depressed to the level of the continuous secretion but not below this. Experiment 23 gives a comparsion of the psychic effects of: a, the sight of a frying steak (ears not stoppered); b, sight and smell; and c, taste of the same food. A distinct stimulation was produced by the sight of the food. One-half hour later the sight and odor of similar food produced a very similar stimulation. After a further interval - of 15 minutes the taste of the food gave a lesser stimulation than sight or sight and smell had previously done. A summary of some of the results obtained on two of our subjects in so far as volumes and ‘total actual acidities” of appetite secretions were concerned, is given in figures 24 and 25. They must, of course, be considered in connection with details of individual experiments. They serve, however, to emphasize the important réle of the sight of food as a stimulus to the appetite. The influence of palatability or unpalatability of a meal on its gastric digestion. ‘Two subjects were given uniform meals prepared and served in the ordinary manner. On a later day they were given the same foods prepared in as unpalatable a manner as possible without altering their chemical composition. The meal used consisted of: cream of wheat, 100 gm.; sugar, 10 gm.; milk, 35 cc.; coffee, 100 cc.; graham crackers, 50 gm.; oranges, 50 gm.; water, 100 cc. On the second experimental day these foods were all mixed together in a conglomerate PSYCHIC SECRETION OF THE GASTRIC JUICE il mass, discolored with small amounts of burnt crackers and charcoal, and the atmosphere at the table saturated with the repulsive odors of valeric and butyric acids. The first of these subjects was of a nervous temperament and from his statements and manner was judged to be easily influenced or dis- turbed by the character and preparation of his food. In fact he posi- tively refused on one occasion to continue eating a meal of the second type mentioned above, although urged to do so in the interest of science. The results on this subject (see figs. 26 and 27) show no delay or inhibition of the acid response of the stomach, although the evacua- tion time was somewhat prolonged. The second subject was accustomed to eating in a laboratory, was of a phlegmatic temperament, claiming and appearing to be very little disturbed by the appearance of food or the condition in which it was served. The results on this subject are given in figures 28 and 29. The unpalatable food showed a rapid, though not quite so rapid, development of acidity and a few minutes quicker evacuation. The first subject was also tried out with a palatable meal 50 minutes after he had violently refused one which he believed to be contami- nated. The result is charted in figure 30. The development of acidity was even more rapid than in the case where the meal was given under normal conditions. If any depression of psychic secretion was carried over through this interval, there were no signs of it. The first subject was also given a meal of unpalatable character similar to the ones already described, but prepared by himself and hence known by him to be innocuous. The results as charted in figure 31 show a rapid development of acidity and quick evacuation. Some information with regard to the gastric response to foods un- palatable in appearance, odor and taste was obtained by experiments on the feeding of Chinese preserved duck eggs called ‘‘pidan.”’ These eggs have dark greenish yolks and yellow-brown ‘“‘whites’ of a firm, gelatinous consistency and possess distinct odors of ammonia and hydrogen sulphide. One subject disliked these eggs but did not know what they were and was not especially prejudiced against them. The other subject, ‘‘Don,” was of a nervous type, and just as he finished eating the eggs he was told in a joking manner by one of the laboratory wits that they were of prehistoric Chinese origin. The subject became clearly suspicious that something had been given him that was not entirely fresh. The results of this test as compared with similar tests. on boiled duck eggs and on raw white and yolk of egg are given in 8 MILLER, BERGEIM, REHFUSS AND HAWK figure 32. They show a depression of gastric secretion after ‘“pidan”’ lasting for an hour and a quarter, the acidity then rising rapidly to normal figures. This delay may have been due to inhibition of appe- tite secretion, gastric activity being finally aroused through chemical stimulation following the solvent action of the slow continued secretion. The failure of this unappetizing food to arouse the secretory or motor activities of the stomach to a normal response is indicated also by our results on the first subject mentioned above. The curves are given in figure 33 and show that while raw hens’ eggs gave an acid response of over 100 in the first hour and left in 2} hours, the preserved eggs at no time gave acidities of over 30 and remained in the stomach 43 hours. It appears that the unappetizing character of these eggs led to a delayed acid response and slow evacuation, perhaps complicated by their failure to show some early digestion with consequent chemical stimulation. Influence of prejudice against a food on its digestibility in the stomach. It is very common to find people who have a prejudice against certain foods generally classed among the most wholesome articles of diet. Certain cases may be due to a food anaphylaxis or sensitivity. Others may be due to defective gastric or intestinal digestion or other causes. Undoubtedly many have no foundation and are the results of wrongly placing the blame of certain digestive disturbances. One of our subjects, ‘‘Ham,” had a strong prejudice against eggs in any form and had not eaten them for years. He was with difficulty persuaded to take eggs prepared in several different ways. The results of these tests are plotted in figure 34, and show that eggs were digested by this subject in a perfectly normal manner, at least as far as the stomach was concerned. Neither did untoward symptoms of any kind develop. Influence of newspaper reading on gastric digestion. Subjects were permitted to read newspapers throughout the course of a meal of palatable foods, the same test meal as used in previous experiments. The gastric responses of two subjects who read newspapers and re- sponses of same subjects with no reading but with usual table talk are charted in figures 26, 28, 35 and 36. No distinct influence of news- paper reading was noted. Responses were quite normal in all cases. The slight differences in acid development and evacuation time were in one case favorable and in the other case unfavorable to newspaper reading. Influence of the unpalatable character of a diet on its ultimate utiliza- tion by the human body. Smith, Holder and Hawk found (9) in a me- tabolism experiment on a normal man that where a uniform diet of a PSYCHIC SECRETION OF THE GASTRIC JUICE aD palatable character was given for several days, followed by a period in which the same foods were jumbled together in dirty dishes and served amidst ill-smelling and otherwise unpleasant surroundings, that the nitrogen utilization in the first case was 86.7 per cent and in the second, 85.6 per cent. The nitrogen balance showed a retention in the ‘first period of 3.0 per cent and in the second, of 6.4 per cent. This in spite of the fact that the subject was only with difficulty persuaded to eat the unpalatable food and that another subject who was given the same kind of food became nauseated and could not continue. Influence of anxiety on gastric digestion (10). The study of the influence of emotional strain on digestion in man offers some difficulties due to the fact that the emotions cannot be readily controlled, nor are the subjects of extreme emotion readily amenable to experimentation. We were, however, able to obtain an interesting illustration of the profound effect of mental anxiety on gastric digestion in the case of one of our subjects. The man was a first-year medical student who had previously served as a subject of gastric tests. He was given 100 grams of fried chicken on the morning of an important examination in chemistry and was asked to write out his answers during the course of the test. He was plainly worried over the outcome of the exami- nation and of his year’s work. The effect upon gastric digestion was the prolonging of the evacuation time to 6} hours. The intra-gastric acidity remained in the neighborhood of 90 for 3 hours. The normal digestion curve for fried chicken on this subject was obtained a week later under the best mental conditions. The time required was 4} hours and the maximum acidity about 65. It is not at all surprising that worry aggravates a condition of gastric ulcer. An interesting experiment on the digestion of milk in the human stomach may be cited in this connection (11). It was found that in the stomach of one of our subjects milk would not curdle. The test was carried out at the end of the year immediately before the final examinations. The subject was one of the most brilliant students in his class and had worked hard. We made several tests on this student and in every case milk left his stomach rapidly and without curdling. He digested all other food normally. The next fall, upon his return to college, we made another milk test upon him and found that his stomach curdled milk in a normal manner. At this time he was in fine physical condition, having had a long, pleasant vacation, whereas in the spring he was in a highly nervous state as a result of his hard study. This serves to illustrate the influence which rigid and prolonged mental application may exert upon the stomach in certain types of individuals. 10 MILLER, BERGEIM, REHFUSS AND HAWK SUMMARY AND CONCLUSIONS The sight alone of a table well set with nourishing foods was found to give rise to a distinct secretion of gastric juice in normal men. The sight of a half grape fruit only resulted likewise in an appetite secre- tion. The sight of the same foods illy prepared and poorly served resulted in no stimulation of appetite secretion. The service of a well prepared meal half an hour after the service of a poorly prepared one gave in some instances a distinct secretion, in others not. The odor alone of frying meat produced in some cases no appetite secretion, in others a slight secretion. Odor alone produced less stimulation than sight alone. The tasting and chewing of food in the absence of smell or sight produced no marked psychic secretion. The combined influence of the tasting, chewing and smelling of food was pronounced and much greater than that of smell alone. The sound and thought alone of a frying steak gave rise to a gastric secretion. The influence of smell with hearing produced little ad- ditional effect. Evil odors depressed sécretion to the level of the continuous secretion. In consecutive tests the sight of food, with and without odor, pro- duced similar degrees of stimulation, while taste alone had less effect. Mixed meals consisting of nourishing ingredients but very un- pleasantly prepared and served gave rise in the case of a phlegmatic individual to no distinct delay in the development of intra-gastric acidity or in evacuation. A more susceptible individual showed a slight delay in evacuation time, but none in acid response. Chinese preserved eggs, unpalatable to our subjects in appearance, odor, taste and belief in their unwholesome character led to delayed acid response and evacuation. In one case the normal acid level was. ultimately attained due to chemical stimulation. In one subject a strong prejudice against eggs was found not to result in any abnormal gastric response when eggs were eaten. The ultimate utilization of the protein of a diet prepared in a most unpalatable manner was not found to be appreciably less than that of the same diet served under the best conditions. Newspaper reading during the course of a meal could not be shown to have any distinct influence on gastric digestion. Anxiety and mental strain were found to markedly delay gastric digestion. The authors desire to thank for their codperation the students of. Jefferson Medical College who acted as subjects of these tests. PSYCHIC SECRETION OF THE GASTRIC JUICE i BIBLIOGRAPHY (1) BrppER AND ScumipT: Die verdauungssafte, etc., 1852. (2) Pavuov: The work of the digestive glands, 2nd English Ed., Philadelphia, 1910. (3) Cartson: The control of hunger in health and disease, Chicago, 1916, pages 236-240. (4) Hompore: Skand. Arch. f. Physiol. 1904, xv, 209. (5) LuckHarpt: Report to Amer. Physiol. Soc., December, 1917, and private communication to one of the authors. (6) Cannon: The mechanical factors of digestion, London, 1911. (7) BickEL AND Sasaki: Deutsch. med. Wochenschr., 1905, xxxi, 1829. (8) FisHpack, SmitH, BERGEIM, LICHTENTHAELER, REHFUSS AND Hawk: This ~ Journal, 1919, xlix, 174. (9) SmiraH, HoLtpER anp Hawk: Reported Soc. Exper. Biol. Med., February 18, 1920; Science, in press. (10) MiLier, BerGerm AnD Hawk: Reported Soc. Exper. Biol. Med., February 18, 1920; Science, in press. (11) LicHTENTHAELER, BERGEIM AND Hawk: Unpublished data. re. 1 a, 7, () HOY ON 00 burke OP hour Ere. 2; Case,- 35-Sokr Case,. 34-0'Brn O fofe) “ho, WOH ee. Volume oO ce. aS > ip) Fia. 3. Fia. 4. g Case, 36-lha |Ca 00. “ho LOH. 02. Volume Sight. Fic. 6. JK 7é ay 4 7 ra a Q - ” ” ” ro) 9 B, my y — Cot S/he | “B77DL Pur Joayy oucp\rpucez mes 5 Se eae oc 3 of 7 fi > a a We | a Ms > My, 4 5 | @) ) OF On, O O O HOY UY, 00 HOY WU, 00 Case, 3/-/cklh. Case,- 30-Cry. aUIngA Vv ‘OY, ‘DO peed! O14 L ly hres le Fre. 10. » @97OFO7RY 702) 9770707044) Case,. 32--Swe ‘Bunge 99 HOW OUpy 00 J hour bres i2s 15 Case. 33-lWiat 87907?) ed. 9024) 979 070 704u / hour Jan), 1183. Case. Z2/ -Mckl wimg buameys 9980) WEES: ‘OUNIA) 00 Fria. 14. 16 Fie. 15. 23-Swe wmmg buameyo 780, 7784S Fig. 16. 17 Case_24-Wat, TGe lige Case,. Cry -25 ‘SSo, lurojousy jeg Be ae) Blun 7QP 00 ZOK/: ‘ Os My 2 O (YY Lhour Fie. 18. 18 eae Fig. 19 -ih Case, 26 -OBra 1724S Ee : ” EAT erottrjO" Oo Teoynuro ty Os/ 0? Og BUNA ‘oo : Pz & za ty, a OM ‘O; 2) Fig. 20. (PUnes Pur) 77 BUS S >: Case 9-/i GF HOW MG, 00 L:AOur ‘Qo , Case,- //-Go. 7B ui SB ee ms S Pu punog Asay Pun[g bu2r9804 L.Aour Fig. 99 9 Rigs Zo: = g S 2980]. WausAal on 6, ; qyhS A VJEOUSA , ] punos HOH 0%, 20 Fic, 24. 21 Case,. Go. ~ -. an We (pq) 3 rh (poob)yauc 4 ae Pures yes ce PLEA ] nee oe SP OLATOALY POP PYE7S Fra. 25. Meal, palatable. S ™ ° ») q S os) HOY A/y 00 ; HOY OY 00 Fig. 26. ce Yi0, KOH ee. Yo, HOH Case,./5-Go, Meal: unhalatahle. Lhour Fra. 27, Case, /4--/Mi Case,-16-%i. z .@) = Meal, unpalatable. s gp “hour FE 1G. 29. = Case-17-Go. \e) e S > 9 © Fra. 30. Case, -/8-%. eal, unpalatable. 00. Vio HOH joey Sle Case. Don-39 - Faw lelks (4o) « -45« Faw Whites fp): eee rd Bovled Turkey £998 (Bo) Vagie ae FPidan (F) Case,. Ham- /08 » Chunese Pidans. (FP) <-=-~. 46° lhole Faw £99: (Few) en BS Hard Boiled *” (Bl) Fia. 33. Case.. rae eo: exh Jreed Sees ee) 59 - : Short ae £og3° J oad Ga Sau 5 a. Yo HOH Case,- /8-Go Meal, palatable hewshaper reading. ee. “ilo, KOH. ‘hour IG;oo: Case -20-Mi hse goreiend ep une aes f i w_ SZ Bf a a GASTRIC RESPONSE TO FOODS! XI. THe INFLUENCE OF TEA, COFFEE AND Cocoa UPON DIGESTION RAYMOND J. MILLER, OLAF BERGEIM, MARTIN E. REHFUSS anp PHILIP B. HAWK From the Laboratory of Physiological Chemistry, Jefferson Medical College, Philadelphia Received for publication February 24, 1920 Many experiments have been made to determine the effects of tea and coffee on peptic digestion in vitro. The object of the series of ex- periments reported here was somewhat different, namely, to determine in what way the responses of the stomachs of normal human subjects to mixed meals would be affected by adding to such meals equal vol- umes of water, tea, coffee or cocoa—hot or cold—and with or without the addition of cream or sugar. Certain observations relative to changes in pulse rate and other symptoms following ingestion of these beverages were also made, but the paca were not complete in this respect. We are not familiar with any previous experiments of the type car- ried out by us. Penzoldt (1), however, studied the evacuation of 200 ec. portions of water, tea, coffee and cocoa on a single subject and found them to leave the stomach as follows: Water, 1 hour, 15 minutes; tea, 1 hour, 30 minutes; black coffee, 1 hour, 45 minutes; coffee with 5 ce. cream, 2 hours, 15 minutes; cocoa with water, 1 hour, 45 minutes; cocoa with milk, 2 hours, 30 minutes. Our experiments were carried out upon normal medical students. They reported at 8:00 a.m. without breakfast, and without removal of residuum they were given the following uniform meal, either alone or with the beverage to be tested which was drunk during the course of ingestion of the other foods. Test meal: Scraped beef, grilled, 75 grams; mashed potatoes, 50 grams; toast, 40 grams; butter, 15 grams; salt, 13 to 2 grams as desired. The tea, coffee and cocoa used were of good quality and prepared of a moderate strength by the dietitian of the hospital, exactly as used in the regular dietary of the institution. Where sugar was used, four 1 The expenses of this investigation were defrayed by grants from Mrs. M. H. Henderson, The Curtis Publishing Company and Dr. L. M. Halsey. 28 GASTRIC RESPONSE TO TEA, COFFEE AND COCOA 29 lumps of sugar per liter were given. In the case of cream approxi- mately 15 cc. of a 20 per cent cream were used. No milk or added sugar were used in preparing the cocoa, which was made up according to the directions on the container. As soon as the meal was finished the subjects were made to swallow the Rehfuss stomach tubes. From 5 to 8 cc. of the stomach contents were aspirated at 15-minute intervals for 13 hours. The subjects were then allowed to attend a class and returned 1% hours later (after a total experimental period of 3 hours). Then two samples were with- drawn at 10-minute intervals, and finally the stomach was completely emptied, this being checked by a lavage of 150 cc. of water. Total and free acidities were determined by titration and pepsin by the Mett method. The volumes of contents at the end of the 3-hour. period were measured. The pulse rate was also determined at inter- vals and subjects reported any unusual symptoms. Tests were carried out in which the following beverages were added to the uniform test meals: (1) Cold water, 1 liter, 15°C. (2) Hot coffee, plain, 1 liter, 45-50°C. (3) Hot coffee, with sugar, 1 liter, 45-50°C. (4) Hot coffee, with cream, 1 liter, 45-50°C. (5) Hot coffee, cream and sugar, 3 liter, 45-50°C. (6) Cold tea, plain, 1 liter, 13°C. (7) Hot tea, plain, 1 liter, 50°C. (8) Hot cocoa, plain, 1 liter, 57°C. (9) Hot cocoa, plain, 3 liter, 57°C. Thirty-seven experiments were carried out on four different subjects, duplicates being made on each subject with the basal diet alone. The results are charted in figures 1 to 34. The results with regard to evac- uation time are summarized in figure 34. INFLUENCE OF BEVERAGES ON THE ACID RESPONSE OF THE STOMACH Influence of cold water on the acid response. The diet alone contain- ing, as it does, considerable meat gives rise to the rapid development of a relatively high acidity, much of which is represented by combined acid (figs. 1, 10, 19 and 26). As no liquid whatever was taken with the meals in these cases and in the early part of digestion the absorp- tion of gastric juice by the meat was pronounced, a uniform mixture 30 MILLER, BERGEIM, REHFUSS AND HAWK was not at once obtained in the stomach; and the samples withdrawn varied somewhat with the exact location of the tip in the stomach. For this reason the duplicate curves do not coincide, in spite of the fact that digestion follows the same course and is concluded after the same lapse of time. Where a moderate amount of liquid is present or the combining power of the food less, it is possible to sample uniformly and to obtain very similar curves with the same food eaten at different times. For this reason the experiments on different beverages are, with regard to acid response, more comparable with each other than with the diet lacking liquid. The curves of acidity, especially total and combined acidity, appeared to rise somewhat more quickly and to a generally higher level with the diet alone than with the added liter of cold water. As the evacuation time was practically unaffected and the acidities still high following the high water ingestion, the water cannot be said to have hindered digestion and must have given rise to a considerable stimulation of acid secretion or been rapidly emptied in large measure, or more probably both. In consideration of the results of these tests it must be constantly borne in mind that unusual amounts of the various beverages were given. This was intentionally done in order that any untoward in- fluences might be accentuated. Influence of coffee on the acid response. Hot coffee, plain, that is without cream or sugar and at a temperature of 45 to 50°C. was given to four subjects. The acid responses are charted in figures 3, 12, 21 and 27. In the first three cases a liter of the coffee was taken; in case 27, however, only half a liter. The charts show in a general way that the acid responses were very similar to the results obtained with cold water, and not very different from those obtained with the basic meal by itself, although not attaining quite the heights of total acidity given by the latter. There was no distinct indication that the coffee in- hibited the secretion of the gastric juice. Sugar was added to the coffee in two cases (see figs. 4 and 22). Comparisons of the acid responses with those of the same individuals taking plain coffee (see figs. 3 and 21) show that the sugar depressed considerably the acid response. It has been clearly demonstrated in this laboratory that sugar depresses gastric secretion; and it is, there- fore, improbable that the low acid values in the earlier period of diges- tion in the cases where sugar was added to the coffee, are due merely to the slower evacuation of the beverage. GASTRIC RESPONSE TO TEA, COFFEE AND COCOA ona The addition of moderate amounts of cream (15 cc. of 20 per cent cream to the liter) was tried out in two cases (compare figs. 12 and 13 and figs. 27 and 28). It will be noted that the acid response was not demonstrably influenced by such addition. Hot coffee with both cream and sugar was given in four cases (see figs. 5, 14 and 29). The results indicate that with the addition of cream and sugar the acid response was higher than where sugar alone was used, and more nearly approached that of the plain coffee. This may perhaps be explained by the fact that the coffee was more appe- tizing with additions of cream and sugar, provoking a more marked appetite secretion and also leading to more rapid evacuation. Influence of tea on the acid response. One liter of cold tea (12° to 13°C.) was given during the course of the standard meal to each of these subjects (see figs. 6, 15 and 30). If the acid responses are com- pared with those of the same subjects to cold water or plain hot coffee, it will be noted that they are very similar, although one subject showed a slightly higher curve with tea than with water or coffee. Hot tea (50°C.) was also given to three subjects (see figs. 7, 16, 24 and 31). With slight variations, generally in favor of the hot tea, the results obtained were not distinctly different from those with water, coffee or cold tea. Influence of cocoa on the acid response. That cocoa when taken in amounts of 1 liter at a meal has a depressing action on the develop- ment of gastric acidity is clearly apparent from the results on our four subjects as charted in figures 8, 17, 25 and 32. In each case the intragastric acidity rises slowly as compared with either cold water, hot coffee or hot or cold tea. This may be due to several causes. Un- doubtedly the sugar content depresses secretion and delays evacuation as sugar solutions have been clearly shown to do. The higher fat con- tent of cocoa may play a part as well as the presence of other con- stituents. In three cases the amount of cocoa drunk was reduced to one-half liter. The depressing action of cocoa on the development of acidity may be noted in these cases (see figs. 9, 18 and 33). In general this depression is about equal to that of one liter of tea or coffee. 32 MILLER, BERGEIM, REHFUSS AND HAWK INFLUENCE OF WATER, TEA, COFFEE AND COCOA ON THE EVACUATION TIME OF THE NORMAL HUMAN STOMACH The influence of the beverages studied on the evacuation time of the stomach is summarized in figure 34, which gives the volumes of the gastric contents at the time of complete removal (approximately 3 hours after ingestion). The evacuation of the stomach was not ap- preciably delayed ‘by the drinking during the meal of 1 liter of cold water; hot coffee, plain; hot coffee with cream; hot coffee with cream and sugar; cold tea, or hot tea; nor were there any distinct differences between the evacuation periods of any of these beverages. The ad- dition of sugar alone to coffee did, however, delay evacuation distinctly in both cases. Most striking, however, was the pronounced delay in evacuation caused by cocoa, this being noted in every case, even where half amounts only were given. For example, in one case the gastric contents at three and one-half hours measured, following the meal alone, 124 ec.; following hot coffee with cream and sugar, 110 ec.; and after cocoa, 465 ce. Temperature appeared to have little influence on evacuation: time, meals including cold tea leaving the stomach in the same time as those including equal volumes of hot tea; nor did cold water delay evacuation. INFLUENCE OF TEA, COFFEE AND COCOA ON THE PEPTIC RESPONSE Peptic activities were determined by the Mett method. In figure 35 are illustrated the values obtained after the uniform meals had been given with water, tea, coffee and cocoa respectively. The only strik- ing differences in the development of peptic activity noted were the low values found after the ingestion of coffee with sugar and especially after the ingestion of cocoa. These low values may be due to the de- pression of the secretory mechanism by the sugar and other substances present in cocoa as well as to retention of the ingested beverages. INFLUENCE OF TEA, COFFEE AND COCOA ON THE PULSE RATE The pulse rate was also followed in each case, although not as sys- tematically as would be necessary for exact comparisons of the effects of the different beverages upon the heart beat. Tea, and particularly coffee, did, however, bring about in liter quantities a pronounced ac- celeration of the heart beat in the first half-hour, one subject showing an GASTRIC RESPONSE TO TEA, COFFEE AND COCOA 33 increase to 150 and 160 beats per minute after hot tea and coffee re- spectively. There was some evidence that coffee brought about a more rapid and pronounced acceleration of the heart beat than the other beverages. From its slow evacuation, cocoa might be expected to produce a less rapid stimulation. An attempt (2) has been made to account for the more rapid action of coffee as compared with tea by showing that in coffee the alkaloid exists as caffeotannic acid soluble in cold water and not readily precipitated, while tea contains a caffeine tannate soluble in hot but not cold water and which is readily precipi- tated by acid (hence presumably by the gastric juice), going readily back in solution in an alkaline medium (such as that of the pancreatic juice). While coffee is usually about a 6 per cent decoction and tea about 14 per cent, it must be borne in mind that tea contains 3 to 4 per cent of caffeine and coffee about 1 per cent. Other systemic effects of tea, coffee and cocoa. Tea and coffee in the amounts taken gave rise in our subjects to nervousness, vasomotor relaxation, sweating, tremors, headaches, dizziness and sleeplessness, in some cases to a marked degree. Three of our subjects were unac- customed to the use of coffee, and the fourth never drank more than one cup at ameal. It is, therefore, probable that the symptoms noted were of a somewhat more aggravated character than would be found in the case of persons habitually drinking much tea or coffee. There is no question, however, that tea and coffee may have a marked effect upon the circulation, and that they are in no wise to be considered as beverages to be used in an unrestricted manner. Cocoa did not give rise to nervous or vasomotor symptoms to any- thing like the same extent as tea and coffee. This may in part have been due to slow absorption. In all cases there was a feeling of fulness after cocoa and a lack of hunger, which may be readily explained by the prolonged retention of this beverage in the stomach. DIURESIS AFTER TEA AND COFFEE DRINKING The urine secretion of our subjects was measured during the period immediately following the ingestion of the test meals with tea, coffee and cocoa. The volumes of urine eliminated are indicated by charts 36, 37 and 38. As much as 866 ce. of urine was excreted within an hour and a half after 1 liter of tea was given, and the values after coffee drinking were of the same order. It is clear, therefore, that these bev- erages left the stomach quickly, were rapidly absorbed, and the excess THE AMERICAN- JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 1 34 MILLER, BERGEIM, REHFUSS AND HAWK water soon eliminated. Coffee with sugar resulted in less rapid elim- ination, and the secretion after cocoa was low in volume, as might be expected from the delayed gastric evacuation. The specific gravities of the urines during the period of most rapid elimination varied from 1.004 to 1.001 or less, indicating a very dilute secretion. SUMMARY AND CONCLUSIONS A study was made of the influence of water, tea, coffee and cocoa upon the gastric digestion of a uniform mixed meal as measured by the acid responses and evacuation times. Evacuation of the stomach was not appreciably delayed by the drinking of 1 liter of cold water, cold or hot tea, hot coffee, either plain, with cream or with cream and sugar. The addition of sugar alone to coffee delayed evacuation. Cocoa in 1 liter quantities markedly delayed evacuation. To a less extent this was true of half-liter volumes. One liter quantities of water, hot or cold tea, hot coffee, plain or with cream, delayed somewhat the rise of the level of intragastric acidity as compared with the basal meal alone. As high acidities and normal evacuation were, however, attained these beverages must have stimu- lated gastric secretion, been rapidly evacuated, or more probably both. Coffee with sugar alone delayed the development of gastric acidity. Coffee with sugar and cream had less effect. Cocoa delayed distinctly the development of intragastric acidity. One liter quantities of tea and coffee gave rise to marked acceleration of the heart beat, to vasomotor relaxation, tremors and other nervous symptoms. Cocoa did not produce these effects but brought about a feeling of fulness at the stomach. Urine secretion during the first 90 minutes after tea or coffee inges- tion varied from 550 to 866 cc., after cocoa from 125 to 372 ce. The authors desire to thank the students who kindly served as the subjects of these tests. BIBLIOGRAPHY (1) Penzotpr: Deutsch. Arch. klin. Med., 1893, xiv, 535. (2) Lancet, 1911, i, 46; ii, 1573. ee. Wo HOH Fig. 1 Diet with /L. Cold ldtter, 15ST. 170.ce ‘removal. lL hour oy Si Case, 24-Co. Diet with ik. Hot €e, lain, 45> 50°C. [GEE TE 3 Case. 48-Ca Diet unth / kh. Hot Coffee, Su 250.ce. removed, Fia. 4 36 Case. 36-Co. Diet with €,Cream and Sugar Case. 44-Ca Diet with thy. Cold Tea,/3 /50.ce removal. hour Fia. 6 37 Case. 28-Ca Diet urth he Ho /52.CC. removed Case.. 32 -Ca Diet with 1k. Cocoa, 57°C. 465.cc. removal. aS ) ol Case.- 53 -G¢. PB: Q Duet Alone. > S Fic. 10 39 Case. /8-Cl Diet uth /k. Cold Water, 15" BS < iS) 8 8 Case, 25.-GL Diet with 1k. Hot Coffee, plain, 45*50T. 3.ce removal. lhour Lb S Case, 49-Gl. Diet with 1h. Hot Coffee,Cream. ec. Yo, HOH /0. C£. removal, Lhour : Si Fia. 13 Case. 37-G Diet with /2 hk, Hot Coffee, Pa) G) hour Fia. 14 41 ce Yo, HOH ee Mio KOH Case. 45-@Cl Diet with 1k, Cold Tea. 13 Fig. 15 Case,.29-Cl ; Diet with 1, Tea, hot S0°C 28, Lee: removal. Case,- 33-Gl. Diet with !k. Cocoa, 57°C. Case. 41-G! S iS S Diet with ‘2k Cocoa Hob. 2 A cS) S) ie, ST, (GE. remove l. /8.Ce. removal. Fia. 19 Case. /9-Fi Thet uy S S : ~ : = Case. 26 -/y i Diet with [L. Hot Coffee Alain, AS-50°C a2 CC removal / hour z. s Case. 50-Ri ; Diet with 1k, Hot Coffeg, 50.¢c. removal lL hour a Fig. 22 45 a|~ Case. 38-Ru Q Diet with / Cream and Sugar / hour y Re. Fie. 23 Case. 30-Fu Diet uith 1K. Tea hot, 50°C ce. Yo, HOH 6.cc removal. Case,. 34-Ri Diet uithih.Cocoa, 57°C 210.CC. removal Lhour Zz 3 S ie Case. 22-vo. S) Diet Alone. > S x 24.ce. removal. Case. 27-Jo. Je Diet with '2k Hot Coffee plan, 45*50C. 3.ce removed. Gl lhour zd x S 2 Case.. 5/-Vo. i Diet with lg kh. Hot Coffee,Cream. Suc rermova l. 3 Case-39-Vvo = Diet with 2k. Hot Coffee Cream and Sugar \Vece removal. Fig. 29 5 3X Case. 47-Jo “Le Duet with /2 kh. Cold Tea,/2C. Zac. removal. Fig. 30 49 THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 1 Case, 3/-JVo Diet with ‘zh, Hot Tea, 50 S = > 5 Ss S) S Fie. 31 = 2 Case. 43.-Vo. o) Duet with Ih Hot Cocca > © © Qcce. removal. Fie. 32 50 TC. _— XS, Diet with 2 kb. Cocoa, 5 /Aour Case. 35-Vo Fic. 33 00007 79 FH+ « CLEARED | Su FS °o Pile i 2090) 9° | + AA a I Vel POH Yr + » OBL PIAD V1 4-7 3 Hours. woons puo WOM aa op) 9 YR 4 y roxtmatel (ured) eaffoF 20,4 “I +. ‘uobag ‘eaffon 72 414 emovalat A wd aaffor 20,7 ay Ca, Gl. Fi. Jo t | 4AY7 PCD "F1 + + Comhlete _ (Set of fou) cts Key. aoe a ouony FOUL SOUg CUP OCR (COL ath) nn a Fie. 34 = Case,. Ft. Be Uniform Thiet with hey) ~ 49 MM eter ny ks) /Liter Water ae O26. | re frlatn Coffee. ' Pa DO. Vou Ao “ with Svgar—-— = 38 % + Ghoe ee eae ¥ ---— OS] ee ee i a el / «“ Qoecoa,. = _ SN iS g Fia. 35 600 Urine Secretion. Case. Gl. ce, Volume. Zo . 40 GO 80 /00 600 4003 Urine Secretion / eae Or Case,. Fre oy o| 9 o] & lol S SS) nS ol si Oo] Y TT £00 - - oer" -* een foo * / hiLautes: ete ig Fig. 37 Urine Secretion. Case,Ca y S 2 ~ § g I. RAPID BLOOD PLASMA PROTEIN DEPLETION AND THE CURVE OF REGENERATION H. P. SMITH, A. E. BELT anp G. H. WHIPPLE From the George Williams Hooper Foundation for Medical Research, University of California Medical School, San Francisco Received for publication February 25, 1920 The published work of Kerr, Hurwitz and Whipple (1) brings out several facts about the blood serum proteins which may be mentioned briefly before we go on to a consideration of the experiments given below. The stability of the serum protein concentration is truly re- markable and obviously of some importance to the body physiology. The normal level is quite constant and considerable deviations from this normal base line are not well tolerated by the body; in fact, pro- found shock may result. When the serum proteins are depleted or washed out by the technique employed the repair or regeneration of these proteins is a slow process requiring from 5 to 10 days, depend- ing upon the amount removed and other factors. It is as difficult to reconstruct these proteins as it is for the body to repair and replace liver cells following an extensive liver injury. It appears that the liver is especially concerned in the normal regeneration of blood serum proteins. Fasting does not prevent serum protein regeneration, therefore it is possible for the body to release these substances or to construct serum proteins from its own protein end products. There is no evidence that increased nitrogen breakdown is responsible for this regeneration of serum protein. The experiments of Kerr, Hurwitz and Whipple were different as to method of blood serum depletion when compared with the experi- ments given below. These earlier experiments were mostly done by interval bleedings followed in each instance by a return of the washed red corpuscles suspended in modified Locke’s solution. Under such cir- cumstances a dog was bled 100 to 200 ec. and after washing by centri- fugalization the same red cells were returned intravenously in Locke’s solution. This procedure was repeated several times during the day until the serum protein depletion was carried to a minimum figure. 54 REGENERATION OF PROTEINS OF BLOOD: PLASMA 55 This method has been used by Abel, Rowntree and Turner (2) and called “‘plasmapharesis.” It is obvious that this experimental pro- cedure introduced. wide fluctuations in blood volume and it was sus- pected that the shock which resulted was to be explained by the repeated hemorrhages and infusions. We shall use the term plasma depletion or plasmapharesis to indicate a removal of plasma proteins by means of repeated hemorrhage followed or accompanied by the replacement of like amounts of red cells sus- ’ pended in a protein-free fluid. The plasma depletion in our experi- ments was effected by a method first introduced by Morawitz (3); bleeding and the removal of whole blood was simultaneous with the replacement of the red cell Locke’s solution mixture The inflow and outflow volume was at all times constant and obviated any fluctua- tion in blood volume. All aseptic precautions were taken in manipu- lation, washing and final preparation of the red cell mixtures. The red cell mixtures were introduced at body temperature and the dog was kept warm during the experiment. The method employed in these experiments enables an investigator to reduce the blood plasma proteins from the normal level of 5 to 6 per cent to a very low level of 1.5 to 0.9 per cent. This can be done in a matter of minutes (2 to 10 minutes) leaving the animal uninfluenced by the large and numerous fluctuations in blood volume and oxygen-carry- ing capacity of the blood which undoubtedly occur in the method used by Kerr, Hurwitz and Whipple. In addition it facilitates observa- tions of that portion of the curve of protein regeneration immediately following a large single depletion and permits observations on an uninterrupted regeneration curve. METHOD The animals used were sound young dogs maintained on a mixed diet. In most cases no food was given the animal for a period of 12 hours pre- ceding the experiment. Free access to water obtained. Under com- plete ether anesthesia and with aseptic precautions an incision was made either into the region of the femoral vessels or of the large vessels of the neck. The artery and vein were exposed and clamped. Into each was introduced a vaseline-coated cannula pointing toward the heart. Plasma removal was effected by withdrawing through the cannula placed in the artery large quantities of blood. This blood was allowed to flow into a graduated bottle. Simultaneously, a suspension 56 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE of washed corpuscles warmed to body temperature was injected under pressure through the venous cannula. This suspension was delivered from a flask which was also graduated. The graduations were used to permit a comparison of the inflow with outflow to be made at any time during the exchange. In this way inflow and outflow were ob- served and kept equal at all times. In order to maintain an even sus- pension of the injection mass, the latter was frequently shaken. The fluid injected consisted of washed dog corpuscles suspended in a modified Locke’s solution in the ratio of three parts packed corpuscles to two parts by volume of the saline mixture. The composition of the Locke’s solution was: sodium chloride, 0.9 per cent; potassium chloride, 0.042 per cent; sodium bicarbonate, 0.02 per cent. The corpuscles in all cases were obtained from the blood of healthy dogs. This blood was drawn into sodium oxalate, centrifugalized, and the plasma removed from the sedimented corpuscles. The corpuscles were then washed twice in the modified Locke’s solution by resuspension and centrifugali- zation. Aseptic precautions were observed in all these manipulations. The exchange was effected in a period ranging in different animals from 2 to 25 minutes. The amount of fluid withdrawn varied in indi- vidual experiments from 60 per cent to 195 per cent of the animal’s blood volume. The amount of blood simultaneously injected corre- sponded within a few cubic centimeters to the amount withdrawn. The blood volume was estimated as 10 ce. per 100 grams of body weight. In a number of experiments the actual blood volume was kindly deter- mined for us by Dr. C. W. Hooper, using a dye method recently de- scribed. This paper (4) shows that the blood volume of active normal dogs as determined by the dye method is approximately 10 ce. per 100 grams of body weight. At the end of the operative procedure the cannulae were withdrawn and the vessels ligated. Vaseline was applied liberally to the wound. In a few instances it was necessary to make a single suture through the subcutaneous tissues at the site of operation. In practically every experiment the wound healed quickly with little or no suppuration. Samples of blood were collected through the arterial cannula at the beginning and at the end of the exchange and again 15 minutes later. Subsequent Samples were taken from the jugular vein by means of a needle and syringe. On each of these occasions two samples were with- drawn. One of these was drawn into a 15 ec. hematocrit tube contain- ing 3 cc. of 1 per cent sodium oxalate. The other sample was drawn into a plain heavy-walled glass test tube and allowed to clot. Both REGENERATION OF PROTEINS OF BLOOD PLASMA 57 of these samples were then centrifugalized at a high rate of speed. The serum from the clotted sample was then used for estimation of serum albumins, serum globulins and the non-protein fraction by the refractometric method of Robertson (5). Percentage corpuscles read- ings were made from the oxalated sample, correction being made for the amount of oxalate solution present. The plasma from this sample was also used for the determination of fibrin. This was carried out according to the method of Cullen and Van Slyke (6). This method consists in diluting 5 cc. of plasma in 100 ce. of salt solution. To this mixture 1.5 cc. of a 2.4 per cent solution of CaCl, was added to supply calcium and to promote clotting, and a Kjeldahl done on the mass of fibrin obtained. The clinical condition of the animal was closely observed. The rectal temperature, rate of respiration, pulse, diarrhea and vomitus, as well as the general appearance of the animal, were noted. In cases in which death resulted, careful autopsies were performed at once. The clinical condition of the animal will be made the subject of a sub- sequent paper with a discussion of the peculiar type of shock which may develop under these conditions. EXPERIMENTAL OBSERVATIONS This paper in general deals with the recovery experiments but in certain tables we include many of the lethal shock experiments (tables 12, 13 and 14). For the sake of comparison we give in table 1 a type experiment which was followed promptly by fatal shock. Many of these experiments will be found in the next paper of this series and in that place the general discussion of this peculiar shock will be pre- sented. It will be noted in table 1 that the reduction of total proteins, albumin and globulin is pretty uniform and is a fall to approximately one-third of normal. The emergency increase of protein during the 15 minutes following the plasma depletion is not as marked as usual (see table 2). There is no further increment of serum protein in the hour following this 15-minute sample and this may be explained in part by the profound shock. The content of red cells in whole blood as shown by the hematocrit is lower than usual and the fall which appears imme- diately after plasmapharesis would indicate the use of a red cell mix- ture containing fewer red cells than intended. This factor does not complicate the remaining experiments and we believe has no signifi- cance. 58 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE Experiment 104. (See table 1). 122 per cent exchange. p Dog 18-48. Young female bull dog. Weight 16.9 pounds. Blood volume on November 20, 1917 (by dye method) was 941 cc. November 20, 1917. Under ether anesthesia 940 ce. blood were withdrawn from the left carotid artery. Simultaneously 1000 ec. of blood corpuscle suspension were injected into the left external jugular vein. The duration of the exchange was 10.5 minutes. Animal showed almost immediately a great fall in pulse pressure and arterial tension. Profound depression with forced irregular respira- tion developed in about an hour. Death 2 hours after the exchange. The autopsy findings are uniform in all fatal experiments and will be described in detail in the following paper. TABLE 1 122 per cent blood volume exchange; dog 18-48; experiment 104 BLOOD SERUM READINGS IN PER CENT FIBRIN. | 2EMATO- TIME (oe. - eee eee IN aoe eee REMARKS Total | Albu- | Globu-| Non- | ?©® CENT) per cent protein| min lin | protein Before exchange..... Neth || Zara OL Gi |) P40) | ves 47 Immediately after...) 1.9 | 1.4] 0.5] 1.6] 0.12 35 Considerable hemolysis 15 minutes after..... Aa \| IUDel| O.Z5 |) Use || Osi 38 Considerable hemolysis 1 hour and 20 min- WLEBechiciier eee tale sale OM Oya eel 30 Fatal shock Table 2 gives the results of an experiment which contrasts with this lethal shock experiment (table 1). The second experiment presents an even large blood exchange in plasmapharesis but this dog is not dis- turbed by the procedure. It will be noted that there are marked indi- vidual differences in dogs as to their tolerance to this plasma depletion. Any given dog will show a considerable uniformity of reaction to a unit exchange but’ must be standardized to ascertain this reaction. Table 2 also shows a fall of total protein, albumin and globulin to about one-third normal following the plasmapharesis. We wish to call atten- tion to the emergency increase in blood proteins which appears within 15 minutes. This is a characteristic reaction which obtains in practi- cally all experiments (tables 11 and 12). The increase in serum pro- teins during the next 24 hours is very marked and exceeds 1 per cent protein—equivalent to more than 20 per cent of the total protein re- placed in the blood serum. Subsequent regeneration of protein in the serum is slow and requires several days for complete recovery. REGENERATION OF PROTEINS OF BLOOD PLASMA 59 Experiment 69. (See table 2). 170 per cent exchange. Dog 18-9. Young female bull dog. Weight 14 pounds. Blood volume on July 19, 1917 (by dye method) was 772 ce. August 2, 1917. Under ether anesthesia 1081 ec. blood were withdrawn from the right femoral artery. Simultaneously 1081 cc. of blood corpuscle suspen- sion were injected into the right femoral vein. The duration of exchange was 12 minutes. Following the exchange the temperature fell about 2 degrees, but returned to the original level within 2 hours. Pulse and respiration were fair at all times. A much smaller volume exchange is shown in table 3, yet consider- able shock resulted. It will be seen that the level of total proteins, albumin and globulin falls to approximately one-half of normal cor- responding to the smaller exchange volume (90 per cent). There is a TABLE 2 170 per cent blood volume exchange; dog 18-9; experiment 69 BLOOD SERUM READINGS IN PER CENT TIME pS |. SS a) TT a REMARKS pretzin | Albumin | Globulin | NOR, Before exchange....... 5.6 3.8 1.8 iL 2 Immediately after..... 2.0 1.3 0.7 1.4 | Moderate hemolysis 15 minutes after....... 2.9 2.0 0.9 1.3. | Moderate hemolysis. No shock 2 Ee ee 4D |) SaeE Bi 1.9 | Slight hemolysis. Normal eee os ste's « 4.5 2.5 2.0 ast namo 98.) =. cis. 4.8 3.2 1.6 Ley moderate increase in serum proteins during the 15-minute period and less than usual during the first 24 hours following plasmapharesis. We cannot explain satisfactorily the remarkable drop in red cell hemato- crit which is present after 11 hours and persists many days. Possibly the red cells used for infusion in this experiment had been seriously injured and went to pieces in the circulation. In confirmation of this suggestion we note the presence of hemolysis in blood samples taken on the first four days following the experiment. It may be suspected that an hemolysin was present in this dog’s blood but there is reason- able doubt whether hemolysins actually do occur in the dog in sufficient amount to destroy large numbers of homologous red cells. Experiment 103. (See table 3). 91 per cent exchange. Dog 18-66. Young female bull mongrel. Weight 17 pounds. 60 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE November 15, 1919. Under ether anesthesia 700 cc. blood were withdrawn from the right femoral artery. Simultaneously 700 ec. of blood corpuscle sus- pension were injected into the right femoral vein. The duration of exchange was 9 minutes. Following the exchange the temperature showed little or no alteration from the original level. The pulse was regular but poor in tension for a number of hours. One-half cubic centimeter of adrenalin subcutaneously was given 4 hours after the operation. After 24 hours the animal was in excel- lent condition. In all these experiments the washing out of plasma proteins is accom- plished by a rapid exchange. The bleeding and simultaneous infusion of the red cell mixture occupies only a few minutes, the limits being TABLE 3 91 per cent blood volume exchange; dog 18-66; experiment 103 BLOOD SERUM READINGS IN PER CENT FIBRIN | HEMATO- TIME IN etc REMARKS Total | Albu- | Globu-| Non- |??? ©PNT| per cent protein min lin | protein Before exchange..... Ge2ialeenon |) wer, 2 NOR Ona? 49 Immediately after...J 3.2] 2.2] 1.0] 1.5 49 15 minutes after..... 3.8} 2.9 | 0.9) 1.6) 0.21 56 SMTVOUES Men sere, setters Ab Sy || Se alsey ike? |) OLA 50 Moderate shock el ANOUTSH reece ae oe AS On 259m et 253 se OAT 33 DOUNGAY Aone ie caer Ara | 268))) 2255) = 2267/0848 35 Dog nomal. Hemolysis 3d day 4284 oh | LL | 225 1 Onay 29 Hemolysis , BUENA Meee eer. Ae) del) 2.4 | $350) Osan 27 Hemolysis Gthivdaye. coraaes ieee ROM OnO ale lea ai tS ilaeC EV Bowdon ooo Ga HEP essere |e UR tty | (0). 745s 32 MO aye teenage s | on tle eee, | eels. 2. OR easel PoGhvday es, eee |) See oo || 1.5) 2.1) 0.56 | 32 7 2 to 25 minutes. Within these limits the speed of exchange, whether 2 minutes or 25 minutes, seems to make little difference. To make this point clear we may contrast tables 4 and 5. The first of these two ex- periments done on the same animal (table 4) shows the reaction follow- ing an exchange of 100 per cent done in 14 minutes. There was definite shock but a rapid recovery. The second experiment done on this dog after an interval of 2 weeks to insure complete recovery, shows the reaction following a very rapid 100 per cent exchange which was com- pleted within 2 minutes. There was if anything less shock on this occa- sion than after the first exchange. It is interesting to note how closely REGENERATION OF PROTEINS OF BLOOD PLASMA 61 the curves of total protein, albumin and globulin in the two experi- ments coincide. The prompt rise in the 15-minute interval is identical and the initial fall corresponds to the other experiments discussed. Experiment 93. (See table 4). 99 per cent exchange. Dog 18-48. Young female bull dog. Weight 13.3 pounds. October 4, 1917. Under ether anesthesia 600 cc. blood were withdrawn from the right femoral artery. Simultaneously 700 cc. of blood corpuscle suspension TABLE 4 99 per cent blood volume exchange; dog 18-48; experiment 93 BLOOD SERUM READINGS IN PER CENT Fiprin | 2EMATO- TIME 2) ee EEE sae eal REMARES Total | Albu- | Globu-| Non- |?2® C®NT| per cent protein} min ‘lin’ | protein Before exchange..... AVON BB 11S} 224) 0.26 58 Immediately after... 2.0] 1.3] 0.7] 1.6] 0.16 58 15 minutes after..... Beer 2:6 | Ontelo dud 1 0.15 58 Definite shock OUAHOUES sts cc ce cj s 2 S.9\) 3.071 O28) Tos) Ora 46 2 Ch ae 0.18 54 Normal 2 i a ARO Vy 3.1 |F 0.9" |) 18S) | Or30 42 PUMEEY Sects ied one AEA Nise 42 dalts|| pele OkS9. TABLE 5 104 per cent blood volume exchange; dog 18-48; experiment 96 BLOOD SERUM READINGS IN PER CENT TIME SS ES ee Le ae eae ae REMARES Total | ajpumin | Globulin.| _No2- protein protein Before exchange....... 6.4 3.5 2.9 ier Immediately after..... Peis | Is 0.8 1.3 15 minutes after....... 3.2 2.2 1.0 1.6 Slight shock GUS: Ateapee RE Ac) 3.9 Ps nirh 1.2 1.6 PROG ate k sciiie)e cies al) i400 2.6 1.9 1:9 Normal were injected into the right femoral vein. The duration of exchange was 14 minutes. Animal showed definite signs of intoxication after about an hour following the exchange, with some bloody feces after about 53 hours. After 24 hours animal appeared to have recovered completely. Experiment 96. (See table 5). 104 per cent exchange. Dog 18-48. Young female bull dog. Weight 13.8 pounds. ‘‘Plasmapha- resis,’ 99 per cent exchange with a duration of 14 minutes, done on October 4 (exper. 93, table 4). Showed definite signs of intoxication. 62 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE October 18, 1917. Under ether anesthesia 650 cc. blood were withdrawn from the left femoral artery. Simultaneously 650 ec. of blood corpuscle suspension were injected into the left femoral vein. The duration of exchange was 2 minutes. Immediately following the exchange there was a slight transient fall in tem- perature. There was no immediate alteration of the pulse; however after about an hour the pulse was of poor volume and the animal appeared decidedly dull. Bloody feces were noted at this time. After 24 hours the animal appeared to be in good condition. Table 6 gives valuable data concerning the speed of exchange in its relation to shock and the curve of protein regeneration. In this ex- periment the blood volume exchange of 75 per cent was completed in 2.5 minutes. There was no shock and we may compare the previous exchanges done on this same animal (Sept. 12, 1917, exper. 84, 80 per TABLE 6 75 per cent blood volume exchange; dog 18-35; experiment 94 BLOOD SERUM READINGS IN PER CENT FIBRIN | ZEMATO= TIME see IN tee REMARKS Total | Albu- | Globu-| Non- | ?®® C®NT| per cenr protein} min lin | protein Before exchange....... Be 4a i Se" Onl ele On a ORzi! 40 Immediately after..... 320) 262, | 058.) 9 a5 OR19 48 15 minutes after....... Bal lhe ack 1.0 13" | 0223 54 Sr INU conensboacense ANG Wards \eelesa| Sa eOs26 61 No shock (OF NUM apagen oss anel “iste lh cise aaj 128 || O80) 44 UP elMOUbeoeeAaevonesooll Coll Boor Welk ass Normal Ded Avanos: sete Cee 4.9| 3.6 1.3 1.9 | 0.44 Sond aye en ten. Sect 4.4} 3.1 1.3 2.8 AT MEO S © Aciaci: 3 Ree AOR |e cece al ee2e2u | ae: cent exchange, shock very slight, time 7 minutes; October 3, 1917, exper. 92, 74 per cent exchange, no shock, time 14 minutes). The curve of serum protein depletion and regeneration is similar to other experiments. The small per cent exchange lowers the total protein to 3.0 per cent and the emergency increase is definite within 15 min- utes, giving a rise to 3.7 per cent. It appears that the emergency re- action by which a considerable amount of serum protein is thrown into the blood stream, can be called out by a large or small exchange using this method. Experiment 94. (See table 6). 75 per cent exchange. Dog 18-35. Young female bull dog. Weight 16 pounds. ‘‘Plasmapharesis,”’ 80 per cent exchange, done in 7 minutes on September 12, 1917, showing very REGENERATION OF PROTEINS OF BLOOD PLASMA 63 slight shock (exper. 84); another 14 per cent exchange done in 14 minutes on October 3, 1917, showing no shock (exper. 92). Blood volume on October 9, 1917 (by dye method) was 671 cc. October 11, 1917. Under ether anesthesia 545 cc. blood were withdrawn from the right carotid artery. Simultaneously 545 ce. of blood corpuscle suspension were injected into the right external jugular vein. The duration of the exchange was 24 minutes. Animal showed a transient drop in temperature of 3 degrees mmediately following the exchange. Animal showed little or no signs of ntoxication. Table 7 shows a remarkably prompt return to normal after a large exchange (109 per cent). The total proteins fell to a level of 50 per cent normal which is a normal reaction. We cannot explain the fig- ures, which appear to show a peculiar reaction on the part of the albu- min and globulin fractions. These peculiar reactions will appear at . TABLE 7 109 per cent blood volume exchange; dog 18-20; experiment 89 BLOOD SERUM READINGS IN PER CENT. FIBRIN BEMATO- TIME a we te ee IN ae REMARKS Total | Albu- | Globu-| Nons | P®® C®NT | per centr protein} min lin | protein Before exchange..... 5.07 3.4) 1.4.0 238) Gas 50 Immediately after...| 2.5 | 0.9| 1.6] 1.4] 0.12 60 15 minutes after..... SuOs) 1.5} hk be |) Ga: Os a1 73 Lo i 42) B20 20 Deep as 73 Very © slight shock Mea oe oh. cs. . SO. Sealy|) 2S 2 le Beal 41 Normal times in spite of every care used in the method, but we are inclined to suspect technical errors as in part responsible. The return of the total protein to normal within 24 hours is unusual and would indicate an unusually large emergency reserve. The peculiar rise in red cell hematocrit will be found in this experiment and in a few subsequent experiments. That it appears in the 15-minute and 4-hour samples but not in the sample taken immediately after the exchange is very perplexing. There is no severe shock to account for any withdrawal of fluid from the blood. We have no convincing explanation to offer. Experiment 89. (See table 7). 109 per cent exchange. Dog 18-20. Young female bull dog. Weight 16.1 pounds. September 20, 1917. Under ether anesthesia 800 ce. blood were withdrawn from the femoral artery. Simultaneously 800 ce. of blood corpuscle suspension were injected into the femoral vein. The duration of the exchange was 14} 64 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE minutes. Animal showed little or no depression. The temperature fell 3 de- grees, however, and the animal shivered considerably for 8 or 9 hours. In good condition after 24 hours. Table 8 shows another typical experiment giving the usual curve of blood proteins following plasmapharesis of moderate amount (90 per cent). The fibrin curve is given in this experiment and we believe this illustrates the usual reaction on the part of this plasma globulin. The method used gives certain opportunities of error when small amounts of plasma are analyzed for fibrin. More work in this field has been completed by Mr. Foster in this laboratory and will soon be published. We do not wish to put too much emphasis on these figures. TABLE 8 90 per cent blood volume exchange; dog 18-68; experiment 106 BLOOD SERUM READINGS IN PER CENT FIBRIN TIME we A ee ee IN REMARKS Total | Albu- | Globu-| Non- | ?2* C2NT protein} min lin | protein Before exchange............ Gay | EON) sal sisy || ea) 2 Immediately after.......... Qe Sa) 2G Oa Sale Oe22 1 baminiutestattere sere Ste) PTW (Er jf alotes ||) 247 22 NOUTBN. 2 tapi each ee 4.6| 3.3] 1.3] 1.8] 0.30 | Moderate shock PBN OUTS Hae aa cre ene 0.19 2d day. 4.8 | 3.8 | 1.0} 2:0 | 0:56) | eNormat 4th day ANO7| 34) Ison lle oa eos Gthidayiacrs nore RA Bee aloe || 22ail |, W428! Experiment 105. (See table 8). 90 per cent exchange. Dog 18-68. Young female bull dog. Weight 15.3 pounds. November 21, 1917. Under ether anesthesia 623 cc. of blood were withdrawn from the right femoral artery. Simultaneously 623 cc. of blood corpuscle sus- pension were injected into the right femoral vein. The duration of the exchange was 9 minutes. The animal showed moderate depression and slight decrease in pulse volume for several hours. Tables 9 and 10 give figures to show the low level of serum pro- teins which may be effected by very large blood volume exchanges (159 and 195 per cent). The usual normal dog will not tolerate such large exchanges without exhibiting profound and often fatal shock. These two dogs were unusually resistant to this experimental procedure and give us the opportunity to study the reaction following such large REGENERATION OF PROTEINS OF BLOOD PLASMA 65 exchanges uncomplicated by shock or notable hemolysis. The low level of total proteins is to be expected and one experiment (table 10) reaches the minimum figure for total protein (0.9 per cent). We have no observation in any of our experiments to show a lower level of pro- tein in the blood stream. The protein regeneration is very rapid in the 15-minute period as well as in the following 24 hours, indicating con- siderable emergency reserve material. TABLE 9 159 per cent blood volume exchange; dog 17-215; experiment 67 BLOOD SERUM READINGS IN PER CENT TIME e- ae i, Se eee ee REMARKS protein | Albumin | Globulin | rotein Before exchange....... 6.2 Ba) 2.6 Wee Immediately after..... 1.3 0.3 1.0 1.6 | Slight hemolysis 15 minutes after....... 2.2 1.2 1.0 1.5 | Slight hemolysis. Slight shock S006 A 4.1 Pint f 1.4 1.4 | Normal oh 4.3 2.1 2.2 2.0 LS GS OS Ce eee 5.2 2.9 2.3 1g TABLE 10 195 per cent blood volume exchange; dog 17-232; experiment 70 BLOOD SERUM READINGS IN PER CENT TIME ae 7) 1 ih 2 MS a. a es REMARKS sien Albumin | Globulin Palen Before exchange...... 6.3 3.7 1.6 1.9 Immediately after....| 0.9 0.3 0.6 1.5 15 minutes after..... : 1.4 0.7 0.7 lave Very slight shock LO ee 4.3 2.3 2.0 165 Normal BARANY isuciiewi ss awe sss 5.5 3.8 LZ 1.9 Experiment 67. (See table 9). 159 per cent exchange. Dog 17-215. Adult female fox terrier. Weight 15.25 pounds. Blood volume on July 19, 1917 (by dye method) was 858 cs. July 31, 1917. Under ether anesthesia 1105 cc. of blood were withdrawn from the right femoral artery. Simultaneously 1105 ce. of blood corpuscle sus- pension were injected into the right femoral vein. The duration of the exchange was 10 minutes. There was slight decrease in force of pulse beat for about an hour. Animal showed little sign of depression thereafter. Experiment 70. (See table 10). 195 per cent exchange. Dog 17-232. Young female coach dog. Weight 13.5 pounds. Blood volume on July 19, 1917 (by dye method) was 714 cc. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, No. 1 66 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE August 3, 1917. Under ether anesthesia 1200 cc. blood were withdrawn from the right femoral artery. Simultaneously 1200 cc. of blood corpuscle suspen- sion were injected into the right femoral vein. The duration of the exchange was 10 minutes. The pulse pressure was poor for about an hour following the exchange. TABLE 11 5 Zz PER CENT : e TOTAL PROTEIN 3 Ae ee mes ea ae eee z - eT El; ils les] 2 (812) 8] 2 | 2 eee 3 2 ® 3~o | 5§ ie A “2 ° | = 3H | 58 = a | ee | Sa Sa Be oS | 8S eee z Ay & le fee = a = a Z = a | 8% = 2/8 | & | Seveodige | 2 | 8 | 8°68 | | eee 67) | *60"|'5.3-)2esH 85 Cesta es ea 0 TAN 920) 55M. 7-188 1.3) | SOrL 0 50 | 58] 58 TADUSRT- | BI2NGF.3.,| S29 29)| BOG 4+-+.-1| 58 [758 ees 75. |'rO4:|$524103.0°) 3-7. | 4990/1245 ONT ee orl eae 40 | 54] 54 75 | 90|5.312.9|3.4]4.3|2.4] 0.511:4| 44+ | 48) 54e esas 80 | 84) 5.6/3.1 |3.116.6/2.5] 0.0) 3.54° + 50 | 63| 58| 33 80 | 82|5.6| 2.4 4.4 | 3.2 2:0"\'at0 50 44 84 | 74|4.9 | 2.6 2.3 * 89 | 101 | 5.5 | 2.1 | 2.4 SoA MAO 2 +--+] 2605) ae ae 90 | 105 | 5.5|2.8|3.5|4.8]2.7] 0.7/2.0] ++] 48] 49] 52 91 | 103 |-6.2:+} 3.2] 3:81 4.3 | 3.0 | 0.6) 4:1) + |- 499) 405 aoe 94. | 62.) 4:9.] 1.8 |'228") 428") 321 2 a0] Stone 96:41 6105-5 |) 227-1 3.6 2 $e (9) 0 98° |- 64.1 6.3 | 3:01 2:8 ye bl ie + OOS Si. | 4.5) 12 Se | e813 eae 90° |\.°98.| 4512-05123 3 51 4208 25. les 2a eee 58 | 58| 58 | 42 100 | 100-| 5.6 | 3.0 |-2.9 | 4.6 | 2.6 |=—0:1] 1.6 | 7 46 | 62| 64] 36 Averages! 5.4 | 2.7| 3.3] 4.8 | 2.7] 0.6] 2.1 50| 55 | 56| 38 - Shock readings: + means slight shock; +-++ means moderate to severe shock; +++ means lethal shock. ; * Death from overdose of ether. A summary of certain factors in many plasma depletion experiments will be found in tables 11 and 12 and the average figures give much in- teresting information. The averages of the experiments which show 100 per cent or less of blood volume exchange (table 11) show an iden- tical emergency increase in the blood serum proteins. The two tables are practically in accord and we note that the average replacement of serum protein during the 15 minutes following the plasmapharesis REGENERATION OF PROTEINS OF BLOOD PLASMA 67 amounts to 0.5 to 0.7 per cent protein—which is an increase of 10 to 14 per cent of the total proteins. The increase during the 24 hours fol- lowing the plasma depletion is considerable and amounts to 2.0 per cent protein which is an increase of 40 per cent total protein, figuring 5.0 protein per cent as the normal for a healthy dog. Further analysis of the blood cell hematocrit figures is of interest. It is unfortunate that we did not obtain hematocrit readings in all our experiments. It is clear that the normal hematocrit before the ex- TABLE 12 me | Zz PER CENT i : TOTAL PROTEIN : Spon ar ee ce Se g 3 ® ' = a oOo ' ue i 2 tear 4 2 8 ea rem is Prelils lislehelflalelalecliels \ g g g a a S 2 Piieebice (2a eeics log, |e oleoled (oe heehee lee ee a | eae Pye SB OO See erie rae |e 5 Baas ee eS eR eS) OE. og B q? ie pe & 102 | 80|5.6| 2.6/2.8 3.0! 0.2 soon i) og 6.41 2.1| 3.21 4.5 | 43) aal 2.4) + bebe lees 38 108 | 98 | 6.3 | 2.2 | 2.7 4 Ons +44] 33| 44] 53 108 beso 5.0 |} 2.5 |:3.0|5.1}2.5 | 0.5|-2.61 + | 50} 60|..73 |.41 110 |. 83} 5.8 | 2.3 | 2.8|4.2|3.5|] 0.5] 1.9) + 46 118 | 63| 7.1 | 4.2 4.8 | 2.9 ere Bee 122 | 104 | 5.71 1.9] 2.3 3.8| 0.4 beter 131 | 66|5.7/1.0|3.1| 4.0/4.1] 1.5) 2.44 +4 | 47] 35] 38| 30 141 | 68|5.5|2.6/2.5!3.9|2.9/-0.1] 1.3] 0 159 | 67|6.2/1.312.2| 4.1/4.9! 0.9] 2.81 + 170 | 69|5.6| 2.0} 2.9| 4.2|3.6|] 0.9] 2.2) +4 191 | 77|5.9| 0.9/1.2 5.0.) 0.8 Heeb de 195 | 70/1 5.3/0.9) 1.4/4.3] 4.4| 0.5] 3.4] 0 Averages! 6.8 | 2.0 | 2.5 '1'4.34.3.81.0.6) 2.1 46| 51 | 55| 39 Shock readings: + means slight shock; ++ means moderate to severe shock; +-+-+ means lethal shock. periment is approximately 50, which is an indication that healthy dogs were used. There is a slight increase in the hematocrit figures at the end of the blood exchange but only to 55 per cent. This as- sures us that a suitable number of red cells was introduced in the red cell Locke’s solution mixture which replaced the blood. There is a trifling increase in the average hematocrit figures for the 15-minute sample but only 10 56 per cent, which may indicate a very slight blood concentration due to loss of Locke’s solution from the circulation. 68 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE There is a distinct fall in hematocrit during the 24 hours following the plasmapharesis—an average of 7 to 12 per cent below the initial figure. This probably indicates a true loss of red cells as we must recall the fact that these red corpuscles which are introduced have TABLE 13 PER CENT a HOS ENGNG ES AC : AE GAERO (PER CENT ELSE RICRETE) oO = i ao 3 “= ele | eVee| 2 | ole | 2 | & logo jogs) 22 elm | a | ctl Heth Gaal Aietmnl pacae Htee eB | S | oe | Sopa a le im |< | 8 | & |e | & || eee 74 | 92] 0.21] 0.16] 0.15] 0.28] 0.05/—0.01| 0.12) 0 50 | 58 | 58 74 | 87| 0.18] 0.08] 0.15 0.10) 0.07 +++] 58] 58] 55 75 | 94| 0.21] 0.19] 0.23] 0.44) 0.02} 0.03] 0.25] 0 40| 48 | 54 75 | 90| 0.25] 0.15] 0.20] 0.35] 0.10} 0.05) 0.20) ++ | 48] 54] 55] 38 80 | 84] 0.17] 0.06] 0.03] 0.40) 0.11/—0.03/ 0.34, + | 50| 63] 58] 33 80 | 82] 0.17] 0.08] 0.13] 0.33] 0.09} 0.05] 0.25) 0 50 44 84 | 74| 0.36] 0.15 0.21 * 89 | 101) 0.24] 0.11] 0.11 0.13) 0.00 +++] 60] 41] 53 90 | 105| 0.42] 0.22] 0.27| 0.56] 0.20} 0.05) 0.34) ++ | 43 | 49] 52 91 | 103] 0.42] 0.19] 0.21] 0.48] 0.23} 0.02} 0.29} ++ | 49| 49] 56] 35 99 | 93] 0.26] 0.16] 0.15] 0.30] 0.10/—0.01/ 0.14, + | 58 | 58] 58] 42 99 | 81] 0.32 0.19 sae 100 | 100] 0.25] 0.16] 0.16] 0.45] 0.09} 0.00/ 0.29} + | 46] 62] 64] 36 102 | 80] 0.42 0.26 =pSR=e 104 | 96] 0.50] 0.13] 0.14] 0.40) 0.37] 0.01] 0.27} + | 52] 66 38 108 | 98] 0.25] 0.12] 0.11 0.13|—0.01 +++] 33] 44] 53 109 | 89] 0.35} 0.12) 0.11 0.23/—0.01 + | 50) GO) -7aeee 110 | 83] 0.75] 0.18] 0.30] 0.44] 0.57) 0.12) 0.26} + 46 122 | 104 0.50] 0.13] 0.14] 0.40] 0.37) 0.01] 0.27|4+++] 52] 66 38 191 | 77| 0.44] 0.13] 0.10 0.31/—0.03 SESEEE Aver- ages. 0.33) 0.14) 0.17| 0.40) 0.19) 0.02] 0.25 49 | 55 |. 57 | 39 Shock readings: + means slight shock; ++ means moderate to severe shock; +++ means lethal shock. * Death from overdose of ether. been submitted to considerable manipulation in the necessary washing previous to the injection. Dogs’ corpuscles too are notoriously fragile. We may assume for the present at any rate that many of the red cells which were introduced had been seriously injured and went to REGENERATION OF PROTEINS OF BLOOD PLASMA 69 pieces in the blood stream during the 24 hours following the blood exchange. We can review the fibrin analyses in table 13 and at once a decided difference appears when we compare the serum protein curve with that of the plasma globulin, fibrinogen. The exchange of blood re- duces the fibrin content to about the same level—that is, we can wash out the same percentage of fibrinogen by the usual plasmapharesis as we do in the case of the serum proteins. The fibrin content is reduced to a little less than one-half normal—from 0.33 to 0.14. During the 15 minutes following the plasma depletion there is no emergency reac- tion on the part of the fibrin as is so constant for the serum proteins. During the next 24 hours the fibrin is restored completely to normal. This may mean that there is no emergency reserve of the fibrin as it can be produced so rapidly in the body in any emergency. We know of many other facts which point to complete dissociation of fibrin and other blood proteins as to production and repair and general usefulness in the body economy. DISCUSSION A theoretical consideration of the factors involved in this protein replacement is difficult at this time. It may be claimed that this in- crease represents, in part at least, not a true increase in the quantity of circulating serum protein, but is the result of the escape from the cir- culation of fluid poor in protein material. However, if any consider- able escape of fluid from the circulation were to occur, one would expect to note a rise in the percentage of red cells in the circulating medium. That no sufficient change in the cell-plasma ratio does occur can be seen from the hematocrit figures given in tables 11 and 12. On the other hand, it may be that the figures represent an actual influx of pro- tein into the circulating medium. Such an influx could conceivably come from some tissue or organ which serves as a storehouse for this type of protein material. Seitz (7) thinks that the liver acts as such a storehouse. Earlier work in this laboratory by Kerr, Hurwitz and Whipple (1) shows a lack of reserve production of serum proteins after plasmapharesis in the Eck fistula dog. This indicates that liver insufficiency may impair the normal emergency reproduction of blood proteins. Of particular interest is the rather remarkable increase in the 15 minutes immediately following the end of the experimental depletion. 70 H. P. SMITH, A. E. BELT AND G. H. WHIPPLE This, if blood volume changes be excluded, appears to be truly a throwing in of ready-formed materials. While the depletion curve of the fibrin fraction of the plasma pro- teins brought about by our experimental procedure compares closely with the curve of depletion of the serum proteins, still a distinctly dif- ferent type of curve of fibrin repletion is revealed. A fairly typical experiment is presented in table 8. This point may also be studied by an examination of the results given in table 13. In these tables it may be seen that the rapid rise immediately following the procedure of depletion which is typical of the serum proteins is absent or at least negligible in the case of fibrin. However, the body seems to be able to supply large amounts of this protein in a space of 24 hours, for as the summary in table 13 shows, the fibrin on the day following the ex- change is already as high as the original figure, or, as occurs in some cases, even higher. When such an over-production does occur the level usually returns to normal in one to two days. This lack of correspondence between the regeneration figures for serum proteins and for fibrin protein in the period of initial regenera- tion, we believe furnishes additional evidence against a theory which would account for all changes in protein concentration in this period by the loss from the circulation of fluids poor in protein. For, in such a case, the concentration of the proteins might be expected to occur to practically the same degree in each. That this does not occur tends to strengthen the evidence given by the hematocrit figures. It may be pointed out that the curve of serum protein regeneration is very different for this type of experiment when compared to the experiments of Kerr, Hurwitz and Whipple. We believe that these differences are to be explained wholly by the differences in the experi- mental depletion of the serum protein. Kerr, Hurwitz and Whipple used interval depletions of smaller amounts but repeated many times during a single day. In this manner they undoubtedly removed much of the large emergency reserve which is so conspicuous in the 24-hour regeneration in the experiments tabulated above. There- fore Kerr, Hurwitz and Whipple observed a curve of protein regen- eration which was much more prolonged before a return to normal was observed. These experiments supplement the earlier ones and strengthen their conclusions. REGENERATION OF PROTEINS OF BLOOD PLASMA 71 SUMMARY A rapid depletion of serum proteins is brought about in these experi- ments by the introduction of normal red blood cells suspended in a modified Locke’s solution, care being taken to keep equivalent the volume of blood removed from the artery and the volume of red blood cell suspension simultaneously injected into the vein. The serum protein depletion is roughly proportional to the size of this exchange and it is noteworthy that the rapid depletion of the total serum proteins can rarely be carried below 1.0 per cent without causing a fatal reaction. An increase in serum protein concentration (serum protein replace- ment) begins immediately following the exchange or plasmapharesis. The increase is very rapid during the first 15 minutes following the ex- change. The increase in serum proteins is more gradual thereafter during the first 24 hours and still more sluggish during the next few days. The normal level may be reached in 2 to 7 days. The rapid replacement of serum proteins during the first 15 minutes following the exchange indicates some reserve supply of this material perhaps held in the body cells. The emergency supply is evidently small and the production of other similar material is difficult and requires time. The blood fibrin reacts in a different fashion. The same initial fall is not followed by a rapid rise in the first 15 minutes. The recov- ery however is complete within 24 hours and probably earlier than this. Fibrin is a very labile protein as compared with the serum albumin and globulin. Bloed volume fluctuations are probably very little concerned in these experimental results. The red blood cell hematocrit ratio shows but little change during the period of initial reaction. BIBLIOGRAPHY (1) Kerr, Hurwitz anp Wurppte: This Journal, 1918, xlvii, 356, 370, 379. (2) ABEL, ROWNTREE AND TURNER: Journ. Pharm. Exper. Therap., 1914, v, 625. (3) MorawirTz: Beitr. z. chem. Physiol. u. Pathol., 1906, vii, 153. (4) Hooprrr, SmirH, Bett AND WuippLe: This Journal, 1920, li, 205. (5) Ropertson: Journ. Biol. Chem., 1915, xxii, 233. (6) CULLEN AND VAN Styke: Proc. Soc. Exper. Biol. and Med., 1916, xiii, 197 (7) Serrz: Arch. gesammt. Physiol., 1906, exi, 309. II. SHOCK AS A MANIFESTATION OF TISSUE INJURY FOLLOWING RAPID PLASMA PROTEIN DEPLETION THE STABILIZING VALUE OF PLASMA PROTEINS G. H. WHIPPLE, H. P. SMITH anp A. E. BELT From the George Williams Hooper Foundation for Medical Research, University of California Medical School, San Francisco Received for publication February 25, 1920 In the preceding communication we have established the curve of serum protein regeneration following a single rapid replacement of whole blood with a red cell Locke’s solution mixture. This plasma depletion (plasmapharesis) washes out more or less of the blood pro- teins and lowers the concentration of the blood proteins in the circu- lating blood. In the preceding article we have submitted many ex- periments which are associated with little or no ‘‘shock,” using this term in the familiar clinical sense. In this paper we wish to discuss more particularly those cases which are associated with severe or lethal shock. A number of such experiments are given in detail below. This intoxication associated with plasmapharesis has been noted by the earlier workers: Morawitz (1), Abel, Rowntree and Turner (2) and Kerr, Hurwitz and Whipple (3). 26 EGHGY 8 as ck ds ches 302: Fave 8) 1 S.0 25 | X-ray intoxication hs a 4.0 | 1.7 | 2:3 | 3.8 | 0.72 | 30 2 eee eee 5.9 | 3.0 | 2.9 | 3.1 | 0.95 Death to a lethal dose of the X-ray—in this instance 200 milliampere min- utes, 90 kilo volts, given over the abdomen. Death on the 4th day with the usual blood-tinged feces and prostration is the usual reaction in these animals given a lethal exposure of the X-ray. Details of this reaction and the post-mortem findings may be found in the publica- THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 1 98 G. H. WHIPPLE, H. P. SMITH AND A. E. BELT tion just noted. The control plasmapharesis which was done 24 hours after the X-ray exposure did not give any symptoms of intoxication and this reaction due to the plasma depletion was not modified by the presence of a great amount of injured epithelium of the small intes- tine. We have many experiments to show that on the second day after X-ray exposure epithelial injury and necrosis can be made out histologically in the small intestine. These cells will undergo rapid autolysis under a variety of conditions and it is quite remarkable that the plasmapharesis should not be modified by this great mass of injured epithelial cells. DISCUSSION In some earlier experiments, Kerr, Hurwitz and Whipple (8), it was noted that the presence of liver injury or liver cell necrosis made a given animal much more vulnerable to the injury and consequent shock which followed a given plasma depletion. Using single rapid de- pletion by the method described in the experiments cited above similar results were observed. A number of such experiments are given in table 21 above and it will be noted that the control experiment in every case shows little or no shock following plasmapharesis, but an identical procedure if associated with slight liver injury was almost always fatal. There is apparently little or no difference in this respect between the liver injury due to chloroform and that due to phosphorus. The liver injury due to hydrazine sulfate was not studied in a sufficient number of experiments. [The interesting fact stands out that a trifling injury due to phos- phorus or chloroform can be tolerated by a dog with no clinical reac- tion. But if at this time (24 to 48 hours after administration of the chloroform or phosphorus) we perform a plasmapharesis of small volume which was previously tolerated by the same dog with little or no in- toxication, we immediately precipitate severe or fatal shock. The combination of slight liver injury and a moderate exchange (plasmapha- resis) will result fatally in almost all cases. How may we explain this observation? There are many possibilities but we favor the following explanation. The chloroform or phosphorus causes an injury to many liver cells and these cells are more susceptible to other injurious agents than are the normal liver cells. A sudden change in the protein con- tent of the blood which bathes these injured cells will react more un- favorably upon them than upon the healthy and more resistant nor- mal liver cells. These damaged (phosphorus) and then shocked (plas- RELATION OF BLOOD PROTEINS TO SHOCK 99 mapharesis) liver cells form substances which are taken up by the blood and carried to all the living cells of the body. If these poison- ous substances are sufficient in amount we observe the development of lethal shock. We may not assume simple intensive injury and paralysis of the liver cells alone because it is known that the body can tolerate complete ablation of the liver cells for a period of 5 to 7 hours (7). When we produce an intensive form of shock (plasmapharesis) we may observe death within 1.5 hour. This cannot be explained by any local reaction limited strictly to the liver cells. We observe in other experiments (table 28) that cell injury of other organs (kidney, pancreas and intestine) does not modify the familiar reaction following a moderate exchange. The control and poisoning experiments give similar reactions. This indicates a peculiar relation of the liver cells to the shock reaction associated with plasma depletion. SUMMARY Bleeding a dog from a large artery and a simultaneous replacement of a red blood cell Locke’s solution mixture may be called “plasma depletion” or ‘“plasmapharesis.”” This procedure will rapidly wash out large amounts of plasma proteins and cause a precipitous fall in the blood plasma protein concentration. The reaction following such procedures may be minimal or it may be lethal. In general the larger the exchange the greater the proba- bility of lethal shock. Repeated plasma depletions carried out at intervals of days or weeks on the same animal will give uniform reac- tions if the volume exchange and other experimental factors are constant. “Plasmapharesis” may be performed with washed red cells suspended in normal dog serum or fresh dialyzed dog serum. When we replace the Locke’s solution in the red cell mixture by dog serum we remove completely the toxic effect of the plasma depletion. This gives con- trol for the experimental procedure but, more important, gives strong indication that the blood serum proteins are stabilizing or protective factors. They are essential environmental factors of the circulating blood in its relation to the body cells. This may be the most important function of these plasma colloids. The presence of injured cells of the kidney, pancreas or intestine does not seriously modify the expected reaction following a uniform plasmapharesis. 100 G. H. WHIPPLE, H. P. SMITH AND A. E. BELT The presence of injured liver cells (chloroform, phosphorus) does profoundly modify the expected reaction following a unit plasmapharesis. A fatal shock reaction is almost constant following even a moderate plasma depletion preceded by liver injury. This would indicate that the liver cells are particularly concerned in the peculiar shock reaction which may follow plasmapharesis and lowering of the blood plasma protein values. It may be that this type of “shock” is not unlike the common ‘‘surgical shock.” The evidence in our experiments gives strong support to the theory that in “shock” there is a primary cell injury which precedes the familiar clinical reaction. BIBLIOGRAPHY (1) Morawitz: Beitr. z. chem. Physiol. u. Path., 1906, vii, 153. (2) ABEL, ROWNTREE AND TURNER: Journ. Pharm. Exper. Therap., 1914, v, 625. (3) Kerr, Hurwitz anp Wuiepte: This Journal, 1918, xlvii, 356, 370, 379. (4) Kerr, Hurwitz AND WHIPPLE: This Journal, 1918, xlvii, 356. (5) ERLANGER AND GassER: This Journal, 1919, xlix, 151. (6) Hatt anp WurppLe: Amer. Journ. Med. Sci., 1919, clvii, 453. (7) WHIPPLE AND Hooper: Jour. Exper. Med., 1913, xvii, 612. III. FACTORS CONCERNED IN THE PERFUSION OF LIVING ORGANS AND TISSUES ARTIFICIAL SOLUTIONS SUBSTITUTED FOR BLoop SERUM AND THE Resuutine Injury To PARENCHYMA CELLS A. E. BELT, H. P. SMITH ann G. H. WHIPPLE From the George Williams Hooper Foundation for Medical Research, University of California Medical School, San Francisco Received for publication February 25, 1920 These experiments were undertaken for the purpose of investigating the formation of the serum proteins in the body. The source of the serum proteins is still a mystery in spite of indirect evidence brought out by Kerr, Hurwitz and Whipple (1) to show that the liver is concerned in the regeneration of new serum proteins as well as in the mainten- ance of the normal serum protein concentration in the circulating blood. It seems too that other organs or tissues as well as the liver must be able under emergency conditions to produce certain amounts of new serum protein. It would appear that the normal wear and tear of serum proteins must be slight as these substances can be formed only with so much difficulty when the normal level has been greatly lowered. Theoretically, perfusion of organs should offer an ideal method of solving these problems. Using a mixture of red cells and whole serum or diluted serum or modified Locke’s solution, the investigator should be able to perfuse satisfactorily the various organs or combinations of organs and tissues. A simple determination of the protein values be- fore and after such a procedure would then give the desired information and enable one to say whether a certain organ did or did not contribute any serum proteins. Information as to whether serum proteins are used by or destroyed in these organs might also be made available. We have been able to convince ourselves that the present methods are not satisfactory to permit of the solution of this serum protein problem just outlined above. One of the conclusions which has been forced upon us is that much of the older experimentation in the field of organ perfusion is of little or no value as regards deductions made from 101 102 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE such experiments which postulated a living organ or organ cells. When parenchymatous organs are perfused with Locke’s solution or some mod- ification of this solution with or without red cells, we wish to suggest the probability that the research worker is dealing not with normal cells but with cells which have been injured or destroyed by contact with the perfusate. The investigator is then perfusing a dying or dead autolysing tissue or organ. Deductions drawn from such experi- ments must be cautious and proper allowance in every case made for this profound injury of the parenchyma cells. The “stabilizing value’’ of the blood serum proteins and the injury done the various body cells by contact with diluted plasma have been emphasized in the preceding communication. We may also note the observation of Guthrie (2) to the effect that organ transplantation is a failure if the transplanted organ is washed out with normal saline be- fore the blood flow through the organ is reéstablished. The trans- planted organ does not resume its function and we may assume that its cells were definitely injured by the short period of contact with the salt solution. The gradual slowing of the perfusion flow through any given organ is a familiar observation and we believe it may be due in part to actual injury of cells, including the endothelium. The simultaneous develop- ment of edema is a part of the same general reaction. From many of our earlier experiments we gained the impression that perfusion or plasmapharesis with the use of Locke’s solution inflicted a destructive injury upon certain body cells and that this injury of pro- tein was responsible for the fatal intoxication and death. Our con- ception was somewhat as follows: the initial injury might damage irreparably liver cells (or other body cells) and from these injured organ cells were derived poisonous substances (protein split products) sufficient to cause death. It has been pointed out that complete par- alysis of all liver cells could not explain this phenomenon as ablation of the liver can be tolerated for 5 to 7 hours, while the shock following a large plasmapharesis may cause death within 2 hours. Postulating the presence in the body of some poisonous substance referable to the perfusion or plasmapharesis, we have attempted in many of our perfusion experiments to demonstrate the presence of a poison in the perfusate at the end of any given experiment. In only one experiment (exper. 5) have we evidence for the presence of any poison under these conditions of perfusion. But we have some evi- dence (expers. 17 and 18) to show that a known poison of protein origin FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 103 added to a given non-toxic perfusate may be in part removed within 20 to 30 minutes’ continuance of perfusion. This is evidence that a poison of colloid nature may be removed from the circulation in such experi- ments—so the absence of a poison in our perfusates does not negative the possibility of poison being formed by the injured cells and contrib- uted to the blood stream or perfusate. We have further been able to show that an enormous dose of proteose-like, toxic material may be wholly removed from the blood stream within a period of 5 minutes after intravenous injection in a normal dog. We have discussed the results of the experiments given below in the light of the plasmapharesis experiments given in the two preceding papers. It may be noted that these perfusion experiments were done before the plasma depletion experiments. The evidence which may be taken from our perfusion experiments is not as definite as that obtained in the later plasmapharesis experiments but these data are all in har- mony. A most important fact is that physiological perfusion of organs is very difficult and slight modification of the blood plasma may have profound effect upon bedy cells. DEVELOPMENT OF PERFUSION METHODS The notion of artificial perfusion was long ago expressed by Le Gal- lois (3). He maintained that by artificial perfusion life might be kept up in any portion of the animal even though separated from the rest of the body. It remained, however, for other workers actually to under- take such experiments. In 1828 Kay (4) showed that artificial per- fusion with blood was capable of restoring irritability to dying muscle. Artificial perfusion of kidneys was first attempted in 1849 by Lobell (5). The work of Brown-Séquard (6) done several years later showed the necessity of oxygenation of the blood used as a perfusate. The oxygenated blood was forced through the arteries by means of a syr- inge. In this manner he perfused various regions including the iso- lated head. He found that he was able in this manner to maintain certain evidences of reflex nervous activity provided the perfusion was commenced promptly after decapitation. Ludwig and Schmidt (7) in 1868 described an apparatus by means of which blood could be forced under constant pressure from a reservoir. Improvements in aeration of the perfusion medium were made by Schréder (8). Fry and Gruber (9) devised an artificial lung by means of which the aeration of the per- fusate could be accomplished without interrupting the flow of blood 104 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE to the region being perfused. Although fluctuation in the pressure supplied to the perfusion medium occurred in the work of the earlier investigators using the syringe injection method, the distinct beneficial effects of such variations in pressure were first recognized by Ludwig and Schmidt (7). Fry and Gruber (9) attached the piston of a syringe supplying the arterial pressure to a motor-driven wheel thus creating by mechanical means a pulsatile pressure. Hamel (10) emphasized the need for pulsatile pressure. He devised an apparatus in which the movements of a pendulum periodically interrupted the flow of the per- fusate to the tissues, thus converting a constant pressure into an inter- mittent one. Jacobj (11) devised an elaborate perfusion apparatus in which pulsatile pressure was created by periodic and forcible compres- sion of a rubber balloon placed in the arterial side of the circuit. He used the principle of aerating the blood by forcing a mixture of air and venous blood through a stretch of tubing at the end of which the blood and air were separated by gravitation. In a later paper Jacob] (12) described a method by means of which the blood was aerated by per- fusion through a lung in which respiration was artificially maintained. In this manner he avoided the direct mixing of the blood with the air. In 1903 Brodie (13) published an account of an apparatus which has subsequently been used by several investigators. With it he was able to perfuse an organ with the use of no other blood than that obtained from the animal itself—a considerable advantage over many of the types previously employed. To create pulsatile pressure he suggests that a fairly distensible piece of rubber tubing placed in the arterial side be rhythmically compressed by a wooden arm. Other forms of perfusion apparatus have been described by Hoff- mann (14), Richards and Drinker (15), Friedmann (16), Mandel (17) and Kingsbury (18). Pulse pressure as a necessary factor in the mechanics has been re- cently reémphasized by Hooker (19), whose apparatus was employed in our experiments. His apparatus can be adjusted in such a manner as to furnish a pulse curve identical in form to that produced in a normal intact animal. Aeration is effected. In addition to the purely mechanical methods of perfusion another slightly different procedure has been employed by some. As far back as 1881 Martin (20) attempted to study the activity of the heart by diverting all of the blood issuing from the aorta back into the right auricle. A heart-lung preparation was thus effected, the circulation being successfully excluded from the rest of the body. This procedure FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 105 or modifications thereof have been used by many workers since that time. In 1914 Bainbridge and Evans (21) susbtituted this living prep- aration for the artificial perfusion machine. The organ to be perfused received its blood directly from the aorta of the preparation. The venous blood issuing from the vein was returned to the right side of the heart. The perfused tissue thus received blood aerated by the lungs and under pulsatile pressure supplied by the heart itself. It should be remarked that in this form of perfusion the study of the per- fused organ is complicated by the metabolism of the heart and lungs themselves. EXPERIMENTAL OBSERVATIONS All perfusion experiments were performed by use of the apparatus designed and described in detail by Hooker (19), (22). Through his courtesy we were able to obtain this machine which was made after the model of his original apparatus. We take this opportunity to ac- knowledge our appreciation for valuable assistance on the part of Dr. D. R. Hooker. The general experimental procedures are covered by the brief description in individual experiments. All experiments were done on dogs under complete ether anesthesia. In all perfusion ex- periments the dog was placed upon a warm pad to keep up the body temperature. Perfusion of hind legs with Locke’s solution Experiment §. Male bull pup. Weight 4.8 kilos. Under ether anesthesia cannulae were inserted and the hind legs were per- fused for 3 hour with Locke’s solution. The temperature of the perfusate varied between 30° and 40°C. The perfusion pressure was between 100 and 110 mm. mercury. The pulse pressure was between 20 and 30 mm. mercury. Pulse rate 130a minute. The 240 cc. of perfusate recovered at the end of the perfusion con- tained 77,000 red blood cells per cubic millimeter. Hemolysis was moderate in amount. To test the toxicity of this perfusate the cells were removed by centrifugali- zation and 100 cc. of the supernatant fluid were injected intravenously into a normal dog. A rise of 1.4° in temperature with vomiting and diarrhea was noted. The pulse was not markedly altered. The perfusate was therefore moderately toxic. Ten cubic centimeters of the centrifugalized perfusate were also injected intraperitoneally into a 100 gram rat. Slight toxicity was evident. Perfusion of hind legs with red corpuscles suspended in Locke’s solution Experiment 7. Male collie mongrel. Weight 16 pounds. The hind legs were perfused with a suspension of blood corpuscles in Locke’s 106 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE solution. The temperature of the perfusate varied between 32° and 38°C. The mean pressure was 50 mm. mercury. The pulse rate was 130 and the rate of flow 180 ce. a minute. Autopsy shows irregular petechial hemorrhages in the muscles and fascia of the hind legs. Slight hemolysis was noted in the 250 ce. of perfusate recovered at the end of perfusion. One hundred and sixty-one cubic centimeters of the centrifugalized end-product were injected intravenously into a normal 13.75 pound dog. A temperature rise of 0.4° was observed. No vomiting or diarrhea occurred. Twenty cubic centimeters of this perfusate were also injected intraperitoneally into a rat weighing 100 grams. There were no evidences of toxicity from the use of this perfusate. These two experiments (expers. 5 and 7) give little information con- cerning the actual perfusion conditions but supply data concerning the production of a poison by the Locke’s perfusion. There is slight posi- tive evidence for a toxic reaction in experiment 5 but a negative reaction in experiment 7. In general we have no distinctly positive evidence that this destructive perfusion of body tissues will give demonstrable evidence of a toxic element in the perfusate. As stated above this may be explained by the capacity of the body cells to remove such poisons from circulating fluids. Perfusion below diaphragm with red corpuscles suspended in modified Locke’s solution Experiment 9. Normal female black and white pup. Weight 6.75 pounds. The oxalated blood from a normal dog was centrifugalized and the corpuscles washed in gelatin-Locke’s solution minus calcium by mixture and recentrifugali- zation. The red cells were then suspended in gelatin-Locke’s minus calcium in the ratio of packed corpuscles 1 to solution 5. This mixture was used as the perfusion medium. Under ether anesthesia the animal was bled. Cannulae, were inserted and the perfusate forced through the aorta and recovered from the right auricle, thus perfusing the area below the diaphragm. The temperature of the perfu- sate varied between 35° and 38°C. The mean pressure was maintained at 50 to 80 mm. mercury; the pulse pressure between 10 and 15 mm. mercury. The pulse rate was 130 a minute. The flow was excellent. Perfusion began 20 minutes after bleeding and was continued for 1 hour. The autopsy was delayed for a few hours after the completion of the per- fusion. The muscles of the hind legs were pale and showed very little edema. No hemorrhages were present in the muscles and connective tissue. The liver was normal except for some air bubbles in its vessels. The capsule of the kid- neys stripped readily. Hyperemia was seen at the cortico-medullary boundary. Suprarenals were negative. The lymph nodes of the mesentery were normal. FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 107 The pancreas showed a considerable amount of edema. The spleen was dark red except for a transverse light band possibly caused by block from emboli. The intestines were filled with red mucoid material. The mucosa was velvety, swollen and deep purplish-red in color. Histological sections of liver and intes- tine give no evidence of tissue abnormality. On centrifugalization a sample of perfusate taken 10 minutes after beginning of perfusion was light pink. A sample taken at the end of 40 minutes was some- what deeper in color, while’ a sample taken at the end of perfusion was dark red. One hundred and fifty cubic centimeters of a centrifugalized sample taken after 10 minutes of perfusion were injected into a small normal dog. The ani- mal vomited once. There was a rise of 1°C. in temperature. The pulse remained good. No marked symptoms of intoxication were present. Eight cubic centimeters of the 10-minute sample were injected intraperitone- ally and 2 cc. were injected subcutaneously into a 75-gram rat. No toxic action was noted. This test was repeated by injecting 5 cc. intraperitoneally into a 38-gram rat. No toxic action was noted. The same amount of the 40-minute perfusate sample was injected intraperitoneally into a 62-gram rat. No toxic action. Perfusion below diaphragm with red blood cells in modified Locke’s solution Experiment 11. Female shepherd pup. Weight 2280 grams. In preparation of the perfusate blood corpuscles were obtained from the blood of a normal dog bled several hours previously into a 1 per cent sodium oxalate solution. The red cells were centrifugalized and washed in Rous’ gelatin-Locke’s solution minus calcium. The packed cells were then suspended in a similarly prepared calcium-free gelatin-Locke’s solution in the ratio of one part of corpuscles to five of the saline mixture. Under ether anesthesia the animal was bled and the cannulae were arranged to perfuse all of the tissues below the diaphragm. Ten minutes were consumed in arranging the cannulae. The temperature of the perfusate was between 35° and 38°C. The pulse pressure was about 10 mm. mercury. Due to clots in the gauze two stops were necessitated over a period of 5 minutes each. One occurred soon after the beginning of perfusion and one some minutes later. The dura- tion of the perfusion was 1 hour. During this period the animal increased 760 grams in weight. Autopsy showed about 75 ec. of pale bloody fluid in the abdominal cavity. Marked edema was present about the pancreas and throughout the mesentery. Hemorrhagic spots were observed over the surface of the kidney and stomach, about the ovary and throughout the muscles and fascia of the hind legs. The liver on section was translucent. On section the kidney showed indefinite dark hemorrhagic spots up to 0.5 cm. in diameter. The whole organ was dark and congested. The stomach contents were normal. Mucus and bloody fluid were present in the small intestine. Congestion and bloody intestinal contents were more prominent in the lower part of the small intestine. The perfusate showed a moderate grade of hemolysis before perfusion but less after 15 minutes of perfusion. Moderate hemolysis existed at the end of the perfusion. Bacteriological cultures showed 500 bacteria per cubic centimeter in samples taken at end of perfusion. 108 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE Two hundred and eight cubic centimeters of the perfusion fluid obtained at the end of the perfusion were injected intravenously into a normal small dog. With the exception of a temperature rise of 1.5° and some shivering there were no signs of intoxication. The two experiments (expers. 9 and 11) show the results of a perfu- sion of all the organs and tissues below the diaphragm by a red cell Locke’s solution mixture. It is to be noted especially that there is marked edema of retroperitoneal tissues and the pancreas. This edema is invariably present in considerable amount except when whole de- fibrinated blood is used as perfusate. We accept the edema as one indication of tissue or ceil injury. The same is true of hemorrhagic areas and ecchymoses, but some of them may be due to emboli. The marked congestion of the intestinal mucosa with the escape of blood- tinged fluid and mucus is also a valuable index of injury. This is a familiar reaction noted in dogs dead from anaphylaxis or large doses of proteose or from surgical shock. The perfusate in both these experi- ments contained no poisonous substance for normal dogs and white rats. Hemolysis is always present in slight or moderate degree in all our experiments. We are inclined to explain a part of this hemolysis by the cell injury in organs or tissues and this cell injury reacts unfavorably upon the red cells with resulting hemolysis. We realize that the dog’s red corpuscles are most fragile and that the red cells are subjected to much mechanical injury in these experiments. Other observers may choose to explain all this hemolysis upon a purely traumatic basis. Perfusion below diaphragm with diluted defibrinated blood Experiment 13. Male bull pup. Weight 2150 grams. Two parts of defibrinated blood obtained from a normal dog were diluted with one part of gelatin-Locke’s solution. The animal was anesthetized with ether, the thorax opened and the cannulae inserted in such a way that the per- fusion medium was forced into the aorta just above the diaphragm and the blood received from the inferior vena cava just below the heart. In this way the entire region below the diaphragm was perfused. The temperature of the perfusion medium was maintained at about 32° to 38°C. The systolic pressure varied from 95 to 120 mm. mercury with a pulse pressure of 20 mm. mercury. The return flow was accidentally occluded for a few seconds at the beginning of the experiment. The perfusion lasted 1 hour. Examination of the region perfused showed hemorrhagic streaks in the dia- phragm and gall bladder. Numerous small areas of hemorrhage accompanied by a considerable amount of edema existed about both kidneys. The liver showed considerable congestion and edema. The lobules of the pancreas were FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 109 distinctly separated by edema. The spleen was quite dark. The stomach showed considerable engorgement with sub-mucous hemorrhages. Externally the small intestines were spotted by numerous small subserous hemorrhages. The intestinal mucosa showed diffuse congestion while the lumen contained a little dark mucus. The colon was more nearly normal in appearance; however, a few small sub-serous hemorrhages were seen. The hind legs showed no hemor- rhage or edema. There was a weight increase of 400 grams. The perfusate showed slight hemolysis before perfusion and a moderate grade of hemolysis at the end of perfusion. The perfusate obtained at the end of per- fusion contained 39 mgm. of non-protein nitrogen and 16 mgm. of urea-nitrogen per 100 ce. TABLE 34 Perfusion below the diaphragm with defibrinated blood. Experiment 16 COLOR OF CARBON HYDROGEN ome Spee all eee eee cpa one PERFUSATE | PER 100cc. TION IBGLOEE Neh Sei os steaeigis 002 [ Pale pink 38.5 05 After 20 minutes..........] Rose 15.7 €23 After 40 minutes..........| Rose 26.1 More blood added At the end................| Deep rose 12.8 7.4 Perfusion below the diaphragm with defibrinated blood Experiment 16. Normal male bull pup. Weight 4.2 pounds. Under ether anesthesia cannulae were inserted for perfusion below the dia- phragm. This region was perfused for 30 minutes with pure defibrinated blood at a rate of 100cc. perminute. The cannulae were then shifted tothe lower ab- dominal vessels and the hind legs perfused at a rate of 25 ce. a minute for 60 min- utes, with the same perfusate. The perfusate was aerated with pure oxygen and was maintained at a temperature varying from 36° to 39°C. The pulse rate was 141 a minute. Intestinal peristalsis was quite conspicuous at the beginning of perfusion but was less noticeable after the perfusion had been in progress for about 5 minutes. During this period the abdominal wall was very sensitive to touch and contracted violently when touched. At autopsy a large amount of clear straw-colored fluid was noted in the abdomi- nal cavity. The animal had gained 200 grams in weight as a result of the per- fusion. The liver was slightly translucent although apparently normal. The spleen was somewhat congested. The pancreas showed no edema. The kid- neys were slightly congested in the pyramidal areas. The stomach was normal. The duodenum was likewise normal but the ileum showed a mucosa congested and dark red with grey mucoid material in the lumen. The hind legs showed no edema and were quite dry in appearance. A few petechial hemorrhages appeared in the fascia. Histological sections: Pancreas and kidney are normal. The spleen and liver show capillary congestion but normal parenchyma cells. The stomach and small intestine are normal except for slight congestion of the ileum. 110 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE In these two experiments (expers. 13 and 16) we used diluted or whole defibrinated blood. The general autopsy picture following the use of whole blood is almost normal and the lack of the edema we believe is to be explained on this ground. Even in the last experiment we note the development of ascites which of course indicates circulatory ab- normality. There is further a distinct acidosis to be explained by inadequate aeration of the blood. Perfusion with whole defibrinated blood—toxic proteose added Experiment 17. Adult male poodle dog. Weight 10.25 pounds. Under ether anesthesia the animal was bled. The arterial cannula was placed just above the bifurcation of the iliacs. The venous cannula was inserted in the inferior vena cava just above the renal veins. Thirty minutes were con- sumed in bleeding, arranging the cannula and starting the perfusion flow. The hind legs were perfused for 32 minutes with defibrinated blood. To the 500 ce. of perfusate then remaining in the apparatus 100 ce. of a proteose solution (lethal dose is 2 ec. per pound body weight, adult dog) were added and the perfusion continued for 12 minutes. The perfusate was warmed to a temperature of 35° to 38°C. The rate of flow was 80 cc. per minute until near the end of the experi- ment, when it decreased to 50 cc. per minute. The aeration of the perfusate was excellent. The proteose solution was prepared as described elsewhere (23) from the material of the obstructed intestine. Autopsy of the perfused dog showed slight icterus of the tissues of the hind legs but no hemorrhages or edema. A few small hemorrhages were found in the right testicle. The pelvic organs were negative. At the end of perfusion there were obtained 300 ce. of. perfusate for analysis. Of this were injected into a 14-pound normal pup, 227 ec., which represented 38 cc. of the original proteose. This would be a fatal dose for an adult dog weigh- ing 19 pounds, provided no proteose had been lost. Pups are more susceptible to proteose intoxication than adults. This dog therefore received a theoretical dose of one and one-half times its lethal dose, assuming that no proteose was lost from the perfusate during the perfusion. The injection of 227 cc. of perfusate caused death in 3 hours with the clinical picture of acute proteose intoxication. Diarrhea and vomiting appeared within 3 hour after injection and continued until death. There was an initial rise in temperature followed by a drop to 36.8°C. half an hour before death. Autopsy findings showed exquisite splanchnic engorgement especially marked in the mucosa of the small intestines, which was a velvety purplish-red coated with mucus—described and pictured elsewhere (24). The results of examination of the perfusate are given in the table below (table 35). FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS Stl Perfusion with whole defibrinated blood—toxic proteose added Experiment 18. Under ether anesthesia a small female mongrel dog was bled. Cannulae were inserted preparatory to perfusing all of the body below the dia- phragm. Defibrinated blood warmed to 35° was perfused through this region under a mean pressure of 110mm. of mercury. The pulse pressure varied between 5 and 15 mm. of mercury. The pulse rate was 120 per minute. Aeration of the perfusate was not quite as satisfactory as usual. Actual perfusion commenced 30 minutes after bleeding the animal. After the flow has been maintained for 20 minutes 100 ce. of proteose solution were added to the 250 ce. of defibrinated blood then remaining in the apparatus. Perfusion was continued for 15 min- utes. The rate of flow was 150 ec. a minute at first, but gradually decreased to 85 cc. a minute toward the end of the experiment. Marked intestinal peri- stalsis was present during the first few minutes of perfusion; blood-tinged feces later. Autopsy showed a moderate quantity of pale blood-tinged fluid in the peri- toneal cavity. The animal had gained 180 grams in weight during the experi- ment. The liver showed edema and small hemorrhages throughout. Numerous small hemorrhages as well as several larger ones were seen in the wall of the gall bladder. The pancreas was negative except for one small hemorrhage. The mesenteric lymphatics contained blood-stained fluid. The adrenals contained several small hemorrhagic areas. The kidneys were negative. The spleen was small, dark and translucent. The stomach showed one fairly large but no small hemorrhages. The mucosa of the duodenum was engorged. The intestinal lumen contained an excess of thin bloody fluid with little mucus. It is pos- sible that some of the hemorrhages noted above are to be explained by short periods of high blood pressure during the periods of perfusion. As was stated above, 100 ec. of the proteose solution (lethal dose is s 2 ce. per pound body weight, adult dog) were added to the 250 cc. of perfusate then in circulation and the perfusion continued for 15 minutes. Of the final perfusate 77 ce. were injected intravenously into a normal 1l-pound pup (no. 17-181). It is evident that these 77 cc. contained 22 ce. of the original proteose solution pro- vided none of this toxic material had been removed during perfusion by the tis- sues of the animal perfused. The reaction of the animal might be expected therefore to be lethal, if no proteose was removed during the perfusion through the tissues of the first dog. The reaction to this intravenous injection was typical for a sublethal toxic dose of proteose. There was vomiting and diarrhea for 2 hours and much prostration. Recovery was evident in 3 hours and the dog was normal in a few more hours. It appears that some of the proteose had been removed as the amount given was more than a lethal dose for a pup of 11 pounds body weight. The final perfusate was further examined and shown to contain definite amounts of hemoglobin (hemolysis). Bacteriological examination (plates) showed 40,000 colonies per cubic centimeter of the perfusate. The non-protein nitrogen at beginning of perfusion was 34.7 mgm. per 100 cc. of perfusate and at the end was 35.8 mgm. 112 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE In these two experiments (expers. 17 and 18) the perfusion was done with whole defibrinated blood to insure a minimum injury of the per- fused tissues. After the initial perfusion a standardized toxic solution was added to the perfusate and again circulated as before. The per- fusate at the end of the experiment was tested on a normal dog and in this way it was demonstrated that some of the poison had been removed. It is easy to show that large doses of this toxic proteose are rapidly re- moved from the circulation of a dog. Following intravenous injection of large amounts of toxic proteose it is possible to demonstrate its presence for 2 to 3 minutes in the blood stream but not after 5 minutes. The presence of bacteria as noted in this and other experiments will not seriously disturb the reaction. If anything, their presence will increase the toxicity of the perfusate mixture. These bacteria probably gain TABLE 35 Perfusion with whole defibrinated blood—toxic proteose added. Experiment 17 PER CENT > NON- CARBON = Brhonrers | i00?| pumocen | mgm. per | mgm. per 100 ce. 100 cc. Before perfusion.......| Moderate AT 19 26 7.5 16 After one-half hour....| Moderate 50 16 31 7.4 16 On adding proteose....| Moderate 47 18 30 7.4 19 At end of perfusion....| Slight 45 20 89 7.4 13 entrance in part through the intestinal tract and in part are added by the manipulation of the perfusate. Efforts were made to preserve the circulating machinery in as near a sterile condition as possible but there are many possibilities for introducing contamination. The ‘proteose solution’”’ used in these experiments is prepared from material obtained from obstructed intestines or closed intestinal loops in dogs. The material is precipitated by five volumes of alcohol, the precipitate dissolved in water and the protein removed by boiling in dilute acetic acid solution. A second precipitation with alcohol is often employed. The final solution is an opalescent fluid which con- tains proteose-like materials. This fluid material is then standardized by intravenous injection in normal dogs, as has been described else- where (23). FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS ba Large infusion of Locke’s solution into portal vein Experiment 21. Dog 17-125. Normal adult mongrel, black and tan. Weight 43 pounds. Under ether anesthesia and with sterile precautions the abdomen was opened and the hepatic-pancreatico-duodenal and pancreatico-duodenal arteries were ligated. Sterilized calcium-free Locke’s solution, 1750 ce., was warmed to ap- proximately 36°C. and injected at a rate of 55 ce. a minute into the portal system through a small venous branch in the mesentery. The abdomen was closed. The animal showed no severe reaction for some time but died 3 days later. Autopsy performed several hours later shows considerable post-mortem change. Dark red softened areas containing bubbles of gas are seen scattered throughout the liver. The serosa of the intestines is quite red. The mucosa shows moderate engorgement and is covered by a buttery exudate. The coil of intestine to which the perfused vein was distributed differs in no way from the other parts of the small intestine. Kidneys are negative. The total non-pro- tein nitrogen of the blood shows but slight alteration from the original value as a result of the experimental procedure. Histological sections: The picture is somewhat confused by post-mortem changes but it is clear that there are scattered areas of liver cell necrosis which are ante-mortem and presumably related to the experimental procedure. These areas include many liver lobules and present a uniform necrosis with scattered leucocytes between the liver cell strands. Bile ducts and blood vessels are nor- mal and no evidence of vascular thrombosis is observed in any sections. Other sections show normal liver parenchyma. Other tissues present nothing of importance. Large infusion of Locke’s solution into splenic artery Experiment 23. Dog 17-200. Young male collie. Weight 29.5 pounds. Normal except for a slight attack of distemper. Under ether anesthesia and with sterile precautions the abdomen was opened. A cannula was inserted into the splenic artery in such a manner that the upper arm of the pancreas, a part of the duodenum and stomach, as well as the liver, were perfused by 2000 cc. of sterile calcium-free Locke’s solution injected at a rate of 50 cc. a minute. The saline was injected at room temperature. Splen- ectomy followed the perfusion. The pancreas showed moderate edema at the end of the experiment, but the dog was not severely shocked. Later a state of intoxication slowly developed and at the end of 36 hours death was imminent. The animal was killed by ether. Autopsy performed at once shows a little blood-stained peritoneal fluid. The liver shows a moderate grade of cloudy swelling. The subserous tissues of the gall bladder are thick and edematous. A number of sub-serous ecchymoses are scattered over the small intestine. Several hemorrhagic areas are present in the mucosa of the small intestine. Hemorrhages and fat necrosis are rather pro- nounced in the upper arm of the pancreas. Histological selections: Organs are normal with the exception of the pancreas and liver. The pancreas shows extensive hemorrhagic necrosis and much nec- THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO1. 114 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE rosis of fat and gland parenchyma. There are no thrombus masses noted in any of the vessels. The head of the pancreas is essentially normal. The liver shows scattered clumps of polymorphonuclear leucocytes and evidence of injury to small clusters of liver cells in various portions of the liver lobules. Large infusion of Locke’s solution into the portal vein Experiment 24. Dog 17-203. Adult male mongrel. Weight 29.5 pounds. Slight distemper. Under ether anesthesia the abdomen was opened and the hepatic artery clamped. One of the larger splenic veins was isolated, a cannula inserted and 1500 cc. of sterile calcium-free Locke’s solution warmed approximately to body temperature were injected into the portal system over a period of 18 minutes. The clamps were then removed from the hepatic artery. Splenectomy was per- formed and the abdomen was closed. The temperature rose to 40°C. shortly after the operation and the animal vomited once; otherwise no clinical disturbance was noted. Complete recovery ensued. The animal was killed 7 days later. The autopsy was negative. No alteration could be made out in the liver. To obviate the mechanical difficulties inherent in organ perfusion and to get information concerning the direct effect of Locke’s solution upon tissue and organ cells we performed a number of experiments of which experiments 21, 23 and 24 are examples. The first experiment (exper. 21) shows a fatal reaction following a large infusion in the portal vein after ligation of the branches of the hepatic artery to limit the blood flow through the liver. We have explained this reaction as due in part to injury of the liver cells by contact with the mixtures of Locke’s solution and blood. It is known that ligation of the hepatic artery will cause no disturbance in the dog. It seems hard to account for these areas of liver necrosis except as due to the action of the portal blood diluted by the large infusion of Locke’s solution into the portal vein. It is to be noted (exper. 24) that a similar experiment was tol- erated without obvious liver injury but the occlusion of the hepatic artery in this experiment was only temporary. The pancreas necrosis was surely caused by the perfusion of the splenic artery (exper. 23). It may be objected that perfusion in this way against the arterial stream will cut off the tissues from oxygen by wash- ing away all available red cells. We are inclined to believe that ar- terial collaterals which are numerous in this region will insure the presence of the necessary number of oxygen-carrying red cells. This objection cannot apply to experiment 21. FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 115 DISCUSSION In attempting an analysis of our own experiments we wish to draw freely on the published work of other investigators. We wish to keep constantly in the reader’s mind that physiological perfusion of any organ is a matter of extreme difficulty and often great confusion is introduced by such methods which are intended to simplify the study of organ function. By physiological perfusion we mean a perfusion adequate to maintain the organ in its normal physiological activity. In the first place let us inquire what criteria of tissue abnormality we have. What sort of evidence is going to lead us to pass judgment concerning the physiological condition of tissues? There are certain conditions under which we may ascertain what is going on in the tissues by a direct observation of the functional activity of the part perfused. For example, in perfusing the kidney the quantity and quality of the urine secreted furnishes some evidence concerning the condition of that organ. The reduction within the organ of oxyhemoglobin to hemo- globin was noted by the earliest observers and is indicative of meta- bolic activity of some nature on the part of the perfused tissues. The nature of the heart beat is indicative of the condition of the perfused heart, although Magnus (25) has shown that if such an inert substance as hydrogen gas be perfused through the coronary arteries heart beats will be stimulated. Sollmann (26) showed that the same result fol- lowed perfusion with cottonseed or paraffin oil. In such cases the per- fusion fluid cannot be thought of as being a nutrient fluid; on the other hand it is not altogether impossible that, as was suggested by Sollmann, the heart beats may be stimulated by purely mechanical factors. For these reasons we must not hastily conclude that, because the gross mechanical movements simulate those occurring in the intact animal, the preparation is in fact an example of normal physiological activity. A criterion as to the condition of the perfused medulla is furnished by observing whether the medulla continues to maintain its normal control over the heart and muscles of respiration. In addition to direct observation of the functional activity of an organ perfused, we have still other kinds of evidence which help us to judge concerning the condition of the tissue. Thus the rate of flow through the vessels is in some cases a valuable indicator for it is a general rule that tissue injury brings about in some way or other a decrease in the rate of flow through the part. We have, in addition, the still more crude signs of tissue injury: edema, congestion and hem- orrhage. 116 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE Of the factors in the procedure of perfusion whose variations might bring about injury to the part, the following may be mentioned as being perhaps the most important: aeration of the perfusion medium, composition of the perfusion medium, interruptions in continuity of flow, temperature of the perfusion medium when it enters the perfused organ, mean pressure and pulse pressure. It should be realized that a perfusion experiment is no better than its weakest point. If any of the above factors react in such a way as to cause injury to an organ, perfection of the other factors will not rem- edy the defect. It is also conceivable that when several organs are being perfused simultaneously, injury to one organ or tissue may react injuriously on others. Concerning the effect on the tissues of composition of the perfusate, the literature contains many references to condition of the tissues as shown by functional activity. The injurious effects of foreign blood have been known since the time of Prévost and Dumas (27) when this fact first began to be recognized through the failure of foreign blood to act normally after transfusion. Though repeatedly shown to be harm- ful in its effects on tissues of another species, foreign blood has been used in perfusions even as late as Brodie (13) who says that ox, sheep or horse blood cannot be used in the perfusion of organs taken from dogs. He finds that as soon as foreign blood is supplied to the perfused heart the beat becomes irregular. The heart next goes into fibrillary twitchings and cannot be recovered from this state even with the ani- mal’s own blood. Although defibrinated blood had been used in the transfusion experi- ments of Prévost and Dumas (27) without the observation of harmful results, Magendie found it incapable of carrying on the normal function of the circulating medium after transfusion. In a series of experiments in 1822 (28) and again in 1838 (29) he presented experimental data to show that the lack of fibrin, reduced through repeated bleeding, de- fibrination and reinjection of the defibrinated blood gives rise to a serous and bloody ransudate into the lungs and intestine with the death of the animal. The weight of Nese s name behind such a statement did much to discredit defibrination in the eyes of other workers, but eventually Bischoff (30), Goll (31), Polli (32), Panum (33), Ponfick (34), and many others began to turn the weight of experimental evidence against a belief in the extreme toxicity of defibrinated blood when used in trans- fusions between animals of the same species FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 117 More recently, by means of perfusion experiments, Stevens and Lee (35) and Brodie (13) present evidence of slight vasoconstriction due to the use of defibrinated blood as a perfusion medium. Their work again points to an injurious action of defibrinated blood when substi- tuted for the normal circulating medium. However, Stevens and Lee believe that the slight vascular contraction which they note can be readily counteracted with pharmacological agents. In 1903 Pfaff and Vejux-Tyrode (36) found defibrinated blood — initely injurious to the kidney of the dog. Quantities of from + to +4 of the total blood were withdrawn from the carotid artery, hidpedt filtered and reinjected into the jugular vein. The repetition of this procedure resulted in the appearance of albumin, hemoglobin and red blood cells in the urine and finally cessation of secretion. However, a rapid return to normal was effected in these animals by bleeding fol- lowed by direct transfusion of whole blood from a normal dog. It would seem that the kidney may be unusually sensitive to this procedure. It is certain that in those of our animals which underwent quite com- plete defibrination (37)—(see expers. 323 and 324)—there were no clinically evidenced signs of injury or toxic manifestations. This may also be said of those experiments of Whipple and Goodpasture (38) in which quite complete defibrination was also effected. The importance of a physiological balance of the normal inorganic salts of the blood is generally recognized. Solutions containing ab- normal quantities of these salts have been shown to be toxic to the per- fused heart. Hooker (22), (39) showed that in perfusion of the res- piratory center a balance of potassium and calcium is essential for a normal function of the preparation. As was early shown by several investigators and more recently reémphasized by Hooker (19) and by others, and as we have found in our experiments, saline solution has the property of setting up such a condition in the tissues that the rate of flow gradually decreases. The effects of variations in composition of perfusion media may also be manifest from the morphological side. Brodie (13) shows that edema results from the use of foreign blood. Hamel (10) shows that edema results in organs perfused with saline under pulsatile pressure. Similarly the kidney when perfused with Locke’s solution exhibits more edema then when perfused with defibrinated blood (19). These re- sults are entirely in accord with our experience. Perfusion with pure Locke’s solution almost invariably produces an extreme grade of edema. Dog’s defibrinated blood diluted with Locke’s solution produced less, 118 A. E. BELT, H. P. SMITH AND G. H. WHIPPLE and pure defibrinated blood produces very little demonstrable edema. It is difficult to explain just why dilution with Locke’s solution should produce edema. In several of our experiments we carried out hydrogen- ion and carbon dioxide determinations on the perfusate and it is inter- esting to note that in cases of marked edema there was a rise in the hydrogen ion concentration and a fall in the carbon dioxide capacity. Whether the edema is the result of the acidosis or not, there still re- mains the question as to what is the cause of the decrease in the buffer substances. Perhaps it may be attributable in part at least to insuf- ficient oxidation in the tissues. Poor oxidation may result from in- sufficient oxygen-carrying capacity of the perfusate or from inadequate aeration of the perfusate in the artificial lung, or may result from a decreased rate of flow through the animal and a stagnation of the blood in the tissues with consequent asphyxia. A factor which has been shown to be of considerable importance in causing tissue injury is that of loss of time in establishing the artificial perfusion after interrupting the normal relations. Most of the earlier workers paid but scant attention to this phase of the problem. In many experiments several hours elapsed before any attempt was made to reéstablish the flow. Grube (40) mentions that in perfusion of the liver with defibrinated blood to which glucose had been added, the gly- cogen content of the liver rises, but only in case the circulation is very promptly reéstablished. The deleterious effect on the kidney of tem- porary anemia is well known. Momentary compression of the renal vessels may cause a cessation of secretion for many minutes. The effect of compression of the cerebral vessels has been known since very early times. The duration of such anemia necessary to produce irreparable damage was long ago studied by Astley Cooper (41). Signs of activity can be restored to the brain of the isolated head provided only that the perfusion is promptly commenced (6), (42), (43) and (44). Skel- etal muscle is capable of surviving much longer periods of anemia than is the case of brain or kidney. Munk (45) holds that in perfusion of the kidney if the flow is not promptly commenced the vessels become nar- rowed, the perfusion flow rendered difficult and a delay in the formation and flow of urinary fluid occurs. Recently Bainbridge and Evans (21) have succeeded in perfusing the kidney without any interruption what- soever in the continuity of flow. A great deal has been said in the literature concerning the value of pulsatile pressure. As has already been noted, Ludwig and Schmidt (7) observed that with constant pressure the rate of outflow from the FACTORS CONCERNED IN PERFUSION OF LIVING ORGANS 119 perfused tissue decreases, but that recovery occurs if the flow is halted for a while. Similar observations were subsequently made by many workers. The recognition of the importance of this factor is evidenced by the numerous forms of apparatus devised to accomplish this end. Hooker (19) holds that in the perfused kidney the amount of urinary filtrate formed varies directly with the magnitude of the pulse pressure. The amount of proteins in the urinary filtrate varies inversely with the magnitude of the pulse pressure. The rate of blood flow through the organ varies directly with the magnitude of the pulse pressure. Recently Gesell (46) has been able in the intact animal to abolish almost completely the pulse pressure in the renal arteries without interfering with the normal mean blood pressure. The result of this alteration was an immediate and practically complete cessation of urinary secre- tion. SUMMARY Physiological perfusion of organs is a matter of great difficulty. Much of the work done with organ perfusion is of little value because a proper appreciation of the limitations of the method does not obtain among laboratory workers. The use of physiological saline, Locke’s solution or various modified solutions with or without red blood corpuscles does not permit of physiological perfusion of organs. The contact of these salt solutions with the tissue cells will result in profound injury or actual cell destruc- tion. Any deduction made from experiments of this nature must be limited by these facts just outlined. BIBLIOGRAPHY (1) Kerr, Hurwitz anp WuipPpLe: This Journal, 1918, xlvii, 356, 370, 379. (2) Guturie: Arch. Int. Med., 1910, v, 232. (3) Le Gatuots: Experiments on the principle of life (Transl. by N. C. and J. G. Nancrede), Philadelphia, 1813. (4) Kay: Journ. des progrés d. sci. et inst. Médic., 1828, x, 67; xi, 18 (cited by Brown-Séquard, see (6)). (5) L6BELL: Diss. Marburg, 1849 (cited by Jacobj, see (11)). (6) Brown-Sf£Quarp: Journ. de la Physiol. de Homme et des Animaux, 1858, i, 95, 353. (7) Lupwic AND ScumipT: Leipziger Ber., 1868, xx, 12. (8) von ScuréperR: Arch. exper. Path. u. Pharm., 1882, xv, 364. (9) von Fry anp GrusBeEr: Arch. Anat. u. Physiol. (Physiol. Abth.), 1885, 519. (10) Hame.: Zeitschr. Biol., 1888, xxv, 474. (11) Jacops: Arch. exper. Path. u. Pharm., 1890, xxvi, 388. 120 A. E. BELT, H. P, SMITH AND G. H. WHIPPLE (12) Jacosy: Arch. exper. Path. u. Pharm., 1895, xxxvi, 330. (13) Bropre: Journ., Physiol., 1903, xxix, 266. (14) Horrmann: Arch. gesammt. Physiol., 1903, c, 242, 249. (15) RicHarps AND DRINKER: Journ. Pharm. Exper. Therap., 1915, vii, 467. (16) FrrmEpMANN: Biochem. Zeitschr., 1910, xxvii, 87. (17) ManpsE.L: Zeitschr. f. biol. Technik u. Methodik, 1908, i, 44. (18) Kineaspury: Journ. Biol. Chem., 1916, xxviii, 167. (19) Hooxnr: This Journal, 1910. xxvii, 24. (20) Martin: Studies from the Biol. Laby., Johns Hopkins Univ., 1881, ii, 119. Reprint in Memoirs from the Biol. Laby., Johns Hopkins Univ., 1895, ib el bp (21) BAINBRIDGE AND Evans: Journ. Physiol., 1914, xlviii, 278. (22) Hooker: This Journal, 1915, xxviii, 200. (23) WuHrIPPpLE AND Cooke: Journ. Exper. Med., 1917, xxv, 461. (24) WuiePLe, STONE AND BERNHEIM: Journ. Exper. Med., 1913, xvii, 286. (25) Maaenus: Arch. exper. Path. u. Pharm., xlvii, 200. (26) SottmMaNN: This Journal, 1906, xv, 121. (27) Privost anp Dumas: Ann. de Chimie, 1821, xviii, 294. (28) Macenpie: Journ. de Physiol., 1822, ii, 338 (cited by Jullien, Transfusion du Sang, 1875). (29) Macrnopie: Legons sur les Phenoménes Physiques de la Vie, 1838, il. (30) Brscnorr: Arch. f. Anat. Physiol. u. Wissensch. Med., 1835, 347. (31) Gout: Zeitschr. f. rat. Med., 1854, iv, 78. (32) Pour: Arch. gén. de Med., Oct. and Nov., 1854 (cited by Jullien, Trans- fusion du Sang, 1875). (383) Panum: Arch. path. Anat. u. Physiol., 1864, xxix, 241. (34) Ponricx: Arch. f. Path. Anat. u. Physiol., 1875, Ixii, 273. (35) STEVENS AND LEE: Studies from the Biol. Laby., Johns Hopkins Univ., iii, 99 (cited by Pfaff and Vejux-Tyrode, see (36)). (36) Prarr AND VeJux-Tyrope: Arch. f. exper. Path. u. Pharm., 1903, xlix, 324. (37) WHIPPLE, SmitH AND BELT: This Journal, 1920, lii, 72. (38) WHIPPLE AND GoopPastuRE: This Journal, 1914, xxxiii, 50, 70. (39) Hooker: Journ. Pharm. Exper. Therap., 1913, iv, 443. (40) Grupe: Journ. Physiol., 1903, xxix, 276. (41) Coorer: Guy’s Hospital Repts., 1836, i, 457. (42) LasorpeE: Cited by Hayem and Barriere (see (43)). (43) Hayem AND Barripre: Arch. de Physiol., 1887, x, 1. (44) GuTHRIE, Pike AND Stewart: This Journal, 1906, xvii, 344. (45) Munk: Arch. path. Anat. u. Physiol., 1887, evii, 291. (46) GesEeLu: This Journal, 1913, xxxii, 70. THE SEASONAL VARIATION IN THE GROWTH OF BOSTON SCHOOL CHILDREN W. T. PORTER From the Laboratory of Comparative Physiology in the Harvard Medical School Received for publication February 28, 1920 I In 1892 I secured the measurement of weight and height, and other physical dimensions, from 34,500 boys and girls in the public schools of St. Louis. In this investigation, as in similar studies upon which our present standards are chiefly based, all the measurements were col- lected at one time, once for all. The children were then distributed by sex and age, and the median weight and height, etc., determined for each year. This is the ‘‘generalizing’’ method. Its economies are great. Thus, in the St. Louis investigation, a little more than a mil- lion data were harvested in eleven weeks, and work in the class room was interrupted for only seven half-hour periods. Far different is the ‘‘individualizing’’ method. This procedure de- mands the measurement of the same child again and again, throughout its period of growth. The measurements must begin at the earliest age with many thousand children, lest death and desertion so thin the ranks that the survivors will be too few for safe statistical treatment. The individualizing method demands, therefore, a formidable expendi- ture in time and effort through many years. Toilsome as this task may be, it cannot be foregone. For the generalizing method conceals a grave flaw; it does not give the growth of the individual child. With this in mind, I asked the Boston School Committee, in 1909, to measure the height and weight of several thousand of the youngest children and to repeat the measurements monthly in the same children throughout their school life. The measurements were made by the school nurses, under the direction of Dr. T. F. Harrington, and after his lamented death they were continued by his successor, Dr. W. H. 1 This defect in the generalizing method and the contrasting advantages of the individualizing method were discussed by me in the Quarterly Publications of the American Statistical Association, December, 1893. 121 122 W. T. PORTER Devine. In this land of criticism, in which our minds are often occu- pied with the real, and still more often with the imaginary defects of our public institutions, the patient, laborious collection of these meas- urements, month after month, year after year, is a monument of devo- tion in which we may all take pride. II These measurements of the weights and heights of Boston children throughout their school life were completed in June, 1919. The sta- tistical analysis began the following month. The first step was to distribute the weights according to age in months. For example, all boys 110 months of age were placed in one group, and the median weight for that group was calculated. It will be observed that the principle of this first distribution is still that of the generalizing method. This principle must be clearly apprehended. In the generalizing method the individual measurements are distributed into groups, by which distribution the personal character of each individual is lost— the individual becomes merely a statistical unit. Whether the meas- urements so distributed have been made monthly, or yearly, or whether all the children are measured but once and the measurements distrib- uted by years of age, is immaterial; the essential mark of the general- izing method is the loss of personal identity—the transformation of the boy or girl into a statistical unit. In this present investigation the first distribution of boys’ weights, described above, is therefore a distribution by the generalizing method. Its fruits are given in table 1, which records the median weight of boys at each month of age from the 60th to the 176th month, inclusive. In this table, the months are not calendar months, but months of age. For example, the weight opposite 110 months is the median weight of 1226 boys born in 1904, 1905 and 1906. Of these, 33 were born in Oc- tober, 1904; 71 in January, 1905; and 10 in April, 1906. The October boys reached 110 months of age in December, 1913; the January boys in March, 1914; and the April boys in June, 1915. When the data in table 1 are plotted, there results the curve shown in figure 1. This curve of monthly growth is in principle that also ob- tained when the median value is calculated for each year, and these annual values are connected by a line—the curve of growth, old style. These curves rise steadily and smoothly in an unbroken line. Such curves have given rise to much loose thought; for a curve ob- tained by the generalizing method is a statistical and not a personal SEASONAL VARIATION IN GROWTH OF CHILDREN 123 TABLE 1 Weights at each month of age from 60 through 176. Boys born in 1904, 1905 and 1906 MONTHS | POUNDS || MONTHS | POUNDS || MONTHS | POUNDS MONTHS | POUNDS || MONTHS } POUNDS 60 | 42.75 72 | 44.00 84 47.68 96 | 52.48 108 | 57.74 | 41.10 3 | 44.25 85 48.30 7 | 52.96 9 | 58.34 2 | 41.30 4 | 44.35 6 48.78 8 | 53.74 110 | 58.64 3 | 42.00 75 | 45.22 7 49.29 9 | 53.90 1 | 59.51 4 | 41.44 6 | 44.97 8 49.80 100 | 54.41 2 | 60.01 65 | 42.26 7 | 45.65 9 50.09 1 | 54.91 3 | 60.54 6 | 42.31 78 =| 45.67 90 50.56 2 | 55.27 4 | 60.94 7 | 42.06 9 | 46.09 1 50.73 3 | 55.87 115 | 61.58 8 | 41.98 80 | 46.34 2 51.25 4 | 55.83 6 | 61.56 9 | 42.56 1 | 47.03 3 51.32 105 | 56.27 7 | 61.83 70 | 43.17 2 | 47.18 4 51.85 6 | 56.76 8 | 62.36 1 44.53 3 | 47.47 95 52.22 7 | 57.23 9 | 62.95 ~ 120 | 63.42 132 | 69.42 144 75.21 156 | 82.19 168 | 92.75 1 64.22 a | aOree 145 76.23 7 | 83.56 9 | 93.13 2 | 64.96 4 | 70.90 6 76.44 8 | 84.43 170 | 93.88 3 | 65.14 135 | 71.60 7 77.48 9 | 85.94 1 | 97.00 4 | 65.69 6 | 72.85 8 78.17 160 | 86.66 2 | 92.67 125 | 66.46 7, | 7.92 9 78.55 1 | 87.56 3 | 94.00 6 | 66.97 8 | 73.13 150 79.82 2 | 88.47 4 | 96.50 7 | 67.49 9 | 73.67 1 80.24 3 | 89.44 175 | 96.00 8 | 67.67 140 | 73.72 2 80.22 BOO FES 6 | 95.50 9 | 68.31 1 | 74.50 3 80.96 165 | 91.58 130 | 68.49 2 | 74.43 4 80.98 6 | 92.25 1 68.97 3 | 74.90 155 81.53 | % et to 60 70 80 90 100 110 120.- ° 130 140 150 160 170 Fig. 1. The median weight of boys at each month of age from the 60th to the 176th month, inclusive (table 1). In this figure the months are not calendar months but months of age. The ordinates are pounds. 124 W. T. PORTER phenomenon. This will be clear, if we consider the personal history of two boys, whom we will call John and James. At age six, John is 30 per cent above, and James is 30 per cent below the average. But John’s family fall upon evil days; illness and poverty pursue them. The family of his comrade prosper; James spends his vacations in the country, where he has a horse to ride; fresh air and good food do their accustomed work. At age sixteen, James and John have exchanged places; James has risen to 30 per cent above the average and John has fallen to 30 per cent below it. Their personal fortunes have altered; but they are still the same distance from the average. The inhuman average is unmoved by the deplorable fate of John and the happy success of James—to the statistical average, John and James are not persons, but statistical units. Thus the generalizing method per se gives no accurate information as to the growth of the individual child. Nor can this defect be removed by measuring repeatedly the same children throughout their period of growth, either monthly or at any other interval, so long as the resulting measurements are treated as statistical units. Thus, table 1 and figure 1 give accurately the monthly increase in the statistical median value, and they give with accuracy the relation between the size of any individual child and the size of other children of the same age; but they do not give, with certainty, the increase in weight of any child. They are not standards of growth, but merely standards of relative size. IIL The extinction of the individual is not the only indictment which can be brought against the generalizing method. This method, still so much employed, is blind to the possibility of seasonal growth. Consider the case of Hyman Katz, drawn at random from the mass of children whose growth histories are now before us. Hyman Katz was born in August, 1905. In his ninth year, he gained four pounds between September and January, whereas he did not gain any weight, to speak of, between February and June. The case of Hyman Katz is the starting point of a series of interesting and important observations. It will be seen that Hyman Katz is not an exception, save that he is a somewhat extreme illustration of a law binding on other boys and girls. Examine table 3 which deals with all boys and girls born in August, 1905. Again the growth in weight is much larger in the second half SEASONAL VARIATION IN GROWTH OF CHILDREN 125 TABLE 2 Growth in pounds of Hyman Katz 1914-16 MONTH WEIGHT ere MONTH WEIGHT eames 1914 SESE pense || © TRE RIA eomess P edaas” September 56.50 January _ 60.50 0.37 October 57.00 0.50 February 60.50 0 November 58.25 1.25 March 60.00 —0.50 December 60.13 1.88 April 60.00 0 May 60.50 0.50 June 61.00 0.50 TABLE 3 Gain in weight of boys and girls born in August, 1905 BOYS GIRLS YEAR gpa to February to June a to February to June pounds pounds pounds pounds 1912-13 3.34 +0.75 1.88 0.12 1913-14 2.17 —0.17 2.50 0.65 1914-15 2.85 +0.67 2.96 0.13 1915-16 1.29 +0.81 1.63 0.71 1917-18 5.85 —0.07 3.19 0.81 1918-19 4.88 +2.90 4.63 1.83 Average...... 3.40 0.82 2.79 0.71 ar SA ay ooh 22) B30 BOG Sheets Ble» 0.82 1 0.71 1 The year 1916-17 is omitted because the schools were closed in September, 1916, on account of an epidemic. 126 W. T. PORTER of the year than in the first half. This conclusion is supported by the examination of the weights of boys born in all the months of 1905, when these weights are distributed by the months of the year. In table 4 the median weight of all boys born in 1905 is given for each month of the year, and in table 5, the increase in the median weight is recorded. Seasonal growth is again demonstrated. The data in table 4 are reproduced in figure 2. 90 80 70 60 50 40 1911 1912 1913 1914 1915 1916 1917 1918 1919 Fig. 2. The weights of boys born in 1905, distributed by months of the year. The ordinates are pounds. (See table 4.) Fig. 2 is the true curve of growth in weight; it shows the seasonal variations. Compare this figure with Fig. 1, the curve of growth “old style,’ in which the seasonal variations are lost. We have deait thus far with the absolute increase in weight. It is desirable to examine also the percentile increase, i.e., the absolute in- crease of each month divided by the weight at the beginning of the month. As would be expected, the seasonal variation again appears. The average total percentile gain in the first five months of the years 1913, 1914, 1917 and 1918 is 1.89; whereas in the last five months of those years, the average total percentile gain is 6.61. SEASONAL VARIATION IN GROWTH OF CHILDREN 127 TABLE 4 The weights of boys born in 1905, distributed by months of the year YEAR JANUARY FEBRUARY MARCH APRIL JUNE SEPTEMBER OCTOBER NOVEMBER DECEMBER > < = pounds | pounds | pounds | pounds | pounds | pounds | pounds | pounds | pounds | pounds 1911 43.00) 42.23) 42.93) 42.83) 42.48) 42.46) 46.20) 45.04) 45.21) 45.31 1912 46.82) 46.70) 46.62) 47.03) 46.99) 47.03) 48.98) 49.22) 49.98) 50.38 1913 50.94) 51.49) 51.73) 52.18) 52.05) 51.97) 53.31] 54.16) 55.00) 55.37 1914 56.45) 56.73) 57.19) 57.38) 56.99) 56.82) 58.34) 59.38) 60.05) 60.79 1915 61.50) 62.08) 62.38) 62.48) 62.50) 62.72) 64.94) 65.02) 65.66) 66.47 1916 67.09) 67.88) 68.34) 68.58) 68.32) 68.53 71.50} 72.18) 72.89 1917 73.39) 73.81) 74.21) 74.43) 74.81) 74.79) 76.49) 77.32) 78.27) 79.23 1918 80.17} 79.88) 81.13) 81.38) 80.85) 81.06) 84.29) 87.00) 87.13) 88.00 1919 89.93) 90.27) 91.17) 92.05) 91.83) 91.55 In table 4, the median weight of all boys born in 1905 is given for each month of the year without regard to the month of age. The seasonal growth is thereby shown. Thus, in the five months from January to June, 1914, the gain was less than 4 pound, but in the five months from June to November, 1914, the gain was 3} pounds. TABLE 5 The monthly increase in weight of boys born in 1905 Fe) a e YEAR ce <5 ae é 8 ea 25 - ae cE pa ae 8 am a a & ee oF BS Bz SS PC ie bine) Seas el nse elias Shea 8 1911 —0.77|+0.70) —0.10) —0.35| —0.02|+3.74 +0.17|+0.10)+1.51 1912 —0.12) —0.08|+0.41) —0.04)+0.04|+1.95) +0.24!+0.76|+0.40)/+0.56 1913 +0.55) +0. 24) +0. 45) —0. 13) —0.08)+1.34|+0.85)+0.84)+0.37|+1.08 1914 +0.28)+0.46)-+0.19) —0.39| —0.17|+1.42/+1.04|+0.67/+0.74|+0.71 1915 +0.58)+0.30)+0. 10) +0.02)+0.22|+2.22)+-0.08)/+0.64 +0.81/+0.62 1916 +0.79| +0. 46) +0. 24) —0.26)/+0.21 +0.68)+0.71)+-0.50 1917 +0.42|-+0.40)+0.22|+0.38| —0.02)/+1.70)/+0.83)+0.95) +0. 96) +0.94 1918 —0.29) +1.25]+0.25) —0.53)-+-0.21)+3. 23) +2.71)+0. 13) +0.87/+1.93 Average ...|-+0.18)+0.47|-+-0.22/—0.16)+0.05 +2.23)+0.96 +0.61/+0.63/-+0.98 For convenience we may divide the total of 2.23 pounds from June to Sep- tember into 0.74 pound for each of the three months. 128 WwW. T. PORTER The errors caused by neglecting the seasonal growth are strikingly shown by comparing a series of weights chosen from table 1 and table 4. Let us take from table 1, twelve weights, beginning with 56.76 pounds, and from table 4, twelve weights, beginning with 56.73 pounds. FROM TABLE 1—BOYS’ WEIGHTS DISTRIBUTED BY | FROM TABLE 4—BOYS’ WEIGHTS DISTRIBUTED BY MONTHS OF AGE MONTHS OF YEAR pounds pounds 56.76 56.73 57.23 57.19 57.74 57.38 58.34 56.99 58.64 56.82 59.51 32 60.01 A 60.54 58.34 60.94 59.38 61.58 60.05 61.56 60.79 61.83 61.50 * July and August; the vacation months. In these two series, the beginning and the end weights are almost the same, but in the middle of the twelve months the divergence is 2 pounds or more. 62 61 60 59 58 57 56 0 1 2 3 4 5 6 7 8 9 10 11 12 Fig. 3. The ordinates are pounds and the abscissae are months. The lower curve records boys’ weights distributed by months of the year (table 4); the upper curve records boys’ weights distributed by months of age (table 1). SEASONAL VARIATION IN GROWTH OF CHILDREN 129 In figure 3, the two series are shown graphically. The enquirer will wish to know how the marked seasonal variation in weight is so completely masked in table 1 and figure 1, and in the upper curve of figure 3, examples of the statistical method hitherto in use. The answer is that in the method hitherto in use—in which the only criterion is the month of age—the month of age for half the boys will fall in a season of rapid growth, while for the other half the same month of age will fall in a season of slow growth. Compare during the months of age from 106 to 118 the growth in weight of boys born in August, 1905, with that of boys born in February, 1905 (table 6). TABLE 6 AGE IN SEASON | weronn | AGE IN SEASON wErGnT Boys born in August, 1905 pounds pounds 112 December, 1914 59.96 118 June, 1915 61.56 106 June, 1914 55.25 112 December, 1914 59.96 1.60 Boys born in February, 1905 112 | June, 1914 55.75 118 December, 1914 61.17 106 December, 1913 54.50 112 June, 1914 55.75 1.25 | 5.42 It is clear that if the age in months is alone considered, the period of rapid growth in boys born in February will coincide with the period of slower growth in boys born in August. The two periods will compen- sate each other if the February and the August boys are treated en masse, as in table 1 and figure 1, and in the standards hitherto so largely used to determine the growth of school children.? LW Rather than accept a dictum so far-reaching, the reader may here suggest that my entire contention rests on a palpable error—a failure to take into account the heavier garments worn in winter. Table 5 2My data do not yet show any seasonal variation in the height of school children. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 1 130 W. T. PORTER shows a net increase of 0.76 pounds in the first five months of the year (+0.18, 0.47, 0.22, 0.05, —0.16), whereas the increase in the last five months of the year is 3.92 pounds (+0.74, 0.96, 0.61, 0.63, 0.98). But it will be urged that in May, which falls in the first five months, the winter clothing is laid aside; and in October, which falls in the last five months, it is again put on. A careful examination of the data will show that the seasonal dif- ferences in growth cannot be explained by seasonal differences in the weight of garments worn. Consider the following groups, taken from table 5. INCREASE FROM INCREASE FROM INCREASE FROM pounds pounds pounds January-February 0.18 | June-September. 2.23 | October-November..... 0.61 February-March. 0.47 November—December. .. 0.63 March-—April...... 0.22 December—January...... 0.98 Average....... 0.29 0.74 0.74 Obviously, the second group is composed of summer months and the third group of winter months. The first group requires, perhaps, a word of explanation. In Boston the month of April offers the promise but not the reality of spring. The average temperature rises from 40° on April 1 to 51° on April 30. In 1919, the maximum temperature on April 16 was 42° and on April 25 it was 38°. In only three days of April was the maximum temperature above 61°. It is unlikely that materially lighter clothing is worn during April. For our present purpose April, therefore, should be classed with the winter months. The figures just presented indicate that clothing is, relatively, a negligible factor. Were clothing an important factor, the weights for June to September would show a decrease, as compared with those for January to April, and the weights for October to January would show a considerable increase as compared with those for June to September. No such fluctuations are apparent. Observe now the figures from March to June (table 5). Increase in weight Pounds March ‘to April. 3 co cciic wis oe vs oe cele e en canoe ice eee +0.22 April to May ii ocdede ies ints Heer Ra ee —0.16 May to. June. . isd ss ced s i ee ee +0.05 SEASONAL VARIATION IN GROWTH OF CHILDREN 1351 If a change from winter to summer clothing be invoked to explain the low average of growth from January to April, just discussed, it cannot be used again to explain the failure to grow from April to May, and a third time to explain the failure to grow from May to June. It will doubtless be found, taking the average of eight or nine years, that the change from winter to summer clothing is made within a short period and bears a definite relation to the temperature curve. Finally, too much emphasis must not be laid upon hypothetical dif- ferences in summer and winter clothing. Careful enquiry shows that the indoor clothing of Boston public school children does not change during the school year as much as might be supposed. Such changes cannot explain away the periodic growth demonstrated in this investi- gation. With every reasonable allowance for error, it seems impos- sible to deny a seasonal variation in the weights of school children. We In figure 2 the seasonal curve departs from the present standard curve by fully 2 pounds, a difference almost three times the average total growth for the first five months of that year. Such deviations justify two deductions: 1. To determine the normal growth in weight, the child must be weighed once a month, or oftener. If the child is weighed less often, the seasonal variation will be missed. 2. True curves of growth in weight demand that the monthly weights be distributed according to the months of the year, and not according to the months or years of age, as is the present custom.’ 3 A substantial part of the cost of this study has been paid by the Permanent Charity Fund, whose assistance is gratefully acknowledged. Iam much indebted also to the Department of Education in Harvard University for the loan of Mr. L. A. Maverick during the summer of 1919. CERTAIN CHANGES NOTED IN ERGOGRAPHIC RESPONSE AS A RESULT OF TOBACCO-SMOKING THEOPHILE RAPHAEL From the Physiological Laboratory of the University of Michigan Medical School, Ann Arbor, Michigan Received for publication March 1, 1920 This report is based upon material secured in the course of an investi- gation undertaken with a view to ascertaining the effect of tobacco- smoking upon voluntary muscular work, as determined through ergo- graphic observation. The study was suggested by Lombard’s (1), (2) findings and the procedure followed was essentially that utilized by him (q.v.). In the summer of 1915 when, for a month, I had the opportunity of association with Lombard in connection with an ergographic study (3) upon which he was engaged at the time, I noted that my own tracings were uniformly free from those periodic recoveries, after the first loss of power, which were consistently characteristic of Lombard’s records (1). These recoveries, or returns of power, failed to manifest them- selves in my curves in spite of any subjective awareness, or even de- sire, and continued physical endeavor. The records showed through- out only a gradual falling off of performance registry (fig. 1) from maximal contraction to zero. After smoking, Lombard noted that the initial loss of power came more quickly, to be followed, however, as in his non-smoking records, by the characteristic periods. In the spring of 1917 I found it possible, personally, to undertake a short study of the effect of smoking upon ergographic response, expect- ing to encounter, if anything, only a diminution in total amount of work performed. Greatly to my surprise, I noted the consistent appearance of periods practically identical with those characteristic of Lombard’s smoking and non-smoking records. Since that time it has been possible to elaborate the results somewhat and, in spite of their apparent slen- derness, they were deemed worthy of report. Prior to the time of undertaking this study, I had been only an occasional smoker, just able to tolerate one or two mild cigarettes with- 132 EFFECT OF TOBACCO ON ERGOGRAPHIC RESPONSE 133 out suffering any of the acute tobacco effects. The specific procedure followed in this work was, as indicated, that employed by Lombard (1). The ergographic records were taken every second hour, daily, from about approximately 9 a.m. to 7 p.m., care being taken to look away from the record as it was being taken and, as far as possible, to avoid 11 A.M. Smoked 9 A.M. 1 P.M. Smoked 3 P.M. Smoked 5 P.M. (Odea & Fies. 1-6 all suggestive influence. After a week of preliminary work to secure adaptation to the schedule and to overcome initial practice effect, the actual smoking records were begun; i.e., a cigarette of medium strength was smoked just prior to the moment of beginning a record, at the ex- piration of the two-hour interval. - Effort was suspended after the first loss of power. 134 T. RAPHAEL At first, aside from an apparently marked decrease in the extent of the record, no essential change was noted. On the third day of the test, however, it was found that, if effort were continued after the first loss of power, It was possible, in contrast to the results of previous trials, to continue long beyond this point, the curve then showing distinet per- iodicity (figs. 2, 3 and 4), seemingly identical with that so manifest in Lombard’s curves. This finding, it might be well to note, was extremely unexpected and wholly unlooked for. It was found, in addition, that this effect would persist for some hours without additional smoking (fig. 5), When electrical stimulation was applied directly to the muscles by means of the induction current, essentially as described by Lombard in his Laboratory Manual (4), the weight having been materially dimin- ished, the record showed both with and without the use of tobacco a distinctly unbroken character, marked by no period formation whatso- ever, thus indicating the causal locus of the periodicity to be further central than the muscles or the nerve endings. ‘The same findings were noted on nerve trunk stimulation also, thus in both respects corrobor- ating Lombard’s personal observations. Four other subjects, only one of whom had previously smoked to any degree, were obtained. Although these subjects were untrained and the schedule much less thorough, the findings seemed to bear out in a general way those determinable in my own records. It might be well to indicate, at this point, that these results seem to be at variance with observations made by Hough (5) in tests carried out upon himself, utilizing a somewhat different apparatus. It appears, therefore, that the essential cause of the periodicity is central nervous system fatigue, or depression, induced in this instance by tobacco products absorbed in the act of smoking. On this basis it might not be impossible to account for Lombard’s periodicity, for which, thus far, apparently no satisfactory explanation has been afforded. He reports having already smoked for a number of years when the periods were first noted and has since observed that although, as indicated, the first loss of power comes more quickly on smoking days, the periods always appear sooner or later, and have apparently the same character at all times. In view of the fact that the periods are present constantly, even when smoking had been dropped for a period as long as a month, it is not inconceivable that the long-continued use of tobacco may have, in this case, exercised some more or less permanent change in the sub- stance of the central nervous system, specifically concerned in the mech- EFFECT OF TOBACCO ON ERGOGRAPHIC RESPONSE 135 anism of ergographic response. On the other hand, it is equally pos- sible that the periods might have occurred even without smoking, had the central nervous system been originally of a type especially readily susceptible to fatigue. Lombard is unable to state whether any of his eight subjects, two of whom manifested periods, were smokers. In the same way it is conceivable that other central nervous system depressants may produce similar changes in ergographic response. It is distinctly unfortunate that time and opportunity did not permit more extensive investigation, and with a larger series of subjects, both smokers and non-smokers, and it is primarily in the hope of stimulating further work that this brief note is submitted. SUMMARY On the basis of the results obtained in this study, it appears that under the influence of tobacco-smoking, a distinct periodicity may be demonstrated in the curve of ergographic response, arising apparently as a result of central nervous system fatigue or depression. BIBLIOGRAPHY (1) Lomsparp: Journ. Physiol., 1892, xiii, 1. (2) Lomparp: Amer. Journ. Pyschol., 1890, 1. (8) Lomparp: This Journal, 1916, xl, 132. (4) Lomparp: Laboratory work in physiology, Ann Arbor, 1914. (5) Houcu: This Journal, 1901, v, 240. DETERMINATION OF THE CAPILLARY BLOOD PRESSURE IN MAN WITH THE MICRO-CAPILLARY TONOMETER C. 8. DANZER anp D. R. HOOKER From the Physiological Laboratory of the Johns Hopkins University Received for publication March 8, 1920 Physiologists have long appreciated the prime significance of the capillary circulation and in recent times clinicians are coming more and more to realize that this part of the circulatory. bed is of importance in connection with purely medical problems. The cardio-vascular sys- tem functions for the distribution of the blood but the effective changes pertaining to the nutriment of the tissues occur in the capillary bed. Here the metabolic interchange takes place; food is delivered to and waste is largely withdrawn from the active cells. It is, therefore, im- portant to know the pressure under which the blood is delivered to the capillaries in different conditions of health and disease and to correlate this knowledge with the functional activity in other parts of the vascu- lar bed. At the present time no method is adapted to such a study either in experimental animals or in the human being. The fundamental observations of Roy and Brown (1) in 1878 on the capillary circulation in the frog were never developed for application to the circulation in the mammal because the method was inadequate to this purpose. Quite recently Cannon (2) has emphasized the signif- icance of the capillary bed in traumatic shock, and Dale and Laidlaw (3) and Dale and Richards (4) in histamine shock. These authors con- ceive that in shock the capillary walls are injured by circulating poisons so that they lose tone with the result that a large volume of blood is pooled in the capillary area. This pooling of the blood in the capil- laries accounts for the primary fall of arterial pressure, which is further accentuated by a transudation of plasma through the injured capillary wall so that there results an actual decrease in circulating blood volume. These two factors are regarded as the basal cause of the circulatory phenomena of shock. Lombard (5) showed a number of years ago that if a drop of oil be placed upon the skin it is possible to see the underlying capillaries, and 136 CAPILLARY BLOOD PRESSURE IN MAN 137 he sought with this information to develop a method of determining the pressure of the blood in these vessels. For this purpose he used a small glass chamber, the floor of which was made of gold beaters’ skin with a small hole in the center. The chamber was filled with glycerine and brought into contact with the skin so that it was possible with a microscope to observe through the glass roof and a small opening in the membranous floor the effects which were developed in the under- lying capillaries when the pressure on that area of skin was elevated by forcing more glycerine into the chamber. Although Lombard made a number of observations with this instrument, the technique was so difficult that it was not possible of wide application. With Lombard’s experience in mind, we have developed a method which we believe is applicable to the study of capillary blood pressure both in animals and in man, and the purpose of this paper is to describe our method and to give the results which we have thus far obtained with its use. There are some twenty-five articles in the literature bearing upon the determination of the capillary blood pressure obtained by the use of various methods, and the results reported in these articles vary from 7 mm. to 70.5 mm. Hg. for the normal capillary blood pressure in man. It is obvious that if the results by the several methods vary to this extent, they cannot be considered as trustworthy. It is not surprising, therefore, that Friedenthal (6) in a recent review of the subject, reaches the conclusion that capillary blood pressure determinations can be of little practical significance. They are, in fact, of less value than in- ferential deductions drawn from the values of arterial and venous pres- sure. Inferential deductions, on the other hand, are certainly not free from objections, as is shown by the following. Thus Fick (7) thought that pressure in the capillaries is almost as high as in the arteries, and that the greatest fali in pressure occurs in the small veins. His conclusions are based on theoretical considerations. Following Poisselle it is generally believed that the principal loss of energy occurs in the capillaries because the blood channel is narrowest in the capillary area. On the other hand Campbell (8) has assumed that the greatest fall in pressure occurs in the small arteries and Levy (9) came to the same conclusion as the result of mathematical calcu- lations. Goldmann (10) believes that the principal loss of energy occurs in the arteries. The lumina of the smallest arteries are not much greater than those of the capillaries. The velocity is much greater in the former, however. The result is the great fall of pressure in the small arteries. 138 Cc. S. DANZER AND D. R. HOOKER Bargolomez (11) has thrown some light on this important question by the following experiments. He cannulated the smallest possible branches of the arteries (middle and posterior auricular arteries) and also the corresponding venous tributaries, and recorded their pressures monometrically. To facilitate the work he dilated the vessels (appli- cation of heat) and introduced needle cannulas. Then he allowed the vessels to return to their original caliber before starting his experiment. Thus he found that the normal blood pressure fall in the capillaries is very slight (about 4 mm. Hg. in the rabbit’s ear). He concludes that the greatest fall in blood pressure (90 per cent) occurs normally in the arteries of moderate caliber. It is perhaps as striking to the reader as to ourselves that a trust- worthy method of actually determining the pressure in the capillaries is wanting. Let us for a moment review the previous methods em- ployed for this purpose, The majority of them rest upon the principle, introduced by von Kries (12), that the paling of the skin is dependent upon the pressure and amount of blood in the superficial capillaries. von Kries employed a small glass plate of known size and applied it to the surface of the skin and the amount of pressure necessary to cause the skin to pale was interpreted as a criterion of the capillary blood pressure. We shall show that this criterion is inadequate in that the paling of the skin is not necessarily accompanied by a cessation of flow in the capillaries and we believe that it is due rather to an empty- ing of one or more of the venous plexuses lying in the dermis. Spalteholtz (13) has published drawings of a reconstruction model of the blood vessels of the skin. These show that there are three venous plexuses in the cutis and one in the subcuticular layer. We believe that as increasing degrees of pressure are applied to the skin, these plexuses are successively more or less emptied of their contained blood with the result that the skin develops varying degrees of pallor. Our observations make it clear that even extreme grades of pallor are not necessarily associated with a cessation of blood flow in the super- ficial capillaries. It follows, therefore, that pallor of the skin cannot properly be regarded as evidence of capil collapse. The principle of skin pallor underlying the method of von Kries was likewise employed by von Basch (14), by von Recklinghausen (15) and by Basler (16). von Basch sealed a small glass capsule on to the skin and determined the pressure necessary to cause paling of the under- lying area. CAPILLARY BLOOD PRESSURE IN MAN 139 von Recklinghausen’s method was essentially the same. It is that which he employed for the determination of the venous blood pressure. Basler used an instrument which he called an ochrometer; this consisted of two closed chambers with glass roofs and floors of gold beaters’ skin which was sufficiently translucent to indicate any changes in color in the underlying skin. Two fingers were used, one to observe the degree of pallor and the other for the control. The microscopic fields were brought up to a single eye piece by means of prisms. One of the chambers was then inflated, causing the membranous floor to press upon the skin until the first evidence of pallor, as determined by the control finger, was observed. von Kries, using his method, gives as the normal capillary pressure 37.7 mm. Hg. This determination was made when the hand was 490 mm. below the crown, a position which we assumed to be approximately at heart level. He likewise reports the capillary blood pressure in the ear as being 20 mm. Hg., and in the mucous membrane of the rabbit’s mouth as 33 mm. Hg. Hough and Ballantyne (17) employed the method of von Kries to study the effect of temperature on the capillary blood pressure. Their readings for the normal ranged from 40 to 50 mm. Hg. at a room tem- perature of 20-21°C. When the temperature of the surrounding air was reduced to 6°C., the capillary pressure was 65 mm. Hg., and when the temperature was raised to 26°, it was from 50 to55mm. Hg. From our own observations we judge that the temperature effects noted by these authors are associated not with actual changes in capillary blood pressure but rather with the behavior of the superficial venous plexus. We believe a significant factor contributing to the pallor of the skin is a decreased amount of blood in this plexus. If the latter is relatively empty and the skin is correspondingly pale, it will require a greater skin pressure to develop a further paling. The paling under these cir- cumstances would then be due to the pressure emptying the deeper- lying plexuses. The resistance to compression of the skin vessels in- creases as the vessels compressed lie deeper and deeper in the several layers of the skin. It follows, therefore, that when the skin is cold and consequently pale, a greater pressure will be required to produce a further paling than would be the case if the skin were of a normal color. Hough and Ballantyne also report a heightened capillary pressure asso- ciated with arterial constriction, hence they thought that capillary pres- sure was dependent on some other factor than that of the arterial tone. These findings like their findings in connection with the effects of tem- 140 C. S. DANZER AND D. R. HOOKER perature do not agree with our results, and we believe that the latter, as in the case of the former, are associated with the inadequacy of the criterion employed. Natanson (18) has published two papers using the von Kries method. He took for his standard the complete blanching of the skin and studied, among other things, the effect of mass compression of the arm on the capillary pressure in the hand. Using the criterion of complete blanch- ing of the skin, he found the normal capillary pressure to be 70.5 mm. Hg. The highest capillary pressures were observed when the con- stricting band exerted a pressure of 42.5 mm. Hg. When the constrict- ing pressure was raised to 52.3 mm. Hg., capillary pressure fell. He assumes that at this pressure both the arteries and veins were being compressed and draws the conclusion that the capillaries cannot be filled with blood except under the influence of arterial pressure. If the latter fails, the capillaries become bloodless and collapse through the effect of a relative rise of tissue tension. We have not been able to confirm the latter observation, and believe that here again the inade- quacy of the criterion employed by Natanson is sufficient to explain his results. Schiller (19) and also Rotermund (20) made use of von Fries’s method slightly modified by the addition of Fick’s ophthalmotonometer (21). This addition served merely to make the pressure readings easy since they could be read off on a scale calibrated against a spring. Schiller found that the highest capillary pressure (about 40 mm. Hg.) occurred when the temperature of water applied to the skin was approximately that of the skin, namely, 35°C. Rotermund, using the same method, found the capillary pressure on the skin of the forehead when the subject was in a recumbent position, averaged about 26.8 mm. Hg. He also applied this method to study the influence of age, nutritional state, dyspnea, nephritis and arterio- sclerosis on capillary blood pressure. von Basch applied his glass capsule method to a study of the capillary blood pressure in human beings and upon experimental animals. The criterion was essentially the same as that employed in the von Kries method. He found that the capillary blood pressure in the rabbit’s ear did not differ materially from that of human subjects. In the rabbit’s ear the pressure ranged from 21 to 25 mm. Hg., and in the healthy human being it was between 25 and 30 mm. Hg. In human beings he observed that a low capillary blood pressure was frequently associated with high arterial tension, but at times the converse of this CAPILLARY BLOOD PRESSURE IN MAN 141 was true, namely, that a high capillary pressure was associated with a low arterial tension. His inference from these observations led him to the belief that the capillary pressure is independent of the arterial pressure. When the chest of an experimental animal was compressed, there occurred a rise in capillary pressure accompanied by a moderate fall in arterial pressure. He interpreted this result as indicating that the rise of capillary pressure followed venous stasis. Similarly in man chest compression produced a moderate rise of capillary pressure with a corresponding fall of arterial pressure. The fact that the capillary pressure rose less in man than in the experimental animal suggested that the venous stasis was less extreme. He also cut the cervical sym- pathetic nerve in the rabbit and found an increase in capillary pressure associated with the slight fall of arterial pressure. This pointed to a dilatation of the arterioles of the rabbit’s ear and was confirmed by direct inspection. The injection of strychnine caused a rise of arterial pressure with a fall of capillary pressure; in other words, the converse of the above experiment. This was interpreted to mean that a constriction of the arterioles had occurred. von Basch formulated the following hypothesis: that the degree of capillary pressure rise may serve to differentiate between venous stasis and arteriolar dilatation. While the former raises capillary pressure markedly without altering arterial blood pressure, a condition of ar- teriolar dilatation produces a moderate increase of capillary pressure with a corresponding fall in arterial blood pressure. He applied these results clinically and obtained data which he regarded as of distinct value in both diagnosis and treatment. Using Hooker’s (22) capsule with the criterion of paling, Briscoe (23), in a study of the Raynaud phenomena, in cases of “irritable heart,” found that when the hand was cyanotic the capillary pressure was 36.9 em. of water as compared with the normal controls which gave a pres- sure of 23.5 cm. of water. In individuals subject to vasomotor changes when the color of the hand was normal, the capillary pressure was 25.3 em. of water, and that when the same hand was blue the capillary pres- sure was increased to 33.3 cm. of water. von Recklinghausen gives as the normal capillary pressure in the finger tip with the hand at heart level, a reading of 52.5 mm. Hg. Basler’s ochrometer has been used by Basler himself and by a number of other observers. The normal capillary pressure as given by Basler is about 7mm. Hg. Goldmann, using this method, concludes 142 Cc. S. DANZER AND D. R. HOOKER that the normal capillary pressure is about 8.5 mm. Hg. A moderate rise in external temperature causes no appreciable change in the cap- illary pressure, but if the external temperature be raised 10°C. or more, the capillary pressure is increased. Landerer (24), using Basler’s ochrometer, found that the normal capillary blood pressure ranged from 17 to 25 mm. Hg. During a cold bath the capillary pressure fell while the arterial pressure rose. In a warm bath the capillary pressure was unchanged while the arterial pressure fell. Landerer, Krauss (25), Friedenthal and others have used this method in clinical cases. Krauss studied the capillary pressure in circulatory disturbances (valvular disease, myocardial disease, venous stasis), pulmonary dis- eases (emphysema, chronic bronchitis, tuberculosis), status asthenicus, severe anemia, carcinoma and cachexia. Besides the ochrometer he also used an apparatus of his own (a microscopic method) and also Weiss’ blood method. He obtained some very interesting results. In brief, they confirmed the older observations of von Basch and Roter- mund and agreed with those of Landerer. The foregoing methods are all based upon the principle of paling of the skin originally introduced by von Kries. A decidedly different method has been employed by Basler and by Weiss. Weiss’s (26) method consisted of pricking the skin and observing the pressure necessary to stop the exudation of blood through the wound. The finger tip was enclosed in a chamber containing fluid in which the pressure could be raised. It was necessary, of course, to insure that bleeding continue after the pressure was removed. Basler (27) pricked the volar surface of the middle finger with a needle; he then sealed a piece of rubber tubing onto the finger (by apply- ing a hot rod against the edge of the rubber). This formed a little chamber in the center of which was a bleeding spot. The chamber was filled with hirudinized saline solution, and closed with a roof. Now the pressure of the exuded blood was registered manometrically. His re- corder was of the lever type and rather delicate. He named this apparatus the “‘Hautmanometer.’”’ The determinations made with this apparatus agreed quite closely with those of the ochrometer. Both of these blood methods are faulty because it is not possible to know the depth of the wound, since the skin thickness is so variable and consequently the size or the depth of the vessels punctured is a matter of chance. CAPILLARY BLOOD PRESSURE IN MAN 143 The theoretical objections to the various methods above discussed together with the fact emphasized by Friedenthal that the readings obtained by these methods vary by more than 100 per cent indicate clearly that they have little or no practical application. In order to be sure that the readings accurately represent the pressure in the capil- laries themselves, a criterion similar to that employed by Roy and Brown, namely, the visual determination of the point at which the capillary flow ceases, is essential. Using a tonometer similar to that which we have employed and the criterion of cessation of corpuscular flow, these authors obtained pressure readings in the capillaries and venules of the web of the frog’s foot of 7.3 to 11 mm. Hg. Lapinski (28) employed the Roy-Brown method to study the effect of nerve section on the capillary pressure in the frog. He obtained no very clear-cut results; however, his determinations of normal capil- lary blood pressure in small frogs was between 15 and 44 mm. Hg. and in large frogs between 30 and 60 mm. Hg. Natanson, in his studies on the effect of mass ligature, controlled his observations on man by experiments on the frog. He found the normal capillary pressure varied from 12 to 24 mm. Hg. in the different capil- laries of the frog’s web. That such a criterion may be applied in the case of the human sub- ject is indicated by the observation of Weiss (29) who, using oil on the skin as suggested by Lombard, first observed corpuscular flow in the human capillaries. We have developed an instrument applicable to man and experimental animals which permits of the accurate use of this criterion. If a drop of oil be placed on the skin of the hand and the area so treated be examined with the microscope under a strong light, as, for example, a 40 Watt electric bulb placed close to the microscope objec- tive, one may readily see the capillary tufts scattered through the field. Most of these tufts come up from the deeper parts of the dermis, fold over and return so that only a small part of the capillary loop can be brought into focus. Under such conditions it is scarcely ever possible to see the corpuscular flow. Occasionally a capillary may be found which, arising from the depths of the dermis, turns at right angles and takes an horizontal course for a short distance, and in such a case, under favorable conditions, the corpuscular flow may be observed. Such capillaries are so rare, however, on the general surface of the skin, that they do not serve for pressure determinations. In the skin overlying the matrix of the finger nail however, many capillaries take such an 144 Cc. S. DANZER AND D. R. HOOKER horizontal course and in a single microscopic field a number of these vessels may be brought into focus at one time. Here one frequently sees corpuscular flow without offering any resistance to the blood stream. If a constricting pressure be applied to the upper arm the corpuscular stream may be seen in many of the capillaries without difficulty. A drawing of the picture seen is given in figure 1. The sharply outlined capillaries are seen as red threads lying on a pink ground. The picture is very much clarified if the skin be first scrubbed with soap and water and afterwards thoroughly dried. This procedure removes the loose epidermis and softens the tissues. The methods based on this principle introduced by Lombard are the Krauss method, Basler’s kapillar-tonometer (30) and Kylin’s method (31). In Krauss’s method the capillaries are visualized through a simple lens (magnifying 10). Krauss takes as his criterion the disappear- ance of the capillaries. We have employed such a method but have not succeeded in making the capillaries disappear at pressures correspond- ing to those of Krauss. We believe that the paling of the skin (which occurs at very low pressures) and the minute size of the capillaries as seen at such a low magnification may give the impression of a dis- appearance of the capillaries when even slight pressure is applied. When controlling this method with our own (in which a magnification of at least 68 is used) we have been able to see distinctly capillaries which were hazy and almost invisible with a magnification of ten. Concerning Basler’s tonometer we may say that except for the fact that the gold beaters’ skin has no circular opening in its center, it is almost identical with that of Lombard. The glycerine reservoir and the fact that the chamber is filled with glycerine, is common to both methods. An added difficulty in the method of Basler may be the collection of air bubbles within the chamber which will disturb the clarity of vision. In our experiments in the early part of 1919 we used chambers which were almost the counterpart of Basler’s, but have found them difficult to work with and for reasons of practicability have discarded them. Basler’s criterion is the disappearance of the capillaries. We believe this to be incorrect in principle and difficult of execution in the majority of instances. In the publication dealing with this method he gives no results obtained by its use. Kylin’s method, which was presented at the Ninth Nordiske Congress of Internal Medicine held at Copenhagen, August, 1919, depends on the visualization of the capillaries. Unfortunately we have been unable to CAPILLARY BLOOD PRESSURE IN MAN 145 obtain any exact information about his method, his criterion or his results. Fig. 1. A drawing of the finger tip to show the area of skin at the base of the finger nail which is used in the determination of the capillary blood pressure. The circle overlying this area represents the miscroscopic field when the skin is observed under a drop of oil with the aid of a strong light. The capillaries are slightly diagrammatic: a single focal plane will not bring them all into focus at one time. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 1 146 C. S. DANZER AND D. R. HOOKER The micro-capillary tonometer. The instrument which we have used is shown in figure 2. It consists essentially of two adjusting devices. Screw / permits of raising and lowering the finger rest so that the area of skin under observation may be brought into an horizontal plane. Screw 2 permits of adjusting the pressure capsule in suitable contact with the skin. The finger rests on plate 3, which rocks so that the finger tip may assume a comfortable position. The instrument is placed on a microscope stage and when the forearm is supported with a comfortable rest, the subject is sufficiently comfortable, so that there a) 4 “Zi 7 VF 9 \\ Fig. 2. The micro-capillary tonometer The instrument is of such a size that it rests on the microscope stage. Description in text. is no movement or tremor of the finger. The pressure capsule is shown in the insert figure, which represents a cross section along the line AB. It consists of a brass ring, 4, into the upper surface of which is sealed a thin glass plate, 5. .Through the latter passes tube 6, which serves for the inflow and egress of air. To the lower surface of plate 4 is attached a brass ring, 7, threaded on the outside to receive the collar, 8. The latter screws air-tight against the washer, 9. Over the brass collar 8 is tied the gold beaters’ skin, 10. The metal tube, 6, is con- nected through a rubber tube with a rubber ball of about 200 ee. capacity, which may be compressed between two plates by means of CAPILLARY BLOOD PRESSURE IN MAN 147 an adjusted screw, a device similar to that which is ordinarily used for calibrating the Hiirthle manometer. A bypass on the tube connecting the capsule with the rubber bulb leads to a single arm mercury man- ometer, by means of which the pressure in the chamber may be determined. After the finger has been brought to a comfortable position on the finger-rest with the area for study in the horizontal plane, the pressure capsule is lowered until the gold beaters’ skin comes softly in contact with the upper surface of the finger. The gold beaters’ skin is tied on suf- ficiently loosely so that when the air is forced into the capsule it exerts pressure upon the skin without loss of pressure due to the tension on the membrane. Preparation of membrane. 'The successful use of this device depends upon obtaining gold beaters’ skin which is free of porous openings, is entirely soft and pliable and is transparent. We have found that gold beaters’ skin prepared according to the following directions fulfills these requirements. ‘The membrane is first washed in tepid water to remove all powder and dust which may be adherent to the surface. This cleansing is facilitated by rubbing both surfaces of the membrane with the ball of the finger or with a piece of absorbent cotton. It is then rinsed several times in clean water, after which it is ready for use. The brass collar, 8, in figure 2 is now removed from the instrument and screwed down over an obturator which, presenting through the collar, has an oval contour such that when the membrane is tied into position it will le somewhat as shown in the detail. With the collar in position on the obturator the membrane is laid over the top and tied securely with several turns of silk thread. The collar is now returned to the instrument, when it may be tested to see whether or not the attached membrane is perfect. If it stands a pressure of 50 mm. Hg. without leak, it is suitable for further treatment. The collar with the attached membrane is now drained free of water and placed in pure glycerine for 24 hours, after which it is removed, drained free of glycerine and placed in castor oil. The preparation should be left in castor oil for at least 24 hours and may be left in the oil indefinitely. It is thus possible to prepare a number of mem- branes at one time and to keep them in castor oil until such time as they may be required. When a membrane is required for use, the ring to which it is attached is removed from the oil and screwed into position on the instrument, a test being again made to be sure that the chamber is air-tight. If a THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, No. 1 148 C. S. DANZER AND D. R. HOOKER leakage occurs at this point it is probably due to an imperfection in the washer. If now the adjustment to the finger be made and the area of skin beneath the member be observed with the microscope, it will be seen that the membrane is entirely transparent so that there is no difficulty in observing the underlying capillaries. Failure to see the capillary field plainly indicates that the membrane is clouded, in which case it should be discarded. This condition is not likely to be found when using a new membrane. A membrane, however, which has been used repeatedly and exposed to air and dirt is lable in the course of time (weeks) to deteriorate and if there is any difficulty in obtaining clear vision of the field it is probable that the transparency of the mem- brane is at fault. When the instrument is not in constant use it is advisable to return the membrane to castor oil. With reasonable care a membrane should last indefinitely. A second technical requirement is that the skin should be clean and free from moisture. It should be scrubbed lightly with soap and water and thoroughly dried. The oil on the skin which makes the underlying capillaries visible probably serves by its intimate penetration of the epidermal layer to do away with light reflection from the uneven sur- face. If then moisture intervenes at the oil-skin boundary, the field is less distinct. It sometimes happens in prolonged observations that the clarity of the field is lost. This may be due to the excretion of sweat. If the oil be wiped off and the finger be thoroughly dried a fresh drop of oil will make the field as clear as before. In using the instrument we have found it desirable to employ a microscope giving a magnification of approximately 70. The essen- tial point is to magnify enough to readily visualize the movement of the red cells when the stream is slowed without undue loss of definition and depth of focus. Leitz objective 3 and ocular 1 fulfill these require- ments. ‘To use such a lens combination with a rather short working distance, the skin with the overlying oiled membrane must be brought quite close to the glass roof of the chamber. Care must be exercised, therefore, that in manipulation the membrane shall not touch the glass. If this occurs one must dismount the collar and clean the glass, other- wise the capillary picture will be clouded. It is of distinct practical help to one using the instrument for the first time, to first become fam- iliar with the location and appearance of the capillary bed under the microscope. When this is done it is a simple matter to locate the proper field with the chamber in position. CAPILLARY BLOOD PRESSURE IN MAN 149 We have experimented with chambers of different diameters but have found that this is a matter of no consequence in obtaining correct read- ings and we now use for routine observations a chamber of such a size that the brass collar is 25 mm. in diameter. In order to insure that the pressure within the chamber is all transmitted to the underlying skin, it is advisable to make several determinations of the capillary pressure with the chamber at slightly different vertical positions. The position of the chamber which gives the lowest capillary pressure read- ing is the correct one. With a little practice one makes this adjustment readily and there is no necessity for such preliminary observations. — The purpose is, of course, to insure that none of the pressure is lost in the resistance offered by the membrane itself. Corpuscular flow. With the chamber in position over the finger, when there is no pressure exerted on the underlying skin, the picture is frequently disturbed by reflections of light from the folds of the mem- brane. This disturbing feature is at once removed if a pressure of 2 or 3 mm. be applied, a procedure which has the effect of smoothing out the membrane over the surface under observation. When the area is first observed it is uncommon to obtain any indication of corpuscular flow. If now the pressure within the chamber be slowly raised by com- pression of the rubber bulb, the corpuscular flow becomes evident and the red cells are seen streaming slowly through the capillary. As the pressure is further raised this corpuscular flow becomes slower and slower until there is no further continuous forward movement of the corpuscles. At this time one frequently sees a to-and-fro movement of the corpuscles without progress. At first thought this phenomenon would seem to be associated with a transmitted pulsation from the underlying vessels. It may however be due to rhythmic contractility of the capillary endothelium. As the pressure is further raised, this to-and-fro movement of the corpuscles ceases. Now a further increase in the pressure acting on the capillaries may cause the corpuscles to travel in a reversed direction, that is to say, from the venous side toward the arterial side. We are able to offer no satisfactory explanation of this fact at the present time. If now the pressure be slowly lowered the reversed flow ceases and the corpuscles again stand still. Then the to-and-fro movement of the corpuscles is seen, and finally with a further lowering of the pressure the corpuscles stream forward in the normal direction. The pressure at this instant is taken to represent the cap- illary blood pressure and our reading is therefore made at this point. 150 C. S. DANZER AND D. R. HOOKER The pressures at which the above events occur in one and the same capillary are depicted in figure 3 and again for a second subject in figure 4. The plotted lines in each figure represent several observations on a single capillary. Note that the skin pales in both subjects at a pressure of about 8 mm. Hg., while the corpuscular stream is not slowed until the pressure has been raised to 18 mm. Hg. in one case (fig. 3) and 27 I pak 0 DA Sues, E50 ¢ eh \ / sig 434-5 45 Corpuscucan RPUSCULAR SLowIne oF ee s Tes Rario Pauinc STREAM BeadeoFowNoPkecress OTAGNATIONNo Procress Beacon FLow SLowSTRean STREAM Fig. 3. To illustrate the pressures at which the several events in capillary blood flow occur. The figures along the abscissa represent pressure in milli- meters of mercury. The ordinates divide the chart according to the observed behavior of the corpuscles with an increasing and then with a decreasing pressure. All the observations were made on one capillary, the several plotted lines repre- senting single observations. mm. Hg. in the other (fig. 4). The slowing of the corpuscular stream undoubtedly represents the condition when the pressure without ap- proximates the pressure within the capillary. It is clear, therefore, that paling of the skin is a wholly inadequate criterion of capillary blood pressure. These figures indicate further that repeated observations of the same capillary agree closely in the pressures at which the events under discussion occur. Experience has taught us, however, that the CAPILLARY BLOOD PRESSURE IN MAN ial slowing of the corpuscular stream which occurs with a rising pressure is much less readily appreciated than is the quickening of the stream which occurs with a falling pressure. Consequently we have found it expedient to raise the pressure until the forward movement of the cor- puscles ceases and then to lower it, giving close attention to the beha- vior of the corpuscles. As the pressure falls the corpuscles at first progress slowly and with a further slight lowering of the pressure there is a sudden sharp acceleration which serves as a sharp criterion and - os” —_—" -— h ; Suez. DRH. ie Car. CorpuscuLar CorPuscuLar ee SLOWING OF : Seacrest ae a adh alah Wh 08.5 SEP SSEASGSSSEHBER RRR GONMER SMe YE Fig. 4. See figure 3. one which is easily recognized. The pressure then represents what we have taken to be the capillary blood pressure. During the procedure above outlined, the capillaries are sometimes emptied of red blood corpuscles, but more often this is not the case, that is to say, when the pressure applied to the skin is sufficient to stop completely the corpuscular flow, the cells are still present in the capillary loop. This observation goes to show that the assumption made by von Kries and other earlier observers that the paling of the 152 Cc. S. DANZER AND D. R. HOOKER skin is associated with an emptying of the lumina of the capillaries is incorrect. Furthermore, one may readily appreciate that the paling of the skin associated with the application of pressure becomes very extreme even before the corpuscular flow is distinctly slowed. Obser- vations of this kind have led us to the conviction that the criteria used by most other observers are wholly inadequate to represent true cap- illary blood pressure. It frequently happens that when one obtains a focus of the capillary field, one or more of the capillaries will stand out conspicuous and large. Such capillaries represent vessels in which the corpuscular flow is stag- nated. The individual corpuscles in such a capillary cannot be made out; apparently they are thickly packed together. The pressure may be raised very considerably without changing in any way the appear- ance of such a vessel. This fact accords with the original observations of Roy and Brown which have been abundantly confirmed by later observers, that all the capillaries in a given vascular bed are not neces- sarily functioning at the same time. The idea that not all of the capillaries are functioning at all times is indeed not a new one. Worm-Miiller (382) in 1873 in studying the influence of blood volume on the arterial pressure comes to the follow- ing conclusion: ‘Under normal conditions (most likely in every part of the body) a large number of empty or poorly filled capillaries stand ready to respond to the needs of the blood stream.’ He also spoke of the dilatation of the capillaries resulting from the transfusion of blood. He found on post-mortem examination that the arteries and veins were not overfilled after transfusion. He therefore assumed that the blood collected in the capillaries. Heubner (33) in 1907 saw many new capillaries open up as the re- sult of the injection of gold sodium chloride into frogs, rabbits, cats and dogs. The experiments on the frogs are described in greatest detail. In these he looked at the frog’s mesentery microscopically and injected an amount equivalent to 0.25 mgm. of metallic gold and in one-half to one minute saw the animal practically bleeding into its capillaries. He estimates that the number of capillaries visible after the injection is three to four times as many as before. Concerning Dale and Laidlaw’s work little can be added to what is so generally known about their researches on “‘Histamine Shock.” By a series of ingenious experiments these workers have attempted to prove a, the active functioning of the capillaries; b, the independence of the capillaries from the rest of the vascular system; c, the capillo- CAPILLARY BLOOD PRESSURE IN MAN tia dilating effect of histamine. They were led to these problems by the discrepancy between the effect of histamine in the intact animal and on the excised arterial or uterine strip. In the former it lowered the blood pressure and in the latter it increased the tone of involuntary muscle strips. Dale and his co-workers explain the marked fall in arterial blood pressure which occurs in the intact animal and which cannot be due to a relaxation of arterial tone since histamine causes the isolated artery to contract, as due to a specific dilator action of hista- mine on the capillaries. Accordingly the capillary beds throughout the body are flooded and the circulating blood is insufficient to fill these areas and at the same time sustain the arterial pressure in spite of a concomitant increase of arterial tone. Whereas the experiments are very suggestive the crucial experiment, namely the effect of hista- mine on the capillaries as observed directly (through the microscope) is as yet lacking. Krogh (34) has thrown considerable light on this subject by counting the number of capillaries in a definite area of muscle tissue in the rest- ing state and immediately after exercise. He examined both fresh and fixed tissues. He found that the working muscle showed many more capillaries than the resting tissue. Hence, the conclusion that many new capillaries open up when necessary (e.g., in the working state). It is obvious perhaps that we avoid the term capillary collapse. We have done this because of a good deal of experimental evidence (partic- ularly that of Roy and Brown) which indicates that capillaries do not collapse when compressed. Why then do we not take capillary empty- ing (i.e., when corpuscles travel back from capillaries into the arterioles) as the criterion? The reason is that in a number of capillaries it is necessary to raise the pressure acting upon them much above that which will cause the corpuscles to stagnate. Hence the readings would be much too high. This is due to the fact that blood is piling up in the patent capillaries as each neighboring one is emptied of its contents (von Recklinghausen). We believe also that as the externally applied pressure rises the arterioles and capillaries are being simultaneously compressed. On this account pressure determination in a great many capillaries may be impossible with the methods of Lombard, Krauss or Basler. Krauss himself admits this difficulty. The impression that we have received from Doctor Lombard personally concerning some of the drawbacks of his method are to the same effect. ee eee ee eee es ee ne ee we SL G's 12 cco neeve ES 61 0°22 0 ¥o ‘8% ‘6E | LI+ rat (PEL > 78o| 88 SI 0°12 GL: 12" Ol face OST u¥0T 7S | Te LI G92 8% (9°2% ‘22 ‘$B } OI- Sc utOL 72 TE QI Le “06 “9g. “ee: ‘2g COLE. SE BS “BZ 0°83 0 1G. | UGi Glace Ge:| =a sn || cOp Leow ey eebamescnl Om qI ; = a 3 GI ‘FI CP [lke ‘ro ‘Te: “Te ‘ca-'6l.) 21 OIT Hie Cale CFs VI a 0°02 Zo— Skee 2g =e leer sl pees joer el 5 GF 8% ‘8% ‘VE ‘9G ‘22 1G c : 0° Go ree IT Pa GI ‘OT Gz Gize ‘Gee ‘Grez, ‘ez rae Ol A 0'1z G'S ‘G'ZZ OT ra 6 a ° GS ‘GG ‘GZ ‘FS a OST-OST “Id “WV USI LT ‘CT £3 Va (BIN Ean Garcon Oc (2 cen aes eSerr Ol | 12 8 ; 2 0°81 LEE Se HN ee TS MS eI ell olf L | 0° FZ Cos GZ ‘ez | 2+ OST ie 8) mes 9 qj G'9Z VO. SG uLGy LGn |e Oze= GST Th Seiad G A 9% “SZ ‘9Z a 0'S T= ‘bS ‘9% ‘GZ ‘9% 2S} 6I- CST Se Weal 1G P i 0°12 Cor SiuGle Cae ike ¢g 0°61 ol ST ‘02 ‘02 | F — OST iS is Te fe (a ZS ‘0G 02 I ro) 8 A “By mua oh 8 me spunod spunod (GOVUAAV) GON AUATITS DAHDITM SUYUVAAY Gwossaud OILY LSOUGCAH SONIGVaU Ivadt THOM LHOIIH wmov Lard as AUVTTIdVO Qa Ly TAOTVO — uo -- ———— S]ONpLarpUr JuULLOW UL ainssaid poojg fiunypdvd ay) fo suorynasasqyd T @TavL “154 155 CAPILLARY BLOOD PRESSURE IN MAN stsouvAd -0.10B JO suishxoied 0} Yoolqng gonbruyooy, ‘sarpnys asay jo qivd AyIvO UT UaYyR} SBurpvod esoyt, “plod B sey yaofqng ploo sey qootqng ‘Aj[Ipwer AIOA MOTA wodj 1veddesip 04 uses ‘dvo ‘YJIM FIOM OF J[MOYWTP yoolqng YFTM FIOM OF 4[NOUBIP a10M syoolqns esoy} pue Avp yey? Apnopd swam ouv«quiayy 0° 9% 0°96 0° &% 0°06 G LT 081 peqoa1log 0 66 0 GE 8+ 8+ 9+ b+ b+ 1D AQ | OT ‘61 ‘OT 02 ‘ST ‘ST LI ‘0G ‘6T GI ‘OT ‘ST IT ‘GIT “GT ‘OL FT Sas teo le: Ohne oh Si oh SI ‘CI Wud ONE 0s ‘og ‘ce 82. Soe Cae 9% 8a FEsCEGE Gove 92 ‘bz ‘8% ‘Gz ‘FZ Ce Bn Ge.n OG AEC Se Blazon bo SIZ 0z Cor ~~ ee) ad id Se | 10 OL i IG OS pdr ’ (a d 156 Cc. S. DANZER AND D. R. HOOKER Because some capillaries exhibit no corpuscular flow and because there is an appreciable variation in the pressure in several capillaries in a given area, we have found it expedient tomake pressure determina- tions in five or more capillaries and to average the results. This will minimize the possibility of striking capillaries of extreme pressures. The majority of our determinations, therefore, represent such an average. Capillary blood pressure in normal individuals. Twenty-five adults and six children were studied. All of the determinations were made with the subjects in the sitting posture, with the hand a little below the heart level. The average capillary pressure in our series was 22.2 mm. Hg. The ages of our subjects ranged from 8 to 47 years. Of the sub- jects studied the lowest average capillary pressure was 17.5 mm. Hg., the highest was 26.5 mm. Hg. These results are given in table 1. Four of our cases (A, B, C, and D in table 1) cannot be grouped with the remaining ones, because of the existence of infections (colds) in two of them and on account of technical difficulties in the other two which occurred at the inception of the work. In some of the cases the distance betwen the level of the hand and the heart level was determined. Like von Kries, von Recklinghausen and Goldmann, we found that capillary pressure varies with the dis- tance between the hand and the heart level. The difference in the pres- sure readings taken when the hand is at heart level and when the hand is above or below the heart level does not, however, correspond to the hydrostatic differrence in these positions. This is clearly brought out by the following table given by von Kries (35). HAND CHANGED FROM HEART LEVEL DIFFERENCE IN CAP. PRESS, DIFFERENCE TO CAPILLARY PRESSURE HYDROSTATIC DIFFERENCE mm. Hg. per cent 205 mm. below heart level 65 33 285 mm. below heart level 116 40 350 mm. below heart level 225 64 The hydrostatic factor is thus a disturbing element in the determi- nation of capillary blood pressure at the present time. It*is therefore advisable that determinations of this pressure should be made at heart level although the inconvenience of such a procedure is considerable. Our readings unfortunately do not conform to this specification, never- theless the range between the determinations on different subjects is remarkably small. In the case of children our readings run appreciably CAPILLARY BLOOD PRESSURE IN MAN L5G lower than in adults but when correction is made for the hydrostatic factor the results correspond closely with those obtained in older indi- viduals. Similarly the results in the group of adults show no relation- ship between age and capillary pressure. We have not noted in the table whether the subjects were men or women because here also we found the sex factor to be of no significance. On the whole, therefore, the data which we have thus far collected point to a very remarkable constancy in the capillary pressure in normal individuals of both sexes regardless of their ages. Krauss also found that the capillary pressure in children was practically the same as in adults. Concerning the readings obtained in the individual capillaries, it may be seen that they run within fairly close limits. There are, however, capillaries whose pressure determinations are definitely outside of these limits. We do not average these extreme ones with the rest. The capillaries of higher pressure may be the ones lying deeper down in the skin or those of larger bore. Possibly they represent the piled-up blood in the patent capillaries resulting from the emptying of the neigh- boring capillaries, as suggested by von Recklinghausen. We discard very low readings when they occur in but one or two capillaries. Theo- retically these are the most accurate. For practical reasons, however, we have arbitrarily eliminated these infrequent (low pressure) capillaries and have adopted instead another criterion which is quite as accur- ate as the former, but more easily carried out. That series of six or more capillaries of the lowest pressure is taken, in which the difference between the extreme capillaries of the series is not more than 6 or 7 (less is preferable). We warn against averaging capillaries of very low or very high pressures with the series. Let us take a concrete example. } eliminate—too low Dee Re Re OW bo 32 eliminate—too high 20 38 | 41 f | Capillary pressure in separate capillaries... | [ eliminate—too high 158 Cc. S. DANZER AND D. R. HOOKER Mhus the, series consists Of.4). 2. scene ee 4 20 -— = 21+ mm. Hg. = average Our figures for normal capillary blood pressure (with the subject in sitting posture, hand below heart level) accord with those of Lombard (18 to 22 mm. Hg.), Landerer and Krauss. The latter two investigators (using Basler’s ochrometer) obtained readings varying from 17 to 25 mm. Hg. on the normal individuals. In our table we have also put down the difference between the actual and the ideal weight of the subject with the idea of noting the effect of over- or under-nutrition on the capillary pressure. Thus far we have nothing definite to say on this subject. The effect of temperature on capillary pressure. Cold. Cold towels with pieces of ice placed between the layers were wrapped around the arm. The cold was not directly applied over the capillary area, hence the results are not to be interpreted as the direct effect of cold on the capillaries themselves but rather as a reflex effect. The duration of the application was 15 minutes. The results are graphically presented in figure 5. ‘Two neighboring capillaries represented by the continuous and the broken lines were studied. Immediately on the application of cold there was a definite drop in the capillary blood pressure. When the cold was removed a very prompt rise of capillary blood pressure, which was maintained for about 13 minutes, occurred. Then there was a marked drop to a point slightly above normal, then another rise and finally a fall. The occurrences after the removal of cold represent the reaction after such a procedure. It is striking how closely the curves of the individual capillaries follow each other. Only at one point do they diverge. The explanation for this would seem to be a change in the lumen of one of the capillaries independent of that of the arteriole since both capillaries were in all probability supplied by the same arteriole. Heat. Heat was studied by means of an electric pad which was wrapped around the arm and kept on for 26 minutes. We made observations on six neighboring capillaries. The electric pad was set so as to create moderate heat. The effect produced by the application CAPILLARY BLOOD PRESSURE IN MAN 159 of heat is also shown in figure 5. Almost immediately there was a rise in capillary blood pressure which was maintained throughout the period during which heat was applied. Upon the removal of the pad the pres- sure in the capillaries fell to a point below the normal value. Soon the pressure average in the six capillaries observed was as before the experi- ment. One is amazed at the striking parallelism of the pressure curves of the whole group of capillaries. Only here and there does a disparity become evident. For example, in the last phase two capillaries show a fall in pressure while two others show a simultaneous rise in pressure, The *ffect of Cold ew Cafe, Preemmnre. The tfpect of Heok on) Cap, (Readurne C;H;03 : C;H.:N;O HO’ Guanine Nucleotide” HO. O=P — O° CsHs03 * CsH4N30 HO’ Cytosine Nucleotide" HOW O=P—O - C;H;03 ° CsHuNs HO’ Adenine Nucleotide!” 7 Journ. Biol. Chem., 1917, xxxi, 337. 8 Journ. Pharm., 1918, xii, 253; 1919, xii, 45. 9 Journ. Biol. Chem., 1917, xxxi, 39. 10 Jones and Richards: Journ. Biol. Chem., 1914, xvii, 71. 11 Thannhauser and Dorfmiiller: Ber. d. d. chem. Gesellsch., 1918, li, 467. Zeitschr. f. physiol. Chem., 1919, ciiii, 65. 12 Jones and Kennedy: Journ. Pharm., 1918, xii, 253; xiii, 45. 196 WALTER JONES HO, O=P —(9)* C;H;O; . C,H;N202 HO Uracil Nucleotide? If yeast nucleic acid is a chemical combination of the four nucleo- tides, then in this union the phosphoric acid groups of the nucleotides must not be disturbed. No additional phosphoric acid linkages can be introduced. The four nucleotides possess together eight replace- able hydrogen atoms; so also, yeast nucleic acid must contain eight replaceable hydrogen atoms. Therefore the nucleotide lnkages of yeast nucleic acid cannot be through the phosphoric acid groups: the linkages cannot involve any one of the four phosphoric acid groups.“ HOW HOW O=P—O ; C;H;0; ; C;H.Ns O—P— Or C;H,02 = C;H.N; HO” | O O | HOW | O=P—O- C;H;0;3 - CsH.N;O O=P—O -: C;H,0, - CsH.N;0 HO” HO” Two nucleotides united through Two nucleotides united but not their phosphoric acid groups through their phosphoric acid groups Until very recently it was conceded by everyone that if the nucleotide linkage of yeast nucleic acid is not through its phosphoric acid groups then it is naturally through the carbohydrate groups. That question is taken up in the following section. II. A COMPARISON OF THE RATE AT WHICH THE PURINES ARE SET FREE FROM YEAST NUCLEIC ACID WITH THE RATE AT WHICH THE PURINES ARE SET FREE FROM THE INDIVIDUAL PURINE NUCLEOTIDES At the time the phosphoric acid studies were made with guanine nucleotide and adenine nucleotide, the rates at which they set free their guanine and adenine were ascertained. It is exceedingly rapid and the same for both purines and for yeast nucleic acid. The rapidity 18 Levene: Journ. Biol. Chem., 1920, xli, 1. 4 The argument here used is essentially that which was employed for the same purpose by Jones and Read (Journ. Biol. Chem., 1917, xxix, 123). Their argument is today as sound as it was when they wrote. YEAST NUCLEIC ACID 197 is such that the time required to dissolve the nucleotide in the hydro- lytic agent has to be considered and in the case of yeast nucleic acid the other products of hydrolysis make the determination of the liber- ated purines very difficult. But one can conclude from the data without hesitation that the purines are set free from nucleic acid and from the individual purine nucleotides with the same rapidity. Again, the exceeding slowness with which the pyrimidines are split by hydrolysis from yeast nucleic acid makes quantitative work impos- sible. But the same slowness characterizes the pyrimidine nucleotides. Therefore the argument which was used above to show that tha nucleotide linkages of yeast nucleic acid do not involve the phosphoric acid groups, may now be used to show that the nucleotide linkages do not involve the purine groups nor probably the pyrimidine groups. This leaves the carbohydrate groups and indicates the formula: HOY O=P — O° C;H;02° CsHuNs HO“ | O HO, | O=P =O -C;H,O*C,H.N;0 HO“ | O HO, | O=P—0 -C;H.0 - CiH-N.O, HO? | O HO. | O=P—O - CsH;02 * CsHiN;O HO? Ill. A COMPARISON OF THE RATE AT WHICH PHOSPHORIC ACID IS SET FREE FROM ADENINE NUCLEOTIDE WITH THE RATE AT WHICH ADENINE IS SET FREE FROM ADENINE NUCLEOTIDE The longest known nucleotide (though not a nucleotide of yeast nucleic acid) is the substance that Liebig!® discovered in meat extract and called inosinic acid. It was afterwards shown contemporaneously by Bauer!® and by Neuberg and Brahn!? that inosinie acid is com- 15 Liebig’s Annalen, 1847, Ixii, 317. 16 Hofmeister’s Beitr., 1907, x, 345. 17 Biochem. Zeitschr., 1907, v, 439; Ber. d. d. chem. Gesellsch., 1908, xli, 3376. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO 2 198 WALTER JONES posed of the groups of three substances, viz., phosphoric acid, pentose and hypoxanthine. Hence inosinic acid may have any one of three structures, i.e., any one of the three groups may be the central group connecting the other two. HO O=P—O : pentose : hypoxanthine (1) HO™ O I Pentose :O—P—O: hypoxanthine (2) | OH Ou Pentose : hypoxanthine : O—P=O (3) OH Liebig knew that inosinie acid is a dibasic acid. This excludes formula (2). Haiser'’ found that by acid hydrolysis inosinic acid loses its hypoxanthine much more rapidly than its pentose.!® This excludes formula (3) and leaves the correct formula (1): HOY O=P—0 - C5303 * CsHsN.O HO/ This method may be used to find the gross structure of any purine nucleotide (but obviously not of a pyrimidine nucleotide) and was ap- plied to guanine nucleotide by Jones and Read?* who found that the substance is a dibasic acid that forms a dibrucine salt and that, by acid hydrolysis, the nucleotide loses its guanine very much more rap- idly than its phosphoric acid. These two facts necessitate the fol- lowing arrangement of the three groups: HO, O=P—O: C;HsOz2 + C;HsN;O0 HOs 13 Monatshefte f. Chem., 1895, xvi, 190. 19 Haiser mistook the pentose for trioxyvaleric acid. Both substances have the formula C;H,O:. 20 Journ. Biol. Chem., 1917, xxxi, 337. a YEAST NUCLEIC ACID 199 The curves of figure 3 are constructed from experimental data ob- tained with adenine nucleotide. The upper curve expresses the rate for adenine and the lower curve the rate for phosphoric acid. In five minutes three times as much adenine is set free as phosphoric acid: or, the liberation of adenine is nearly complete in thirty minutes, that of phosphoric acid only after two hours. mei ‘ I 1 I / | | t ! My ' { | IAT ER eas > dP EAT ci Hours ' 2 3 <. Fig. 3. The upper curve shows the rate at which adenine is set free from adenine nucleotide. The lower curve shows the slower rate for phosphoric acid. Adenine nucleotide forms a dibrucine salt and conducts itself toward alkalis like a dibasic acid. Hence its groups must be arranged as indicated in the formula HO. O—P— Or C5H:O3 .. C;H4sN;O HO” IV. THE ACIDITY OF ADENINE NUCLEOTIDE In the article that follows it will be assumed that the nucleotides conduct themselves like acids toward indicators and alkalis. One cannot work long with the nucleotides without knowing that this is true of them all. When a lead salt is decomposed with sulphuretted hydrogen the filtrate from the lead sulphide is strongly acid to litmus. As the matter is of considerable importance, however, an exact exami- nation of adenine nucleotide was made. A weighed portion of the nucleotide (50 mgm.) was covered with water and titrated with 0.104 N sodium hydroxide using phenolph- thalein for the indicator; 2.69 cc. were required. A second portion of 50 mgm. was titrated using methyl orange for the indicator. Half as much alkali was required, 200 WALTER JONES - STANDARD ALKALI USED FOR 50 MGM. OF THEORETICALLY REQUIRED FOR ADENINE NUCLEOTIDE : = y ival f O ivalent of Using phenolphthalein Using methy] orange Spode ig e Be hodrogen! ¢ ce. ce. ce. ce. 2.69 2.68 1.35 1.31 When free phosphoric acid is titrated with sodium hydroxide using methyl orange as an indicator exactly one equivalent of hydrogen is neutralized; but if phenolphthalien is used as an indicator, twice as much alkali is required, which of course corresponds exactly to two equivalents of hydrogen. Adenine nucleotide therefore conducts itself toward alkalis exactly like free phosphoric acid. This is what one would expect a substance of its structure to do. HO HO. 0=Pp—OH O=P—OR HO” HO” phosphorie acid adenine nucleotide Experimental Weighed portions of adenine nucleotide, placed in flasks provided with condensing tubes and treated with twenty parts of 5 per cent sulfuric acid, were heated in a boiling water bath. At various inter- vals from 15 minutes to 3 hours, a flask was removed from the water bath and both free phosphoric acid and free adenine were quantita- tively determined as follows. The hot fluid was made alkaline with ammonia and treated with a slight excess of magnesia mixture. After standing 5 hours, the crystalline magnesium ammonium phosphate was filtered off, allowed to dry in the air and weighed. The weight was divided by the weight of the nucleotide used in the experiment to obtain the weight per gram of nucleotide, so that the results of various experiments could be directly compared with one another. It was also found convenient not to calculate the corresponding amount of phos- phorus but to express the phosphoric acid throughout in terms of mag- nesium ammonium phosphate (MgNH,PO..6H.0). The ammoniacal filtrate containing adenine was treated with a solu- tion of silver nitrate in ammonia and the precipitated adenine-silver YEAST NUCLEIC ACID 201 compound was filtered off, thoroughly washed, suspended in water and decomposed with hydrochloric acid. After filtration from silver chloride, the acid fluid was evaporated to dryness (with the usual pre- cautions) for the expulsion of all but a trace of free hydrochloric acid, and a solution of the residue in a little warm water was treated with a slight excess of picric acid. The precipitated crystalline adenine picrate was filtered off, allowed to dry in the air and weighed. This weight was divided by the weight of the nucleotide used in the experi- ment and the corresponding amount of adenine was calculated. In addition, the total amount of phosphoric acid obtainable from the nucleotide after completely burning was determined and, for com- parison, was expressed in terms of magnesium ammonium phosphate per gram of nucleotide. The results are given in table 1. Results obtained in a similar way by Jones and Read with guanine nucleotide are given in table 2, and for comparison the results obtained with yeast nucleic acid by Jones and Riley are given in table 3.7 TABLE 1 Adenine nucleotide MAGNESIUM AMMONIUM PHOSPHATE ADENINE PICRATE MgNH:POsk. 6H2O PER CENT NUCLEO- aera CALCULA- OF TIDE Pe ag 5, 1) 7 Os. ead = REO aT a” TED THEO- USED = : per ee 2 : ADENINE pena : er gm. of| of theo- . ergm.o (0.3700) Obtained nucleotide rence Obtained nucleotide (0.671) 0.2161 5 min. 0.0304 | 0.1409 | 21.0 | 0.1253 | 0.5800 | 0.2146 | 58.0 0.6204 15 min. 0.1915 | 0.3087 | 46.0 | 0.5569 | 0.8976 | 0.3330 | 90.0 0.6167 30 min. 0.2566 | 0.4160 | 62.0 | 0.5980 | 0.9696 | 0.3587 | 97.2 0.6302 tly, 0.3552 | 0.5637 | 84.0 | 0.6285 | 0.9973 | 0.3700 | 100.0 0.6111 2 hrs. 0.4084 | 0.6683 | 99.5 | 0.6017 | 0.9844 | 0.3652 | 98.7 0.6123 3 hrs. 0.4088 | 0.6678 | 99.7 | 0.6154 | 1.0051 | 0.3729 | 100.8 0.3033 4 hrs. 0.2023 | 0.6669 | 99.4 | 0.1114 | 0.9930 | 0.3674 | 99.3 0.3247 Total 0.2159 | 0.6646 | 99.5 0.3103 Total 0.2060 | 0.6639 | 98.9 Theoretical 0.6710 | 100.0 1.000 | 0.370 100.0 21 Two values are the results of later experiments. 202 WALTER JONES TABLE 2 Guanine nucleotide” MAGNESIUM AMMONIUM PHOSPHATE GUANINE NUCLEOTIDE | TIME OF MgNHsPOs. 6H20 USED HYDROLYSIS Obtained [Berm af |Pes cent of] Oprainea | Pere: of [Pen oant of 0.4234 5 min. 0.0499 0.1178 18.8 0.0835 0.1972 50.0 0.5030 15 min. 0.1348 0.2680 42.8 0.1747 0.3473 88.1 0.5000 30 min. 0.2004 0.4008 64.0 0.1901 0.3802 96.3 0.3487 1 hr. 0.1881 0.5395 86.2 0.1396 0.4002 101.5 0.3360 2 hrs. 0.2017 0.6000 95.8 0.1310 0.3899 98.9 0.3448 3 hrs. 0.2158 0.6259 100.0 0.1360 0.3944 100.0 0.4112 4 hrs. 0.2539 0.6174 98.6 0.1617 0.3932 99.8 TABLE 3 Yeast nucleic acid? MAGNESIUM AMMONIUM PHOSPHATE MgN HaPOs. 6H20 Koon Gams ene From I II Obtained | Per gram Dyan ane nance freee eee alestides total cleotides ey (0.295) 1.0031 | 15 min. 0.1374 0.137 0.003 0.85 0.134 40.4 41.25 1.0001 | 30 min. 0.2090 0.209 0.005 1.70 0.204 Giles 63.0 0.8642 if lowes 0.2315 0.269 0.010 3.40 0.259 87.8 91.2 0.9837 2 hrs. 0.3125 0.318 0.020 6.80 0.298 101.0 | 107.8 0.9251 3 hrs. 0.3038 0.329 0.030 10.20 0.299 LOI Osea: 1.0305 4 hrs. 0.3488 0.338 0.040 13.60 0.298 101.0 | 114.6 0.8333 5 hrs. 0.2876 0.345 0.050 17.00 0.295 100.0 | 117.0 0.9927 6 hrs. 0.3565 0.359 0.060 20.40 0.299 101.0 | 121.4 0.8179 Total 0.5112 0.590 half the total = 0.295 TABLE 4 Pyrimidine nucleotides*4 NUCLEOTIDES USED MAGNESIUM AMMONIUM PHOSPHATE MgNH:zPOs. 6H20 TIME OF HYDROLYSIS Obtained Fer en. oo Per cent of total 1.0362 3 hours 0.0647 0.0624 10.0 0.3587 Total 0.220 0.624 * Table of Jones and Read. Journ. Biol. Chem., 1917, xxxi, 337. The four- hour period is added. *8 Data taken from article of Jones and Riley, Journ. Biol. Chem., 1916, xxiv, i, except 15-minute and 30-minute periods. ** From data of Jones and Read, Journ. Biol. Chem., 1917, xxxi, 39. THE ACTION OF BOILED PANCREAS EXTRACT ON YEAST NUCLEIC ACID WALTER JONES From the Laboratory of Physiological Chemistry, Johns Hopkins Medical School Received for publication March 12, 1920 The earliest investigations of nucleic acid had shown that by mild acid hydrolysis, the purines are set free with part of the phosphoric acid and part of the carbohydrate; but that violent hydrolytic processes are required to set free the pyrimidines with the remainder of the phos- phoric acid and carbohydrate. A great amount of ingenuity was there- fore not required to formulate a hypothetical structure for nucleic acid. The substance must be composed of four complexes, all of which con- tain a group of phosphoric acid and a group of carbohydrate but each complex contains a different one of the four nitrogenous groups. Change the word “‘complex”’ to the word ‘‘nucleotide” and the above becomes essentially the modern nucleotide theory of the constitution of yeast nucleic acid. But one important matter concerning the constitution of yeast nu- cleic acid was not touched upon by the early investigators. At what points are the nucleotides joined together to form yeast nucleic acid or, in other words, what is the mode of nucleotide linkage in yeast nu- cleic acid? Without any experimental evidence it was finally agreed . that the nucleotide linkage is through the phosphoric acid groups and this assumption remained undisturbed until Jones, Germann and Read furnished the experimental evidence to show that this mode of nucleo- tide linkage is not correct. The principal object of the present paper is to describe an experiment which proves this in a very simple and strik- ing way. Pig’s pancreas contains a variety of active agents (ferments) which decompose nucleic and further act on its decomposition products. When an aqueous extract of pancreas is boiled, all of these active agents are destroyed but one, viz., the one that decomposes yeast nucleic acid into its nucleotides. ; 203 204 WALTER JONES An aqueous extract of pig’s pancreas was boiled and filtered. Yeast nucleic acid was added to the clear extract and the mixture was allowed to digest for 20 hours at 40°, when the nucleic acid had disappeared! and a mixture of the four nucleotides could be isolated. This ferment (if it is a ferment) acts rather rapidly at first but more slowly afterwards. It is exceedingly more active at 40° than at 20° and exhibits about the same activity whether the solution be amphoteric, faintly alkaline or acid to litmus. It is not present in the liver nor the spleen and does not decompose thymus nucleic acid.2_ In the decompo- sition of yeast nucleic acid by this ferment neither phosphoric acid nor purine bases are set free and deaminization does not occur. But when nucleic acid is converted into nucleotides by this active agent there ts not the slightest change in the acidity of the solution. Yeast nucleic acid is represented below by two formulas which differ from one another in only one respect. Formula I has its nucleotide link- ages through the phosphoric acid groups; formula II has its nucleotide linkages through the carbohydrate groups. If formula I is correct, the conversion of nucleic acid into its nucleo- tides should be attended by a marked increase in acidity: but if formula II is correct, there should be no increase in acidity. If formula I cor- rectly represents the structure of nucleic acid, then the increased acidity due to the decomposition of 2 grams of nucleic acid into its nucleotides should require about 8 ce. of 0.1 N sodium hydroxide for neutralization. But as a matter of fact it was not possible to demonstrate any change in acidity with sensitive indicators when 2 grams of nucleic acid was decomposed into its nucleotides. This is a crucial experiment that decides against the phosphoric acid linkage and, I think, in favor of the carbohydrate linkage since the other possibilities are only of academic interest. I cannot feel responsible for, nor even interested in the impossibility of two carbohydrate groups uniting with one another, and presume that such a union will be conceded without any proof if the other conceivable possibilities are excluded. 1 The merest trace of nucleic acid can be detected in such a solution by the ad- dition of sulphuric or hydrochloric acid. * This is curious. An active agent is present in animal pancreas which is specifically adapted to plant nucleic acid. It suggests evolutionary matters. YEAST NUCLEIC ACID 205 I II HOW HOW O=P—O° C;HsO3 : C;H4N;O O=P—O> C;H;02 r C;H.N;O | HO | an en Lace hp pagkaged nan MMI ARS Live ais Ho OH | HO. | OH O=P—O- C;Hs0; * CsHiN;0 ne *CsH,O - CsH4N;0 | HO | pda a oy: Pie 2. Ped MO. Aiter scisifor aye Olina nO!" ( He OH | HOW | OH O=P—O- C;H3:03 ° CsHiNs Oe —O-C;H,O - CsH.N; | HO one oe ges Piaeb ness” icant a OH | HOW | OH O=P— O ms C;H;03 = C.H3N202 O—P— Or C;H,02 - C4H;N.02 HO” HO” th i) HO, HO, O=P—O- C;H;03 ° CsH4N;O O=P—O : C;H;0; ° CsHiN;O HO” HO” HO. HO. O=P—O: C;H;03° C,H.N3;0 (OP —(9)9 CsH;03 : C,H.N;0 HO” HO” HO, HOV O=P — 0: CsHs303 * CsHiNs O=P — O- C;H;03 ° CsH.Ns HO” HO” HOV HOW OSS —=(0)- C;H;0; i C,H;N.0>2 (QS —=10) < CsH;03 ri CsH;N202 HO” HO” EXPERIMENTAL A mixture of 2 kilos of carefully trimmed and ground pig’s pancreas, 2 liters of water and 30 cc. of chloroform was allowed to digest for 12 hours at the room temperature in a tightly closed vessel with frequent and violent agitation. After the tissue had by this means become thoroughly penetrated with chloroform, the mixture was placed in a thermostat and allowed to digest at 40° for 2 days, when it was cooled and filtered. The clear, pale yellow filtrate was then boiled, filtered from a small coagulum and after cooling was preserved with chloroform for use in the following experiments. The data given are selected from a large amount of similar data. a. 50 ce. of boiled extract + 0.250 gm. yeast nucleic acid? 3 In all experiments the digesting material was preserved with chloroform. 206 WALTER JONES . 50 cc. of boiled extract + 0.375 gm. yeast nucleic acid 50 cc. of boiled extract + 0.500 gm. yeast nucleic acid 50 cc. of boiled extract + 0.750 gm. yeast nucleic acid 50 cc. of boiled extract + 1.000 gm. yeast nucleic acid 50 cc. of boiled extract + 1.250 gm. yeast nucleic acid . 50 ec. of boiled extract + 1.500 gm. yeast nucleic acid All were digested at 40°. After 24 hours a, b, c, d and e gave no cloud with H.SO.; a and b gave nothing with HCl, but ¢, d and e gave a slight cloud. After 48 hours a, b and c gave nothing with HCl; d and e only an opalescence; f and g a faint cloud. a. 50 cc. of boiled extract + 0.750 gm. yeast nucleic acid at 20° b. 50 ce. of boiled extract + 0.750 gm. yeast nucleic acid at 40° After 12 hours sulphuric acid gave a dense precipitate with a but nothing with b. a. 25 ec. of boiled extract + 0.3875 gm. yeast nucleic acid b. 25 cc. of boiled extract + 0.375 gm. yeast nucleic acid c. 25 ce. of boiled extract + 0.375 gm. yeast nucleic acid d. 25 ec. of boiled extract + 0.375 gm. yeast nucleic acid a was made amphoteric to litmus; b was made faintly alkaline; c was markedly alkaline; d was left acid. All behaved about alike. After 20 hours digestion at 40° all gave a faint cloud with H,SO, but nothing after 36 hours. a. 50 ce. of boiled pancreas extract + 0.750 gm. yeast nucleic acid b. 50 ce. of boiled pancreas extract + 0.250 gm. thymus nucleic acid c. 50 cc. of boiled spleen extract + 0.250 gm. yeast nucleic acid d. 50 cc. of boiled liver extract + 0.250 gm. yeast nucleic acid e. 50 ce. of phosphate mixture (pH = 6.4) + 0.250 gm. yeast nucleic acid. Digested at 40°. After 19 hours a failed to give a cloud with H.S0O, but b, c,d and e all gave dense precipitates with H.SO, even after 48 hours. Three cubic centimeters of boiled pancreas extract were diluted with 10 cc. of water and treated with 4 drops of a solution of brom-cresol purple. On comparison of the color with a set of standard colors its acidity (pH) was found to be between 6.0 and 6.4 (about 6.2). A larger quantity of the extract was then titrated with 0.1 N sodium hy- droxide to an acidity (pH) of 6.4. Changes in the acidity of such a solution can easily be detected with brom-cresol purple. One and one-half gram of yeast nucleic acid were dissolved in 75 ce. of the above extract and the solution was titrated with 0.1 N sodium hydroxide to an acidity of 6.4, i.e., until the color which it gave with 4 drops of brom-cresol purple exactly matched the color similarly given by the extract. The extract and the solution were then digested at QWs as YEAST NUCLEIC ACID 207 40° for 48 hours. The nucleic acid had entirely disappeared and the color produced by brom-cresol purple with the solution exactly matched the color produced with the extract. The acidity of both solution and extract had not changed from 6.4. A trace of 0.1 N hydrochloric acid caused a very perceptible change in the acidity of the solution. This experiment confirms many less accurate ones that were made. In the tests for the influence of acidity on the activity of the ferment no faintly alkaline solution ever became amphoteric to litmus and no amphoteric solution ever became acid. In fact, no change at all in the tint given to litmus was ever noticed as digestion proceeded. 50 ce. boiled pancreas extract 50 ce. boiled pancreas extract + 0.750 gm. yeast nucleic acid The two solutions were digested at 40° for 48 hours, when the nucleic acid had entirely disappeared. Each solution was made alkaline with ammonia and treated at the boiling point with an excess of magnesia mixture, when perfectly white crystalline ammonium magnesium phosphate was precipitated. This was filtered the next day, allowed to dry and weighed. MgNH,zPO;.6H20 from the experiment 0.5474 MgNH,P0O,.6H20 from the blank 0.5469 An experiment was made with boiled extract which involved 125 gm. of yeast nucleic acid. After the nucleic acid had disappeared by diges- tion at 40° the product was heated to boiling and treated with neutral lead acetate as long as the reagent gave a precipitate in the hot fluid. This precipitate which consists principally of lead phosphate was fil- tered off with a pump and the pale yellow filtrate was treated in the warm with more lead acetate. At first no precipitate was produced but after the addition of a sufficient excess of the reagent a copious granular precipitate was thrown down. After cooling, the precipitated lead salts of the nucleotides were filtered off, washed, suspended in warm water and decomposed with sulphuretted hydrogen. The filtrate from lead sulphide was evaporated at 45° under diminished pressure and the nucleotides were thrown out and dried with absolute alcohol. This product is undoubtedly identical with the mixture of nucleotides for- merly obtained by Jones and Richards* from yeast nucleic acid by treat- ment with an extract of pig’s pancreas that had previously been di- gested for a long time at 40°. It contains all four nucleotides and forms a mixture of crystalline brucine salts. As the separation of the nucleo- tides from one another is foreign to the principal point of this paper, the matter will be taken up later in a separate article. ¢ Journ. Biol. Chem., 1915, xx, 25. wD | ca) eye te ee 1m 4 hye a es - 4 ‘ ead iy Pie ce | oe 5 ; J s a 2 i ; , | » hon title hua-s¥ sit a iin’ ] | ‘n lohie ba ra My +. ee) A yd Tay (feed | ) ‘ v*] An (ee ft Ke ey: iggy SE) Tae AL, Li i Z | ee THE AMERICAN JOURNAL OF PHYSIOLOGY VOL. 52 JUNE 1, 1920 No. 2 A STUDY OF FORCED RESPIRATION: EXPERIMENTAL PRO- DUCTION OF TETANY SAMUEL B. GRANT ann ALFRED GOLDMAN From the Physiological Department of Washington University Received for publication March 15, 1920 INTRODUCTION Tetany is a disease which has been subjected to a rather large amount of experimental study in recent years, the fact that its symptoms may be so readily produced by parathyroidectomy in animals, and the ques- tion of the relation of parathyroid tetany to idiopathic tetany in man, making it a particularly inviting field for experimental work. In the course of the work about to be described, it was found that all the essential symptoms of tetany could be produced in the human subject by forced respiration. It was noted during the course of a series of experiments in which the subject was required to go through a period of increased respiration, that the urine collected at the end of this period was cloudy, and that the cloudiness was due to the fact that the urine was alkaline and con- tained precipitated phosphates. In explanation of these observations it was surmised that the overventilation, by washing out carbon dioxide of the blood, tended to make the blood more alkaline, and that in order to maintain a constant hydrogen ion concentration in the body, alkali was excreted by way of the kidneys. The present investigation was undertaken with the idea of determining this question. METHODS The experiments were carried out on human subjects according to the following plan. During the period of forced respiration the subject, with thoracic and abdominal pneumographs attached, was lying on his 209 THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 210 SAMUEL B. GRANT AND ALFRED GOLDMAN back, and breathed as deeply as possible in time with a metronome at the rate of about 14 per minute. This was continued until symptoms of tetany developed, usually for from 15 to 60 minutes. The pneumo- graphs recorded the depth of respiration on smoked paper, within sight of the subject, who was thus better able to maintain the deep breathing. In the later experiments, when the subjects had become accustomed to forcing respiration, the pneumographs were discarded. ‘The metronome, however, was used throughout to prevent the rate of breathing from becoming involuntarily too slow. The alveolar carbon dioxide tension was measured at the beginning and at the end of each experiment, the Plesch (1) bag method of col- lecting the air and the Marriott (2) colorimetric method of determining carbon dioxide tension being used. In some of the experiments the alveolar carbon dioxide tension was determined at intervals throughout the period of forced respiration, and afterwards until it had returned to normal. The urine secreted during a measured period of time just before the experiment was collected, and also that for the period during which the deep breathing occurred, and in some eases at fixed intervals subse- quently until the urine had returned to normal. Each specimen was analyzed as to hydrogen ion concentration, titratable acidity and ammonia content. The hydrogen ion concentration was determined by the method of Henderson and Palmer (3). The acidity was titrated with 0.1 N NaOH, using phenolphthalein as the indicator. Ammonia was determined by the Folin permutit (4) method. Twenty cubic centimeters of blood were drawn with a syringe from the veins at the elbow just before and at the end of forced respiration, and the hydrogen ion concentration, carbon dioxide capacity and cal- cium content were determined. The hydrogen ion concentration was determined by the dialyzing and colorimetric method of Levy, Rown- tree and Marriott (5). Samples of blood immediately upon being col- lected and before coming in contact with the air, were put into the col- lodion sacs and dialyzed. Specimens of oxalated whole blood were also used in several cases. The carbon dioxide capacity of the blood was determined by the method of Van Slyke (6), and calcium was deter- mined by the Lyman (7) method, both whole blood and serum being used. TETANY FOLLOWING FORCED RESPIRATION IN MAN 211 RESULTS Tetany in forced respiration. Symptoms of tetany developed in each of twenty-four experiments, with the exception of the one control. The diagnosis was based on the presence of carpopedal spasm, Chvostek’s sign, Trousseau’s sign, Erb’s sign of increased electrical irritability, and in one instance a tetanic convulsion. Attention was drawn to this condition in the first experiment. The subject, S. G., breathed as deeply as possible for sixteen minutes, at the rate of 15 per minute. About ten minutes after the start his fingers began to tingle. At the end of fifteen minutes the muscles of his face felt stiff, and there was some difficulty in articulation. In the second experiment, A. G., subject, tingling in the fingers was noted. When at the end of forty minutes an attempt was made to determine the blood pressure, inflation of the arm band caused the hand imme- diately to go into a typical spasm of tetany. The fingers were flexed at the metacarpophalangeal joints; the second and third phalanges remained extended at the phalangeal joints; all the fingers, especially the thumb, were adducted and the hand was flexed at the wrist. The spasm could be voluntarily overcome by active motions, but as soon as these ceased the hand returned to the former position. Generally the spasm relaxed almost immediately after cessation of the forced breath- ing, but sometimes it persisted for as long as three minutes. In sub- sequent experiments the other signs and symptoms of tetany noted above were brought out. Two subjects were used throughout the course of the experiments, and it was found that they differed somewhat in their reaction. Sub- ject S. G. developed, after breathing deeply for about ten minutes, spasticity of the facial muscles. Chvostek’s sign for tetany could then be easily elicited, i.e., tapping a branch of the facial nerve caused spasmodic contraction of the facial muscles. As the deep breathing continued this spasticity increased, until finally the muscles of the face, especially those about the mouth, contracted spontaneously, causing puckering of the lips and decided difficulty in articulation. In the meantime the hands had begun to tingle and feel numb, and finally, after the spasm of the face was well marked, they spontaneously went into the typical spasm described above. Chvostek’s sign remained positive after forced respiration had ceased, and a slight reaction could be obtained at least twenty minutes after the end of the experiment. In subject A. G. the facial phenomenon was not nearly so marked, but the spasm of the hands was greater. He often noted a slight headache 212 SAMUEL B. GRANT AND ALFRED GOLDMAN soon after deep breathing started. In about fifteen minutes tingling of the hands began, and this gradually became more intense until the hands suddenly went into a very marked spasm. ‘Trousseau’s sign could usually be obtained several minutes before the onset of the spon- taneous spasm. When the facial nerve was tapped only a slight twitching of the muscles occurred and this was usually not obtained until the carpal spasm was intense, and not always then. No spon- taneous facial spasm occurred as in subject S. G. Spasm of the feet took place in two experiments in which A. G. was subject. The feet were extended, the first phalanges of the toes flexed and the second and third extended. Tingling in the abdominal and thoracic muscles was noted by both subjects on several occasions, when the period of deep breathing was prolonged. The electrical irritability of the muscles of the forearm, or Erb’s sign, was tested in two experiments, and in both it was found that con- traction occurred with a much smaller current at the end of deep respi- ration than before. The indifferent electrode was fastened to the upper part of the right arm, and the stimulating electrode was applied over the ulnar nerve just above the elbow, the same point of stimulation being used before and after forced breathing. The number of milli- amperes of current necessary to cause contraction with the cathodal opening, cathodal closing, anodal opening and anodal closing stimuli before and at the end of deep respiration in experiment 18 were as fol- lows: Experiment 18 coc ccc AOC acc Before deep respiration...............0.0+: 55 2.5 2.5 2.0 CC TETANY At end of deep respiration.................| 1.25 0.5 1.25 1.25 In experiment 5 they were: Before deep respiration................ SED 1.8 At end of deep respiration............. 0.5 0.+ A complete tetanic convulsion occurred in experiment 18, in which A. G. was subject. No note of the exact rate of respiration was made but it was faster than usual. After breathing very deeply for thirty minutes he suddenly went into a complete tetanic convulsion. At the onset he involuntarily gave a loud, high-pitched scream, probably due TETANY FOLLOWING FORCED RESPIRATION IN MAN 213 to contraction of the muscles of respiration and forcing of air out through contracted vocal cords. The entire body was rigid, all the muscles being contracted in tetanic spasm. The back was arched some- what, and all extremities extended completely. Relaxation occurred within thirty seconds, and there was no further spasm. There were no ill effects following the convulsion. The usual sequence of symptoms in the experiments is given in the following chart of a typical experiment, together with some of the experimental data. Experiment 11. A. G., subject TIME Alveolar CO: tension, 40 mm. Urine pH = 4.9 Blood pH = 7.4 Plasma CQ: capacity, 63.6 vol. per cent 0 minute | Starts deep breathing—12 per minute 10th minute | Slight headache 13th minute | Tingling in hands 16th minute | Trousseau’s sign obtained. Slight Chvostek 17th minute | Both hands tingle markedly 20th minute | Spasm of both hands 21st minute | Alveolar CO, tension 20 mm. 22d minute | Blood collected DEH —7.5 Plasma bicarbonate 45.8 vol. per cent 23d minute | Spasm of feet. No facial spasm. Stops deep respiration 25th minute | Urine collected. pH=7.9 Before | Changes in the blood, urine and alveolar air. Decided changes took place in the alveolar air, blood and urine as a result of the forced respi- ration. Alevolar carbon dioxide tension fell, the blood became slightly more alkaline, the urine became decidedly alkaline, the plasma bicar- bonate was reduced, the ammonia excretion was diminished, and there was a slight increase in the calcium content of the blood. Alveolar carbon dioxide. In fifteen experiments in which the alveolar carbon dioxide tension was measured it fell from an average of 42 mm. before to 21 mm. of mercury at the end of the period of deep breathing (see table 1). The fall was rapid at first and then became slower. With the termination of the period of forced breathing the tension rose at first rapidly and then more and more slowly, to almost reach normal in twenty minutes (see figs. 1, 2, 3, 4). 214 SAMUEL B. GRANT AND ALFRED GOLDMAN In experiment 17 the alveolar air was collected both by the Hal- dane (8) method, which gives alveolar air as determined by the tension of the gases in arterial blood, and the Plesch method, which gives alveolar air at the tension the gases are contained in venous blood. It was found that the difference in carbon dioxide tension as deter- Z TABLE 1 ALVEOLAR CQ2 TENSION LENGTH OF PERIOD EXPERIMENT SS ae | OR DEEP RESPI~ SUBJECT Before deep At end of deep RATION respiration respiration minutes 1 40 22 16 8. G 2 35 18 50 A. G 3 38 18 32 A.G 4 43 22 15 eG 5 40 22 25 A.G 6 40 25 25 8. G 8 45 22 25 8. G 9 41 24 26 A. .G 10 45 22 20 8. G 11 40 20 23 A.G 12 45 25 20 8. G 13 37 20 28 A.G 14 45 21 37 ALG 15 45 20 23 oy 16 45 22 23 8. G Average..... 41.8 21 25.9 TABLE 2 ALVEOLAR CO2 TENSION (MM.Hg.) TIME AFTER START OF DEEP RESPIRATION DIFFERENCE IN TENSION Plesch Haldane minutes mm. mm. mm, 0 45 37 8 5.5 31 23 8 11.5 26 19 7 16.0 23 18 5 PANES 22 18 4 mined by these two methods was less at the termination of the period of exaggerated respiration than it had been at the beginning, as shown in table 2. Hydrogen ion concentration of the blood. This was determined in five experiments, and the average fall was from pH 7.41 before to pH 7.57 TETANY FOLLOWING FORCED RESPIRATION IN MAN Bid at the end of deep breathing (see table 3). These determinations were made on fresh blood directly from’the vein, before clotting occurred. In three experiments the hydrogen ion concentration of oxalated blood was determined. Loss of carbon dioxide to the air could not be so readily prevented when this was done, but the blood was placed in a narrow test tube from the syringe, corked, gently mixed with oxalate crystals, and then dialyzed at once in the deep narrow collodion sacs. The results were as in table 4. The hydrogen ion concentration of blood may be calculated from the carbon dioxide capacity of the plasma TABLE 3 pH oF BLoop LENGTH OF PERIOD EXPERIMENT eS | SOR DEEPIRESPI“ SUBJECT Before deep At end of deep RATION respiration respiration ei (oad Tie Oo eT TOT 9 7.40 7.60 26 A. G 10 7.45 7.65 20 8. G 11 7.40 7.50 23 A.G 12 7.35 7.60 20 8. G 13 7.45 7.50 28 AacG Average..... 7.41 7.57 23.5 TABLE 4 pH oF OXALATED BLOOD LENGTH OF PERIOD EXPERIMENT SS nn Lg DEBP RESPI- SUBJECT Before deep At end of deep RATION respiration respiration minutes 11 7.45 7.65 23 AG: 12 7.35 7.60 20 Sa G: 13 Minoo 7.50 28 A. G. and the alveolar carbon dioxide tension, using Hasselbach’s (9) formula. This was done in three experiments in which the required data had been obtained, and in each of these the blood was shown to have be- come more alkaline during deep respiration. In experiment 2 pH rose from 7.44 to 7.61, in experiment 12 from 7.34 to 7.48, and in ex-. periment 13 from 7.40 to 7.55. Plasma bicarbonate. The carbon dioxide combining power of the plasma was determined in four experiments. It fell from an average of 59.5 v.p.c. before to 44.9 v.p.c. at the end of deep respiration. The- greatest fall was 17.8, the smallest 13.0 (see table 5). 216 SAMUEL B. GRANT AND ALFRED GOLDMAN TABLE 5 CO: CAPACITY OF PLASMA LENGTH OF PERIOD EXPERIMENT SS a Ln ae ||| GORYDEEP RESPI= SUBJECT ee ies v.p.c. v.p.c. minutes Thi 63.6 45.8 22 A. G. 12 56.0 43.0 17 S. G. 13 52.9 Dont 27 A. G; 21 65.6 O2e2 21 S. G. Average..... 59.5 44.9 21.75 Hydrogen ion concentration of urine. The urine specimen collected after forced respiration was alkaline to litmus in twelve of thirteen ex- periments. The determination of the hydrogen ion concentration showed, in the average of thirteen experiments, pH 5.2 in the speci- mens voided before the period of exaggerated respiration, and pH 7.4 in the specimens voided after this period (see table 6). TABLE 6 pH oF URINE LENGTH OF PERIOD EXPERIMENT |< eae oe cana 1 en ML ORDERED Ear SUBJECT Before deep After deep RATION respiration respiration minutes 1 6.4 6.9 16 S.-Ge 2 4.7 Ufeal 50 Ay Ge 3 5.5 7.4 32 A. G. aa 4.8 7.3 15 5.G, 5 4.7 ose 25 A, G: 6 5.0 Coil 25 er 8 4.7 Ut 25 2G: 9 5.5 7.2 26 A. G. 10 5.2 7.8 20 S.A, 11 4.9 7.9 23 ype: (oe 12 4.8 Tel 20 S.'°G: 13 5.8 atl 28 AoGt 16 (595 Cf 7.2 23 8. G Average..... 5.2 | 7.4 25.2 Titratable acidity of wrine. In twelve experiments the acidity of the urine as titrated with 0.1 N sodium hydroxide, using phenolphthalein as the indicator, fell from an average of 7.54 cc. 0.1 N acid to 2.23 ce. per half-hour (see table 7). TETANY FOLLOWING FORCED RESPIRATION IN MAN | 217 TABLE 7 TITRATED ACID LENGTH OF PERIOD EXPERIMENT j—aET aaa GES. sal en a | ON DEEP RESPI- SUBJECT Of before urine; cc.| Of after urine; cc. RATION 0.1 N per } hour 0.1 N per } hour cc. 0.1 N cc.0.1 N minutes 2 10.93 2.80 50 A. G 3 8.70 1.75 32 ne 4 9.37 1.68 aS} Ss. G 5 10.54 2.36 25 A.G 6 10.05 1.46 25 Ss. G 8 4.59 0.75 25 S. G. 9 6.34 2.14 26 A. G. 10 3.53 1.24 20 8. G 11 5.40 1.67 23 A. G 12 5.80 2.07 20 8. G 13 7.88 1.74 28 A. G 16 9.72 3.05 23 8. G Average..... 7.54 23) 26 Ammonia excretion. The ammonia content of the urine was deter- mined in twelve experiments. The average decrease was from 13.18 mgm. ammonia nitrogen per half-hour before, to 5.57 mgm. ammonia nitrogen after forced respiration (see table 8). TABLE 8 AMMONIA LENGTH OF PERIOD EXPERIMENT pnt «viee OF DEEP SUBJECT mgm Nis nitrogen|mmgus. Nii mitrogen| RESPIRATION per half-hour per half-hour mgm. mgm. minutes 2 16.45 7.85 50 A. G. 3 13.40 4.74 32 A. G. +f 14.20 5.32 15 S. G. 5 16.50 5.85 25 A. G. 6 9.00 3.90 25 8. G. 8 11.74 4.28 25 S. G. 9 17.40 1013. 26 ANG: 10 8.00 3.93 20 S. G. 11 12.00 220 23 A. G. 12 10.20 4.50 20 SaiG. 13 14.25 4.13 28 Aa Gt 16 15.00 5.00 23 s:,.G. Average..... 13.18 5.57 26 218 SAMUBL B. GRANT AND ALFRED GOLDMAN Calcium content of blood. Calcium was determined in whole blood in one experiment and in serum in three experiments, and in each case a small but definite increase in calcium content was found in the blood collected at the end of the period of deep respiration over that found in the blood before deep respiration. All determinations were run in duplicate, some in triplicate, and a blank was run in each case. The figures are given in table 9. TABLE 9 Ca CONTENT OF BLOOD LENGTH OF PERIOD OF EXPERIMENT SUBJECT Before At end of DEEP RESPIRATION deep respiration | deep respiration mgm. per 100 cc. | mgm. per 100 cc. minutes 18 (Whole blood) 7.00 7.29 67 AniG: 19 (Serum) 11.81 13.19 51 Ss. G. 20 (Serum) 12S 13.31 62 ASG: 21 (Serum) 12.84 13.44 21 S. G. The blood taken from the veins at the end of deep respiration had a much brighter red color than normal venous blood. Control experiments. It was necessary to show, a, that the intake of food was not influencing the results; and b, that the results could not be attributed to any factor other than the overventilation. In experi- ments 9, 10, 11 and 12 the subjects omitted breakfast on the morning of the experiment. There was no appreciable difference between the results in these experiments and any of the others except that the titratable acidity of the control urine specimens was less than in the . other experiments. In experiment 7 means were taken to prevent the subject from low- ering the carbon dioxide content of the alveolar air. He breathed through a tube connected to a fifteen-liter bottle, while the observer at the same time breathed through a second tube connected to the bottom of the bottle. The resulting high carbon dioxide content of the air in the bottle prevented the subject from washing out the carbon dioxide from his blood, in spite of the forced respiration. Under these condi- tions the alveolar carbon dioxide tension was slightly higher at the end of the experiment than it had been before. The acidity and ammonia content of the urine increased slightly, rather than decreasing as in the other experiments, and the hydrogen ion concentration of the urine remained about the same. The figures are given in table 10. TETANY FOLLOWING FORCED RESPIRATION IN MAN 219 No symptoms of tetany occurred in this experiment. The subject became very slightly cyanosed, and after the experiment was dyspnoeic for a few minutes. Breakfast was omitted on the morning of the ex- periment. TABLE 10 A. G. subject. Deep breathing 17 minutes URINE pH URINE ACID (cc. 0.1N) | NH3 NrTROGEN (mMeEM.) ALVEOLAR CO:2 Before After Before After Before After Before After 4.6 4.8 7.79 8.27 in {eye 13.3 38 40 Four of the experiments may now be taken up in detail in order to illustrate more concisely the exact method of procedure and the uni- formity of the results. Experiment 12. S. G., subject TIME Alveolar CO; tension=45 mm. Urine pH=4.8 0 minutes | Blood pH=7.35 Plasma CO, capacity =56 v.p.c. 10 minutes | Starts forced respiration 20 minutes | Chvostek’s sign obtained. Alveolar COz,28 mm. Tin- gling in fingers 23 minutes | Spasm of facial muscles—Chvostek’s sign marked. Trousseau’s sign positive 27 minutes | Blood collected pH=7.60 Plasma CO, capacity =48 v.p.c. 30 minutes | Alveolar CO, tension=25 mm. Stops forced respiration 40 minutes | Urine collected pH=7.1 45 minutes | Alveolar CO, tension=33 mm. 55 minutes | Alveolar CO, tension=37 mm. 1 hour 10 minutes | Urine collected pH=5.0 lhour 15 minutes | Alveolar CO, tension=39 mm. 2 hours 10 minutes | Urine collected pH=5.0 2 hours 20 minutes | Alveolar CO2 tension=40 mm. The graph of this experiment (fig. 1) shows the synchronous changes in alveolar air, blood and urine. Time in minutes is plotted on the horizontal axis, and alveolar carbon dioxide tension in millimeters of mercury, plasma carbon dioxide capacity in volume per cent, hydrogen ALVEOLAR CO, TENSION 05 CAPACITY ss 220 OF BLOOD PLASMA PH OF URINE cc .IN ACID PH 0 Mleh F BLOOD TETANY FOLLOWING FORCED RESPIRATION IN MAN 221 ion concentration of blood and urine in terms of pH, and ammonia excretion and titrated acidity of urine per half-hour in ce. .IN solutions on the perpendicular axis. Deep breathing starts at 10 and ends at 30 minutes. As the alveolar CO, tension falls, the CO, capacity of - the plasma falls, hydrogen ion concentration of blood and urine fall, excretion of ammonia decreases, and titrated acidity of urine falls. After deep breathing ceases, alveolar CO, tension rises, and correspond- ing changes occur in the urine. TIME Experiment 138. A.-G., subject Alevolar CO, tension=37 mm. 0 minute Urine pH=5.8 Blood pH=7.45 Plasma CO, capacity =52.9 v.p.e. 10 minutes | Starts forced respiration 25 minutes | Slight headache 30 minutes | Alveolar CO, tension=23 mm. 31 minutes | Tingling in right foot. No Chvostek 32 minutes | Hands tingle 33 minutes | Tingling increased 34 minutes | Trousseau’s sign markedly positive 35 minutes | Both hands in spontaneous spasm 36 minutes | Slight spasm of feet 37 minutes.| Blood drawn pH=7.50 Plasma CO, capacity =38.7 v.p.c. Spasms of hand more intense 38 minutes | Stops forced respiration 40 minutes | Alveolar CO, tension=20 mm. 42 minutes | Urine collected pH=7.7 Temporary dizziness and trembling of hands on standing up after forced respiration 45 minutes | Alveolar CO, tension=28 mm. 50 minutes Alveolar CO, tension=32 mm. 1 hour Alveolar CO; tension=37 mm. 1 hour 23 minutes | Urine collected pH=6.9 1 hour 50 minutes | Urine collected pH=5.0 In the graph of this experiment (fig. 2) the same changes are shown as in figure 1, with the exception that the titrated acidity of the urine is not charted. Deep breathing starts at 10 and ends at 38 minutes. The third specimen of urine after forced breathing is more acid than the control urine specimen. 0, ALVEOLAR C TENSION J = « > uu ° <= a. AMMONIA EXCRETION t 13. Forced respiration begins 222 Sa Fig. 2. Experimen 38 minutes 1 hour 1 hour 1 hour 1 hour 1 hour 1 hour 1 hour 1 hour 1 hour 1 hour TETANY FOLLOWING FORCED RESPIRATION IN MAN TIME 30 minutes 5 minutes 7 minutes 12 minutes 18 minutes 22 minutes 23 minutes 26 minutes 28 minutes 30 minutes 2 hours 2 hours 45 minutes 3 hours 45 minutes 4 hours 15 minutes TIME 20 minutes 25 minutes 30 minutes 38 minutes 40 minutes 41 minutes 42 minutes 44 minutes 45 minutes 55 minutes 1 hour 7 minutes 1 hour 15 minutes 1 hour 25 minutes 1 hour 35 minutes 2 hour 25 minutes 3 hour 25 minutes 223 Experiment 16. S. G., subject Urine collected pH=5.0 Urine collected pH=5.7 Alveolar CO, tension=45 mm. Starts forced respiration Alveolar CO, tension=31 mm. Alveolar CO: tension=26 mm. Tingling in hands Marked tingling in hands. Carpal spasm Alveolar CO, tension =22 mm. Stops forced respiration Urine collected pH=7.2 Urine collected pH =6.8 Urine collected pH=5.8 Urine collected pH=4.8 Urine collected pH=5.5 Alveolar CO, tension =23 mm. Experiment 4. S. G., subject Alveolar CO:=43 mm. Urine collected pH=4.8 Starts forced respiraton Tingling and numbness in hands Headache Rigidity of facial muscles. right and left Carpal spasm both hands Abdominal and thoracic muscles begin to tingle. spasticity felt by observer Stops forced respiration Aveolar CO: tension=22 mm. Urine collected pH=7.3 Chvostek weakly positive Alveolar CO, tension=32 mm. Chvostek negative Alveolar CO, tension=40 mm. Urine collected pH=4.8 Alveolar CO, tension=43 mm. Urine collected pH=4.7 Urine collected pH=5.2 Chvostek markedly positive, No 224 0 10 203040501 10 2030 40502 10 2030 40503 10 Fig. 3. E 53 minutes. SAMUEL B. GRANT AND ALFRED GOLDMAN aun CO_ TENSION Pu OF URINE SZ SZ SZ \/ e rus OOD Xv xperiment 16. Forced respiration begins at 30 minutes and ends at TETANY FOLLOWING FORCED RESPIRATION IN MAN 225 Figure 3 is a graph of experiment 16. Deep breathing starts at 30 minutes and ends at 53 minutes. The fall in alveolar CO: tension is very marked at first and then gradually becomes smaller as the low Pu OF URINE AMMONIA EXCRETION | OBS KS sreacemeeeeeeese POO OOO SOOO OKI Fig. 4. Experiment 4. Forced respiration begins at 30 minutes and ends at 45 minutes. level is reached. - The urine pH reaches its control level 1 hour and 15 minutes after the forced respiration, while the specimen collected 2 hours 15 minutes after is more acid than the control one. The other changes are similar to those shown in figures 1 and 2. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 226 SAMUEL B. GRANT AND ALFRED GOLDMAN The changes in experiment 16 (fig. 4) are similar to those described in the previous ones. The deep breathing begins at 30 and ends at 45 minutes. The return of the alveolar CO, tension to the control level is shown, 60 minutes elapsing before this occurs. The ammonia does not reach the control level in 2 hours and 30 minutes. DISCUSSION Forced respiration, by washing out carbon dioxide from the alveoli, reduces the carbon dioxide content of the blood and so tends to make the blood more alkaline than normal. This reduction of carbon dioxide H,CO; concentration depends, and on the reéstablishment of this ratio, NaH CO; passes out of the blood plasma into the tissues, and a con- siderable portion is excreted by way of the kidneys. Apparently, CO: may be washed out of the blood more rapidly by overventilation, than NaH COs; is decreased by excretion or other means, for on this assump- tion alone is it possible to account for the increased alkalinity actually found in blood during deep respiration. The diminution in ammonia formation is a further evidence of the effort of the body to compensate for the increased alkalinity of the blood. The lowered urine acidity is due partly to the increased alkali excretion, but probably largely to a retention of acid radicals. In view of these considerations it seems logical to conclude that there is a true condition of ‘alkalosis’ in the body during overventilation; the body fluids are actually more alkaline than normal. It is realized that the term “alkalosis” is a poor one, but we are using it strictly in the sense of an actual change in hydrogen ion concentration. In this con- dition of alkalosis we have an actual reduction of the ‘alkaline re- serve,’ due to the elimination of NaHCO; from the body fluids in an effort to maintain a constant hydrogen ion concentration. The coex- istence of a more alkaline blood and a reduced alkaline reserve is another example of the inadequateness of the terms “alkalosis” and “acidosis.” The question now arises, is it possible to establish an etiological re- lationship between this condition of alkalosis and the tetany which develops during forced respiration? This may be discussed in con- nection with the views as to the cause of tetany that are held at the present time, the two principal ones of which are, first, that it is due to a calcium deficiency in the blood; and second, that it is due to a disturbance of the acid-base equilibrium in the body. in the blood disturbs the ratio upon which the hydrogen ion TETANY FOLLOWING FORCED RESPIRATION IN MAN 227 The literature contains some references to this problem. Thus, Wilson, Stearns and Thurlow (10) and Wilson, Stearns and Janney (11) have found that in parathyroidectomized dogs, just before an attack of tetany, a period of “alkalosis” develops, which is neutralized by the acid substances that are produced during tetany. They base their conclusion that an ‘‘alkalosis’”’ obtains on the study of the dissociation constant of the oxyhemoglobin, the CO. content of alveolar air, the hydrogen ion concentration of blood, the ammonia content of urine and the total acidity and hydrogen ion concentration of urine. In- jections of acid and also of calcium salts relieve the tetany. They state that possibly part of the beneficial action of the calcium salts may be due to a relative increase in acid radicals caused by their administration. This might be brought about by the formation of Ca3(PO,)2 from the carbonates, thus liberating HCl. McCann (12) finds after parathyroidectomy (and following operations on the stomach which exclude the acid secreted from the duodenum) that there is a marked increase in the CO, combining power of the plasma, coincident with the development of tetany and concludes that tetany is a condition of alkalosis in which a disproportion between the rates of secretion of acids and alkalies by the gastro-intestinal tract may be a factor. He points out that clinically gastric tetany is most apt to occur in those cases in which there is some pyloric obstruction, so that the acid gastric secretion does not pass into the duodenum to excite the alkaline intestinal secretion, and that this might be expected to leave the blood with an excess of basic substances. MacCallum (13) and his coworkers in a paper just published state that when the pylorus of a dog is occluded, and only distilled water is introduced into the duodenum, the gastric mucosa continues to secrete hydrochloric acid. At the same time the alkali reserve increases and tetany-like symptoms appear. This can be prevented by the introduction into the duodenum of sodium chloride or the symptoms may be made to disappear by the intravenous injection of sodium chloride. Injection of hydrochloric acid is not so successful. Calcium determinations by a dialyzing method showed no change in the calcium content of plasma. The authors think that possibly the disturbed equilibrium of acids and bases in itself is the cause of the symptoms. Cases of tetany following therapeutic administration af sodium bi- carbonate have been reported. Howland and Marriott (14) cite three cases in children, all of which were associated with a marked diminution of the calcium content of the serum. Harrop (15) reports a case in an 228 SAMUEL B. GRANT AND ALFRED GOLDMAN adult who developed tetany after the intravenous injection of sixty grams of sodium bicarbonate. There was no associated diminution in calcium. Binger (16) has shown that when sodium phosphate in sufficient quantity and at the right reaction is injected into dogs intravenously, symptoms of tetany result, associated with a marked decrease in the calcium content of the serum. But if the solution injected has a hydro- gen ion concentration greater than 10~*, no tetany occurs even though there is a diminution in calcium. Ammonia poisoning produces symptoms strikingly similar to those of parathyroid tetany. Marfori (17) found after intravenous injections of the carbonate, lactate and tartrate of ammonia in dogs, tremors, muscle twitchings, even tetany and opisthotonus, and irregular respi- ration, vomiting, etc. MacCallum and Voegtlin (18) found increased ammonia content of the blood in parathyroid tetany as did also Carlson and Jacobson (19) in their earlier work. Later, however, using im- proved methods, Carlson and Jacobson could find no change in ammonia content of blood in parathyroidectomized dogs. The fact that increased respiration is associated with changes in the reaction of the urine had been previously noted by Hasselbach (20). He found that when the alveolar CO, tension was raised a few milli- meters by the influence of morphine on the respiratory center, the urine became very slightly more acid, and that when the alveolar CO: tension was decreased by stimulation of respiration by light baths, the urine became barely perceptibly more alkaline. After the present work had been started, Leathes (21) published a paper in which it was stated that increasing the respiration voluntarily for from one-half to one hour increased the alkalinity per cent of the urine from about 35 per cent to about 90 per cent. There was no men- tion made of symptoms of tetany or other changes during deep respi- ration. Yandell Henderson and Haggard (22) found a great reduction of CO, capacity of blood plasma of dogs in ether hyperpnoea and in ex- cessive pulmonary ventilation by artificial respiration, together with a concomitant fall in blood pressure. They state that when the CO, capacity of plasma falls below 33 to 36 v.p.c., the process becomes irre- versible, and the animal dies in shock. The lowest point to which CO, capacity of plasma fell in our work, in those experiments in which it was determined, was 43 v.p.c. Symptoms of tetany are not mentioned in their paper. TETANY FOLLOWING FORCED RESPIRATION IN MAN 229 Stein (23) in discussing tetany as a sequel of gynecological operations cites fifteen cases in which it occurred. In four of these the symptoms of tetany occurred during the operation, and disappeared quickly with- out recurrence later. In three cases the tetany occurred during light ether anesthesia, one of these as narcosis was being induced, while the patient was still conscious. In the fourth case tetany occurred after the removal of eight liters of fluid from the peritoneal cavity by para- centesis, and lasted for fifteen minutes. It seems not unlikely that tetany in these instances may have been due to overventilation. McClendon (24) has studied the effect of hydrogen and hydroxy] ions on the pulsation rate of the jelly fish and the heart of the conch, and finds that the activity may be inhibited by increasing H-ions in the surrounding fluid or perfusing liquid, as the case may be, and augmented by increasing the OH-ions. Biedermann (25) found that the height- ened excitability of voluntary muscle produced by NaCl solution is greatly increased by the addition of Na2COs, the muscle then contracting rhythmically and exhibiting a striking increase of response to artificial stimuli. Strong solutions of NaSO, and dilute NaOH act in a similar manner but to a less degree. He seems to attribute this to a specific effect of the sodium ion, but it may be that part of the action is due to the alkalinity of the solutions in the case of NaszCO; and NaOH. Recently a number of investigators have come to regard tetany as being due to a decrease in the calcium content of the blood. Thus MacCallum and Voegtlin (18) and MacCallum and Vogel (26) have shown that in the tetany of parathyroidectomized dogs there is a marked diminution of blood calcium. Howland and Marriott (27) have found that in infantile tetany the calcium content of the serum is greatly reduced, the average being about 5.6 mgm. per 100 cc. compared to 10 or 11 mgm. in the normal child. The administration of calcium promptly relieved the symptoms of tetany. On the other hand we have found in tetany of forced respiration no diminution of calcium content of the blood or serum. ‘There is, as a matter of fact, a slight increase in calcium in every instance in blood specimens taken at the onset of tetany over those obtained under normal conditions of respiration. This may not necessarily mean that calcium plays no definite part in the production of the tetany of increased respiration. Thus Barille (Wells, 28) states that a decrease of CO: in the blood results in precipi- tation of calcium. And Marriott (29) has shown that when artificial blood is brought into equilibrium with CO, at 30 mm. Hg. tension, calcium is precipitated in a perceptible cloud. The possibility must 230 SAMUEL B. GRANT AND ALFRED GOLDMAN therefore be considered that when the CO. of blood is reduced by over- ventilation a portion of the calcium is “precipitated” or in some way rendered inactive, though still present in the circulating blood. On the basis of this supposition the increase of calcium which we have found might be interpreted as being due to a compensatory mechanism on the part of the body to make up for a loss of active calcium. The calcium in our subjects was not precipitated in the blood, however, for centrifuging the serum specimens did not affect the analysis. It is conceivable, however, that the calcium is “precipitated” as very minute particles which are held in solution by the colloids of the serum, in which case centrifuging might have no effect on the calcium present. A fur- ther attempt to determine whether calcium plays a part in the produc- tion of tetany of overventilation will be made in a future investigation. It may not be inappropriate to refer in this connection to the hypoth- esis advanced by Stoeltzner (30) attributing tetany to an accumulation of calcium in the blood. He found support for his views in some obser- vations on the electrical reactions of peripheral nerves in infants after the administration of calcium. But these observations of Stoeltzner’s have failed of confirmation. Tetany occurs under so many circumstances that it probably cannot be attributed to a single etiological factor. It is more likely that it must be regarded as a symptom complex, and possibly any condition that heightens the irritability of peripheral nerves may cause it. Obser- vation seems to indicate that alkalosis may be such a cause, and it probably is the underlying factor in the tetany that develops as a result of forced respiration. SUMMARY 1. Forced respiration causes symptoms of tetany to occur in the human subject; these include carpopedal spasm, Chvostek’s sign, Trousseau’s sign, Erb’s sign, and in one instance a tetanic convulsion. 2. As a result of the fall of alveolar CO, tension produced by over- ventilation, there is a reduction in the hydrogen ion concentration of the blood, a reduction of the CO. capacity of plasma, a change in the reaction of the urine to the alkaline side, a decreased excretion of am- monia, and a slight increase in the calcium content of the serum. 3. The underlying factor in the tetany of forced respiration is the alkalosis. TETANY FOLLOWING FORCED RESPIRATION IN MAN 231 We wish to express our deep appreciation of the invaluable help and interest in this work given us by Dr. Joseph Erlanger. We are also indebted to Dr. P. A. Shaffer and Dr. W. McK. Marriott for their kindly interest throughout the work. Note. After the completion of this paper, several important refer- ences bearing upon our results were found in a journal which had been temporarily lost from the library. H.M. Vernon (31) in some obser- vations on the production of prolonged apnoea in man says, “I found that after about 6 minutes of forced breathing the muscles of my hands passed into a condition of tonic rigidity, and for the first one and one- half minutes of the subsequent apnoea they were completely paralyzed. My other muscles were apparently unaffected.” In no case in our experiments were the hands paralyzed, for we have always been able to voluntarily overcome the spasm, and this is also true of the spasm of idiopathic tetany. Yandell Henderson (32) had a large number of men undergo voluntary hyperpnoea for short periods of time (usually 45 to 90 seconds) and makes the following statement. ‘Several of the subjects after vigorous hyperpnoea for 1 or 2 minutes, experienced in varying degrees the phenomenon mentioned by Vernon. . . . Even more common in our experience is a prickling sensation in the legs and arms, or in some cases in the entire body and face, somewhat similar to a hand or foot ‘asleep.’”’ In one case, after vigorous hyperpnoea for two minutes a “shivering fit’? came on, similar to that seen in a chill, and involving apparently all the muscles. Hill and Flack (33), in determining the effect of oxygen inhalation on muscular work, had men undergo voluntary hyperpnoea for short periods. They observed feelings of numbness in the limbs, with a spastic state of the hands. One subject felt “twitching of the facial muscles.” In another the “mouth was felt pursed up in the form of an O, so that speaking became difficult,” and the muscles surrounding the eyes seemed contracted, so that the eyes were difficult to open fully. ; Apparently none of these investigators realized that the phenomena observed in forced breathing are identical with those seen in tetany, and no effort was made to determine the factors upon which they depend. Dae ‘ SAMUEL B. GRANT AND ALFRED GOLDMAN BIBLIOGRAPHY (1) Puescu: Zeitschr. p. exper. Path. u. Therap., 1909, vi, 380. (2) Marriott: Journ. Amer. Med. Assoc., 1916, Ixvi, 1594. (3) HENDERSON AND Pater: Journ. Biol. Chem., 1912, xiii, 393. (4) Fotin anp Bex: Journ. Biol. Chem., 1917, xxix, 329. (5) Levy, RowntTrREE AND Marriott: Arch. Int. Med., 1915, xvi, 389. (6) Van Stryke: Journ. Biol. Chem., 1917, xxx, 347. (7) Lyman: Journ. Biol. Chem., 1917, xxix, 169. (8) HaLDANE AND PriesTLEY: Journ. Physiol., 1905, xxxii, 225. (9) HasseLBatcH: Biochem. Zeitschr., 1916, Ixxviu, 112. (10) WiLson, STEARNS AND THURLOW: Journ. Biol. Chem., 1915, xxiii, 89. (11) Witson, STEARNS AND JANNEY: Journ. Biol. Chem., 1915, xxiii, 123. (12) McCann: Journ. Biol. Chem., 1918, xxxv, 553. (13) MacCatuium Et Au: Bull. Johns Hopkins Hosp., 1920, xxxi, 1. (14) Hownanp AND Marriorr: Quart. Journ. Med., 1918, xl, 289. (15) Harrop: Bull. Johns Hopkins Hosp., 1919, xxx, 62. (16) Brncer: Journ. Pharm. Exper. Therap., 1917, x, 105. (17) Marrort: Arch. f. exper. Path. u. Pharm., 1893, xxxiii, 71. (18) MacCaLLuM AND VorGTLIN: Journ. Exper. Med., 1909, xi, 118. (19) CarRLson AND Jacopson: This Journal, 1910, xxv, 403; 1911, xxviii, 133. (20) HasseLBatcu: Biochem. Zeitschr., 1912, xlvi, 403. (21) Leatues: Brit. Med. Journ., 1919, no. 3056, 165. (22) HENDERSON AND Hacearp: Journ. Biol. Chem., 1918, xxxiil, 333. (23) Stern: Interstate Med. Journ., 1916, xxiii, 1078. (24) McCuenpon: Journ. Biol. Chem., 1916, xxviii, 135. (25) BIEDERMANN: Electro-physiology, New York, 1898, 104 (26) MacCaLium aNp VoGEL: Journ. Exper. Med., 1918, xviii, 618. (27) HowLanp AND Marriott: Quart. Journ. Med., 1918, 1, 289. (28) Weuts: Chemical pathology, Philadelphia and London, 1918, 443. (29) Marriorr: Personal communication. (30) SroriTzNerR: Jahrb. f. Kinderh., 1906, Ixiii, 661. (31) Vernon: Journ. Physiol., 1909, xxxvili, xx. (32) HenpEeRsoN: This Journal, 1909, xxv, 310. (33) Hi~tu anp Fiack: Journ. Physiol., 1910, xl, 360. A STUDY OF THE CARBOHYDRATE TOLERANCE IN ECK FISTULA AND HYPOPHYSECTOMIZED ANIMALS (POSTERIOR LOBE REMOVAL) Liver FUNCTION IN THE METABOLISM OF SUGARS CONRAD JACOBSON From the Peter Bent Brigham Hospital, Boston, Mass., and the Hunterian Labo- ratory of Surgical Research, Johns Hopkins Medical School, Baltimore, Md. Received for publication March 17, 1920 In the course of some studies in the Hunterian Laboratory on the sugar tolerance of animals deprived of the hypophyseal gland (1) it was shown that there was a temporary lowering of the assimilation limit for saccharose (the sugar which was primarily used for reasons then given), often with spontaneous mellituria—usually dextro-rotatory, occasionally laevo-rotatory—followed by a subsequent rise to above normal. It was observed at the same time that this artificially raised tolerance as well as the normal tolerance of animals (cat, dog and rabbit) could be lowered occasionally to the point of spontaneous gly- cosuria by the administration, intravenously or by mouth, of hypophys- eal posterior lobe extracts. It was assumed and has been since proven that the administration of the extract had a glycogenolytic effect and was accompanied by a hyperglycemia. In the further investigation of this subject, some efforts were made to determine the source of this gly- cogen. This was first attempted by staining methods—Best’s Carmine of the liver—before and after discharge of the glycogen either by Ber- nard’s Piqtire, by stimulation of the hypophysis or by giving hypo- physeal or adrenal extracts. Though a lessening in the glycogen con- tent of the cells could be demonstrated, the cells were never found entirely free from glycogen and the method had to be abandoned. Marked fatty changes in the liver had also been noted in some of the hypophysectomized animals and the suggestion arose that these changes might produce modifications in the storage capacity of the liver for glycogen and for carbohydrate metabolism. It was therefore decided to use an Eck fistula animal and to note what effect the shunting of the 233 234 CONRAD JACOBSON liver from the portal circulation would have upon the sugar tolerance both before and after hypophysectomy. The Eck fistula operation was first successfully performed on animais by von Eck in 1877 (2) and repeated with equal success by Stolnikow in 1882 (3). It consisted essentially of making an anastomosis of the portal vein to the vena cava with subsequent ligation and more re- cently, double ligation and transection of the portal vein near the liver hilus. In such a condition the blood supply of the liver is maintained by the intact hepatic artery, but the blood coming from spleen, pancreas, stomach and intestines, carrying with it the products of gastro-intestinal and splenic activity, is carried into the vena cava instead of to the liver and thus directly into the venous circulation. The purpose of the origi- nal operation was a clinical one, that of furnishing drainage for the abdominal fluid, ascites, incident to a cirrhosis of the liver, and neither of these two original investigators did any reliable work on changes in the metabolism. The operation has never proven successful when applied to human subjects both because of the operative difficulties and of the danger of subsequent toxic effects. It has, however, on repeated occasions been of extreme value in physiological investigations regarding carbohydrate metabolism and liver function. In 1893 Hahn, Massen, Nencki and Pawlow (4), working with Eck fistula dogs, showed that such animals fed on a meat diet developed marked toxic effects, usually ending in the death of the animal. There was noted a definite decrease in the urinary output of uric acid and ammonia and a large increase in carbamate. On injecting the latter substance into normal and Eck fistula animals, they were able to pro- duce similar toxic effects and concluded that accumulation of carbamate was the causal factor of the toxic symptoms. Rothberger and Winter- berg in 1895 (5) verified the work of Pawlow et al, but could not satisfy themselves that carbamate was the toxic agent, concluding indefi- nitely that the toxic effect was caused by some substance which was removed by the liver under normal conditions but which, under the altered conditions of Eck fistula, accumulated in the blood. Nencki, Pawlow and Zaleski in 1896 (6) from experimental evidence concluded that ammonia was the cause of the toxic symptoms and not the car- bamate. However in 1897 Nencki and Pawlow (7), by further experi- mentation with Eck fistula animals in which large portions of the liver were either removed or thrown out of function by ligation of the hepatic artery, concluded that the extreme toxic symptoms were far out of pro- portion to the increase of ammonia and could not consider the latter LIVER FUNCTION IN METABOLISM OF SUGARS 235 as the main causal factor. Popelski in 1898 (8) showed that the Eck fistula dog had a low carbohydrate tolerance and that an alimentary glycosuria resulted when such an animal was fed large quantities of glucose. Sachs in 1898 (9), working with frogs whose livers had been extirpated, showed that they had a lower tolerance for laevulose than normal intact control animals, a finding which could not be confirmed with dextrose, galactose or arabinose. In 1901 Strauss (10), working with Sachs’s findings as a rationale, introduced the laevulose test for hepatic insufficiency. In 1902 Schultz and Miller (11) had the oppor- tunity of closely observing a clinical case of portal embolus at the liver hilus and were unable to find any definite untoward symptoms except that of ascites, reporting no pathological findings in urine. In 1907 Filippi (12) showed that there was a lowered tolerance for carbohydrates in dogs subsequent to Eck fistula operation. He showed that such animals possess an increased muscle glycogen content characteristic of over-nutrition and a low liver glycogen content characteristic of inanition. He inferred that the muscles could form glycogen inde- pendently and that the liver was neither specific for nor indispensable for carbohydrate metabolism. Hawk in 1908 (13), from feeding experi- ments, showed that quite extreme toxic effects were produced by feeding meat extracts (especially Liebig’s extracts of meat) in conjunction with the meat diet. He was, however, unable to produce any such toxic symptoms by the introduction of carbamate nor was he able to produce glycosuria on carbohydrate feeding. Fischler in 1910-11 (14) noted the acute degeneration in the liver and fat necrosis following Eck fistula such as could be produced by pancreatic injury. These could be pre- vented by preliminary injections of trypsin. The meat intoxication, according to his view, was caused by an alkalosis due most probably to ammonia unneutralized by the liver and this condition could be relieved by acids. Michaud in 1911 (15) showed that Eck fistula prevented adrenal glycosuria but that when such an animal is fed sugar by mouth, the blood sugar either remains normal or increases within normal limits, as in any intact animal. In 1911 Voegtlin and Bernheim (16) showed that the bile pigments and bile acids were decreased in an Eck fistula dog to such an extent that the ligation of the common duct was not followed by jaundice and by only a trace of bile in the urine. They noted also the occasional attacks of convulsions, somewhat similar to those occurring in tetany, produced by meat feeding and showed that these could be prevented by feeding plenty of calcium in the form of bone. They inferred that the Eck fistula condition was compatible 236 CONRAD JACOBSON with life if care was used in feeding. In 1911 Towles and Voegtlin (17) showed that there was no essential difference in the metabolism of creatin and creatinin in a normal and an Eck fistula dog and concluded that the liver is not an organ of prime importance in the metabolism of these substances and that the creatinin could hardly be held account- able for the toxic symptoms evoked on meat feeding. In thus excluding the portal circulation from the liver and diverting the portal stream directly into the systemic venous circulation, there is seen to be a re- duction in the liver function: a, in the synthesis of the urea from ammonia salts; 6, in the production of bile salts and bile acids; and c, in the lowered carbohydrate tolerance. It is with this latter feature that this paper is mainly concerned. Conduct of experiments. Normal healthy dogs were used throughout the experiment—good-sized animals and those only who proved adapt- able to sugar feeding on the pre-operative tolerance determinations. To have comparable results, weighed quantities from the same lot of sugar were used on all the animals. Sugar feedings were under as uni- form conditions as possible,—namely, one sugar feeding a day, on empty stomach, 4 to 5 hours after meals, with definite amounts of water for dissolving the sugar and avoidance of all food or drugs with the feed- ings. Sugar in the urine was tested for and determined by Fehling, Nylander, fermentation and the polariscopic methods. The animals were fed on a mixed diet of boiled meat and bread, and were kept in cages during the tests. The urine was collected from beneath the cages. The Eck fistula operation consisted of making a typical lateral anas- tomosis between the portal vein and the inferior vena cava using small enterostomy clamps (Jerusalem-Jeger) and silk for the sutures. On completion of the anastomosis, the portal vein was doubly ligated and transected between the ligatures, near the liver hilus.!_ Posterior 1 Following Eck fistula operation, the animals remained apparently well and in all outward respects normal. There is an early but decreasing tendency to loss of weight and at first bounteous feedings are necessary to maintain the general body weight. When fed on a meat diet for a great length of time or by the addi- tion of meat extract to a diet for several days, convulsive attacks somewhat suggestive of tetany are likely to occur. These are characterized by disorien- tation, indifference to surroundings, muscular twitchings, paralysis of the limbs, rigidity of back, ataxia and fine tremors. These may last as long as 48 hours and may lead to the death of the animal. The attacks are usually preceded by a period during which the animal seems below par, may be dull and drowsy or may be unusually restless or nervous. At this stage the acute attack may occasionally LIVER FUNCTION IN METABOLISM OF SUGARS 237 lobe hypophysectomy was performed by the intracranial method de- scribed by Crowe, Cushing and Homans (18). This consists of con- tralateral temporal openings through bone and dura which allows the brain to be lifted and dislocated upwards through one decompression opening. The posterior lobe is then enucleated from the dangling gland from below, through the temporal opening of the opposite side. Protocols Animal A. Well-nourished, normal male dog, weight 11.2 kilos April 10-16. Pre-operative cane sugar tolerance, 130-140 grams April 17-28. Eck fistula operation—uneventful convalescence April 28-May 15. Post-operative sugar tolerance. Weight 10.8 kilos Cane sugar, 14 grams. Laevo sugar in urine Glucose, 55-60 grams Laevulose, 7-8 grams May 25-June 3. 2nd post-operative sugar tolerance Cane sugar, 14-25 grams Glucose, 55-60 grams Laevulose, 7-8 grams June 18-25. 3rd post-operative sugar tolerance. Weight 10.8 kilos Cane sugar, 14-15 grams Glucose, 55-60 grams Laevulose, 7-8 grams. Mild convulsive attacks July 7-15. 4th post-operative sugar tolerance. Weight 10.2 kilos Cane sugar, 14-15 grams Glucose, 55-60 grams Laevulose, 7-8 grams be aborted by feeding the animal plenty of bone, a substance for which he seems particularly desirous. Injections of calcium chloride or feedings of calcium lactate do not seem to have any effect when the acute attacks have already begun. After the attack has worn away the animal seems to be in normal condition again. One of the animals died during a severe second attack while the other two dogs had two and five attacks respectively. If care is taken to keep the animal on a mixed diet and excess of meat is avoided and if there is plenty of cal- cium in the form of bone, there seems to be no reason why an Eck fistula animal should not live for a long period of time. The animals reported here were ob- served for 139, 340 and 408 days respectively after the Eck fistula operation; the two latter finally succumbed to conditions unassociated with the fistula operation. On autopsy the only pathological findings of interest are those associated with the liver. This organ is small, hard, firm and rather tougher than normal. The. gall bladder and bile ducts appear to be normal. On section, the liver lobules seem to be decidedly smaller than normal with central atrophy and some fatty degeneration, conditions which have been produced experimentally by a reduction of the blood supply of the liver itself. 238 CONRAD JACOBSON July 17-30. Hypophysectomy—posterior lobe removal—ether Urinary output from 1520 ec. to normal in5 days. No permanent poly- urla Slight trace of sugar in the first specimen; no permanent glycosuria August 3-10. {st post-operative tolerance. Weight 11.2 kilos Cane sugar, 14-15 grams Glucose, 45-50 grams Laevulose, 7-8 grams August 18-24. 2nd post-operative tolerance Cane sugar, 14-15 grams Glucose, 50-55 grams Laevulose, 7-8 grams September 6-13. 3rd post-operative tolerance. Weight 11.3 kilos Cane sugar, 14-15 grams Glucose, 55-60 grams Laevulose, 7-8 grams September 20-28. 4th post-operative tolerance Cane sugar, 14-15 grams Glucose, 60-70 grams Laevulose, 7-8 grams October 5-13. 4th post-operative tolerance. Weight 10.8 kilos Cane sugar, 14-15 grams Glucose, 65-70 grams Laevulose, 7-8 grams October 20-26. 6th post-operative tolerance Cane sugar, 14-15 grams Glucose, 65-70 grams Laevulose, 7-8 grams November 6-15. 7th post-operative tolerance Cane sugar, 14-15 grams Glucose, 65-70 grams Laevulose, 7-8 grams. Mild convulsions November 24-30. 8th post-operative tolerance. Weight 10.8 kilos Cane sugar, 14-15 grams Glucose, 65-70 grams Laevulose, 7-8 grams Summary of animal A NORMAL AFTER ECK AFTER HYPO- TOLERANCE FISTULA PHYSECTOMY grams grams grams CAN eCTSUP ATE. ie cee te ee eee 130-140 14-15 14-15 CGIICOSE sees cio ee tee 55-60 65-70 IeakevAllOsezie ant. ous eine ee kee 7-8 7-8 Animal used for pancreatectomy—died May 24 of pneumonia. LIVER FUNCTION IN METABOLISM OF SUGARS 239 In this, the first of the Eck fistula animals in good condition, the saccharose tests even in small amounts gave sugar which proved to be laevo-rotatory. As the saccharose is necessarily broken up into dex- trose and laevulose in the process of its absorption and utilization in the body, it seemed necessary to resort to separate tests for the tolerance of the two sugars before the operation. It showed also that the assimi- lation tests under the circumstances with a complex sugar like saccha- rose, such as had been used, was capable of misinterpretation; for with hepatic insufficiency there would always be a low assimilation limit and a laevo-rotatory substance in the urine whereas the glucose tolerance itself might be little modified. Animal B. Well-nourished, normal male dog, weight 13.3 kilos May 11-June 6. Pre-operative sugar tolerance. Weight 13.3 kilos Cane sugar, 170-175 grams Glucose, 110-115 grams Laevulose, 85-90 grams June 6-19. Eck fistula operation—uneventful convalescence June 19-July 10. Post-operative sugar tolerance Cane sugar, 10-15 grams Laevulose, 7-8 grams Glucose, 75-80 grams. July 12-22. Hypophysectomy—posterior lobe removal—ether. Weight 12.3 kilos Urinary output 1360 cc. to normal in 7 days. No permanent polyuria No glycosuria July 23-August 2. 1st post-operative tolerance Laevulose, 8-9 grams Cane sugar, 10-15 grams Glucose, 70-80 grams August 10-18. 2nd post-operative tolerance. Weight 12.6 kilos Laevulose, 8-9 grams Cane sugar, 10-15 grams Glucose, 70-80 grams August 28-September 8. 3rd post-operative tolerance Laevulose, 8-9 grams Cane sugar, 10-15 grams Glucose, 100-110 grams September 17-24. 4th post- operative hae: Weight 12.7 kilos Laevulose, 8-9 grams. 1st series of convulsions Cane sugar, 10-15 grams Glucose, 100-110 grams October 4-10. 4th post-operative tolerance Laevulose, 8-9 grams Cane sugar, 10-15 grams Glucose, 100-110 grams 240 CONRAD JACOBSON October 17-22. 6th post-operative tolerance. Weight 12.9 kilos Laevulose, 8-9 grams Cane sugar, 10-15 grams Glucose, 100-110 grams. 2nd convulsive attack Animal died on October 24—2nd series of convulsions Summary of animal B NORMAL AFTER ECK AFTER HYPO- TOLERANCE FISTULA PHYSECTOMY grams grams grams @anevsugar {F200 21 Me dake ee Sat 170-175 10-15 10-15 GI COSC eis ace 110-115 75-80 100-110 VME VAllOSC)vaciiioee uc Be De eee Eee 85-90 8-9 8-9 Animal C. Well-nourished, normal male dog, weight 14.6 kilos October 12-November 1. Pre-operative sugar tolerance Cane sugar, 170-175 grams Glucose, 120 grams Laevulose, 90 grams. Weight 16 kilos November 2-10. Eck fistula operation—uneventful convalescence November 10-December 1. Post-operative sugar tolerance Cane sugar, 20-25 grams Laevulose, 7-8 grams Glucose, 100 grams. Weight 12.7 kilos December 2-12. Hypophysectomy—posterior lobe removal—ether Urinary output 1970 ce. to normal in 10 days. No permanent polyuria No glycosuria December 12-January 1. 1st post-operative tolerance Laevulose, 7-8 grams Glucose, 80-90 grams Cane sugar, 20-25 grams. Weight 12.7 kilos January 12-19. 2nd post-operative tolerance Laevulose, 7-8 grams Glucose, 80-90 grams Cane sugar, 20-25 grams. Ist attack of convulsions January 29-February 6. 3rd post-operative tolerance Laevulcse, 7-8 grams Cane sugar, 20-25 grams Glucose, 80-90 grams. Weight 12.9 kilos. 2nd attack of convulsions February 17-24. 4th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 80-90 grams March 14-23. 4th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 90-100 grams. Weight 13 kilos LIVER FUNCTION IN METABOLISM OF SUGARS 241 April 9-20. 6th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 110-120 grams. 3rd attack of convulsions May 15-26. 7th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 140-150 grams June 8-14. 8th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 140-150 grams. Weight 13.1 kilos July 6-14. 9th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 140-150 grams. 4th attack of convulsions July 28-August 4. 10th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 140-150 grams. Weight 13.3 kilos August 11-16. 11th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 140-150 grams August 24-30. 12th post-operative tolerance Laevulose, 7-8 grams Cane sugar, 20-25 grams Glucose, 140-150 grams. Weight 13.3 kilos Summary of animal C NORMAL AFTER ECK AFTER HYPO- TOLERANCE FISTULA PHYSECTOMY grams grams grams , 22 ULiE se 2, 5 re eee =r 170-175 20-25 20-25 SUPRORE 6 oh oe Bie 120-130 90-100 140-150 LESrTloOge:.. (gS ee 90-95 7-8 7-8 Animal used for pancreatectomy—died December 15. The tolerance of an animal for sugar is the lowest dose of that sugar, following the ingestion of which, sugar just appears in the urine—the threshold dose producing glycosuria. It is dependent upon a number of factors such as the variety of sugar, the general state and condition of the animal, the manner and method of administering the sugar. This threshold dose is an indefinite amount varying between rather wide limits. Of the several methods of introducing sugar,—intra- THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 242 CONRAD JACOBSON venous, subcutaneous and alimentary,—the two former give the lower values and the latter, the most frequently used, the highest values. For alimentary glycosuria the absorption rate from the gastro-intestinal tract must exceed the storage and utilization power of the tissues so that a hyperglycemia is produced with a blood sugar content high enough to overcome the normal impermeability of the kidney. It is evident that anything that concerns the speed of absorption, the utili- zation rate, the storage power of the tissue and the permeability of the kidney will result in changes in the tolerance. It is also evident that small amounts of sugar, through rapid absorption, will produce an abrupt rise in the blood sugar content with resultant glycosuria while larger amounts with slower absorption may be unaccompanied by the appearance of the sugar in the urine. The rate of absorption of sugar from the gastro-intestinal tract is not a constant factor and, as such, influences markedly the amount of sugar brought to the blood and tissues for utilization. The higher the tolerance, the more inaccurate the determination is and the grgater the limits of variance. With the lower tolerance sugars, the determination is less difficult and the limits more readily defined. In many cases it is extremely difficult, and in some animals impossible, to obtain the tolerance with high tolerance sugars on account of the intolerance of stomach and intestinal tract for the huge and abnormal masses required for one feeding. The tol- erance is usually stated in grams per kilo of body weight. This is open to grave error as young animals have a much lower tolerance per kilo than adult animals and obesity, as is well known, is not accompanied by a proportionate increase in the sugar tolerance. The indefiniteness of alimentary sugar tolerances is readily seen when one peruses the literature for tolerances determined by different investigators. Hof- meister’s (19) tolerances for glucose in dogs ranged from 1.3 to 5.8 grams with 2 to 3 grams per kilo as the average. This is an unusually low tolerance, in fact much lower than that after intravenous admin- istration and is rather out of the question. Pflueger (20) found some ° 11 to 16 grams per kilo. That the method of administration had a definite effect upon the tolerance is seen by his figures of 11 to 12 grams when the dextrose was given with meat stew and of less than 8 grams per kilo when it was given with soup. Schlessinger (21) found the tolerance to be about 10 to 11 grams per kilo and that small young animals had a lower tolerance per kilo than the larger adult animals. Boeri and De Andreis (22) found values of 4 to 6 grams for fasting animals and 10 to 13 grams for well-nourished animals. Pratt and LIVER FUNCTION IN METABOLISM OF SUGARS 243 Spooner (23) found values of 11.5 grams for dogs. Quarta (24) found an average of 4 grams for male and 10 grams for female dogs. Sex has been found to have no appreciable effect upon the sugar tolerance and his value of 4 grams per kilo is entirely too low. Filippi’s values were 8 to 10 grams for fasting animals and he lays great stress upon the fact that to obtain comparable results in alimentary feeding, sugar should be given under certain definite conditions. The tolerance values given for laevulose are few in number. Quarta gave the value as between 3 to 4 grams per kilo and de Filippi as about 1.5 gram. The tolerance for saccharose shows the same wide variation. Hoeppe—Seyler (25) sets it at 20 to 30 grams per kilo, Hofmeister at 3.6 grams and de Filippi at about 4 grams per kilo. The osmotic irritation, the gastric and intestinal intolerance for large and abnormal masses of sugar, the lack of any control over the varying rates of absorption, make the alimentary method, though practically the only feasible one, far from an ideal one for investigative purposes. With an adaptable animal, under uniform conditions, comparative results on the same animal may however furnish data of some value. Table 1 gives the tolerances obtained in this investigation. The normal values differ considerably from those given by de Filippi. The saccharose tolerances agree fairly closely with those obtained by Goetsch, Cushing and Jacobson who found in their series of animals weighing 9 to 11 kilos, a tolerance of 9 to 13 grams per kilo of body weight. Tolerance after Eck fistula operation. Following Eck fistula opera- tion there are marked changes in the carbohydrate tolerance ascertained by alimentary feeding. Subsequent to this surgical procedure there is usually a marked loss in body weight, and some of the post-operative tolerances given in the above table are undoubtedly high since they are calculated upon the lowered body weight. The glucose tolerance is modified to a less degree than the laevulose or cane sugar tolerance. This seems to indicate that the muscle and the body tissue generally can manage very well the storage and utilization of the glucose that is shunted to them by the Eck fistula operation and leads one to infer that the liver is an organ not absolutely necessary for the utilization of that sugar. The most marked changes occur in the laevulose tolerance. In an Kek fistula animal a small amount of laevulose ingested is followed by a laevulosuria, there being almost an intolerance for that particular sugar. Muscular tissue can undoubtedly utilize a small amount of 244 CONRAD JACOBSON laevulose as is evidenced by muscle perfusion experiments. It is questionable whether the muscles can change laevulose into glycogen and to any large extent thus utilize it. It seems quite probable that the liver acts as an organ of storage for the laevulose and alters it in some manner or other to make it more utilizable by the tissues. The liver evidently is indispensable for laevulose metabolism. TABLE 1 Sugar tolerance; grams per kilogram of body weight NORMAL AFTER ECK FISTULA eee Ae Dog A WWerphitr Par | Ginger snaps. ................... Petal a pieced Cookies:....5.0....05..5.0% 19-Lot | Spiced cookies................... 20-Owe | Chocolate layer cake................ ot -eHar | Mary Ann: cookies.............. re) @innamon bun...............--- Pepe | CmMMamon DUNa.:.......5 52+... Pew e Ni GyInper DEA...) . . 2... ee elect e eee ao—huden|Gineer bread. <: .............-55+ 26—Rud |} Short cake, strawberry.............. 27—Ree | Short cake, strawberry..............- 28-Spe | Bread with peanut butter...... 29-Owe | Bread with corn syrup.......... Oe NE Ve ES ee HIGHEST orn or aeseie [Caer WSs Slow 4:30 97.5 xine Rapid 3:45 100.0 Oe: Rapid 2:45 99.0 Riane Slow 4:00 92.0 eae te Rapid 3:00 83.0 eee Rapid 2:45 73.5 Rapid 3:15 73.0 Rapid Salis 80.5 Seay Rapid 3:30 78.0 ae Rapid 3:30 17.9 See Rapid 3:15 75.5 at ie Slow 4:15 88.0 ee Rapid 25 91.5 Dy eet Rapid 2:30 83.0 ec Slow 3:00 106.0 eS cet Slow 2:30 107.0 a hea Slow 3:30 101.5 eet Rapid 2:15 89.0 sees: Rapid 2:15 104.0 Rapid 2:30 26.5 SL ate Slow 2:45 70.5 See Rapid 2:30 72.0 ce sae Slow 2:45 127.5 Rapid Bald 65.0 abe Rapid 3:00 106.0 Rapid 3:00 124.5 Slow ar15 130.0 ME se Slow 3alp 120.5 Merten, Rapid 215 55.0 WV” Bs gas Oe 3:02 90.0 the slow emptying type required 2$ hours as compared with 2} hours for raisin and rhubarb pies (see fig. 10). A subject of the rapid empty- ing type (see fig. 7) required only 2 hours for pumpkin pie as compared with 3 hours for cherry and 22 hours for apple pie. 252 MILLER, FOWLER, BERGEIM, REHFUSS AND HAWK Mince, custard and lemon méringue pies. The mince, custard and méringue pies are higher in protein than simple fruit pies and might, therefore, be expected to undergo greater change in the stomach. Mince pies are generally considered to be the least easily digested of pies, due to their content of meat and spices. This is borne out by our experiments (see figs. 7 and 10). Mince pie required 3% hours to leave TABLE 3 The response of the human stomach to puddings NUMBER KIND OF PREPARATION _ TIER OF. | Seago “TOTAL f=Dali\pindian. 2)... Meso See eee eee Rapid 2:00 113.5 2-Gle ‘|'Gorn'’starchy ) oo... ee eee eee eee Slow 2:30 92.5 3-Dal |’Chocolate Ae) 2 Lee eee eee Rapid 2:00 110.0 A-Spe) Chocolate: f/f s..20 4c eeeee eee Slow 2:00 91.0 5=Dalalp Brown) Bettys. sesso eee eee Rapid 2:15 69.5 6-spe- | Brown ‘Betty ii | onthe EER eee eee Slow 2:45 110.0 (Haar «| Cabinet, no iren tact eee tr ere Rapid 2:00 87.0 S=ope ir Cabinet :.%.' ste o-5 see cee eee Slow 2:45 82.0 O2Thea: |MRICE Ss oi. ee hte ce ee ee 2:00 67.0 1 OSVine (RICE 22k oa See + eye eee Rapid 2:00 101.0 iicbeam|eRiceswitheraisins. see ern Rapid 1:45 76.5 =a ERCup custardes seep eee eee eee Rapid 2:00 107.0 (3=Gle~ |ECup custardee soso soe ace eee Slow 4:15 112.0 14— Dale PP lM ee ser ok eee oe eee Rapid 2:15 113.0 15=Gilee |SPlumes 4 eee: ct ae aoe Slow 4:30 96.5 16=Kar™ | Bread?:;. 5 Seo) 52s os ad aac 2:15 79.5 17-Dalk. |WBread ip 58 eth Acs aac oo eee ee eel 2:00 108.0 18-Gole|Pbread | ota cca tetas he Ae als Oe ee Rapid 1:30 96.0 10= Kar | Bread ¢,cdc.c Ctaee meen ek biel epee eee Rapid 1:45 104.0 Z0=Mic) | haplocac. cates eee eee eee Slow 2:30 74.0 ZA=NicDilpApplestaplocarem.qaeae eee eee ree Slow 2:15 106.5 22-Col. | Gelatines.. 2 pesca eee eee Rapid 2:00 61.5 23=Kar|i'Gelatine), 224 ees eae eee Rapid 1:30 62.0 AVCT AE CX :.calis Beek Sines a Mes atin oo Aeros eee 2:18 92.0 the stomach of an individual who required 2? hours for apple pie and 3 hours for cherry pie. The other subject required 2? hours for mince and 2} hours for either raisin or rhubarb pie. The highest acid re- sponse was given with mince pie, and the acid combining power was also markedly greater than for simple fruit pies. One experiment each was made on plain custard and cocoanut cus- tard pie, the same subject receiving both (see fig. 8). The plain custard GASTRIC RESPONSE TO PASTRY AND PUDDINGS 253 required 15 minutes longer to digest and developed a higher acidity due probably to the slightly higher protein content. The high combined acidities of both would be expected from their known content of milk and egg proteins. The custards required a longer digestion than rhu- barb pie but were by no means handled with difficulty inasmuch as cherry pie took 15 minutes longer than either. Related in composition to the custards are the lemon méringues which, when properly made, are fairly high in eggs. They leave the stomach in about the same time as the custards (23 hours) and show a less rapid development of acidity than other pies (figs. 7 and 10). COMPARATIVE RESPONSE OF THE STOMACH TO PIE CRUST, CONTENTS AND WHOLE PIES The popular opinion that pie crusts are the least digestible portions of ordinary pies would also appear probable from their known com- positions, possessing as they do most of the protein and fat of many pies, especially fruit pies. The absolute digestibility of pie crusts must, of course, depend a good deal on their texture and composition. The response of the stomach to pie contents would be expected to approxi- mate that of fruit sauces or puddings of similar composition. Thus we have found (4) that fruits in general leave the stomach very rapidly with the development of considerable free but little combined acidity. Pie crust alone was given in 100-gram portions to four subjects (figs. 2, 5, 6 and 11). This was evacuated in from 2? to 3} hours. In the case of one subject the contents of a peach pie left the stomach in 14 hour as compared with 23 hours for the crust. For apple pie and crust the required periods were 23 and 3; hours respectively. Pie crust thus remains in the stomach distinctly longer than an equal weight of the contents of apple or peach pie. The acid responses of the two were not so different, although crust usually gave a slower rise with a more sustained curve. The whole pie naturally is intermediate between pie crust and con- tents in the response it evokes. Thus subject “Mul” (figs. 1 and 2) retained peach pie 2 hours and contents 14 hour. Another subject, “Rud” (fig. 9) required a few minutes longer for the rhubarb alone than for the pie, while a third subject, “Wil,” took 2? hours in each case (fig. 6). Thus these fruit pies require little longer to digest in the stomach than the interiors alone. 254 MILLER, FOWLER, BERGEIM, REHFUSS AND HAWK If crust is compared with whole pie, we find in the case of “ Mil” the pies requiring 2 to 23 hours and crust 22 hours; in the case of “Son” 21 hours for rhubarb pie and 2? hours for crust alone. Sub- ject “Wil” required 23 to 3} hours for pies (mince being the slowest) and 3} hours for crust. Crust thus takes somewhat longer to digest in the stomach than the contents of all but the least digestible pies. It will be seen, therefore, that the response of the stomach to whole ples approximates that of the pie contents, in spite of the fact that the crust alone requires a distinctly longer digestion period. We did not, however, find that crust was in any sense indigestible, although 100 grams of it were given at a time. PIE WITH ICE CREAM AND WITH CHEESE To the usual 100-gram portions of apple and cherry pie were added in each of two cases 50 grams of vanilla ice cream. The responses are charted in figures 4,5 and 8. The cherry pie alone required 2} hours. With ice cream it remained in the stomach half an hour longer. Apple pie, however, showed the same evacuation time (24 hours) in each case. In one case the total acidity was slightly higher after adding ice cream; in the other the reverse was true. It is clear, therefore, that ice cream, though greatly increasing the food value of the dish, need not necessarily increase the burden of the stomach to any marked extent. Twenty grams of cream cheese were added to apple pie in one case and compared with apple pie alone (see fig. 6). It will be noted that the acid responses in the two cases do not differ greatly, although nat- urally cheese increases the acid combining power of the food ingested. The addition of the cheese delayed evacuation about a quarter of an hour, and led to a slight secretion of gastric juice after the stomach was otherwise empty. 4 THE RESPONSE OF THE HUMAN STOMACH ,TO CAKES Among the cakes and related products tested out by us were angels’ food, devils’ food, fruit cake, chocolate layer cake, lady fingers, straw- berry short cake, crullers, doughnuts, cream puffs, cinnamon buns, ginger bread and cookies of different kinds. One hundred grams were given in each case. GASTRIC RESPONSE TO PASTRY AND PUDDINGS 255 Angels’ food, devils’ food, fruit cake and short cake. Angels’ food was given to three men. A few days later an equal amount of devils’ food was given to each of the same three subjects for comparison (see figs. 12, 13 and 15). It was found that angels’ food cake remained on the average 20 minutes longer in the stomach than devils’ food, the evacu- ation time of these cakes ranging from 2? to 44 hours. This may well be related to the greater protein content of the angels’ food which is particularly high in eggs. The distinctly higher acidities developed in the stomach in the case of angels’ food cake, and the much more pro- nounced acid combining power must also be due to the high protein content of this cake as compared with devils’ food which, however, is higher in calories due to its sugar and chocolate content. Chocolate layer cake (fig. 15) left the stomach of a subject of the rapid-emptying type in 23 hours or a few minutes sooner than devils’ food cake. The acidity developed was low, due perhaps in part to the depressing action of the sugar icing on secretion. Old and fresh fruit cake were compared, using two subjects on each (see figs. 138, 14 and 17). It is a popular belief that old fruit cake is much more easily digested than the same cake in the fresh condition. Portions of freshly baked cake were given to the men, and three weeks later the same subjects were given other portions of the cake which had been kept in tin at room temperature for that period. In each case it was found that while the fresh cake required 3% hours for gastric digestion the older cake required 34 hours. The acid re- sponses were practically identical on the fresh and older cake. No distinct difference in gastric response to fresh and old fruit cake could, therefore, be detected. Two subjects were given strawberry shortcake (75 grams cake and 25 grams of berries) without cream. This cake left the stomach in moderate time (3 to 34 hours) and gave rise to the development of high intragastric acidities, in part probably due to the acidity of the berries themselves. Lady fingers were given to one subject (see fig. 13). The evacuation time of these was 31 hours as compared with 3 hours for devils’ food and 33 hours for fresh fruit cake. Doughnuts, crullers, ginger bread, cinnamon buns and cream puffs. Ginger bread was given to two subjects (see figs. 9 and 16). It re- mained in the stomach from 2+ to 3 hours or a few minutes less than strawberry shortcake or chocolate layer cake and a shorter time than devils’ food or angels’ food cake. Ginger bread brought about a mod- 256 MILLER, FOWLER, BERGEIM, REHFUSS AND HAWK erate stimulation of acid secretion with, in one case, a rather slow devel- opment. Evidently the ginger did not induce any increased secretion of gastric juice. Cinnamon buns were tested out on two subjects (figs. 14 and 21), and required from 23 to 23 hours. In one subject these buns left the stomach sooner than fruit cake, angels’ or devils’ food cake or lady fingers. The other subject evacuated cinnamon buns a few minutes sooner than doughnuts. In acid response the buns were found to follow the general trend of the cakes just mentioned, not attaining, however, to the level of angels’ food cake. Neither is the acid response as high as that of bread with peanut butter. Probably the sugar of the buns depresses secretion somewhat. Bread with peanut butter (100 grams bread and 25 grams of peanut butter) remained in the stomach 3} hours as compared with 23 hours for cinnamon buns. The acidity developed was considerably higher in the case of bread with peanut butter (see fig. 22). Bread with corn syrup (80 grams of bread and 20 grams corn syrup) left the stomach in 2} hours (fig. 16) or in the same time as ginger bread and a little sooner than chocolate layer cake or devils’ food. The acidity developed was only moderate, being slightly lower than for ginger bread, the syrup probably depressing secretion. Cream puffs were given to one subject (fig. 19). They developed an acidity of over 100 and left in 23 hours or much sooner than crullers. Crullers and doughnuts were compared on one subject (fig. 20). The doughnuts remained in the stomach 15 minutes longer than the crullers or 23 hours. The maximum acidities developed were about the same in each case, but the acid development was slower in the case of crullers due, perhaps, to their higher sugar and fat content. One subject required 44 hours for crullers. His digestion did not, how- ever, appear to be entirely normal on the day of this test. A third sub- ject required for doughnuts a digestion time of 3 hours as compared with 2¢ hours for cinnamon buns (see fig. 21). It seems, therefore, that doughnuts are handled a little less readily than crullers, but that neither of these are unusually difficult for the normal stomach to handle. The response of the stomach to cookies. Spiced cookies, ginger snaps and Mary Ann cookies were studied in a few cases. The Mary Ann cookies remained in the stomach of asubject of the slow-emptying type for 27 hours as compared with 4 hours for devils’ food and 44 hours for angels’ food cake, and this in spite of the higher content of the cookies in dry matter. Ginger snaps required in this subject 3° hour ~ GASTRIC RESPONSE TO PASTRY AND PUDDINGS 257 longer than the Mary Anns. A high acidity was developed on ginger snaps, but this was very slow in development, indicating that the ginger depressed gastric secretion. Spiced cookies left the stomach in 2+ hours in both cases where they were tried out. They remained the same time as crullers but not so long as doughnuts. They left an hour sooner than fruit cake. Fairly high acidities were developed by these cookies. It appears, therefore, that cookies, in spite of their high content of dry matter, leave the stomach sooner than many cakes, and in really equivalent amounts would be handled more readily by the stomach. This may in part be due to their more granular texture. THE DIGESTION OF PUDDINGS IN THE HUMAN STOMACH The following puddings were studied: vanilla corn starch, chocolate corn starch, tapioca, apple tapioca, gelatin, Indian pudding, bread pudding, Brown Betty, cabinet, rice, rice with raisins, cup custard and plum pudding. One hundred grams of pudding were given in each case. Starch and tapioca puddings. Five experiments were carried out on corn starch and tapioca puddings (see figs. 22, 23, 24 and 26). Vanilla and chocolate corn starch and plain tapioca and apple tapioca were tested. All of these puddings left the stomach in from 2 to 23 hours. Apple tapioca was evacuated 15 minutes sooner than the plain pudding. It developed a higher acidity with less combined hydrochloric acid. The starch puddings left the stomach rather quickly and developed a moderately high acidity (91 to 110). Cereal puddings, cabinet, Brown Betty, bread pudding, rice pudding and Indian pudding. Bread pudding quickly developed in the stomach a high total and moderate combined acidity and was rapidly digested, leaving the stomach in from 13 to 2 hours in the case of subjects of the rapid-emptying type (see figs. 17, 24 and 27). Brown Betty pud- ding remained in the stomach 2} hours as compared with 2 hours for bread pudding. A subject of the slow-emptying type required 2% hours for this pudding (figs. 22 and 24). Cabinet pudding left the stomach of a subject of the rapid type in 2 hours as compared with 13 hours for bread pudding (see fig. 27). A subject of the slow type (see fig. 22) required 2} hours or the same time as for Brown Betty. Cabinet developed a lower acidity than bread pudding. ; THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 258 MILLER, FOWLER, BERGEIM, REHFUSS AND HAWK Indian pudding gave a response practically identical with that of bread pudding, leaving the stomach in 2 hours and developing a high acid point of 113 at-an hour and a half, the high protein content of this pudding being undoubtedly responsible for the high combined acidities found (see fig. 25). Rice pudding was tried both with and without raisins (see figs. 1 and 4). The acid responses were in each case practically the same. The evacuation times on plain rice pudding were in each case 2 hours. Rice pudding with raisins required in one case 2 hours and in the other only 13 hours. The combined acidities were practically the same in each case. The rice puddings are thus among the more easily digested puddings. Gelatin and plum puddings and cup custards. A widely used gelatin dessert of strawberry flavor was given to each of two men (see figs. 27 and 28). It left their stomachs very rapidly (in 1} to 2 hours) and gave rise to very little stimulation of acid secretion. The low com- bined acidities also indicate that this product leaves the stomach too soon and with too little stimulatory effect to be markedly altered in the stomach or to throw any considerable burden upon it. Plum pudding was found to be one of the less readily digested pud- dings (see figs. 23 and 27). Its relatively high food value must, how- ever, be borne in mind. A subject of the rapid-emptying type required 21 hours to evacuate this food as compared with 13 hours for gelatin and 1? hours for bread pudding. A subject of slow-emptying type required 43 hours for plum as compared with 23 hours for corn starch pudding. Moderately high total and combined acidities were devel- oped. Cup custards remained in the stomach about the same time as Brown Betty pudding and a few minutes longer than bread pudding, corn starch or Indian puddings, but not so long as plum pudding (see figs. 23 and 24). Being very high in protein, custards would be expected to show a rapid gastric stimulation with much combined acidity. This was found to be the ease. COMPARATIVE RESPONSES OF THE STOMACH TO PIES, CAKES AND PUDDINGS From tables 1, 2 and 3 the average acid responses and evacuation times of all subjects on pies, cakes and puddings may be obtained. They show that the grand average evacuation times were for puddings, GASTRIC RESPONSE TO PASTRY AND PUDDINGS 259 pies and cakes, 2:18, 2:27 and 3:02, respectively. The average highest total acidity was practically 90 in each of the three cases. Perhaps a more significant comparison may be obtained from table 4, in which are summarized the results from six subjects, each of whom besides receiving a number of pies or cakes or puddings was also given certain foods belonging to one of the other two classes. It will be seen that here also pies left the stomach sooner than cakes, and puddings still more rapidly. TABLE 4 EVACUATION TIME AND ACIDITIES SUBJECT Puddings Pies Cakes (Cll. oi Sn Aan 1:30 96.0 3:00 80.2 ‘SIE: Abin an Sh Ag Se eae 2:30 94.3 3:00 118.0 Wears 954+ Rel Risen she aed 1:52 83.1 2:00 97 .0 [et 2 ok ies Does Seen Bed 1:67 las 2:30 76.0 IW. = Glia AS oe a 2:00 104.0 22s 100.5 [REG | oo Fe Ee ie ee 2:41 101.4 3:00 115.0 FNVONA GERM h fests cle coade| 7200 89.8 2:24 93.7 3:00 104.4 SUMMARY AND CONCLUSIONS A study was made of the acid responses and evacuation times of nearly fifty pies, cakes and puddings in the normal human stomach. The average evacuation time on puddings for all subjects was 2 hours _ and 18 minutes as against 2 hours and 27 minutes for pies and 3 hours and 2 minutes for cakes. Averaging the highest total acidities values were obtained for puddings of 92, for pies of 90 and for cakes of 90. Direct comparisons of the three types of foods on the same individuals indicated also that pies were handled more readily than cakes, and pudding somewhat more readily than either. Fruit pies, such as apple, pumpkin, raisin and peach, left the stomach in from 2 to 2? hours and developed a moderately high acidity (90 to 100). Most of the acidity is due to free HCl, the acid combining powers of these pies being low. Cherry pie, high in sugar, remained in the stomach a few minutes longer than the above. Rhubarb pie was treated in the same way as fruit pies, leaving the stomach in from 2 to 2} hours. Custard pies left the stomach in moderate time (24 to 23 hours) and possessed a fairly high acid combining power. Lemon méringue 260 MILLER, FOWLER, BERGEIM, REHFUSS AND HAWK showed a similar response except that the acidity developed more slowly. Mince pies required a rather long time to leave the stomach (23 to 3+ hours) and developed high total and combined acidities. Pie crusts alone remained in the stomach distinctly longer than most whole pies or the contents alone of such pies, and gave a lower but more sustained acid curve. On the other hand, differences in evacuation time of whole pies and the contents of such pies were usually very slight so that pies with crust, if properly made, could by no means be classified as difficult for the stomach to handle. The addition of 50 grams of ice cream to a small piece of pie did not increase the burden of the stomach to any marked extent. The addi- tion of 20 grams of cheese to apple pie increased the digestion time only a few minutes. Angels’ food cake remained distinctly longer in the stomach than devils’ food cake and developed a higher total and combined acidity. Chocolate layer cake left the stomach in moderate time, acid secre- tion being depressed by the sugar of this cake. Fresh and old fruit cakes showed almost identical acid responses and evacuation times in the human stomach. Strawberry shortcakes left the stomach in moderate time (3 to 34 hours) and developed high intragastric acidities. Lady fingers left the stomach in 3 hours or in about the same time as other cakes. Ginger bread evoked a rather slow acid response and left the stomach in moderate time. Cinnamon buns left the stomach sooner than most cakes but with a similar acid response. Bread with peanut butter remained in the stomach longer than cinnamon buns and developed a higher acidity. Bread with corn syrup left the stomach in 2} hours, the syrup de- pressing secretion somewhat. Doughnuts remained in the stomach a few minutes longer than crullers. Acidities developed a little more slowly in the case of crullers, due perhaps to the sugar and fat content. These fried cakes required a digestion time but little longer than the average for cakes. Cookies, in spite of their high content of dry matter, were found to leave the stomach sooner than most cakes, probably on account of their granular texture. Ginger cookies required a little longer than those less highly flavored, and the acid development was less rapid. GASTRIC RESPONSE TO PASTRY AND PUDDINGS 261 Chocolate corn starch pudding, rice pudding with or without raisins, Indian pudding, bread pudding and gelatin left the stomachs of indi- viduals of the rapid-emptying type very quickly (in 13 to 2 hours). Rice pudding with raisins left a little sooner than plain rice pudding. Cabinet pudding, Brown Betty pudding, cup custard and apple tapioca required only a few minutes longer. Plain tapioca remained in the stomach a little longer than apple tapioca. Plum pudding left the stomach slowly. The highest total and combined acidities were caused by Indian pud- ding, bread pudding and cup custard, all of these being high in protein. The other puddings developed moderate acidities except gelatin, which produced little acid stimulation and left the stomach very quickly. BIBLIOGRAPHY (1) Beaumont: Physiology of digestion, 2nd ed., Burlington, Vt., 1847. (2) FisHpack, SmitH, BerGEIM, LICHTENTHAELER, REHFUSS AND Hawk: This Journal, 1919, xlix, 174. (3) MinuerR, BeRGEIM AND Hawk: Unpublished data. (4) Smitu, FisHpack, BerRGEIM AND Hawk: Unpublished data. ith Fei sins GP z G3 Frist: GA S: 3 oot (Ap = Ae He See ddin ng = Fia. 1 Case, MLCR Wil oes inside fie: 1Sgms2) Case. Har 75 AppleRe Ry) ce Yo, HOH Fia. 3 Case Lea 4 Apple Fre with Toe Cream vee PY ¢ Gort OF ud 9, with Aacsins iF) oe. “ho, HOH Ne Fig. 4 263 eS Case. Lea S| Bee @ S Xt Case. Lhkl Ss : Us? Bye weth Cream Cheese eo Ss bs Bis de, Apple Pee. h S LS) Fie. 6 nee wae mphttin Fre es Apgringue Fee G erry A co Yo, HOH Fig. 8 265 4 Paeakt it Clraarbngad Ge A) oe Mo, KOH Fre. 9 Rw Case S Ze Phare eB ie Ad e &-) 44 eigen Pee geo g 449 Pumpkin ae (Fn) Fre. 10 : Case,-Son S 8f Beersl © GB Fig. 11 a) (ase_ Far 30. An e Food In) 4g 5 S18 oy 2a Mo, HOH. /20 /00 80 /.Aour Fig. 12 267 ee ner xt G e ood 17) S We << 2257 rere Op Qpte fresh [EF a Ss S Fie. 13 Case. Mor ti Cong, (REE ap ce Wo, KOH Fie. 14 Case. Owe ws BA Food Cake JE ‘ eCrea a Chocolate Layer ( ake es ea aN S 8s Z S SS Case,- Qwe 30 einger bread, gentgr iuith Conn Syrup Br) oe Yio, HOH Fic. 16 269 Case. Gol /3 Shiced.Cookées ( 4 Eras Fido ‘ U u 70 vitCake, old 9 Pp ldicer 117 Case,. Fee 88 Straw berry Short Ca Fig. 18 270 Case,- Cor BS eailers na, ah Fie. 19 Case, Lot ceed Cooke Bh Bocspmuds = BB 3% Urvutlers 0c “to HOH /Z2O Fig. 20 271 Case She Cinnamon Bun 42 Wh, {No ughnuts ee No, HOH ince, Pal Case, ie x | g , Geis ‘bien 2B. SS Se 1, es ead OSs er wrth ee Br!) Ss Ss Fra. 22 Case. Yle Tiger Tdding tr) toc =e) 3 £8 # ees Ball S : y Fie. 23 Case, Bap z 9 rowr) she x - a Lind mA S AS 8 Ge ie Fia. 24 273 ce Vio, KOH oe “io, HOH (ase. Dal of ae ee ae Sundae Fia@. 25 Case- Ce 2) Bhi | 58 Apis IT) ey ft z um f-vddtng x S i Beings ved P RIG) 27 Case,. Col 35 Strawberry Gelatin (2) ce “ho, HOH SEGMENTAL ACTIVITY IN THE HEART OF THE LIMULUS D. J. EDWARDS From the Physiological Laboratory, Cornell University Medical College, New York City Received for publication March 22, 1920 In the structural make-up of the Limulus heart the arrangement of the muscular elements into a long tube with well-defined segments, and the grouping of the nervous elements into a median ganglion and lateral nerves, are features which have a significant bearing upon its functional activity. Attention is especially directed to these features, since it appears from the work of Carlson (1) and others that activity in this heart is initiated and conditioned by this intrinsic nervous system, and that contraction in all segments is practically simultaneous. These considerations have suggested the present study, which is an analysis of the contractile process in the different parts of the muscular tube, and of the sequence of activity in the different segments in an attempt to discover the exact temporal relations of the different parts. In order to make a detailed study of the contraction event it is neces- sary to employ instruments of unquestionable efficiency. The short- comings of the suspension method of recording heart action are too well known to need repeating; it suffices to note that records made from the heart of this form by arranging in various ways, so that a thread hooked into one of the lateral arteries may be attached to a cumbersome record- ing lever, gives a very inadequate representation of the contraction. Small myocardiographs recording by means of Frank capsules were used in the experiments here described, and the arrangement of the apparatus was essentially the same as described by Wiggers (2) and used for analyzing the contraction process in mammalian auricle. The method of adapting it to the Limulus heart was as follows: The myo- cardiograph arms were adjusted to a given segment so as to exactly span the width of the heart at the end of diastole. A fine silk thread was then passed through the superficial layer of the heart at the lateral borders and the arms of the myocardiograph tied thereto. In this ‘manner there was slight traction on the heart until it began to contract. 276. Land SEGMENTAL ACTIVITY IN LIMULUS HEART 277 Moreover, there was very little resistance offered to the shortening as the myocardiograph and also the Frank capsule were covered with very thin rubber membrane. The record obtained is that of the mid-segmental shortening, the part of the heart that normally exhibits greater contractility than the inter-segmental portions where the ostea are located and in the anterior part the lateral arteries are attached. The myogram of different segments. In considering the relation of the heart of the Limulus to its circulatory system it will be noted that the five anterior segments are each connected with outgoing pathways for the blood. If one reasons from analogy to the mammalian heart, a greater contractility should mark the myogram of these parts as compared with that of the posterior segments. Fig. 1. Simultaneous myograms from heart segments 2, S2, and 5, S;, showing differences in the contour of the curve from these parts. A comparison of the myogram from the two ends of the heart, for example, segments 2 and 8, shows differences in the contour, but the curve from segment 5, which has lateral openings, does not exhibit significant difference from that of segment 6, which has no such open- ings. It appears, therefore, that specialization in contractility, such as is shown in the ventricle of the mammalian heart, is not sharply delimited in the heart of the Limulus to that part from which the circu- lating fluid is outgoing. The contraction amplitude is nearly always greatest in segment 2. The myogram of this segment, in figure 1, shows a small primary rise, b to c, upon which is superposed the main contraction, c to d. The upstroke of the principal rise is rapid and the summit is relatively THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 278 D. J. EDWARDS pointed, indicating that the state of maximal contraction occurs in all of the elements of this segment at practically the same instant. All of these features are exhibited in the myogram of segment 3, except- ing that the degree of contraction is usually a little less. The myograms of 4 to 8 all show a smooth rounded contour and, barring differences in the amplitude, the curves from these different segments are practically superimposable. The shortening process in these segments develops slowly, as is shown fairly typically in the myogram of segment 5, figure 1, and moreover the maximal shortening is reached gradually, giving the curve a flattened summit. These features indicate that the con- traction takes place in a wave-like manner; some of the elements at the beginning of systole are contracting while others are quiescent, while at the end of systole some are undergoing relaxation while others are still contracting. This form of contraction is suggestive of the type found in the mammalian auricle (2). In the vigorously acting heart the myograms of segments 2 and 3 are characterized by a slowly developing state of tension preceding the beginning of contraction. This is shown in the myogram of segment 2, figure 1, in the part of the curve from a to b. The diastolic phase con- tinues so that the curve passes somewhat below the level at the begin- ning of systole. Then follows the state of developing tonus during which the curve gradually swings back. In a slowly beating heart the curve reaches the level before the next systole begins, but with a more rapid rhythm the tonus increase extends right into the next systole. These presystolic tonus changes have not been found in hearts exposed for an hour or more to experimental procedures; they have also not been exhibited in the records from the posterior segments. The sequence of segmental activity. Inasmuch as the experimental evidence for the Limulus heart favors the view that the impulse is initiated in the median ganglion, the precise time of activity in different segments is a matter of some interest. A casual inspection of a pulsat- ing heart gives the impression of a simultaneous event throughout the nine segments. Carlson (3) states that ‘when an empty heart has been beating for several hours, or till nearly exhausted and the rate of the pul- sations is in consequence much reduced, it can be made out . . . . that the contractions start in the posterior third of the heart and travel anteriorly. . . . . This is evidently the case in the fresh and vigor- ous heart.” In a later communication the same writer (4), discussing the degree of automatism of different parts of this heart, attributes this function primarily to the middle third, in a less degree to the posterior third, and least in the segments of the anterior third. SEGMENTAL ACTIVITY IN LIMULUS HEART 279 From the foregoing statements it is evident that the question of segmental sequence can be answered only by exact measurements of initial activity in different segments. This has been done by simul- taneous myograms of selected parts of segments, and by myograms related to the wave of excitation as detected by a string galvanometer. The usual procedure of taking simultaneous myograms was to attach one myocardiograph to segment 2 and a second to segment 3, then after a short interval records were taken. One myocardiograph was then changed to the next segment and the process repeated, and in a similar manner records were taken from all segments. Usually only one series was taken from a heart and sometimes only a few segments were com- pared, but each record was taken in duplicate and the mean of these was used in making up the data for that experiment. In general the results show that the middle segments, that is, segments 4 and 5, exhibit activity in a given cycle earlier than the posterior or anterior segments. A comparison of the beginning of contraction of segment 5 and of segment 2 in sixteen experiments shows that segment 5 preceded by an average interval of 0.046 second. In nine experiments comparing segment 2 with segment 4 there is shown an average time interval of 0.047 second, with segment 4 preceding. This indicates that the beginning of contraction in segments 4 and 5 is practically simultaneous and that it normally precedes the contraction of segment 2. On the other hand, a comparison of segment 2 with segment 6, for example, shows in eight experiments an average difference of 0.023 second, indicating that while segment 6 slightly precedes the beginning of contraction of segment 2, it follows by an average interval of 0.024 second the contraction of segment 4. A comparison of the initial activity in segments 7 and 8 by this method has given less satisfactory results because of the usually low degree of contractility which makes it difficult to determine precisely the beginning of shortening. It is possible, therefore, to make only a general statement concerning them. In six experiments comparing the beginning contraction of segments 2 and 7, it is shown that segment 7 precedes that of segment 2 in four instances and follows it in the remaining two. In a similar way, the data for seven experiments comparing segments 2 and 8 show that the contraction of segment 8 preceded the contraction of segment 2 in three instances and followed it in the remaining four. This indicates that the contraction of the posterior segments normally takes place at practically the same time as the anterior ones. 280 D. J. EDWARDS In order to test the matter of segmental sequence in another way, the excitation process of different segments was taken with a string galvanometer and in turn related to the myogram of a particular seg- Fig. 2. Electrocardiogram recorded from segments 2, I; 4, IT; 6, III, and 8, IV, to show time relation of excitation in these parts to the contraction of seg- ments 2 and 5. Position of points indicated by P. Ordinates in bottom portion of curves designate time intervals of 0.20 second. ment. The current was led off by placing one electrode on part of the dorsal muscle and the other beneath the heart so that the segment under observation rested upon it, SEGMENTAL ACTIVITY IN LIMULUS HEART 281 The characteristic form of the electrocardiogram for this heart has been very well described by Hoffmann (5); little attention, therefore, was given to this feature. In the experiments reported here the begin- ning of deflection was taken as the index of initial activity, and in figure 2 is shown a series of records which illustrate the time relations for different segments in a fairly typical way by this method. Curve J shows that excitation in segment 2 precedes the beginning of shortening in this segment by 0.08 second, and by 0.05 second the initial recorded shortening in segment 5. In curve JJ of this figure it is shown that excitation of segment 4 precedes the contraction of segment 5 by 0.09 second. In a similar manner it is determined from curve J/J that excitation in segment 6 precedes the contraction of segment 5 by 0.06 second, and from curve JV, that excitation in segment 8 precedes contraction of segment 5 by only about 0.03 second. Relating these data, it appears, therefore, that the excitation of the median segments takes place about 0.04 second before excitation of the anterior segments and about 0.03 second before excitation of the posterior segments. In this connection it is interesting to note that the distance separating the electrode contact on segment 4 and that on segment 2 was 29 mm., and the time difference in the sequence of activity of these segments, as related to the myogram of segment 5, was 0.04 second. It is evident from these data that the rate of conduction was approximately 72 cm. per second, a figure that is significant in view of the average rate of 40 em. per second for the conduction rate in the Limulus heart nerves, as obtained by Carlson (6) with less sensitive methods. Mention should be made of the condition met with in some experi- ments in which the anterior segments showed unmistakably an earlier state of activity than the middle segments. We have in figure 3, for example, a record of contractions of segments 3 and 5 showing the contraction of section 3 preceding by 0.05 second. This proved to be a constant feature of the contraction of this particular heart, since other records in this experiment comparing segments 1 and 6, and 1 and 5, show the same general result. Of the thirty-three experiments in which records have been made this phenomenon has been observed to occur with a degree of constancy in only four hearts. It is significant, how- ever, that records taken from the same heart late in the experiment, after a considerable period of exposure, show initial activity in the median segments. A careful study of initial activity in the different segments of the Limulus heart makes quite evident that the distribution of the excita- 282 D. J. EDWARDS tion process is a rapid event, and that it normally encounters no marked areas of resistance which slow the excitation wave in its spread from the middle region of the heart. Garrey (7) has very clearly demon- strated, however, that this heart is susceptible to the usual types of blocking influences. Moreover, it is a fact well known to all who have worked with this heart that an adequate stimulus applied to any part starts a contraction wave which passes in either direction. The evidence presented here supports the conclusion, I believe, that normally the contractions of the middle segments precede those of both anterior and posterior segments. The initial tonus changes often Leh si20 FAVRE EEG RST} iit itiisil i thehttis tis tate t atta ities tii Fig. 3: Myograms from heart segments 3, S;, and 5, S;, showing an instance in which contraction in the anterior segments preceded that of the middle seg- ments. Position of points indicated by P. Time curve designates 0.02 second for each full vibration. exhibited in segments 2 and 38 are significant. It is believed that the increased tonus effects a more quickly acting mechanism in this part of the heart and, coupled with the high state of irritability in a vigorous heart, the recorded activity of this part may precede the initial activity of the relatively less contractile middle segments. The observation of Carlson that a heart, after beating for some time under experimental conditions shows a marked lagging behind of the anterior segments, has been corroborated. But this is taken to mean that the exposure, with the resulting decrease in the nutritive supply from loss of the cir- culating fluid, favors an earlier appearance of fatigue and blocking in- SEGMENTAL ACTIVITY IN LIMULUS HEART 283 fluences in the anterior part because these segments exhibit normally a greater degree of contractility and, therefore, a relatively more rapid using up of the stores of energy. It has been pointed out in the study of the myogram from different segments that the function of contractility is exhibited in a diminishing degree in the anteroposterior direction; that the heart as it approaches exhaustion shows initial activity in the posterior segments, is interpreted not as the normal sequence of activity but as an expression of the change in irritability which accompanies muscular exhaustion. SUMMARY 1. Myograms, made with small myocardiographs, from different segments of the Limulus heart, show greatest contractility in segments 2 and 3 and a diminishing degree posteriorly. The median and posterior segments give evidence of a more gradual development of the shortening of the muscular elements than is shown for the anterior segments, and similarly the end of the contraction process is less marked. 2. The sequence of activity in different segments, by the method of simultaneous myograms and by the method of relating the excitation wave to a standard myogram, indicates initial activity in segments 4 and 5 from 0.04 to 0.05 second before that of segment 2, and in seg- ment 6, about 0.03 second before that of segment 8. 3. In afresh and vigorous heart there is exhibited a presystolic tonus increase in segments 2 and 3 which appears to hasten the contractile event in these parts so that at times it occurs simultaneously with or even precedes the contraction of the middle portions. I wish to acknowledge my indebtedness to the Director of the Marine Biological Laboratories for working facilities and to the Director of the New York Aquarium for supplying a part of the material. BIBLIOGRAPHY (1) For literature compare Cartson: Ergebn. d. Physiol., 1909, viii, 373. (2) Wiecers: This Journal, 1916, x1, 219. (3) Caruson: This Journal, 1904, xii, 70. (4) Caruson: This Journal, 1904, xii, 471. (5) Horrmann: Arch. f. Physiol., 1911, 142. (6) Caruson: This Journal, 1905, xv, 101. (7) Garrey: This Journal, 1912, xxx, 283. THE ARTERIAL PRESSURE CURVE AS INFLUENCED BY THE OCCLUSION OF CERTAIN VASCULAR AREAS AND BY HISTAMINE D. J. EDWARDS From the Physiological Laboratory, Cornell University Medical College, New York City Received for publication March 22, 1920 It has been shown in a number of researches by Wiggers (1) that the contour of the pressure curve recorded by the membrane mano- meter affords reliable qualitative data on the dynamics of the circu- lation. An outstanding feature of this method is that it gives infor- mation concerning the failure of normal dynamic relations earlier than the blood pressure results. It has made possible, therefore, a fairly precise designation of the onset of circulatory failure (2). In the course of studies on blood flow (3) the use of a recording strom- uhr made it necessary to close for periods of as much as twenty minutes the main blood vessels supplying visceral organs and in other instances those supplying the extremities. Roy and Brown (4) were probably the first to call attention to the fact that occlusion causes in the frog a dilatation of the arterioles to twice their normal size in two minutes, the effect extending also into the capillaries and veins. Recently attention has been directed to the possible harmful tendency of occlu- sion measures by the experiments of Mann (5) in which he shows that a shock-like condition can be produced by ligating all of the limb struct- ures excepting the main artery, and Erlanger and Gasser (6) have shown that a similar condition results from a partial obstruction of the inferior cava for about a 2-hour period. In view of these results it is obviously desirable to know whether the less extreme measures used in blood flow experiments produce early changes in the dynamics, also whether there are differences in the character as well as the time of appearance, which distinguish obstructions of central areas from those involving peripheral organs. There has been included in the program a study of the influence of histamine on the pressure curve, since the observed capillary toxicity of this product suggested a causal relation to the effects following oc- clusion of vascular areas. 284 PULSE CURVE AFTER -HISTAMINE AND VASCULAR OCCLUSION 285 METHOD The pressure curve in all experiments was taken by introducing the cannula of the optical manometer into the carotid artery as low down in the neck as possible. A record of mean blood pressure was taken by connecting a mercurial manometer to the opposite carotid artery. This necessitated the closure of two principal arteries but it has been shown that the free anastomosing of the vertebral and spinal arteries permits of compensation through these channels. Dogs weighing 8 to 10 kilos were used, anesthetized with ether and maintained by closed method with especial care to keep the degree of narcosis very light throughout. Four different methods of occlusion were made use of: a, Ligating of three extremities by placing the ligature about the appendage in such a manner as to exclude the main artery; 6, clamping of the principal veins to three extremities; c, clamping of the inferior mesenteric vein; and d, clamping of the inferior cava above the level of the renal veins. In applying the clamp to the inferior mesenteric and the inferior cava a small right abdominal incision was made through which these vessels were exposed and especial care was taken to disturb the viscera as little as possible. Ligation of the extremities and clamping of the principal veins to these parts: In the experiments where ligation effects were studied the ex- tremities were enclosed in the ligatures for a 20-minute period, then for 5 minutes the ligatures were released and in this interval records were usually taken at the beginning and end. The ligatures were then applied again and the procedure repeated. In other experiments clamps were placed on the principal veins and these were manip- ulated according to the procedure given for the ligatures. The methods differ, therefore, partly in that clamping of the veins does not block certain anastomosing pathways close to the trunk, but mainly in that clamping does not produce tissue injury as does alternate application and release of the ligatures. The immediate changes when the circulation is reéstablished in the extremities are not essentially different in ligating from that when clamps were applied to the principal veins. Usually there is a fall in pressure of not more than 10 mm. mercury with gradual recovery during the 5-minute interval; sometimes there is shown a sharp initial rise and then a fall. The optical tracings taken at the beginning and end of this period show no significant changes. 286 D. J. EDWARDS The records, when viewed as a whole for the seven experiments of the series and the successive stages in each compared, show moderate changes after about 3 hours of the alternate occlusion and deocclusion. All of the experiments were continued for approximately 4 hours. The final contour of the pressure curve usually takes the form of a marked primary oscillation followed immediately by a sharp decline of the curve replacing the systolic plateau of the normal curve. These features are quite characteristically shown in the three segments reproduced in figure 1, in which corresponding points in the curves are indicated by similar letters. These curves were taken as follows: J at the outset of the experiment; // at the end of 2 hours and 20 minutes during which the alternate 20-minute occlusion and 5-minute release of two femorals Qa f 2 Fin f Fig. 1. Segments of the optical arterial curves during the progress of successive occlusion and deocclusion of the veins of three extremities. and one brachial veins were carried out; J/7 at the end of 3 hours and 40) minutes of alternate occlusion and release. The change in mean blood pressure taking place in this period is slight, but significant in view of the fact that the directional change is downwards and that the optical curve is indicative of lowered peripheral resistance. The following table will assist in bringing out this feature of the blood pressure change and also the time of appearance of a marked change in the form of the pressure curve. This table shows that averaging the interval before a marked change was apparent in the contour of the pressure curve there was required about 33 hours of alternate clamping of the veins or ligating of the extremities and that much longer time was necessary to give unmistakable signs of decreased peripheral resistance. Definite indications of circulatory failure and PULSE CURVE AFTER HISTAMINE AND VASCULAR OCCLUSION 287 a shock-like condition were shown, however, in only one instance, that of experiment 8, but the data for this experiment give ground for the view that the occlusion effects per se played a relatively less part than trauma of the sensory nerves from application of the ligatures. There were marked sensory signs with each application of the ligatures and mean blood pressure in the early part of the experiment rose from 172 to 206 mm. Hg.; then occurred a fairly sudden break and a fall in press- ure. The optical curve does not give indication of increased resistance as the main cause of this rise; it is probably due, therefore, to an in- creased cardiac output. _ Thus, however viewed, the alternate occlusion and deocclusion of the circulation of three extremities fails to effect changes in the dynamics of a more pronounced character than have been shown to occur in the TABLE 1 : ONSET OF BLOOD PRESSURE LEN GTH OF CHANGE IN STATE OF SEUSS EXPERIMENT EXPERIMENTAL Ss RIAL RESISTANCE : PERIOD PE ESS AT FINISH At the start At the finish CURVE hours |minutes| hours |\minutes 1 106 110 5 50 3 45 Decreased Be 74* 118 3 45 3 15 Decreased 8 172 88 + 30 3 25 Very low 10 132 114 3 2 10 Decreased 11 152 132 3 50 5 20 Decreased 14 166 2. 5 3 45 Decreased ‘15 126 112 4 No change initial stages of circulatory failure of shock (2) and a period well over 3 hours is required for this change to appear. Clamping of the inferior cava: The moderate degree of change in the pressure curve resulting from occulsion of the circulation of the extremities suggested extending the procedure to include larger vascular areas. It is interesting to notice in this connection that the minute volume flow to the posterior extremities, as calculated from the flow in the femoral vein (7), is45 cc. There are no data available concerning the supply to the anterior extremity but it is usually assumed to be approximately equal to that of the posterior extremity. The three extremities, therefore, would represent a volume flow of 135 cc. per minute. On the other hand, it has been shown that the kidneys have a minute volume flow of 195 ec. for the two organs (8). These data represent average normal values and at best are only approximate, 288 D. J. EDWARDS but they serve to indicate that the clamping of the cava affects a many times greater vascular field than was included under the experiments on three extremities. The pressure changes following cava obstruction indicate in a striking way that there is a greatly diminished venous return which causes some cardiac embarrassment. There is a reduction in the amplitude of the curves with a retardation in the gradient of the upstroke and the end of systole is not sharply marked by a distinct incisura. With the first deocclusion of the cava there is evident in the pressure curve a resumption of the main characteristics of the original, but after the second period of clamping little tendency to recovery is shown; further- more, there is indicated a marked relaxation in peripheral tone. The total time of occlusion of the inferior cava necessary to give these indi- cations of a break in the dynamics was never more than 1 hour. It may have been somewhat less since the periods of occlusion were from 25 to 30 minutes. The results of one experiment in which the clamp was applied just posterior to the inlet of the renal veins is of interest, since in this manner the kidney circuit is left intact. The blood pressure fell about 30 mm. Hg. when the clamp was on but showed practically complete recovery each time it was removed. This was repeated four times, covering a period of 2 hours and 10 minutes, and at the end of this time the optical curve showed the characteristics of initial failure. There is suggested, therefore, a degree of proportionality between the time of appearance of these circulatory changes which influence the form of the pressure curve and the extent of the vascular field occluded Occlusion of the inferior mesenteric vein: The inferior mesenteric vein drains an entirely visceral area in contrast to the conditions considered above. Moreover, it represents a vascular field as indicated by the figure for blood flow of 164 ec. per minute (9) quite comparable to that studied under occlusion of the extremities. The outstanding features in the pressure curve during the progress of this type of occlusion are the relatively sudden appearance of the change in the contour of the curve and the slight indication of progressive development after this initial change. The first period of clamping was about 40 minutes and the pressure curve shows (fig. 2, IZ) at the end of this time a sharp primary oscillation (a—b) without plateau (b-c) and with low diastolic limb (e-f). Hach record was taken at the end of the 5-minute period of deoeclusion. Segments III to VI, while differing somewhat in amplitude, show slight changes in their essential PULSE CURVE AFTER HISTAMINE AND VASCULAR OCCLUSION 289 characteristics and give evidence of a comparatively fixed state of dilatation after the first period of stasis. Furthermore, these succeeding periods of occlusion show only moderate depression in blood pressure and nearly identical recoveries during the periods of deocclusion. These changes lend support to the view that the cardiac mechanism is not markedly affected by the degree of reduction in the venous return. The significant features in these results appear, therefore, as the rapid initial decrease in blood pressure with comparatively small changes from the alternate periods of occlusion and deocclusion, and the slight tendency to a progressive type of change in the pressure curve. Many factors are concerned probably in the production of the effects fol- lowing occlusion of the venous return from such vascular areas, but the ia 5 | | | ; any 2 1 Mi | 4 | | ay c | i} H fea ; d \ ig € : | ae ad \ | Yi ¥ . 4 Cc | ‘ nee 292 | é Ti | d a | hg " 4 a aes b Bk ‘i GL | l ka. Ly meas a J v y ‘ ee yy + P| SS a2 te 2 F = # Fig. 2. Segments of optical arterial curves during the progress of successive occlusion and deocclusion of the inferior mesenteric vein. following warrant special consideration: (a) Mechanical distention of the blocked venous system; (b) a toxic action of products formed in the stagnant areas which exert further injurious effects when periodically flushed into the circulation; and (c) the damaging of the vascular walls with accompanying alterations in permeability. The present experi- ments do not permit of direct conclusions concerning which of these factors plays the essential réle in the production of the effects studied. Indirectly they give evidence, however, in favor of the view of mechani- cal dilatation as the main cause of these primary changes, since it has been shown that occlusion of an intestinal area which affords by virtue of its plastic character a favorable site for mechanical stretching of its vascular walls, is immediately affected and the change appears almost in full magnitude at the outset. The occlusion of a peripheral 290 D. J. EDWARDS field of comparable vascular magnitude, which presents a comparatively rigid vascular area, gives results essentially different in that corre- sponding changes in the pressure curve appear much later in the course of events. Thus far, no account has been taken of the possible toxic effects of substances formed during the periods of stagnation. In order to throw light directly upon the probable degree of influence which substances of this character exert the experiments below with histamine were performed. Successive injections of histamine: The results of Dale and Laidlaw (10) show quite conclusively that the administration of histamine may produce a condition simulating traumatic shock. This phenomenon & JE dE IEE mee | | J e ; ey é FM d f 2 Fig. 3. Segments of optical arterial curves at certain stages during the blood pressure change from injection of 0.15 mgm. histamine. The numbering of seg- ments corresponds to positions indicated on figure 4. they assign to increased capaciousness through the relaxation of capillary tone from the toxic action of this substance. In their experiments cats were used for the most part and a dosage of a milligram per kilo body weight was found necessary to produce circulatory collapse. In the present experiments on dogs it was sought to test not only the immediate effects but also the accumulative action of histamine on the dynamics of the circulation; accordingly, the injections were. usually 0.04 or 0.05 mgm. (the ergamine of Burroughs, Wellcome & Co.) repeated at 10-minute intervals over a period of 3 to 4 hours. Such a dosage never fails to give at each injection a fall of over 50 per cent in blood pressure at the start. The recovery is rapid at first and usually complete in less than 3 minutes but as the experiment progresses and PULSE CURVE AFTER HISTAMINE AND VASCULAR OCCLUSION 291 also following larger injections a somewhat longer time is required before the pressure returns. In figure 3 are shown portions of the optical records taken during the changes produced by the injection of 0.15 mgm. histamine. These curves, taken in connection with the tracing of the blood pressure change shown in figure 4, portray the immediate effects of this substance upon the pressure relations. Coincident with the fall in mean blood pressure the optical curve shows (fig. 3, segment JJ) a large primary oscillation it UH eh Fig. 4. Tracing showing blood pressure change as related to optical curves in figure 3. Numerals correspond to segments in that figure. The intervals A, Band C correspond to 2 minutes, 5 minutes and 5 minutes, respectively. Time curve indicates seconds. (a—b), indicative of a collapsed state of the arterial system (c-d), and a flattened diastolic limb (e-f), signifying a low state of peripheral resist- ance. There is a rapid restoration of peripheral tone indicated in segment JI/, e-f, by the immediate rise in this portion of the curve. This change in resistance suffices to restore mean blood pressure in a short time but it will be seen from segment JV that the recovery of the dynamic relations is complete before mean blood pressure has reached its original level. The extent to which cardiac output is influenced by 292 D. J. EDWARDS these changes does not appear from the present data. Further experi- ments dealing with this feature are in progress. Turning now to the influence of the successive injections, it was demonstrated at the outset that small doses, e.g., 0.01 mgm. per injec- tion, have little accumulative effect. With individual injections in amounts of 0.05 mgm. it has been possible generally to produce in dogs of about 8 kilos signs of initial circulatory failure after a total of 0.5 to 0.7 mgm. had been injected in the manner indicated. The stages in the failing dynamics are very well brought out in figure 5. Segment I was taken at the outset of the experiment and serves as a basis for comparison; segment JZ was taken after seven 0.04 mgm. injections had been given; and segment J//, after thirteen such injections. Close SS aed 32S Wile AC oe eee & | 2 4b | e | € Bias | Cc) | | " | ay | 43 | | L) Plead | | . | a : hemes 1 aT BPAY B.PIL3E AP-/42 | RPE" | BER Fig. 5. Segments J, IJ, III and IV, optical arterial curve during progress of successive injections of histamine. Segment V, taken 25 minutes following injection of 5 ec. adrenalin (1: 5000). Description in text. inspection of these curves shows a sharper contour in JJ and III and a slight diminution in the height of the primary oscillation. These features are significant in that they anticipate the more marked changes characteristic of failing peripheral resistance and that they take place without significant changes in mean blood pressure. At this stage of the experiment the dosage was increased. There followed two 0.2 mgm. injections at the usual 10-minute intervals and one injection of 0.5 mgm. Whereas the smaller amounts had usually caused a fall in blood pressure to 50 to 60 mm. Hg., the latter dose depressed the press- ure to 36 mm. Hg. Segment JV was taken 20 minutes following the 0.5 mgm. injection and at a time when blood pressure had practically recovered. It will be seen that the introduction of the larger amounts augmented the features already noted and that the curve now exhibits PULSE CURVE AFTER HISTAMINE AND VASCULAR OCCLUSION 293 a certain likeness to that shown under the occlusion of the veins to the extremities and characteristic of initial failure. These results present quite typically the effects on the optical curve of repeated injections of histamine. It is significant to note, also, that all of the experiments were continued for a period of about 4 hours without a single instance of circulatory collapse appearing, notwith- standing injections at the conclusion of some of the experiments of a milligram of histamine following closely upon successive doses only slightly less, with a total given during the experiment amounting to. 3 mgm. or more. These observations lend support to the view that the dilatation resulting from the amounts of histamine used, and adminis- tered in the manner indicated, causes only moderate dynamic changes. Furthermore, it appears that the change is one that may be due to the reduced resistance from simple capillary dilatation. With the possi- bility in mind that a moderate vasoconstriction might counteract the effects of the periodic histamine dilatations and thereby restore the failing dynamics, injections of 5 ce. of adrenalin (1:5000) were given at the rate of about 1 cc. per minute. The usual constrictor response followed but passed off in ten to fifteen minutes and a vascular equili- brium was again established. The pressure relations at this stage are very well shown in segment V of figure 5, taken 25 minutes following the administration of the adrenalin. It shows that, notwithstanding a good state of peripheral resistance signified by the height and slope of the diastolic limb, e-f of the curve, there is a diminished distention of the arterial system evidenced by the increased amplitude and sharp contour of the primary oscillation, a—b. The marked vasoconstriction of the adrenalin appears to have the effect, therefore, of augmenting the condition started by the histamine. In this connection it is interesting to consider briefly the possible mechan- ism of action of adrenalin whereby it accentuates the dilatation initi- ated by histamine. Hartman and McPhedran (11) state that large doses of adrenalin dilate the intestinal vessels. On the other hand Erlanger and Gasser (12), while not able to agree with this conclusion, point out that portal pressure is greatly increased by giving large amounts. The results presented above give unmistakable evidence of a decrease in peripheral resistance after repeated small doses of histamine. It is well to bear in mind that several factors may contribute to this end, important of which are decreased arterial tone, dilatation of the capillary areas and decreased viscosity of the blood. While it is not possible THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 294 D. J. EDWARDS to state with certainty regarding which of these factors plays the greater role in the reduction of peripheral resistance with histamine, it is reason- able to assume that the action of adrenalin described above is due to a further dilatation of the weakest link in the vascular circuit. The usual absence of reduction in arterial tone from adrenalin, and also the excellent results of Dale and Richards (13), indicating that his- tamine is a capillary dilatant; support the conclusion that the capillary system is the part less resistant and is the area which compensates for the intense constriction of the adrenalin. The data at hand do not indicate whether this effect is mainly on the intestinal area or is gener- ally present in the capillary blood-bed of all organs. The observation of a great increase in portal pressure lends support to the view that the area supplied by this vessel may play the larger part in these results. One point still remains to be discussed; that is, the bearing of the results with histamine upon the effects following occlusion of different. vascular areas. The optical curves show features in common, each presenting changes indicative of initial failure of normal dynamic relations. It might appear, therefore, that the occlusion effects were caused by metabolic products with histamine—like action formed during the periods of stasis. There seems little doubt that such prod- ucts may contribute to the effects of occlusion, but that they are not the main cause is evidenced, I believe, by the relatively large amounts of histamine necessary to initiate failure changes and by the observa- tion recorded above that during the periods of reéstablished circulation through the occluded part there are never indications of a dilating sub- stance being carried into the general circulation. SUMMARY 1. Alternate periods of occlusion and release of the circulation of three extremities, when continued for about 4 hours, produce changes in the optical pressure curve comparable to those shown in initial circulatory failure. There appears to be little difference whether the occlusion is produced by a, ligating all of the limb structures except the main arterial supply; or b, clamping the principal venous channels from these organs. 2. Clamping the inferior cava anterior to the renal veins shows indications in the pressure curve of cardiac embarrassment from di- minished venous return. Two 30-minute periods of occlusion of this vein interrupted by 5-minute intervals of reéstablished flow produce PULSE CURVE AFTER HISTAMINE AND VASCULAR OCCLUSION 295 marked changes in the dynamic relations. By lessening the vascular area blocked, the time of appearance of these changes is somewhat delayed. 3. Blocking the venous return from a visceral area brings on a failure type of pressure curve much sooner than occlusion of a peripheral area of comparable vascularity. 4, The observations of a slow, progressive development of the occlu- sion effects of peripheral structures, the rapid onset with little evidence of progressive development for visceral areas, and the character of the changes portrayed in the pressure records, support the interpretation that the cause is essentially one of mechanical distention of the blood- bed in the occluded parts. 5. The optical pressure curves during the blood pressure change from a small dose of histamine show a sudden decrease in the peripheral resistance with a greatly diminished arterial distention, and an immedi- ate recovery of peripheral tone. Injections of 0.04 to 0.05 mgm. repeated at 10-minute intervals produce after many such doses a small degree of decrease in peripheral resistance. Larger doses serve to accentuate these initial effects of the periodic dilatations. 6. Adrenalin augments the failure changes initiated by the repeated histamine injections. BIBLIOGRAPHY (1) Wieeers: Modern aspects of the ciculation in health and disease, Philadel- phia, 1915; Arch. Int. Med., 1915, xv, 77; This Journal, 1914, xxxili, 382. (2) WiceeErs: This Journal, 1917, xlv, 485. (3) Epwarps: This Journal, 1914, xxxv, 15. (4) Roy anp Brown: Journ. Physiol, 1879, 1, 323. (5) Mann: This Journal, 1918, xlvii, 248. (6) ERLANGER AND Gasser: This Journal, 1919, xlix, 151. (7) Burtron-Opitz: This Journal, 1903, ix, 161. (8) Burron-Orpitz: Arch. f. d. gesammt. Physiol., 1908, exxiii, 553. (9) Burron-Oprrrz: Arch. f. d. gessammt. Physiol., 1908, exxiv, 495. (10) Date anp LarpLaw: Journ. Physiol., 1918, lu, 355. (11) Harrman anp McPueprRan: This Journal, 1917, xiii, 311. (12) ERLANGER AND Gasser: This Journal, 1919, xlix, 345. (13) DaLe AND Ricuarps: Journ. Physiol., 1918, li. 110 THE ARTIFICIAL PRODUCTION OF MONSTERS DEMON- STRATING LOCALIZED DEFECTS AS THE RESULT OF INJURY FROM X-RAYS W. M. BALDWIN From the Union University (Albany) Medical College Received for publication March 22, 1920 In a recent paper (1919)! the author has detailed the results of a series of experiments upon frogs’ ova where, by the use of X-rays generally applied to the surface of the ovum, a uniform defect was produced in the embryos. Inasmuch as a fixed quantity of energy was concentrated upon the ovum in the early stages of its development, practically all of the embryos presented the same gross and micro- scopic structural defects. In this present series of experiments, how- ever, the X-ray energy was permitted to act upon a relatively small amount of the egg substance through the utilization of a perforated lead screen interposed between the tube and the specimen. The energy was derived from a large Coolidge tube which carried a milli- amperage of 50 at 50 kilovolts. The interposed lead diaphragm meas- ured about 24 mm. in thickness, through which a hole 0.3 mm. in diameter had been bored. Fertilized eggs of the frog, in a develop- mental stage no later than the two-celled, were placed behind the hole in the diaphragm in line with the center of the target of the tube and at a distance of 22.5 em. from it. By this arrangement a maxi- mum effect of X-ray energy upon a relatively small amount of the pro- toplasmic mass of the egg was brought about. An effort was made to so place the eggs that the energy entered at a point in the animal hemisphere and emerged at a corresponding point in the vegetable hemisphere. A columnar mass of protoplasm, 0.3 mm. in diameter, extending from the upper to the lower surface of the 1.7 mm. ovum, was subjected thereby to a maximum amount of direct radiant energy. In each instance an exposure varying from two to four minutes was 1 The artificial production of monsters conforming to a definite type by means of X-Rays. Anat. Rec., xvii, no. 3, November, 1919. A bibliography list is appended. 296 PRODUCTION OF MONSTERS BY X-RAY 297 made. The eggs were then transferred to a specimen jar containing 1000 cc. of tap water and permitted to develop at room temperature, the water being changed frequently. It was ascertained after repeated experimentation first, that the amount of X-ray energy passing through the thickness of the 24 mm. screen at the distance given, and under the conditions of exposure men- tioned, exercised no injurious effect whatever upon normal ova. Like- wise, it was demonstrated that the secondary radiation arising from the edges of the hole in the diaphragm was so small in amount as to be a negligible factor. The effects recorded in these experiments were owing apparently, therefore, to the direct action of the X-ray energy upon the protoplasm of the egg using this term in its broadest sense, in the restricted region mentioned. A study of the external features of the developed embryos reveals but little in comparison with the microscopic findings. In some instances it may be detected that the development of the eye or of the ear or both upon the same side is defective. Occasionally, as well, the gill tuft upon one side is absent. Furthermore, a lateral curvature of the body at the junction of the head with the trunk or at the level of the trunk, or at the level of the junction of the trunk with the tail may be observed, but this curvature is no greater than that which occurs with many normal tadpoles which are fixed together in large numbers. Some show an asymmetrical position of the dorsal fin at the level of several adjacent segments of the trunk. On the other hand, the color of the epidermis is normal throughout. The length, breadth and thickness of the tadpoles have not suffered to any appre- - ciable degree over the controls. A study of the serial sections of the tadpoles demonstrates, however, a peculiarly abnormal condition of the anlagen of several segments of the trunk and neck regions. As the sections are traced in sequence from the head to the tail, at a level beginning ordinarily at the optic chiasm and extending caudally to the region of the stomach, a marked difference is to be observed when the corresponding portions of the embryo to the right and to the left of the notochord are compared. Upon first glance the impression gained is that of unusual asymmetry of the embryo in which the neural tube, the notochord and the enteron have left their customary median position and migrated together toward one or the other side. _The space ordinarily occupied by the mesodermal tissue and its derivatives is considerably smaller in area upon the side toward which the notochord has migrated than upon the 298 WwW. M. BALDWIN opposite side. At such levels the neural tube and the notochord, while occupying the real median plane of the embryo so far as anlagen are concerned, actually lie nearer one lateral trunk surface than the other by reason chiefly of the above mentioned inequality of development of the mesodermal derivatives. This inequality of mass distribution is the result of a retardation of development of certain consecutive hemi- segments of the embryos. In no instance is this phenomenon restricted to one segment. Most often it involves those located immediately caudal to the point of origin of the optic stalk from the brain, and usually the defect extends as far caudally as the stomach. A close microscopic examination of the cells in this defective area demonstrates the point that the various organ anlagen are not absent, but the processes of differentiation in the cells constituting them have been so greatly checked by the X-ray energy in contrast with those of the corresponding anlagen of the opposite half, that they present an appearance of the greatest degree of developmental retardation. The individual cells present certain features of both cytoplasm and of nuc- leus, moreover, which are unmistakable indications of deviations from the normal morphological characters of these cells. The abrupt shifting of the neural tube, of the notochord and of the enteron from their normal median position in the mass of the embryo gives rise to the appearance of a lateral flexure of the trunk at the level of injury. In the caudal portion of the affected area a simi- larly abrupt reverse flexure brings these structures back to their normal median position. A singular feature of the abnormal embryos consists in the absence of any indication whatever of inflammatory reaction to injury in the affected area. Similarly the absence of ex- truded cells and of protoplasmic degenerations, involving either nucleus or cell body, may be interpreted as indicating a type of injury not of the most severe grade, as that noted by the author in his previous paper. The cytoplasm of the cells in the affected area possesses a large, more-coarsely granular appearance than is normal. Pigment granules which, in the later stages of development, are more completely restricted to the epidermal cells, are to be observed in those cells the substance of which lay in the path of the rays through the ovum. Their presence in numbers enables one to mark out readily the direction of this path in the sectioned embryo. The nucleus of these cells presents, in the main, two general departures from normal; first, the chromatin is clumped either into large, deeply-staining masses restricted for the PRODUCTION OF MONSTERS BY X-RAY 299 most part to the periphery of the nucleus, leaving the center in contrast comparatively clear; or, secondly, the chromatin is more finely gran- ular, more uniformly distributed, but invariably much less intensely staining than normal. It might be inferred that these two features represent different degrees of the same type of injury. There is no difficulty encountered, however, in readily differentiating between normal and rayed cells through these features. Furthermore, the contrasted appearance of these cells to normal cells in these partly- rayed specimens is all the more readily appreciated since both types may be seen in the same microscopic field. In the previous paper on type embryos, the presence of nuclear and of protoplasmic detritus in the body cavities, such as the enteron, neurocele, optic vesicle, etc., was interpreted, as had been done by Hertwig previously with radium embryos, as a certain indication of such a degree of severe injury as to eliminate these cells from partici- pation in the subsequent developmental stages of the embryo. Such features in this present series are, however, as was mentioned above, entirely absent. The injury produced by the rays was not sufficiently intense, apparently, to destroy the vitality of any of the cells. There ensued, on the contrary, a temporary suspension of differentiation fol- lowed by a lengthened tempo in the developmental rate of these affected cells. The absence of a reactive inflammation is, at present, difficult of interpretation unless we assume that possibly the damage wrought by the rays has been wholly intracellular. As the different systems of the embryo are taken up in order and studied through the serial sections, several facts assume significant importance dependent upon the degree of reaction to X-ray influence. The ectoderm in the areas not in the path of the X-rays appears normal in thickness and in the disposition of its cells. Where the pencil of rays has passed through the embryo, however, those ectodermal cells. in line with it depart from the normal in structure most markedly. In this area the entire thickness of the ectoderm in general is markedly increased. The individual cells are larger and the nuclei more deeply staining. The cell borders and the nuclear outlines are, however, distinct. This increase in thickness is, however, not uniform through- out the area affected, there being produced as a result an unevenness in the external surface of the ectoderm which amounts to a corrugation or wrinkling of this layer without involving, however, the basement. membrane. In many instances the ridges so produced are hollowed out through the presence of channels which run a greater or less dis- 300 W. M. BALDWIN tance in a direction generally parallel to the underlying surface of the ectoderm. This reaction to X-ray influence resembles strongly that referred to by the author in his previous paper and noted as well by Hertwig in his study of radium embryos. There is no evidence of actual death or of desquamation of any of the ectodermal cells, neither are there indications of inflammatory reaction. In no specimen has the thickness of the layer been reduced. As the brain and spinal cord are followed through the sections, there is to be noted at the level of injury a marked dwarfing of the brain vesicle upon the affected side. What appears to be a compensatory hypertrophy of the vesicle wall upon the unaffected side has resulted in a shifting of the median ventral furrow in the floor of the neurocele toward the affected side. This accentuates the appearance of asym- metry. The neuroblasts upon the unrayed side are normal in size, in shape and in the appearance of their nuclear and cytoplasmic material. Stratification ordinarily has progressed to such a degree that the ex- ternal fiber layer may be distinctly differentiated from the nerve-cell layer. Upon the affected half of the vesicle there is a remarkable dwarfing of the neuroblast layer. The size of the individual cells is remarkably reduced, the nucleus is small and but faintly staining, and the protoplasm is but slightly granular. The sectional cross-area of the cell-mass constituting the floor of the vesicle is less than one-half that of the unaffected side. The dorsal median ridge in the roof of the brain occupies, however, its normal position. But very few bits of extruded nuclear or protoplasmic material are to be found in the neuro- cele. As the tube is traced cephalically or caudally, it is found to make a sharp, almost right-angled flexure away from the affected hemiseg- mental area, this flexure being so sharp when seen in the cross-sections as to appear as almost a longitudinal section. The purpose of these bends is naturally to bring the tube back into its normal median posi- tion in the unaffected segments of the body. Dependent upon the level of the segments involved, either the optic vesicle or the otic vesicle, or both, upon the same side may be involved in the injury. Where the former has been encountered by the X-rays the cells constituting the optic stalk and the vesicle demonstrate the same abnormal features of nuclear and of protoplasmic structure as those presented by the neuroblasts upon the affected side of the brain. Stratification of the lining cells is absent; the lens is ordinarily not formed. These same features may be detected as well when the otic vesicle is involved in the injury. Where, however, a greater amount PRODUCTION OF MONSTERS BY X-RAY 301 of X-ray energy has acted upon these two organs, they are then entirely absent. The optic vesicle may be reduced to a mere stump attached to the ventral surface of the brain. There is no indication given through the thickening of the ectoderm of the formation of a lens. As was noted in the brain cavity, there is but little indication of the pres- ence of extruded cells or of nuclei in these two vesicles. The shifting of the notochord from its normal position toward the affected side is brought about by the same abrupt angularity of flexure as was noted in the neural tube. By reason, however, of a greater degree of migration, its normal ventral relationship to the tube is somewhat altered. There is nothing, however, to be noted as ab- normal either in the appearance of its cells or in their arrangement. Since the injury to these embryos is restricted to the neck and cephalic segments, the only portion of the enteron demonstrating abnormalities is the pharynx. In those segments where the injury has been severe but half of the pharyngeal wall is present. No effort has been made by the organism apparently to remedy the defect through the completion of the defective half of the tube-wall. The lining cells upon the affected half remain small, spherular and isolated. The number of them is greatly reduced and an orderly stratified grouping is completely absent. The general arrangement of these cells corre- sponds to that found in a very early stage of development. Differen- tiation of these cells, as was true also of the neuroblasts and of the cells of the optic vesicle, has been practically completely suspended. Upon the unaffected side, however, the pharyngeal wall is distinctly clothed by well-differentiated cells corresponding to the period of devel- opment. The contrast between these two halves of the pharyngeal wall is one of the most striking features presented by the embryos. Similarly, the pronephros in the affected area is reduced greatly in the number of its tubules and in their size. The individual cells lining the walls of these tubules present features corresponding to those seen in the affected half of the pharynx, 1.e., irregularity in arrangement together with spherulation and isolation of the cellular elements. Cephalic to and caudal to the affected area the muscle segments are normally symmetrical in point of differentiation and of growth. In the affected segments, however, the myoblasts may be sharply con- trasted from the standpoint of differentiation. The pigmentation of these cells, present normally only in early stages of development, is retained and, as well, the spherular, isolated, non-differentiated char- acters of the cells are distinctive. Upon the unaffected side the myo- 302 W. M. BALDWIN fibrillae may be readily distinguished. The myoblasts, which at this stage are elongated, have assumed their definitive arrangement with respect to the long axis of the embryo. These cells are sharply con- trasted developmentally with those of the affected side. The cross- sectional area of the affected half of any myomere approaches ordinarily but less than half that of the unaffected side. But few or no abnormal features may be discerned either in the heart or the pericardium. The great vessels passing through the rayed seg- ments, on the other hand, demonstrate certain abnormal features. Most prominent among these is the retarded, embryonic condition of the lining endothelial cells and the undifferentiated condition of the tunics of the vessel walls. In only the largest of these have they made their appearance. Moreover the smaller blood vessels are entirely absent. The cellular elements which should assist in the formation of them remain small, spherular and isolated. There are no coordinated attempts at organization. As a direct result of these defects the gill tufts upon the affected side, normally dependent for their size chiefly upon the presence of these blood vessels, owing to their almost complete absence are markedly dwarfed or completely absent. There is a great difference in size between the mesenchymal mass upon the affected half when compared with that upon the unaffected half of the same segments. That upon the affected half is much smaller owing to the defective size, shape and number of its cellular elements. The affected cells retain their earlier, embryological, spherular features. Isolation of the individual cells is most marked. The absence of mitotic figures in these cells is characteristic. Development of them has been strongly inhibited. There is no indication, however, of an edematous condition which might be anticipated in view of the absence of small blood vessels, thereby bringing about a separation of these cells as was found to be the case where the heart was markedly affected when the whole ovum was rayed. Microscopic study of these mesen- chymal cells shows a lack of features of specialization in their pro- toplasm. It is significant that evidences of reparative reaction to the injury sustained by the cells are entirely wanting. Assuming that the posi- tion of the neural tube and of the notochord indicates the median plane of the body, the cross-sectional area of the affected hemi-seg- ments is reduced to less than half that of the normal segments. The. several factors responsible for this condition are to be found, as was PRODUCTION OF MONSTERS BY X-RAY 303 noted above, in the undifferentiated condition and lack of growth of the myomeres, the pharynx, the blood vessels, the otic and optic ves- icles and in the connective tissue-forming cells themselves. These specimens demonstrate, therefore, the possibility of the pro- duction of localized injury by means of X-ray energy. This localiza- tion is sharply focused upon certain regions but involves all cells within the rayed area. This may be marked out most distinctly from the unaffected area. It is significant, in addition, that the injury received by the ovum at the two-cell stage should be transmitted through so many cell generations without any greater manifest effort at restitu- tion. Sharp localization of organ-forming substances at this early stage of development is brought out as well by the experiments. The nature of the change whether physical or chemical is of less importance from the embryological standpoint than is the demonstration of sym- metrical distribution of these substances at so early a developmental stage. A consideration of the nature of the change wrought, whether an upset in the phase of an emulsion globule or a distinct oxidation or reduction of protein, lipin or carbohydrate is less striking than the phenomenon of the transmission and distribution to subsequent cell- generations of this same altered sukstance. Here again this feature is overshadowed by the manifest inability either of the cells or of the organism as a whole to rectify the damage done. The absence of cell extrusions, of protoplasmic and of nuclear detritus, all argue for an invariably uniform quantity of energy applied to the embryos, and so regulated as to avoid these evidences of more severe injury as were noted in the previous paper. The absence of blood-cellular features of inflammatory reaction may be referred to two possible sources, due either to a destruction of the anlagen of these elements but for which no definite evidence exists, however, or because of the intra-cellular and non-destructive character of the alteration produced. THE RELATION OF THE EPINEPHRIN OUTPUT OF THE ADRENALS TO CHANGES IN THE RATE OF THE DENERVATED HEART G. N. STEWART anp J. M. ROGOFF From the H. K. Cushing Laboratory of Experimental Medicine, Western Reserve University Received for publication March 26, 1920 INTRODUCTION It is well known that-adrenalin causes acceleration of the heart after section of the vagi and excision of the stellate ganglia. v. Anrep (1) showed in the dog that stimulation of the peripheral end of the splanch- nic nerve caused acceleration, associated with the peculiar features of the splanchnic blood pressure curve which Elliott (2) explains as due to the augmented epinephrin output known to be elicited by stimula- tion of that nerve. Pearlman and Vincent (3), working with the heart only partially isolated from the central nervous system by section of the vagi, have confirmed Elliott’s interpretation of the peculiar form of the blood pressure curve caused by splanchnic stimulation and observed manifest augmentation of the heart. We have ourselves shown (in the cat) that the epinephrin given off at the normal average rate under the conditions of our experiments causes acceleration when the adrenal vein blood collected in a cava pocket for a minute or two is released. The acceleration begins a little time after release of the pocket, the length of the interval depending upon the state of the circulation, and is coincident or nearly so with the dilatation of the pupil if the superior cervical ganglion has been previously excised. The beginning of the dilatation has been seen to coincide approxi- mately with the “dip” associated by Elliott with the action of the liberated epinephrin (4). Also when the splanchnic is stimulated with the adrenal veins open it can sometimes be seen that about the same time as the reactions of the eye (sensitized by removal of the superior cervical ganglion) appear, the denervated heart begins to become accelerated. From all this it seems clear that epinephrin 304 EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 305 liberated from the adrenal in response to stimulation of the splanchnic plays a part in the acceleration of the heart. We shall show later on that the whole acceleration is not due to the epinephrin, but that other factors are involved, since a good acceleration, even as great as that obtained with intact adrenals, can be observed when they have been removed or when discharge of epinephrin has been prevented in other ways. It will be pointed out in discussing these results that, with such a reaction as acceleration of the heart, there is no inconsistency in attributing a share in the reaction to epinephrin and yet asserting that sometimes as great a maximum acceleration may be attained in its absence as when it is being given off. It is possible that the position of the maximum acceleration or of the beginning of the acceleration on the blood pressure curve may be shifted when epinephrin action is excluded. But however this may be, it will, we believe, be evident when our results have been displayed that it would be a very unprom- ising venture to attempt to found a method of estimating the rate of output of epinephrin upon such a reaction, even under conditions, as in splanchnic stimulation, in which it is known that an increase in the epinephrin output can contribute to it. Where adrenal blood is col- lected in a cava pocket and then released there is no question that the acceleration is produced by the epinephrin and practically by that alone, because the experiment has been so simplified that only one factor is acting, namely the admission into the circulation of the epi- nephrin-containing blood. But even here the reaction is of such a char- acter that it could hardly lend itself to anything like an exact assay of the epinephrin in the blood from the pocket. Nevertheless in a recent paper Cannon (5) has described experi- ments on cats in which, from the acceleration of the so-called denervated heart caused by stimulation of the central end of the sciatic nerve, by asphyxia and by emotional excitement, he professes to prove that these conditions produce a marked increase in the rate of output of epi- nephrin from the adrenals. Apparently admitting that the catheter method (6) is a difficult one to obtain positive results with, he intro- duces this as a relatively simple method which can be carried out by ‘any competent experimenter,” and he states that the results of these experiments confirm ‘‘in every particular’’ his previous conclusions as to the influence of emotional excitement, asphyxia and sensory stimula- tion upon the adrenal secretion. We have never quarreled with the catheter method because of its difficulty, but because it cannot yield the data necessary to determine the rate of output of epinephrin or to 306 G. N. STEWART AND J. M. ROGOFF measure the changes in that rate. We shall take another occasion to point out again the reasons why we cannot accept Cannon’s con- clusions based on experiments with the catheter method. The tech- nique of obtaining the cava blood is surely not beyond the reach of ‘““any competent experimenter’”’ in physiology, and the assaying of the epinephrin in the blood would be easy if such concentrations existed there as, by implication, we must conclude that Cannon assumed to exist during or after the action of the factors studied by him. The denervated heart reaction now adopted by Cannon “‘as an indi- cator of adrenal secretion” is also easy enough to carry out. But it labors under even more serious defects, when employed as a quantita- tive method of measuring the epinephrin output and of estimating changes in the rate of output, than does the catheter method. For in the latter an attempt was at any rate made to obtain blood containing a portion of the epinephrin given off by the adrenals and to test it by a method of bio-assay which, if properly applied, does permit the epinephrin concentration in the sample of blood to be estimated. When Doctor Cannon stimulates the central end of the sciatic numer- ous reflex effects may be caused, the consequences of which upon the rate of the heart cannot be easily controlled. Among these the vaso- motor reflex changes are conspicuous, and it is obvious that in this way great variations may be produced in the pressure in the cavities of the heart, the rate of blood flow through the coronary system and, therefore, the amount of epinephrin passing through the coronaries, any of which may lead to an acceleration of the heart beat without any change having occurred in the rate of output of epinephrin. Cannon, however, states that after removal of the adrenals or their ligation en masse or after removal of one adrenal and section of the splanchnic of the other side, acceleration of the heart is no longer caused by sciatic stimulation. In one experiment he obtained some acceleration after division of both splanchnics in the thorax, but much less than before. He concludes that the acceleration following sciatic stimulation is due entirely to a reflexly increased output of epinephrin. And obviously assuming that no other factors are involved, he states that ‘comparisons of the increased rate due to sciatic stimulation with effects of adrenalin (quantitated as base) injected intravenously indi- cate that the range of reflex adrenal secretion lies between 0.001 and 0.005 mgm. per kgm. per minute, 1.e., from 5 to 25 times the amount regarded by Stewart and Rogoff as the normal output.’’ In view of our own results proving conclusively that epinephrin, if a factor, cannot EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 307 be the sole factor in the heart reaction elicited by stimulation of the sciatic, it would scarcely be worth while to spend much time in exam- ining such data. But it may be pointed out that Cannon does not state by what control experiments he has established a quantitative relation between the maximum acceleration reached and the dose of adrenalin, and not, for instance, between the dose of adrenalin and the duration of the acceleration or the total surplus number of beats in the period of acceleration. Also if such data are to have real quanti- tative value the adrenalin ought to be administered while the blood pressure and therefore the coronary blood flow are increased to approx- imately the same extent as during stimulation of the sciatic. As regards Cannon’s statement that the reaction is abolished by removal of the adrenals, it must be noted that, if it were granted that the only change caused by removal of the adrenals or section of the splanchnics is the suppression of the epinephrin output, his result would simply show that the acceleration previously obtained had been due essentially to epinephrin. It would not show that any increase had occurred in the rate of output, unless it were demonstrated that a redistribution of the blood, due to the vascular reactions evoked by the stimulation and necessarily associated with the passage through the coronary circulation of an increased proportion of the epinephrin already being given off, was insufficient to account for the reaction. But it cannot be granted that the operations practised to eliminate the epinephrin output have no other consequences. It is astonishing with what indifference both splanchnics are cut merely in order to interfere with the epinephrin output, as if all or most of the splanchnic fibers innervated the adrenals. The same is true of the removal of the adrenals. Some writers seem to assume that it is practically impos- sible to injure any important nerves when the adrenals are removed or tied off and that all the consequences which follow their removal are necessarily due to the loss of epinephrin. Gley and Quinquaud (7) have pointed out that the opposite is the case. In the dog according to them it is practically impossible to remove the adrenals without severely injuring the splanchnics. They say, however, that with their large experience they have succeeded in operating on the dog also, so as to eliminate the adrenals without injury to the splanchnics. Pearlman and Vincent (3) take exception to Gley’s statement that the difference in the splanchnic blood pressure curve observed by v. Anrep (1) in dogs after and before removal of the adrenals is due to injury to nerve fibers, and no doubt in Vincent’s hands the operation is as little harmful as it 308 G. N. STEWART AND J. M. ROGOFF is possible for it to be. If the operation can be more easily done in the cat, it still needs care and experience to reduce this cause of error to a minimum, particularly in the case of the right adrenal, the liga- tion of which is lable to injure the splanchnic. For this and other reasons the removal of the adrenals, as frequently done in acute experiments, may be attended with a marked drop in the blood pres- sure not related to loss of function of the glands. It must be remem- bered that in experiments on the denervated heart, the adrenals are removed after a considerable preliminary operation and the extrinsic regulatory nerves of the heart are eliminated. This may make it more difficult to excise or tie off the adrenals without a fall of blood pressure than in experiments on normal animals, such as those of Young and Lehman (8), of Hoskins and McClure (9) and of Bazett (10). The diminution in the heart rate remarked by Cannon, and apparently attributed by him entirely to the loss of epinephrin, is according to our observations not unrelated to the fall of pressure, although since there is reason to believe that the epinephrin liberated at the ordinary rate may affect the heart, some part of the slowing may be due to loss of epinephrin. Cannon has not given any data by which one can judge how great the change of blood pressure was in his experiments, but as he injected gum salt solution in one of them the fall may be assumed to have been sometimes considerable. Now, whatever inter- pretation one chooses to put upon the heart acceleration produced by sciatic stimulation, a reflex or more than one reflex action must be essentially involved in it. Any operation which impairs the conduc- tivity of the reflex arcs must, therefore, tend to diminish or abolish the effect. And if a negative result obtained after removal of a certain organ is attributed solely to the specific effect which the operator intended to produce, and not at all to the general effects which he did produce, then, of course, positive results elicited before the operation in any way whatever will seem to depend entirely upon the specific activity of the organ removed. In what has been said above we desire to state distinctly that we do not imply that Doctor Cannon did not remove the adrenals skilfully, and with full knowledge of the importance for his control experiment of injuring the nerves in the vicinity as little as possible and of main- taining the animal in a good general condition. All we know is that we obtained positive results after elimination of the adrenals where he obtained completely negative results. It is clearly the positive results which are significant for the decision of the question of the relation of EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 309 the epinephrin output to the heart acceleration caused by sciatic stimulation and not the negative ones. Our experiments were all made on cats. The greater number of them were performed two years ago. Only a brief notice of a portion of them has been published (11). They were mentioned, so far as could be done in the few minutes allotted to us, in the discussion of Doctor Cannon’s paper at the meeting of the American Physiological Society last December. We studied the effect upon the acceleration produced by stimulation of sensory nerves of eliminating the epineph- rin secretion: a. By clipping the adrenal veins, either with or without simultaneous ligation of the renal vessels. b. By removing one adrenal (the right) and denervating the other and allowing the animal to recover completely from the operation. In a number of these animals the denervated adrenal was also removed in the final experiment, and the heart reactions obtained before and after its removal compared. c. By removing one adrenal (almost always the right) and allowing an interval for the animal to recover, before performing the experiment upon sensory nerve stimulation with removal of the other adrenal. In this way it was supposed that the condition of the animal after removal of the remaining adrenal would be better than if both were removed at the time of the experiment. d. By removing both adrenals during the experiment on the heart reaction. In all cases in excising the glands every precaution was taken to avoid injury to nerves, by making a careful dissection between the capsule of the gland and the cortex, tying the vessels with fine ligatures. EXPERIMENTS IN WHICH THE ADRENAL VEINS WERE CLIPPED OFF OR TIED In principle this is the most satisfactory way of eliminating the epine- phrin output, since no other organ than the adrenals is interfered with and there is no damage to important nerves and no injurious fall of blood pressure. The procedure has been extensively employed by Gley. v. Anrep (1) also used it in some of his experiments on splanchnic stimulation and satisfied himself that the effects attributed by him to epinephrin were not obtained when the suprarenal vein was clipped and the corresponding splanchnic stimulated. And Pearlman and Vincent (3) state that they have usually obtained quite satisfactory THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 310 G. N. STEWART AND J. M. ROGOFF results by simply clamping and unclamping the adrenal veins. Can- non takes exception to clipping because it does not eliminate the heart reaction to sciatic stimulation or asphyxia which he interprets as indi- cating increased epinephrin output, and he, therefore, assumes that there must be leakage of epinephrin by anastomotic venous channels. He quotes, for instance, an experiment in which before ligation of the adrenal veins asphyxia of a certain duration caused an acceleration in the heart rate of 40 beats. After ligation of the veins the accelera- tion was precisely the same. We should have thought the only pos- sible interpretation of such a result would have been that epinephrin had nothing to do with the reaction, in the particular experiment, at any rate. For who will believe that after the adrenal veins were tied just as much epinephrin passed out of the adrenals by some difficult collateral path, in the minute for which asphyxia was maintained, as would have passed out by the adrenal veins plus these hypothetical anastomotic channels? And how can the heart acceleration be a quantitative reaction for epinephrin if blocking the adrenal veins does not at least diminish the reaction? When a reaction known to be caused by epinephrin is studied the result is quite different. For example, we sometimes observe a small dilatation of the pupil (after previous removal of the superior cervical ganglion) on stimulation of the peripheral end of a splanchnic nerve, with the corresponding or even with both adrenal veins clipped. But this is much smaller than the reaction obtained before or after by similar stimulation with the veins open, and usually on release of the clips there is an additional and greater reaction indicating that epinephrin had been pent up in the adrenal veins by the clips. We suggest that the reason why Can- non gets such positive results with the adrenal veins tied, and must have recourse to removal or ligation of the adrenals to obtain negative results, is that in tying the veins he does not inflict such injury as causes a general deterioration of the animal incompatible with a posi- tive heart reaction, whereas in removing the adrenals he appears to do so. In our experiments with clipping of the adrenal veins we compared the heart acceleration and rise of blood pressure caused by stimulation of the sciatic when the blood flow from the glands and, therefore, the epinephrin output were going on unhindered, with the acceleration and blood pressure rise obtained when the flow from one or both adrenals was obstructed. It was supposed that if the epinephrin is an impor- tant factor in the heart reaction, the reaction would be distinctly EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 311 smaller with the adrenal veins clipped off, that is to say, provided that the reaction can be used at all as a quantitative test for epi- nephrin. The protocols show that this expectation was not realized. If the epinephrin liberated from the adrenals is an appreciable factor, it is not easy to disentangle its influence from that of the other factors which can affect the heart rate. It must be remembered that even if it were clearly demonstrated that the acceleration caused by sciatic stimulation was diminished by interference with the output of epi- nephrin, the other potential factors, such as rise of blood pressure, not being interfered with, this would only prove that epinephrin takes a share in the reaction, not that its rate of output is increased by the stimulation. The peripheral end of a splanchnic nerve was also stimulated with the corresponding, or both adrenal veins clipped or open, in order to compare the effect of a procedure which is known to increase the epi- nephrin output on the heart rate with the effect of sciatic stimulation. It will be seen that it may be difficult to demonstrate, by comparing the maximum acceleration caused with the adrenal veins clipped and open, that the epinephrin undoubtedly liberated by stimulation of the splanchnic takes any sensible share in the heart reaction. However, this is rather an illustration of the deficiencies of such a reaction as a quantitative test for epinephrin than a proof that the epinephrin liberated with the adrenal veins open is without effect upon the rate of the heart. As already mentioned, when the epinephrin is allowed to accumulate in a cava pocket or even when the epinephrin pent up in the adrenal vessels by clipping of the adrenal veins is released a dis- tinct acceleration is produced, and it may be assumed that the epineph- rin coming steadily off from the adrenals, without being accumulated, will tend to exert a similar action, especially when the amount of epinephrin passing per unit of time through the coronaries, or its con- centration, is abruptly increased by the vasomotor changes associated with splanchnic or sciatic stimulation. It seems, however, improbable that with such a reaction as acceleration of the heart the acceleration produced by simultaneous action of two factors, singly effective, should be the sum of the separate effects, whether the reaction is measured by the maximum acceleration attained, or by the duration of the acceleration or by the total surplus number of beats. It seems more likely that when the heart is keyed up to a certain point by the action of one factor, whether this acts upon a local accelerator mechanism or not, it will break loose, so to speak, from the relative stability of rate 312 G. N. STEWART AND J. M. ROGOFF imposed upon it by removal of its extrinsic nerves, and execute a run of quicker beats, the duration and maximum acceleration of which may have no simple relation to the absolute magnitude of the exciting influence, and which may not be greatly modified by a concomitantly acting influence, itself capable of independently producing a similar reaction. The results of these experiments are illustrated by some protocols. Where it is mentioned that a nerve was stimulated with a vein clipped, the clip was applied, unless otherwise stated, a few seconds before the stimulation and was removed as soon as the portion of the curve to be used for counting the heart rate had been completed... When it is simply mentioned that a nerve was stimulated, without any reference to clipping, it is implied that the veins were open. Under ‘“‘rate” is always given the number of heart beats per minute; under “‘pressure”’ the blood pressure in millimeters of mercury. Protocol. Cat 175; male; weight, 2.2 kgm. 11:30 a.m. Urethane, 5 gm. by stomach tube. 1:10-1:48 p.m. Cut vago-sympathetiecs and excised stellate ganglia;! pre- pared central end of left sciatic for stimulation. Rate 1:50 p.m. Before stimulation of sciatic. -.°.--......+..---- eee 188 25 seconds after beginning stimulation........................ 210 1:55 p:m. “Before stimulation! oft sciaties:+. 5. -3-)J.5- eee eee eee 187 20 seconds after beginning stimulation......................-- 202 40 seconds after beginning stimulation........................ 212 Now opened abdomen and purposely manipulated intestines 2:03;p.m. ‘Before stimulation of sciatic®.....---25-2 +> eee eee 178 30 seconds after beginning stimulation.......................- 212 2:10'p:m~ Before stimulation of sciatic. <).9.0 51-5. 2- 75. ee aoe ee eee 186 30 seconds after beginning stimulation......................-- 214 2:14p.m. Before stimulation of sciatic; left adrenal vein clipped........ 184 30 seconds after beginning stimulation........................ 206 2:16 p.m. Prepared peripheral end of left splanchnic in abdomen 2:20 p.m. Before stimulation of splanchnic; left adrenal vein clipped.... 184 15 seconds after beginning stimulation......................-- 192 30 seconds after beginning stimulation.....................--- 208 45 seconds after beginning stimulation......................-- 224 1 With the cat in the supine position an incision is made in the axilla and a part of the second rib exposed by separation of overlying muscle. A portion of the rib anterior to the angle is carefully separated from its periosteum and resected. The lower part of the stellate ganglion is usually thus exposed and the excision of the ganglion easily completed without injury to the pleura. In all the experiments, except when otherwise stated, the period of stimula- tion of the sciatic was 30 to 40 seconds. When the heart rate is given after a certain number of seconds from beginning of stimulation this means that a suffi- cient portion of the curve beginning at that point was counted. The corre- sponding blood pressure was that at the end of the portion counted. EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 313 2:26 p.m. Before stimulation of splanchnic; left adrenal vein clipped.... 180 15 seconds after beginning stimulation........................ 196 30 seconds after beginning stimulation........................ 208 45 seconds after beginning stimulation........................ 220 * _ 60 seconds after beginning stimulation........................ 220 gs2yem.; Before stimulation-of splanchnic ... 302: 4:20). : 5123 [sd oe eve on: 182 During first 15 seconds of stimulation.....:................-- 180 During second 15 seconds of stimulation...................... 200 During third 15 seconds of stimulation....................... 204 During fourth 15 seconds of stimulation...................... 216 2:47 p.m. Before stimulation of sciatic; both adrenal veins clipped...... 184 During first 15 seconds of stimulation. .:.......0sc0s0..6..--- 188 During second 15 seconds of stimulation...................... 188 During third 15 seconds of stimulation....................... 200 During fourth 15 seconds of stimulation...................... 204 2:50.) ieee erone spnmuUlation Of SCIAbie...,. 008.022. no fianc ea ctiet ea ehs ae... 176 During fret-t5 seconds of stimulation... 2.202.202 2....0.02... 180 During second 15 seconds of stimulation...................... 196 During third 15 seconds of stimulation............5.......... 208 During fourth 15 seconds of stimulation...................... 216 3:05 p.m. Before stimulation of sciatic; both adrenal veins clipped...... 188 Paning ars: lb seconds ot stimulation 1) 5465 see6 eeu; . 188 During second 15 seconds of stimulation...................... 196 During third 15 seconds of stimulation....................... 204 Durimns fourth 15 seconds of stimulation: ..j42+ 22s -eae ee. - 204 Rate Pressure S40. meer berore stumulaizon of sciatic... 2... oc alesse sae oe mee 176 82 24 seconds after beginning stimulation................ 203 140 3:50 p.m. Before stimulation of splanchnic; left adrenal vein CELI ay OYE16 (Eee AeA Ne eg ME eT RENE tree Sees ied Ne, 174 58 28 seconds after beginning stimulation................ 215 166 3:57 p.m. Before stimulation of splanchnic; both adrenal veins EL 01 0121 A a oe em a Sepa) eh bin A ae ae 185 62 24 seconds after beginning stimulation ............... 206 131 40 seconds after beginning stimulation................ 221 i-0o p.m.) Betore stimulation of sciatic............--see8--<-5-: 174 46 12 seconds after beginning stimulation................ 197 118 40 seconds after beginning stimulation................ 200 4:10p.m. Before stimulation of sciatic; both adrenal veins Glippedsn: oc 2 Saree ce cade o< sa eaonee aes terres 189 52 15 seconds after beginning stimulation................ 200 112 30 seconds after beginning stimulation................ 204 4:15p.m. Before stimulation of splanchnic...................... 184 55 20 seconds after beginning stimulation................ 209 130 Excised left adrenal 4:30 p.m. Before stimulation of splanchnic....................-- 215 70 22 seconds after beginning stimulation................. 229 90 314 G. N. STEWART AND J. M. ROGOFF Some of the curves from these experiments are used in the next paper to demonstrate that the acceleration of the heart on stimulation of the central end of the sciatic, contrary to Cannon’s statement, is easily obtainable after the abdomen has been opened. The protocols themselves and other protocols coming later in the present paper also afford abundant evidence that the statement is baseless. In figure 1 portions of the blood pressure curve from cat 175 before and during stimulation of the left splanchnic (A and B, taken at 3:50 p.m.) are AB BRARB VIELE LIERDPLALEL Litititis Fig. 1. Parts of blood pressure tracings from cat 175. "Before'sciatie stimulation®=+—. «- 526s e oe fee eee 210 84 211 84 Counts of successive portions of the curve after begin- | 215 102 ning of stimulation (with progressive increase in 4 230 112 strength). i. jays cateisteei more eereieen bee enone ke 128 244 128 10:05)a.m> Before'sciatie stimulations -- +4 snes ee eee eee 214 100 Counts of successive portions of the curve with stimu- Fe st lation as aboves ieee ee eee 240 118 10:10 a.m. Prepared adrenal veins for clipping (extraperitoneally) 10215 a.m=_ (Beforeisciatic stimulationses. seen eee eee eens 217 84 : Gt : A 223 100 Successive counts after beginning of stimulation as aDOVe: Staten. ae ae eee es se Ds TE tre ee ER Sitca RE ONES Bi tS RR Bs oo G3 248 110 10:20 a.m. Before sciatic stimulation; both adrenal veins clipped. 206 70 212 88 Successive counts after beginning of stimulation as | 222 89 abOVEs 12). 2 IO BEE ng sis Sh ee 223 88 225 86 Just after release of adrenal veins..................... 228 95 15 seconds after release of adrenal veins............... 245 104 10225)a.m-) Beforesciatic suimulations.e-eeenee ee er eerie 214 80 216 88 Successive counts as above during stimulation......... i 104 | 253 117 10:30 a.m. Before sciatic stimulation; both adrenal veins clipped. 215 77 : : F 227 84 During stimulation as abOVEN Us. s23 2. a. ae. ore eee ee \ 228 98 Just after release of adrenal veins...................-- 236 96 12 seconds after release of adrenal veins............... 239 100 In the last experiment (cat 193) to be quoted in this section of the paper, only the vago-sympathetics were cut at the beginning of the experiment, and the abdomen was opened. ® The effects on the eye were those described in the footnote to protocol of eat 190, p. 319. EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 325 Protocol. Cat 193; male; weight, 2.7 kgm. Left superior cervical ganglion excised 10 days previously. 12:08 p.m. 12:10 p.m. 12:15 p.m. 12:20 p.m. 12:40 p.m. 12:50 p.m. 1:00 p.m. 1:02 p.m. 1:20 p.m. Under urethane (5 gm.) opened abdomen; prepared peripheral end of left splanchnic for stimulation; cut vago-sympathetics. Rate Pressure Before stimulation of splanchnic; left adrenal vein Slanted cee tt AS Jes 5 0 DA SO hk 157 11 seconds after beginning stimulation................ 175 18 seconds after beginning stimulation................ 180 No eye reactions during stimulation; after release of clip good pupil and nictitating reactions in 9.2 seconds Before stimulation of splanchnic....................-. 162 10 seconds after beginning stimulation................ 209 Good pupil and nictitating reactions in 11.5 seconds Before stimulation of splanchnic; left adrenal vein NESE a Ay s/o etalk ia mga clin, 3-4 ocak ean i ee 172 17 seconds after beginning stimulation................ 190 No eye reactions during stimulation; on release of clip good pupil and nictitating reactions in 8.2 seconds Excised stellate ganglia Before stimulation of splanchnic; left adrenal vein PUNDIT eS tne SEO rs a 8 are Fe oP te ges AS 185 18 seconds after beginning stimulation................ 201 No eye reactions during stimulation; on release of clip good pupil and nictitating reactions in 11 seconds Stimulated splanchnic with left adrenal vein and coeliac and superior mesenteric arteries clipped. Pretare Shiri nOne co: 2 A ep erinese es aks pa hisaainseye 203 17 seconds after beginning stimulation................ 202 No eye reactions during stimulation; on release of adrenal vein good pupil and nictitating reactions in 10.6 seconds Before removal of adrenal clip...................--55: 202 12 seconds after release of adrenal vein................ 238 Removed clips from coeliac and superior mesenteric Prepared central end of sciatic for stimulation Before stimulation of seiatie:(i2.0:2 85h t.ho asin. eee. 174 11 seconds after beginning stimulation................ 184 20 seconds after beginning stimulation................ 194 34 seconds after beginning stimulation................ 200 Before’ stimulation of Sciatic. sc. ee ee Shs oe ee eee oe 184 16 seconds after beginning stimulation................ 200 98 145 96 136 120 162 114 144 134- 136 121 128 78 102 104 68 108 The effects of stimulating the splanchnic, with the corresponding adrenal vein open and clipped, were not obviously different from those in the other experiments, where the stellate ganglia had been excised at the beginning, nor did removal of the stellate ganglia cause any essential change. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 2 The failure of the heart reaction with splanchnic 326 G. N. STEWART AND J. M. ROGOFF stimulation when the coeliac and superior mesenteric arteries were clipped is clearly associated with the absence of a rise of pressure when the splanchnic area is thus eliminated. On removal of the clip from the adrenal vein the usual good eye reactions were obtained and also an acceleration of the heart due to release of the epinephrin pent up in the adrenal vessels. It may be tedious to point out again that an hour and a half after the abdomen had been opened, with one splanchnic cut, after all the clip- ping and unclipping of the adrenal veins and of the coeliac and superior mesenteric arteries (and a good many observations actually made have been omitted from the protocol to save space) stimulation of the sciatic still gave a good acceleration of the heart, although, according to Can- non, no reaction ought to have been obtained. In view of such facts what becomes of Doctor Cannon’s assertion, unsupported by any evidence that when we collect adrenal vein blood from a cava pocket, the ‘peculiar’ conditions of our experiments do not permit of the demonstration that with stimulation of sensory nerves there is a vast outpouring of epinephrin from the adrenals, a demonstration which he obtains by a misinterpretation of a heart reaction equally well elicited whether the adrenal veins are open or clipped off, whether the eye (sensitized by removal of the superior cervical ganglion) is giving negative or positive reactions for epinephrin? Since, however, Doctor ‘Cannon maintains that the mere clipping or tying of the adrenal veins need not interfere in the least with the passage of epinephrin to the heart, and has even convinced himself that adrenalin, injected into the lumbar vein crossing the adrenal after it has been ligated on each side of the gland, passes sofreely into the circulation that it causes a large rise of blood pressure, evidence of another kind will now be given that the acceleration of the heart relied upon by him to prove a markedly augmented rate of epinephrin output when the sciatic nerve is stimu- lated has no such significance. EXPERIMENTS ON ANIMALS WITH ONE ADRENAL REMOVED AND THE NERVES OF THE OTHER CUT We have previously shown (13) that after this operation the epinephrin output is either greatly reduced or abolished, within the limits of sensitiveness of the test objects (rabbit’s intestine and uterus segments) employed to detect and estimate it. It was there- fore of interest to see whether the acceleration of the heart would EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 327 be obtained in these animals by stimulation of sensory nerves. The result was positive in all the cats, a good acceleration being elicited by stimulation of the sciatic. For example, in cat 201, a female weighing 2.575 kgm., the right adrenal was removed, the nerves of the left including the splanchnic cut. A portion of the left semilunar ganglion was removed. The left superior cervical ganglion was excised at the same time. Eighteen days thereafter, the animal being in good condi- tion, the vago-sympathetics were cut and the stellate ganglia removed under urethane and the central end of the left sciatic nerve prepared for stimulation. A blood pressure tracing was taken as usual from’the right carotid. Stimulation of the sciatic caused the heart rate to increase from 177 before stimulation to a maximum rate of 211 beats per minute. The blood pressure increased from 54 mm. before stimu- rn Nf ON LON gt tN ff ww ata alana ap! Fig. 4. Parts of blood pressure tracings from cat 201. A, before and B, a portion commencing 20 seconds after beginning of sciatic stimulation. Reduced to three-fifths. lation to a maximum of 132 mm. of mercury. The acceleration and the blood pressure increased together. Successive portions of the curve starting from the point of stimulation yielded the following heart rates and blood pressures 180 (72,) 193 (97), 202 (107), 209 (118), 209 (132). The blood pressures are in parentheses. A sample of the curve is reproduced in figure 4. The effect of sciatic stimulation on the pupils did not differ materially from that already described in cats in which the epinephrin output had not been interfered with (see foot- note to protocol of cat 190). The nictitating membrane did not move. The abdomen was afterwards opened, a short cava pocket made and adrenal vein blood collected, which was assayed on rabbit segments. It was shown that the epinephrin output could not have been more than one-fiftieth of the average output per kilogram per minute, under ooo) G. N. STEWART AND J. M. ROGOFF the conditions of our experiments. The preliminary operation had, therefore, effectively severed the epinephrin-secretory nerves of the left adrenal. In spite of this and also in spite of the fact that one splanchnic had been divided, and with both splanchnics intact a greater reflex rise of blood pressure might have been obtained, a maximum acceleration of the heart of 32 to 34 beats was elicited by stimulation of the sciatic. According to Cannon, the whole of the acceleration must have been due to a great outpouring of epinephrin from the left adrenal, reflexly stimulated along efferent nerve paths which had been divided 18 days before, and which certainly had not regenerated in that time. In another cat (202), a female weighing 2.545 kgm., the right adrenal was removed and the nerves of the left cut. The left superior cervical ganglion was excised at the same time. Eighteen days thereafter the heart was denervated under urethane and blood pressure tracings taken from the right carotid. Stimulation of the central end of the sciatic caused an acceleration of the heart rate from 145 per minute before stimulation to 161. The blood pressure rose from 92 to 134 mm. of mercury. Successive portions of the curve, from the beginning of stimulation, yielded the following heart rates and blood pressures: 152 (100), 155 (110), 157 (130), 160 (126), 161 (134). The pupil reactions were as previously described in the footnote to the protocol of cat 190. There was no movement of the nictitating membrane. After the sciatic stimulations the abdomen was opened, and 3 speci- mens of adrenal vein blood collected—the first, about 1 gram (dis- carded), the second, 2.9 grams in 4 minutes (0.75 gm. per minute), the third 5.1 grams in 9 minutes (0.57 gm. per minute). No evidence was obtained with rabbit segments that the blood contained any epinephrin. It was shown that the third specimen could not have contained 1: 70,000,000 epinephrin, i.e., the output could not have been 0.000007 mgm. per minute for the cat, or 0.0000027 mgm. per kilogram per minute. In other words it could not have been one- hundredth of the average normal output, and there was no evidence that any epinephrin was being given off. The interval after the pre- liminary operation was quite short in this animal, 8 days, which pos- sibly may be an unfavorable condition for obtaining a large heart reaction, although we have no evidence as to this. In the next experiment (cat 215) the animal was allowed to survive for 2 months. EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 329 Protocol. Cat 215; female; weight 2.38 kgm. May 17, 1918. Right adrenal excised, left denervated and right superior cervical ganglion excised. July 15, 1918. Under urethane (5 gm.) cut vago-sympathetics, excised stel- late ganglia, prepared central end of left sciatic for stimulation. Rate Pressure 11:15 a.m. Before sciatic stimulation for 38 seconds (8 em.)*®...... 219 134 During first 14 seconds of stimulation................. 232 162 During next 13 seconds of stimulation................ 238 170 Buming next) 1S SG@OMG see.) s go cjchece nis % sree arms ee ape 244 172 DUTIneMext 20 SCCONUS +25... <1. «..2 0 s.< cose seam aieseeete 228 147 11:20 a.m. Before sciatic stimulation for 33 seconds (6 cm.)....... 222 144 During first 14 seconds of stimulation................. 240 180 During next 14 seconds of stimulation................ 248 196 iDurine next ils secondse:: Acs. o> «Ants Sei ee S See eee 249 180 11:25 a.m. Before sciatic stimulation for 26 seconds (6.5 em.)..... 219 152 During first 8 seconds of stimulation.................. 226 178 During next 7 seconds of stimulation.................. 230 182 DUTIMNe NeXt lS SCCONUSH 1... on. coc ccs ne cete eae eee 246 190 11:30 a.m. Before sciatic stimulation for 40 seconds (6 cm.)....... 224 148 During first 13 seconds of stimulation................. 239 177 During next 14 seconds of stimulation................. 252 180 During next 13 seconds of stimulation................. 251 182 Die Menext MP SCCONGS: 3c cco'ss cc econ eee ee ents 248 168 11:35 a.m. Before sciatic stimulation for 37 seconds (4 cm.)....... 222 148 During first 12 seconds of stimulation................. 239 173 During next 11 seconds of stimulation................. 244 180 During next 11 seconds of stimulation................. 252 180 Wurinesnext |4secondsees 4. a ease eee ee eee 249 168 Later on the abdomen was opened and adrenal vein blood collected. Assays of the adrenal blood released from a cava pocket were also made by the pupil reaction. The protocol shows that stimulation of the sciatic caused good acceleration of the heart with corresponding changes in the blood pressure. The rabbit intestine assay of the adrenal blood demon- strated that a substantial output of epinephrin was still going on. The concentration of the second specimen was about 1: 6,500,000 and that of the third specimen about 1: 3,750,000, corresponding in either case to an output of 0.00008 mgm. per minute for the cat, or 0.00003 mgm. per kilogram per minute, one-seventh or one-eighth of the aver- age normal output. It is impossible to say whether an unusually large proportion of the secretory fibers of the left adrenal had escaped 8 Eye reactions, as described in footnote to protocol of cat 190, p. 319. The distance between the coils is given in brackets. 330 G. N. STEWART AND J. M. ROGOFF section at the preliminary operation, or whether some regeneration had occurred. In any case the innervation of the one adrenal remain- ing must still have been seriously crippled. Yet the heart reaction on which Cannon relies as an “indicator of adrenal secretion’? was as well obtained with sciatic stimulation as in normal animals, corre- sponding with the general excellent condition of the cat two months after the primary operation. In two cats operated on in the same way but only a relatively short time before (8 days and 7 days), the experiment was completed by excising the remaining (already denervated) adrenal. It was not thought advisable in these cats to complicate the experiment by at- tempting to estimate the residual epinephrin output, if any, on the adrenal vein blood. But it was shown that the epinephrin store of the adrenal constituted a full load, a good indication that protection of the gland from depletion during the experiment was relatively complete. Protocol. Cat 444; male; weight, 2.9 kgm. Right adrenal excised and left denervated 8 days, left superior cervical ganglion excised 14 days previously. Under ether cut vago-sympathetics, excised stellate ganglia, prepared central end of right sciatic for stimulation. Rate Pressure 10:40 a.m. Before sciatic stimulation (8 cm.)...................-- 133 132 2 seconds after beginning stimulation................ 130 128 30 seconds after beginning stimulation................ 132 130 10:45 a.m. Before sciatic stimulation (5 em.)..................-:- 133 112 6 seconds after beginning stimulation................ 140 40 seconds after beginning stimulation................ 152 92 60 seconds after beginning stimulation................ 145 106 10:50 a.m. Before sciatic stimulation (8 em.)).........-4.-.seeeeee 127 108 3 seconds after beginning stimulation................ 132 30 seconds after beginning stimulation................ 136 100 45 seconds after beginning stimulation................ 131 104 10:50°.a.m. “Before'sciatie stimulation’(Giems kee e eee eee eee 121 120 5 seconds after beginning stimulation................ 130 126 11:10 a.m. Excised left adrenal (extraperitoneally) 11:15.a.m.. Before sciatic stimulation) (Siem) a.-ee eee eee eee 130 103 3 seconds after beginning stimulation................ 133 96 23 seconds after beginning stimulation................ 140 60 seconds after beginning stimulation................ 144 114 11:48 a.m. Opened abdomen, tied coeliac axis and superior mes- enteric artery 11'50)a:m, Before/sciatie stimulation (Giem)) see eee eee eee 139 74 8 seconds after beginning stimulation................ 144 74 3 seconds after stopping stimulation................. 142 72 40 seconds after stopping stimulation................. 134 70 EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART Ses 11258 a.m. Before sciatie stimulation (4 cm.)..................+0- 123 72 2 seconds after beginning stimulation................ 130 77 15 seconds after beginning stimulation................ 133 80 40 seconds after beginning stimulation................ 137 81 60 seconds after beginning stimulation................ 138 78 12:05 p.m. Before sciatic stimulation (8 cm.)....................- 130 76 2 seconds after beginning stimulation................ 133 80 30 seconds after beginning stimulation................ 135 82 In cat 444 the greatest acceleration caused by sciatic stimulation before excision of the remaining adrenal was 19 beats. With repeated stimulations the acceleration effect declined. The last stimulation before excision gave only 9 beats. A weaker stimulation just after excision gave a maximum acceleration of 14 beats, and another stronger stimulation gave a maximum acceleration of 15 beats. Even after the abdomen was opened and the superior mesenteric artery and coeliac axis tied, a maximum acceleration of 15 beats was obtained in the absence of the adrenals. In cat 441, also, while the greatest acceleration before excision of the remaining adrenal (17 beats) was not reached afterwards, the acceleration immediately after excision was the same as that obtained just before it (9 beats), although the stimulation in the latter case was the stronger. Protocol. Cat 441; female; weight, 1.66 kgm. Right adrenal excised and left denervated 7 days previously. Under ether cut vago-sympathetics, excised stellate ganglia, prepared central end of right sciatic for stimulation. Rate Pressure 10:55 a.m. Before sciatic stimulation (10 cm.).................... 219 100 6 seconds after beginning stimulation................ 229 116 20 seconds after beginning stimulation................ 228 10:58 a.m. Before sciatic stimulation (8 em.)..................22. 217 115 6 seconds after beginning stimulation................ 223 122 11:02 a.m. Before sciatic stimulation (6 em.)......../...........-. 208 106 10 seconds after beginning stimulation................ 215 118 30 seconds after beginning stimulation................ 213 11:06 a.m. Before sciatic stimulation (6 cm.)..................... 211 109 5 seconds after beginning stimulation................ 217 18 seconds after beginning stimulation................ 228 120 28 seconds after beginning stimulation................ 227 11:10 a.m. Before sciatic stimulation (10 em.).................... 207 113 5 seconds after beginning stimulation................ 215 116 11:15 a.m. Before sciatic stimulation (8 cm.)...................-. 206 109 5 seconds after beginning stimulation................ 222 139 332 152320 "a-m- 11:25 a.m. 11:40 a.m. 11:42 a.m. 11:45 a.m. 11:51 a.m. 11:55 a.m. 11:58 a.m. 12:00 m. 12:05 p.m. 12:10 p.m. 12:15 p.m. G. N. STEWART AND J. M. ROGOFF Rate Pressure Before sciatic stimulation (7 em.).............-.------ 216 5 seconds after beginning stimulation................ 224 Before sciatic stimulation (6)cm.).......-.5- 4.055 sees 212 5 seconds after beginning stimulation................ 221 Excised left adrenal (lumbar route) Before sciatic stimulationy(Sicms) peer eE ee ele ae eee 214 5 seconds after beginning stimulation................ 220 20 seconds after beginning stimulation................ 223 Before sciatic stimulation (10 ¢m_).-.......-....:..-<2 211 5 seconds after beginning stimulation................ 214 23 seconds after beginning stimulation................ 215 Before sciatic stimulation (9’em:)2.72.2....20. cles 210 6 seconds after beginning stimulation................ 214 Before sciatieystimulation (ems) ree oe eee eee 208 6 seconds after beginning stimulation....... Re inti. 210 Before sciatic stimulation (8\cm.)22.42 24. eee eee 211 16 seconds after beginning stimulation................ 218 Prepared central end of left sciatic for stimulation Before stimulation of left sciatic (9 cm.).............. 214 30 seconds after beginning stimulation................ 217 Before stimulation of right sciatic (10 em.)..,......... 215 5 seconds after beginning stimulation................ 216 Before stimulation of right sciatic (8 em.)............. 236 5 seconds after beginning stimulation................ 242 11 seconds after beginning stimulation................ 245 116 148 98 119 The left adrenal weighed 0.226 gm. and contained 0.23 mgm. of epinephrin at end of experiment. In the last cat (450) of this series to be mentioned a longer period (34 days) was allowed to elapse between the primary operation and the experiment, in order that the animal might have more fully recovered. Protocol. Cat 450; male; weight, 2.96 kgm. Right adrenal excised and left denervated (with section of left splanchnic as usual) 34 days previously. Under urethane (5 gm.) cut vago-sympathetics; excised stellate ganglia; prepared central end of left sciatic for stimulation. 11:10 a.m. 11:14 a.m. Rate Pressure Before sciatic stimulation (40 em.)..................-- 173 During first 18 seconds of stimulation................. 178 During next 20 seconds of stimulation...............-. 182 Just afteriend ofistimulations ees eee eee Eee 185 30 seconds after end of stimulation.................... 180 Before sciatic stimulation (9 em.)..................08- 174 During first 12 seconds of stimulation................. 178 During next 16 seconds of stimulation................. 183 Just after end of stimulation... 5-:.0seee heen eee 191 102 114 114 EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 11:18 a.m. 11:40 a.m. 11:43 a.m. 11:48 a.m. 11:58 a.m. 12:10 p.m. 12:12 p.m. 12:15 p.m. 12:21 p.m. 12:26 p.m. 12:35 p.m. 12:36 p.m. 1:01 p.m. Rate Before sciatic stimulation (Sem) ..255 seem eres ne oe: 171 During first 16 seconds of stimulation................. 178 During next 17 seconds of stimulation................. 187 Just aitemenGlotistimml ation cess ce eee eee = 192 20 seconds after end of stimulation.................... 190 Excised left adrenal (extraperitoneally) Before sciatic stimulation: (9'em:).s4.2 esse ee eee sc 174 During first 15 seconds of stimulation................. 179 During next 20 seconds of stimulation................. 189 Before sciatic stimulation (Siem, isc... see elaceia« ore 174 During first 14 seconds of stimulation.............. aaa Wis: During next 20 seconds of stimulation................. 191 DUSt Aver CNG Ots i Mnl atlONe sca seiers astelatererslereieiaie eee 200 15 seconds after end of stimulation...........:........ 198 35 seconds after end of stimulation.................... 194 Before sciatic stimulation (7 cm.)......... Sie oe bad 177 During first 17 seconds of stimulation................. 181 During next 20 seconds of stimulation................. 192 Just abverenG Of stimulation. ..«ssctacele ceive sieeeiae 200 20 seconds after end of stimulation.................... 195 Prepared peripheral end of right splanchnic (in thorax) Before splanchnic stimulation (10 cm.)...... mt a 167 During first 20 seconds of stimulation................. 176 During next 20 seconds of stimulation................. 184 DUSC aiberenaol SoMa tlOMs sence eaeiaetice serrate ieee 186 Before splanchnic stimulation (8 cm.)...............4- 177 During first 20 seconds of stimulation................. 175 During next 20 seconds of stimulation................. 185 Just after’ end of stimulation, (3... oan sodas eae = ose 189 Before splanchnic stimulation (7 cm.)................. 178 During first 16 seconds of stimulation................. 179 During next 20 seconds of stimulation................. 186 Justatterena otstimimllatlonees eee ieee 187 Beforeisciatie stimulation: (7 ems )pesees eee cs sae. 4. 179 During first 20 seconds of stimulation................. 182 During next 20 seconds of stimulation................. 180 Cut left splanchnic in thorax Before sciatic stimulation (7 em.)..............---++-- 181 During first 18 seconds of stimulation................. 182 During next 20 seconds of stimulation................. 181 VUStHaberenaLon stimulation once cine amar cece 184 Before splanchnic stimulation (7 cm.)................. 183 During first 16 seconds of stimulation................. 184 During next 20 seconds of stimulation................. 196 Just atterendson stimulations =.nceteeies oe se sealers .7.0.....5 222 eeeee 212 124 5 seconds after beginning stimulation................ 242 15 seconds after beginning stimulation................ 255 156 23 seconds after beginning stimulation................ 250 156 37 seconds after beginning stimulation................ 228 138 60 seconds after beginning stimulation................ 218 134 11:05 a.m. Before sciatic stimulation) (6\cms)—-- eee ee 213 119 6 seconds after beginning stimulation................ 252 154 28 seconds after beginning stimulation................ 247 11:09 a.m. “Before sciatic stimulations (4.cmy)—=--s eee eee eee eee 212 118 5 seconds after beginning stimulation................ 230 133 22 seconds after beginning stimulation................ 248 153 11:20 a.m. Before sciatic stimulation G'em)o. 2 eee 221 112 6 seconds after beginning stimulation................ 236 130 22 seconds after beginning stimulation................ 247 140 11:25 a.m. — Before sciatic stimulation (6.em))-...... ees ee ee eee 230 118 8 seconds after beginning stimulation................ 241 130 11:28 a.m. Before sciatic stimulation (4 em.).............-.-+----- 225 133 7 seconds after beginning stimulation................ 231 146 43 seconds after beginning stimulation................ 227 130 11:45 a.m. Exposed cord in midcervical region 1 Es US 0 MD Sans OM ONES rs” is Sa cinco ool: 208 90 EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 337 Rate Pressure 11:55 a.m. Before total asyphxia (for 45 seconds)...............+. 207 84 3 seconds after beginning asphyxia................... 204 10 seconds after beginning asphyxia................... 225 88 55 seconds after beginning asphyxia................... 210 88 12:05 a.m. 3 minutes after transection of cord between 4th and 5thicervicall sezments).....) Joo. iis eh. oon eee ase 193 51 As will be seen from the protocol, excellent heart reactions were obtained both before and after the removal of the remaining adrenal in cat 440. It is to be particularly noted that excision of the left adrenal left the blood pressure and heart rate practically unchanged. Fig. 7. Parts of blood pressure tracings from cat 440. A, before and B, a portion commencing 15 seconds after beginning of sciatic stimulation, before excision of remaining adrenal; C, before and D, 15 seconds after beginning of sciatic stimulation, after excision of remaining adrenal. Zero line moved up 41 mm. The maximum acceleration for the last sciatic stimulation prior to removal of the left adrenal was 33 beats per minute, and for the first sciatic stimulation after removal of the adrenal 42 beats, the maximum pulse rate reached being practically identical in the two cases (253, 250). Samples of the curves used in counting the pulse rate are given in figure 7. The greater portion of each of the two curves, much reduced, is reproduced in figure 8, to show that excision of the second adrenal has not in any way essentially changed the vascular reaction. Here was a cat, then, without adrenals in which the acceleration pro- duced by stimulation of the sciatic was actually greater than that 338 G. N. STEWART AND J. M. ROGOFF produced by a similar stimulation while one adrenal was still intact. Yet, according to Cannon, the last acceleration, like the first, must have been due solely to augmented epinephrin output from the adrenals. It should be noted that the exposure of the cord in the midcervical region led to a drop of blood pressure to 90 mm. of mercury and a corresponding slowing of the heart to 208 beats a minute. Subsequent transection of the cord caused the blood pressure to fall to 51 mm. of mercury and the heart rate to 193. It must be remembered that the adrenals were out, and the decrease in the heart rate can have nothing to do with absence of adrenal epinephrin. Cannon has emphasized the fact that after removal of the adrenals the pulse rate drops. Our observations show that this is always true provided the blood pressure ae Fc ia Ua a Se B Pca OL pee A 220 £53 264 249 Fig. 8. Blood pressure curves from cat 440. ) eae ae 90 6 seconds after beginning stimulation................ 171 103 25 seconds after beginning stimulation................ 170 12°32 p.m. Before sciatic stimulation (7 em:)....-.....-.-0-.-5 anes 87 10 seconds after beginning stimulation................ 181 132 26 seconds after beginning stimulation................ 184 128 35 seconds after beginning stimulation................ 182 118 12:40 p.m. Sciatic now stimulated twice with an interval of only 2 minutes, to fatigue the reaction, and then at...... 12:50 p.m. Before sciatic stimulation (6 cm.)......%=......->-.+.een ade 92 10 seconds after beginning stimulation................ 181 126 35 seconds after beginning stimulation................ 183 1:03 p.m. Sciatic again stimulated twice in eae succession and then at.. ioe weuloie teeGek eee 1:14p.m. Before Serie sEatulations (6 cm. ei mi d7a:6a ein) axe eee ee 171 76 10 seconds after beginning stimulation................ 175 107 40 seconds after beginning stimulation................ 177 1:20 p.m. Opened abdomen, tied renal arteries ral WelNS5 cscs 1:24p.m. Before sciatic sernlation (Gtems) 403.28 2S De eee 88 4 seconds after beginning stimulation................ 181 110 30 seconds after beginning stimulation........ Boe cil 98 1:33 p.m. Stimulated sciatic with abdominal worte clipped: Before clipping aorta... -.6 s--e-. ep oe ee ae ef Before sciatic stimulation: .2: 52... +2. ¢.- eeee 177 100 10 seconds after beginning stimulation................ 184 30 seconds after beginning stimulation................ 186 114 After release of aorta.. Sig Bisco seidieis oi 8b ee eae 30 seconds after cslenees of: ini SA eee a Ri. Sasi se 185 ' 86 EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 349 1:37 p.m. Stimulated sciatic with abdominal aorta and vena cava clipped; Rate Pressure Before stimulation (7 em.).. ny Bea TS PI eee Paella 76 8 seconds after beginning ean i cveueteretae Taay SL OO 30 seconds after beginning stimulation........ 2 SZ 100 10 seconds after removal of clips from ae Ans cava. 176 68 2:21 p.m. Before sciatic stimulation (7 cm.).. £1). See Bi 159 47 10 seconds after beginning cieeonintier shamneetseee.. ets 159 60 Nistiatiewend-ob stimulation... . 2; «406.5 + oecc een | tals i2ap.m. betore sciatic stimulation (4 ¢m.)<... 0 6ic-.eccmeapere 155 43 10 seconds after beginning stimulation................ 161 55 40 seconds after beginning stimulation................ 161 Accelerations as great as 20 beats per minute were obtained after removal of the adrenals. Before removal the maximum acceleration seen in two observations with different strengths of stimulus was 15 and 32 beats respectively. The blood pressure fell from 135 to 90 mm. of mercury after the adrenalectomy. When the sciatic was then repeatedly stimulated with only short intervals between the successive stimulations, the heart reaction, as estimated by the maximum accel- eration, diminished but about the same absolute rate (180 to 183) was reached at the height of the acceleration. Later on in the experiment, when the abdomen had been opened and the renal vessels tied, an acceleration of 7 beats per minute was caused by clipping the abdom- inal aorta just above the bifurcation, as is done in the collection of blood from the cava pocket. This acceleration was accompanied by a rise of blood pressure from.77 to 100 mm. of mercury. When the sciatic was now stimulated with the aorta still clipped (the clip was put on only a short time before stimulation of the sciatic so that the excitation of the nerve should not be interfered with) a further accel- eration of 9 beats a minute occurred. Sciatic stimulation gave a similar acceleration with both abdominal aorta and cava clipped, as in making the pocket. At the end of the experiment, when the blood pressure had fallen to 40 to 50 mm. of mercury, no sensible accelera- tion, or with stronger stimulation only a small one, could be elicited through the sciatic and the rise of blood pressure was also small. But it would surely be absurd to attribute this belated failure of a reaction which had been well obtained after the adrenalectomy to the absence of a reflexly excited outpouring of epinephrin. As to varying the strength of stimulation, it should be stated that we did not consider there would be any point in merely comparing the reaction obtained before and after the removal of the adrenals with 350 G. N. STEWART AND J. M. ROGOFF practically the same nominal strength of stimulus, since the excitability of the reflex mechanisms cannot be counted upon to remain the same, especially where considerable changes of blood pressure have occurred. What we tried to do in all the experiments was to elicit as large a reaction as possible, both before and after elimination of the adrenals, using for this purpose the strength of stimulation which seemed most effective. The strongest stimuli were not generally the best. Not infrequently we found that stimuli of the same strength as gave the maximum acceleration before removal of the adrenals did so after their removal also. But sometimes it was necessary to increase the stim- nA A " fy\ W an Nan NV iF WWW ‘yy, ; VS Vy, Ni. VW W . i will \ Wy th “hyyll 4 Wy \ ay " Niwa- Wan\ tis Myyitst yy gt Fig. 14. Parts of blood pressure tracings from cat 448. A, before and B, a portion commencing 26 seconds after beginning of sciatic stimulation, after excision of both adrenals. Reduced to four-fifths. ulus. This is mentioned because if the same strength of stimulus is employed before and after removal of the adrenals, a small, or even no acceleration might be obtained after the operation, which would not mean that the reaction had disappeared because of the loss of the adrenals, but that the excitability of the mechanisms concerned in it had diminished. The same is, of course, true of the vasomotor reflex, which we took as an indicator of effective stimulation. Portions of a curve from cat 448, showing acceleration on sciatic stimulation after removal of the adrenals, are reproduced in figure 14, and the greater part of the curve on a reduced scale in figure 15. EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 351 Protocol. Cat 437; female; weight, 1.61 kgm. Left superior cervical ganglion excised one month previously. Under urethane (3 gm.) cut vago-sympathetics, excised stellate ganglia, prepared central end of left sciatic for stimulation. Rate Pressure Howes-m. Before sciatic. stimulation®........:.. 02.40. ccsecescee 260 116 5 seconds after beginning stimulation................ 274 150 11:50 a.m. Opened abdomen, tied off renal arteries and veins..... : ff-55 a.m) \Before’stiatice stimulation... ... 225...) 2202. 253 60 10 seconds after beginning stimulation........ 264 84 12:00 m. Before stimulation of sciatic with aiderae le aaa FLEE A el i re ee ae et OSS 250 64 2 seconds after beginning stimulation................ 270 88 15 seconds after beginning stimulation................ 273 12:05 p.m. Before stimulation of sciatic with abdominal aorta and PAVARC BBE 32:2 3.07.(4- mo o:Jen1z5-4< ob 02s oeak Ptson ated 2 255 60 4 seconds after beginning stimulation................ 278 84 Pep its Pexeined DOtm@ AMTENSIS, 12... os a =e cys te wena ouae nee e'e Pep Berore seinvic stimulation. .....:.....2<-.2.2+s250 eek 203 44 4 seconds after beginning stimulation................ 232 59 2:15 p.m. Before sciatic stimulation. . net EOS Sete F214 42 10 seconds after beginning Sida b sashes iefey p20 53 rh wnat att patil aet iat ial il Hg gt tig 164 184 Fig. 15. Blood pressure curve from cat 448. Sciatic stimulation after excision of both adrenals. Reduced to one-half. In cat 437, in which a superior cervical ganglion had been excised a month previously and in which, therefore, eye reactions were also available, the adrenalectomy was performed at a different stage of the experiment, and after the abdomen had been opened for some time and observations made on the results of sciatic stimulation with the abdom- inal aorta and cava clipped, as in the collection of blood from the adrenals. The demonstration that the heart reaction is obtained after removal of the adrenals was, if anything, more striking than in many of the other experiments. Sciatic stimulation yielded a maximum 8 The eye reactions were those described in the footnote to protocol of cat 190, p. 319. By G. N. STEWART AND J. M. ROGOFF ~ acceleration of 14 beats per minute before the abdomen was opened. After opening the abdomen and tying off the renal vessels, the blood pressure had fallen to 60 mm. of mercury as compared with 116 mm. at the beginning of the experiment. An acceleration of 11 beats was given by stimulation of the sciatic. The abdominal aorta was now clipped and a short interval allowed for any effect on blood pressure and heart rate to develop, and then the sciatic was stimulated. The heart rate compared with that immediately before stimulation was increased by 23 beats per minute. An equal acceleration was given in another observation, in which the abdominal aorta and inferior cava (above the junction of the iliac veins) were clipped in the same way shortly before stimulation. In such observations the clips were not removed till the portion of the curve to be used for counting the heart beats had been completed. Both adrenals were then excised and there- after an acceleration of 29 beats was caused by stimulation of the sciatic. It is surely impossible to reconcile the results of such an experiment with Cannon’s statements: a, that after opening the abdo- men it is very rare to obtain any acceleration of the heart by sciatic stimulation; b, that preparing the cava pocket for collection of adrenal blood, as practised by Biedl, Hoskins and McClure (14), ourselves and other investigators, renders it impossible to detect the great increase in epinephrin output which according to him is indicated by the heart reaction; and c, that removal of the adrenals abolishes the reaction because it is due entirely to a reflexly excited increase in the secretion of epinephrin. The last experiment to which we shall allude (cat 447) was one in which sciatic stimulation caused only a small acceleration at the begin- ning of the experiment, accompanied by insignificant changes of blood pressure. The cat was only half grown, but it is not known whether this had anything to do with the relatively small effects. The reaction was not essentially modified by removal of the adrenals, or subse- quently by clipping the abdominal aorta. The blood pressure was diminished from about 100 mm. to little over 50 mm. of mercury after the adrenalectomy, and the heart rate was also diminished from over 200 to 175 beats a minute. As the blood pressure fell stil! lower toward the end of the experiment, the heart rate diminished further to 160 beats a minute. EPINEPHRIN OUTPUT AND RATE OF DENERVATED HEART 353 Protocol. Cat 447; female; weight, 1.24kgm. Under ether cut vago-sym- pathetics; excised stellate ganglia; prepared central end of left sciatic for stimulation. Rate Pressure 1-O0.acm., Before seiatie stimulation (@'em:)...........2s..+..s+.. 20s 115 Just-atver endiof stumulations. 2277/5552 52.4 Sees. os 218 110 11:02 a.m. Before sciatic stimulation (10 cm.).............2-e000- 210 100 12 seconds after beginning stimulation................ 207 100 MEOs asm Before sciatic stimulation (9iem.)s 10.02 <- 24 eae esee 205 110 10 seconds after beginning stimulation................ 209 112 ff-1t am. Before sciatic stimulation (7 em.)....../...2-.028-..... 206 110 12 seconds after beginning stimulation................ 209 118 20 seconds after beginning stimulation................ 212 11:18 a.m. Before sciatic stimulation (9 cm.)..................62. 202 104 15 seconds after beginning stimulation................ 208 108 11:40 a.m. Opened abdomen, excised both adrenals MeAsiame Bberorersciaticistimulation (7 cms): sth. .2.s'ss anne 175 53 15 seconds after beginning stimulation................ 181 63 40 seconds after beginning stimulation................ 184 iiEAGea.m. Before sciatic stimulation (9 cm.)).2..........-. asses | L&Q 55 15 seconds after beginning stimulation................ 182 55 30 seconds after beginning stimulation................ 183 55 1i=48%a.m. - Before’sciatie stimulation (6 cm.).:.....+....-+e8esc.. | LMe 54 Just after beginning’ stimulation.:.¢.0........-0cs.6.. Ltd 10 seconds after beginning stimulation................ 172 54 11:53 a.m. Before stimulation of sciatic with abdominal aorta Clap pede (S:Cmiy)!}tys kiecc- <0 Gerejs sis Heese aearees eee Peon 172 55 6 seconds after beginning stimulation................ 171 59 AttermengOr ScmWlAtlONs...0- Bes Ose JEU 50,000; AE., 0.0020; Ry, 10,000. Tuning fork struck by hand just before observa- tion. C, transformer amp. method. Tension, Scene — oO 10m? amie 2 Tix, 125° Bo (same as Ep) 163; Ec, 0; AE, 0.0020;Rn, 10,000. 432 ALEXANDER FORBES AND CATHARINE THACHER with sufficient accuracy to keep the string of the galvanometer in the field. And after a balance had been found with difficulty, it was also difficult to maintain it, for the string kept travelling slowly sidewise as a result either of the running down of the A-battery or the B-battery, or of the change of temperature in the tube. A few records were taken with this method, and as was expected, they showed no great difference from those taken with the condenser method. The chief difference is that there is no falling off during the record of a continued current, such as occurs in the condenser method. On the other hand the gradual shift of base line due to the disturbance of balance mentioned above, though not as rapid as that due to the discharge of the condenser in the other method, is more troublesome in practice. The amplification was also appreciably less with this method than with the condenser. This is probably due to the fact that the experiment was made with the Fig. 16. Comparison of bridge and low resistance string whose resist- condenser methods, showing in bridge ance was so low that it approxi- meio geen of tetne rest" at a shortcteit of the bridge Srinete Figure 16 shows records made with A, condenser method. Tension, thecondenser method and the bridge 1 em. = 5.50 X 1077 amp. Jy, 1.25; method under approximately com- £, 250; He, 0; R, 50,000; AEc, 0.0010; parable conditions, excepting speed Ry, 10,000. Speed of film, 19.5 em. cera per second. : Be ridocact nodal These experiments led us to the 1 cm. = 2.75 X 10-7 amp. J,, 1.18; conclusion that the condenser E, 223; Ri, 40,000; R2, 30,000; Rs, method was by far the most satis- 10,700; Ee, 0; AHe, 0.0010; Fn, 10,000. factory. We proceeded to install a Speed of film 10.7 cm. per second. permanent wiring plan whereby we could rapidly shift from the standard arrangement to the electron tube arrangement, wired in accordance with the condenser method, by merely throwing a few switches. This served a twofold purpose. It enabled us to make rapid comparisons between the excursions obtained with the standard arrangement and those obtained with the electron tube, thereby reducing to a minimum the error introduced by change ant ‘Bets ye AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 433 of tension of the string. It also rendered the method permanently available at a moment’s notice for the amplification of any action cur- rents found, in the course of a physiological experiment, to be too small to record with the string galvanometer unaided. This wiring plan is shown in figure 17. In the compactness obtained by installing Fig. 17. Complete wiring diagram showing switches for rapid shift from con- nection of galvanometer directly with tissue or artificial source of current to connection with electron tube. G, galvanometer (core grounded). £, connection to earth. PS, protective shunt. DS, switch for connecting galvanometer with tissue or substitution resistance in standard arrangement; in electron tube ar- rangement this switch is always thrown to right. 2, in standard arrangement, substitution resistance; in electron tube arrangement, by-pass resistance. N_ non-polarizable electrodes for connection with tissue. C, condenser in series with string. A, filament-heating battery. B, plate battery in three sections connected with switches for charging them in parallel and discharging them in series. J, wires for connection with power mains. Ze, grid bias battery. PC, pole-changing switch in source of current used for test or compensation. D, single-pole double-throw switch for use in electron tube arrangement. When artificial current is used for test or calibration this switch is thrown to left to short-circuit tissue leads; when recording action currents it is thrown to right and switch 2? is open, thereby disconnecting test current circuit from grid circuit. In standard arrangement switches / and 3 are open, switch 2 is closed upwards. In electron tube arrangement / and 3 are closed, DS is thrown to right, 2 is closed downward (or open if D is to right). Cf. reference 4, fig. 1. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOU. 52, NO. 3 434 ALEXANDER FORBES AND CATHARINE THACHER all the apparatus on shelves of the same table which supported the resistances and switchboard of the old equipment we found the double advantage of convenience and of minimizing the disturbances due to induction in the grid circuit. THE SEARCH FOR MAXIMUM AMPLIFICATION General considerations. Having chosen the condenser method as in- dicated in figure 5 as that which was most worth developing, it re- mained to determine if possible how to find those values of the four vari- ables, filament current, grid potential, plate potential, and by-pass 0) E Es i Fig. 18. Simplified schema for analysis of condenser method (see text). Z£, plate battery; Rp, plate-to-filament resistance of tube; R, by-pass resistance (designated R, in figs. 5 and 11); Es, battery susbtituted for condenser; S, string of galvanometer; O, junctional point assumed for convenience to be at zero po- tential in analysis; Z, junctional point actually put to earth, here assumed to be point of variable potential. resistance, Rs, which would yield the maximum possible amplification. Our aim was to find the maximum current in the string which could be made to result from a given change in the difference of potential between grid and filament. For purposes of analysis we may look on the system as simplified in accordance with the schema shown in figure 18. In this the electron tube and the string are shown simply as resistances, and the condenser is replaced by a battery. In reality a difference of potential is main- tained between the plates of the condenser, equal during the resting state, to the voltage drop in R (representing FR, in fig. 11), and the AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 435 capacity of the condenser is so large that for such brief times as we are dealing with, this difference of potential remains practically constant; for purposes of argument we may look on the capacity as infinite. It acts like a battery whose voltage exactly equals the fall of potential in R in the resting state, and it is therefore so represented in the analysis. In practice the zero point which is grounded, and therefore at zero potential, is that point in the system which is connected to the string, the positive end of R, the filament and one side of the input circuit. But in the following discussion it will simplify matters to assume that the point of zero potential is that which is connected with the nega- tive terminal of the B-battery, the negative end of R and the negative plate of the condenser (0 in fig. 18). The potentials of all other points in the system are positive with respect to this point, and we may assign to them absolute instead of relative values if we consider the potential of this point fixed at zero. When the difference of potential between grid and filament is raised the resistance of the tube is de- creased, and the portion of the total plate battery voltage used up in the tube is decreased. Consequently arise in potential occurs at that point in the system at which the filament, the positive end of R, and the string are connected together (Z in fig. 18). Since the negative plate of the condenser is taken to be fixed at zero potential, the positive plate, (in the present schema, the positive terminal of the battery E,) may be taken to be at a fixed potential H,. When the potential of the point Z is raised, a current will flow through the string from this point to the positive condenser plate. If the potential difference in the grid circuit is lowered the reverse process occurs, everything happening in like manner but in the opposite sense; but to avoid confusion of terms it will be simplest to consider in every case the current resulting from raising the grid potential. The quantities we must deal with may be designated as follows: = total plate battery voltage. E, = difference of potential between plate and filament. E, = difference of potential between grid and filament, grid potential. E, = difference of potential between plates of condenser. R, = plate-to-filament resistance of tube. R = by-pass resistance. (Designated as Rz in fig. 11.) S resistance of string in galvanometer. 1, = plate-to-filament current within tube. 1, = current flowing i in R, i, = current in string galvanometer. 436 ALEXANDER FORBES AND CATHARINE THACHER In the resting state a constant current flows through the tube and the resistance, R, in series, while no current flows in the string; in other words 7, = 7,, and 17; = 0. It is our aim to make the current in the string, resulting from a given change in grid potential as great as possible; in other words, to make aa a maximum.® The immediate cause of the string current, 7;, is the rise of potential at the point Z, or the difference of potential between this point and the positive plate of the condenser; this results directly from the di- minution of voltage drop within the tube, AH#,. This in turn results directly from the change in resistance of the tube. It is therefore our aim to make the change in tube resistance as great as possible without introducing factors which will offset the advantage thus gained. With these considerations in mind we may attempt to find in the tube char- acteristics a-‘clew to the best method of fulfilling the desired conditions. The problem resolves itself into two phases; (a) how to make a given change of grid potential produce the greatest proportional change in tube resistance; (b) how to make a given change of tube resistance pro- duce the greatest current in the string. We shall consider (b) first, since the result of this consideration will influence our mode of attack upon (a). Reduced to its simplest terms the fall of potential at the point Z is that which occurs at the junctional point of two resistances con- nected in series with a battery of fixed voltage, when one resistance changes and the other does not. This may be illustrated simply in figure 19 (R, and R being transposed for convenience). Let the nega- tive pole of the battery be kept at zero potential, and the positive pole at the potential HE. Then if R remains fixed, the potential of the junctional point, /,, will change when the resistance R, changes. It can be shown that for any given ratio between R and the initial value of R,, the change in /,, (dH,) depends directly on the proportional or dR, Rs that for a fixed value of # the change in F, for a given proportional change in R, is greatest when R is made equal to R,. In other words percentage change of R,; 1.e., on or dlog R,. It can also be shown ® For those unfamiliar with this notation it may be explained that A signifies a definite increment (or decrement if negative) of the quantity designated; Pun ee dt sy eles : ayy ee : : d signifies an infinitesimal increment; Ee signifies the rate at which a variable dx y is changing per unit change of a related variable x at any given time. AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 437 dE, d log Ry E E dition obtains, since E; is half of H, the value rn will equal = attains its maximum value - when R = R;. When this con- If we are free to make the battery voltage, H, as large as we like, and to make R correspondingly large in order that the actual plate voltage in the tube (£,) may be kept within proper limits, we can increase the value of —_——— Fe nanee of eel at Z for a given pro- dE, d log Ri portional change in Ry): aa more than is possible if H# is fixed and R is made equal to the initial value of R,, as was assumed above. E = Se) E, = 0 Fig. 19. Schema to illustrate change of potential at junction of resistance in series (see text). When £ and R are thus increased, 7 eee BR: approaches the value E, as E becomes infinitely large. In other words if we operate the tube at its proper plate voltage we can make the desired change of potential almost twice as great by making the battery voltage, ZH, very much larger than the plate voltage #,, as by making it only twice as large as E,. On this account it would appear desirable to make the battery voltage as large as possible and to increase R correspondingly. In practice the difficulty and danger of having in the laboratory a bat- tery of much more than twice the 200 volts at which the plate of one of these tubes should operate, would be so objectionable as to offset any such small increase in potential as could be had in this way. 438 ALEXANDER FORBES AND CATHARINE THACHER From the above we have seen that for our purpose we should seek the maximum proportional change of resistance in the tube for a given change of grid potential. This leads us to study the tube char- acteristics. Van der Bijl (1, p. 180) gives us the general equation correlating plate-to-filament current with plate and grid potentials, i, = a(yE, + E. + ©), where a, y and e are constants depending on the structure and properties of the tube, and the other symbols have the meanings already assigned to them. Langmuir (5) has given a similar formula chiefly differing from van der Bijl’s in that the expo- nent is 3/2 instead of 2. It is generally held that some portions of the characteristic curve obey more nearly the formula of van der Bijl, and other portions the formula of Langmuir. The relations embodied in these formulae are expressed graphically in figure 10. Referring to figure 10 B, showing the relation between plate current and grid potential, and bearing in mind that at a given plate voltage plate current varies inversely with the resistance of the tube, we see that the porportional change of resistance for a given change of grid po- tential becomes greater and greater as the grid potential becomes more and more negative. It can be shown mathematically that if either van der Bijl’s or Langmuir’s equation holds, the maximum value d log R, 24 dE. change of grid potential) is obtained at the point in the curve where the current 7, becomes zero: i.e., where the resistance becomes infi- nite. If we go too far toward this point on the curve we shall defeat our ultimate aim, for the current in the string 7; is always less than a, and if this is made excessively small a corresponding limit is placed on7i;. We must seek that point in the characteristic curve of the tube as determined by cathode temperature and by E, and E, which, every- thing considered, will give us the greatest current in the string for a given change in E,. Two ways are open to seek this end, complete mathematical analysis based on the formulae expressing the tube characteristics, and actual experiment with a large variety of systematically chosen values of the independent variables,—filament current, battery voltage, grid poten- tial, and by-pass resistance. Both methods were tried, and the results will be described in turn. Before leaving the general features of the problem it should be noted that with a given grid potential, H., the current increases more rapidly as the plate voltage, Hy, is increased than it would if the tube (i. e., proportional change of tube resistance for a* given AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 439 resistance were fixed. That is, the increase in plate voltage actually lowers the resistance of the tube. The effect of this is to reduce the amplification which would otherwise be obtained. The reason is as follows: Taking our standard case in which the grid potential is raised by a small amount, this change, AZ’., causes a diminution of the tube resistance, R,. But since the plate voltage is not fixed, but depends on the ratio between R, and R in accordance with the relation pee, Ry E R+R because of the above noted characteristic of the tube, increases the tube resistance. Thus the change in resistance which a change in EF, tends to produce is in part wiped out by the resultant change in the opposite sense caused by the resulting change in £;. Mathematical analysis. In this section we are indebted to Dr. E. L. Chaffee for developing a method of analysis and for many helpful criticisms and suggestions. We shall continue to assume the condenser replaced by a battery and the point 0 (fig. 18) to be maintained at zero potential. The obvious method of searching for the law of maximum amplification would be to find an expression for = the diminution of R, serves to lower E;. This in turn, in terms of all the variables in- c volved (except filament current which we may regard in the mathe- matical treatment as fixed), and then by differentiating for each vari- able in turn to find at last the value for each which will give the absolute maximum. Following this method we may begin with the formulae E, + Si, R ig = t — 1, and 7, = derived from Kirchhoff’s laws. Combining these with van der Bi}l’s law for plate current, 7, = a(yH, + £.)*, (neglecting the quantity which is insignificant), we obtain the formula i, = a(yE, + E,)? — BE, + Sis R from which by a simple process of algebra we obtain the formula ea Ro(yEy + E.)? — Ey ; R+S 440 ALEXANDER FORBES AND CATHARINE THACHER This expresses the current in the string in terms of the controllable quantities. But when we attempt to differentiate this expression with respect to L., making allowance for the fact that E, is a function of E., we encounter insoluble expressions, and the method proves im- possible. An ingenious method has been proposed by Doctor Chaffee whereby dis a very close approximation to the true value of can be expressed dE. in comparatively simple terms. The method is as follows:— Referring to figure 18 and applying Kirchhoff’s laws we have the following fundamental equations: EL is —ar = 0, (1) 1p = ly = tes (2) ER ee (3) Combining (1) and (2) and multiplying through by R, we have E,R + i,SR — i,k (S + R) =O substituting (3) and rearranging, we have zit (f= E;) (R + 8) as ER ns (4) % dt, = 1 des Differentiating, a Te 5) ifferentiating iE, RS ( Changing (2) to read 7, = 7, — 7,, and combining it with (1), we have, E, +i8 — ik +7,R = 0; from which = — + St Gest S) 5 (6) RK Differentiating, z 2 uate . (7) di, R } Thus far the equations are rigorous. At this point we make an assumption which introduces a slight inaccuracy, but greatly simpli- fies the treatment. We assume that the tube is operated under con- ditions such that the curve of 7, plotted against /; is a straight line, AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 441 and that the slope of this line is not altered when it is displaced to right or left by a change of E.. For the small changes of HE, and E, involved in this work the assumption is a close approximation to the truth. The operation of the system on this assumption is shown graphically in figure 20. Two curves of 7 plotted against FE, for slightly different values of E., are shown side by side. The slope of diz 1 : these curves in the region used is such that an Oe Tr, may thus be b tice EY _Fig. 20. Graphic representation of simplified mathematical analysis by Doctor Chaffee (see text). regarded as the virtual resistance of the tube. The expression for the curve now becomes — ~ (Es ea mice sony» | (8) In this expression 7, u, and £, are constants. ‘ Then = = a and | eA ii dHE,/£-. or dE,/#b =r dis dE.’ di, _ di,., diy dE, di, dE, the desired quantity; We may now seek an expression for di di ; 5 The expression =e means ae when £, is kept constant. 442 ALEXANDER FORBES AND CATHARINE THACHER Substituting (7) in the above we have di, R y, Uy | gE. 2 Rees sale The next step is to find diy > under the conditions in the system. dE. d 1p = Les Ons” Dye OF, OL, Gin = oer dE, OF, E, dE. OE, Ey Substituting the partial derivatives above this becomes: diy _1 dB, dE, r dk, (9) Now in this system any change in EF, causes a change in Ey. From the analysis of the system in accordance with Kirchhofi’s laws as indicated in equations (1) to (5) any change in E, must be correlated with 7, in accordance with the formula ee Clea (5) Therefore any change in /, produced by a change in F, must cause a corresponding change in 2, which when plotted as in figure 20 will fall on a line whose slope is indicated by = , 1.e., the cotangent of — . Let dE; be the differ- ential of H, when 7 remains constant, 1.e., the change of FE, re- quired to restore 2, to its original value when it has been altered by a change in F; dE, will then represent the horizontal distance between the two curves. We shall use dk; express the actual differential of E, in the system as resulting from a change of E,. whose angle with the horizontal is — By definition Ot, bh dE, _ j&) Ao Bpae Ae dE, \dE Die Minis OE, r AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 443 Therefore = — uy. 10 7B, fe (10) By inspection of figure 20 it is evident that seul, Sie (11) di, Combining this with (5) we get dE, — dE, di, _(R+S)r dip dE, RS Simplifying this equation and adding 1 to each side, dE, _ (R+S8)r+R8S 12 dE, RS a2) Combining this with (10) we get dE, uRS —_— = — —__.. 13 do, (@+8)\r+ RS 8) Substituting this expression in (9) and reducing we get dE, (R+S8)r+RS Finally by combining this with (7) we obtain the desired formula di, uk (15) dE, (R+S)r+-RS dis dE,’ amplification depends, in terms of two known resistances and two constants which may be obtained from the characteristic curves of- the tube, and these curves may readily be obtained by direct experi- ment. Evidently to make this expression large we should choose large values of » and R, and small values of r and S. But whereas the value of the expression is directly proportional to y, its increase with increase of RF is relatively slight. This equation enables us to calculate the quantity upon which 444 ALEXANDER FORBES AND CATHARINE THACHER Experimental. Most of the experiments made in order to determine the method of obtaining the maximum amplification were performed before the mathematical analysis by Doctor Chaffee had been devel- oped. The end result being a function of four independent variables, it was difficult to isolate the effect of varying each quantity by itself. A large number of experiments was necessary. In order to make the basis of comparison uniform we kept the resistance R,, representing the nerve, constant at 10,000 ohms. It has already been mentioned that with the tubes first used, when an initial negative grid potential amounting to 1.5 volts or more was applied, the resulting excursion of the string became independent of the value of this resistance. There- fore the insertion of R, under this condition became unnecessary. Early in the experiments we found that considerably greater amplifi- cation was obtained with a negative grid potential than without. Thereafter all experiments bearing on maximum amplification were made with at least 1.5 volts negative grid potential. For the same reason of uniformity of comparison it was necessary to use the same string throughout a series of experiments, or at least to keep the string resistance constant. Most of the earlier quantitative comparisons were made with string G, having a resistance of 20,000 ohms. Later this string lost its conductivity, and we were left to complete the experi- ments with only string B, having a resistance of 910 ohms. Therefore in order to make the experiments quantitatively valid we introduced a resistance of 19,000 ohms in series with string B to make the resist- ance of that part of the circuit the same as it had been with string G. In a few experiments we placed 20,000 ohms instead of 19,000 ohms in series with string B for convenience. In comparing results, ampli- fication was defined as the ratio between the excursion of the string when connected with the electron tube, caused by placing a given change of potential (one or two millivolts), on the grid, and the excur- sion obtained by placing the same difference of potential in series with the string and the resistance R, arbitrarily chosen to represent the resistance of a nerve. Clearly the amplification with a given arrange- ment depends on the magnitude of resistance selected to represent the nerve, since as already stated this magnitude has no effect on the excursion obtained with the electron tube under the best conditions of operation, whereas it makes a great difference in the excursion obtained without the tube. In giving figures for amplification the assumed value of R, must be stated. AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 445 In order to simplify the problem as far as possible we decided to select a standard value of filament current and to use it in all quan- titative experiments in order that this should not enter as a variable into the result. As already stated we found that the plate current in- creased rapidly with increase of filament current, until the latter reached a value of about 1.1 amperes, and after this, much less rapidly. Van der Bijl has shown (1, p. 175) that a tube should operate in the latter part of its curve; that is, under the condition that the plate current varies little with change of filament current. It is also easier to maintain uniform working conditions with the conductivity of the tube as insensitive as possible to changes of cathode temperature. On the other hand it is desirable not to operate the tube with too high a cathode temperature, since to do so shortens the life of the tube. For these reasons 1.20 amperes was selected as the standard filament cur- rent and used in a great majority of the experiments. In a few cases this was varied in order to determine the effect of doing so. The values EL, R and EL, then remained to be studied with a view to choosing the optimum value of each. In our earlier experiments we varied E,, keeping E and R constant, and then varied R, keeping and E, constant. The results were difficult to interpret, since by changing either one of these variables we changed indirectly the value of E,, the actual voltage drop within the tube, thereby changing the tube resistance and producing a complex instead of a simple change of conditions. For instance, if #. was changed, keeping EH and F& con- stant, the tube resistance was changed, and that portion of the total battery voltage, EL, used up in the tube was correspondingly changed. The consequence was that we were not only studying a different part of the 7, EH. curve but a different part of the 7, E, curve as well. In like fashion a change of R was complicated by a similar change of E;,. We finally decided to carry out some experiments keeping £, constant at approximately 200 volts, the value of the plate potential recom- mended by Doctor Arnold as giving the best results with this type of tube. In this way we sought to determine under comparable condi- tions the best values of grid potential and total battery voltage, and as a corollary of the latter the value of the by-pass resistance, R. As nearly as we could judge from an effort to determine the influence of plate potential as distinguished from that of other factors in the earlier experiments, the best amplification was obtained in the neighborhood of 200 volts, although this value was by no means critical: amplification almost as great was obtained with values higher and lower than this. 446 ALEXANDER FORBES AND CATHARINE THACHER Adherence to approximately 200 volts as a standard therefore ap- peared advisable. If the plate potential, E,, is to be kept constant the problem of choosing the total battery voltage, #, becomes second- ary to that of choosing the value of the by-pass resistance, R. That is, the choice of F is the means to the end of obtaining the desired ratio between by-pass resistance, R, and tube resistance, Rp. Preliminary experiments had shown the greatest amplification to be obtained with values of E. between —1.5 and —4.0, and with values of the ratio lying between 0.4 and 0.9. In order to carry out a b complete experiment to show the effect of varying the ratio Hi at dif- b ferent values of grid potential, and of varying the grid potential at different values of the resistance ratio, we made a careful series of measurements of the tube resistance with the plate potential kept at 200 volts and the following values of grid potential: —2.0, —2.5, —3.0, —3.4 and —3.9. We selected the following values of the bat- tery voltage, H: 278, 302, 322, 342, 362, as giving with the desired plate potential a series of resistance ratios ranging from 0.4 to 0.9. Each series of observations was to begin with a battery voltage of 278. We therefore calculated from the above measurements of tube resistance the value of by-pass resistance, R, which for each value of grid potential would make the plate potential 200 volts with a battery voltage of 278. In this way each series of observations would start with the same value of F, the same value of E,, and consequently the same resistance ratio - namely 0.4. With these calculations made b beforehand the experiment was begun, and on the first observation of each series, the plate current was read with the milliiammeter. There- after each time the battery voltage was increased the resistance, R, was correspondingly increased just enough to maintain the same value of the plate current. In this way we could be certain that we were maintaining the same value of the plate potential, Z,, so long as the grid potential and filament current were not altered, and these were carefully controlled. Having completed such a series with a grid potential of —2.0, we began the second series with a grid potential of — 2.5, using the same steps of battery voltage and beginning with the previously calculated value of R, noting the resulting value of the plate current and again keeping this constant throughout the series. In this way five series were carried out, but the last three were incom- AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 447 plete, since we had not resistances enough to maintain the higher ratios when the tube resistance became very high as a result of using a large negative grid potential. At frequent intervals during the experi- ment the apparatus was rearranged according to the standard wiring plan, and control observations were made withthe current sent directly through the string. From these observations we plotted a curve showing the gradual change of string tension due to heating of the galvanometer, which enabled us to apply the necessary corrections to make the comparison valid. Owing to slight inaccuracies in the previous measurements of tube resistance and possibly to slight TABLE 1 R/Rp Ee 0.40 0.52 0.62 0.75 0.88 A —2.0 10.9 11.4 LER: 10.8 10.7 —2.5 10.8 11.0 tet 10.9 10.5 —3.0 10.9 11.0 TE 10.6 —3.4 10.6 10.5 10.5 —3.9 9.6 10.1 B (corrected) —2.0 11.7 12.4 12.8 13.0 13.4 —2.5 11.6 12.0 Ted ibs mal 1 —3.0 ja baye 12.0 tee ier! —3.4 11.4 (Lats 12.0 —3.9 10.3 11.0 changes in the behavior of the tube, we were unable to keep the resistance ratios precisely the same in each series of observations. But by means of the milliammeter readings of the plate current we were able to be sure that the plate potential remained constant dur- ing each series, and we were able to maintain it at nearly the same value through all of the series. The results are tabulated in table 1 A, the figure for amplification being given on the basis of 10,000 ohms as the value of R,. From this experiment it would appear that the greatest amplifi- cation occurred with a grid potential of about —2.0 volts and with a value of the by-pass resistance about half as great as that of the 4A8 ALEXANDER FORBES AND CATHARINE THACHER tube. On the other hand it is evident from this and from the prelimi- nary experiments that these values are not at all critical; in fact very nearly as great amplification is obtained with values of all the quantities varied over a wide range; in the case of grid potential all the way from —1.5 to —4.2; in the case of by-pass resistance all the way from equality with the tube resistance down to less than half of this value. After the completion of this experiment the mathematical analysis given above was developed, and the results of experiments were there- upon compared with the calculated values of amplification based on the digen uRS dE, (R+S)r+RS making these calculations the values of » and r were obtained by making careful measurements of the tube characteristics, plotting curves of these on codrdinate paper and measuring the slopes of the curves at the proper points. The comparison of observed and calculated values showed in general a striking agreement. In those experimentsin which no initial nega- tive grid potential was applied and in which therefore the grid cir- cuit absorbed current, there was a wide discrepancy between the calcu- lated and observed values. But inasmuch as the mathematical analy- sis takes no account of such action in the grid circuit this particular discrepancy is to be expected. It is furthermore unimportant, since the amplification is much less under this condition than with a nega- tive grid potential. Therefore only those results in which there was a negative grid potential, and the system operated as an electrometer, need be considered. In all such cases the greatest discrepancy found amounted to only about 15 per cent; in a majority of observations it was less than 5 per cent; in some observations there was practically none. On the other hand in the experiments just described in which conditions were most carefully controlled with reference to constancy of plate potential, we found an inverse correlation which, though slight, was constant, and therefore led us at first to think that the formula failed to deal adequately with all the factors. According to the formula the 66 per cent increase in by-pass resistance between the second and fifth observations in the top row of table 1 A, in which grid potential was kept at —2.0 volts, should have caused an increase in amplification amounting to 8 per cent. In reality it caused a decrease in amplifica- tion amounting to 7 per cent. Similarly in the second series when the grid potential was kept constant at —2.5 volts a 60 per cent increase in formula deduced by Doctor Chaffee, AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 449 by-pass resistance, instead of causing the calculated 6 per cent in- crease in amplification, did cause a 4 per cent decrease in amplification. As stated above, in each of these comparisons grid potential remained constant, and plate potential was kept constant with the aid of the miliammeter in the plate circuit. Therefore, since we were working at precisely the same point in the characteristic curve of the tube, the greatest source of inaccuracy, viz., the determination of r and yu depending on the slopes of the characteristic curves, was completely eliminated. An explanation of the apparent discrepancy was found in the leak- age of the condenser. It has already been noted that leakage occurred and that the resistance of fifteen one-microfarad condensers connected in parallel, (the arrangement regularly used) amounted to 32 megohms. When the voltage drop in R is increased, as in this experiment, by in- creasing the resistance ratio and with it the total battery voltage, the difference of potential between the condenser plates is increased and with it the current in the string due to leakage. This current is great enough to. place the string under appreciably increased tension on account of displacement. This increased tension reduces the excur- sion resulting from a given current, and the greater the displacement the greater will be the reduction. In this way it seemed probable that the increase in amplification which should have resulted from increas- , Aap : : ing the ratio = was more than nullified by the concomitant increase b in string tension. An experiment was made to test this point. The string was adjusted to the same initial tension that was used in the experiment described, and then deflected by a current of approximately the same magnitude as that resulting from the condenser leakage in the ease of the lesser voltage drop. A small difference of potential was then applied and the excursion measured. Then the string was further deflected by a current as great as that leaking through the condenser under the greater voltage, and the excursion resulting from the same difference of potential was measured. The decrease of sensitiveness caused by the smaller displacement amounted to 17 per cent; the de- crease caused by the larger displacement, 26 per cent. A correction for this difference applied to the amplification as observed in the ex- periment under consideration, sufficed to change the apparent loss of amplification to a gain of almost exactly the right amount to correspond with the calculated value. The agreement between observed and cal- culated values was thus brought well within the limits of observational THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 450 ALEXANDER FORBES AND CATHARINE THACHER error. The experimental verification of Chaffee’s formula thus became as complete and satisfactory as the experimental conditions rendered possible. In table 1 B is shown the result of the experiment as modified by the corrections for change of string tension due to displacement. This shows the amplification which presumably would have occurred in this experiment had there been no leakage in the condenser. In table 2 are shown a number of calculated and observed values of dis from experiments with different values of u,7, R and S. The c TABLE 2 di, dE. x 104 iW r R S Calculated Peon 35.7 71,500 50,000 910 4.85 4.95 40.0 67,000 50,000 20,000 3.51 4.05 37 .0 62,000 50,000 20,000 3.47 3.30 41.4 69,000 50,000 20,000 Seo) ; 3.83 40.0 71,000 100,000 20,000 3.80 3.80 40.0 71,000 50,000 20,000 3.35 3.42 39.0 65,000 50,000 20,000 3.51 3eap 39.0 65,000 100,000 20,000 3.98 4.02 39.0 65,000 120,000 0,000 4.07 4.11 40.5 69,000 60,000 21,000 3.95 3.99 40.5 69,000 100,000 21,000 3.88 4.30 41.7 71,000 72,000 21,000 3.70 3.85 41.7 71,000 114,000 21,000 3.97 4.23 observed values are corrected for string displacement. They illus- trate the agreement between prediction and result; and considering the difficulty in obtaining accurate values of uw and r, they are quite satisfactory. PRACTICAL DEDUCTIONS Since the formula for amplification dis Bh is ex- dE, (kR+S8S)r+ R&S perimentally proved valid, it should be a simple matter to lay down practical rules for obtaining maximum amplification. Four distinct quantities are involved and may be dealt with in turn. If r could be considered by itself we should want to make it as small as possible since it appears only in the denominator; that is, we AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 451 should want to choose the steepest part of the curve correlating 7, with #, (in fig. 10 A). But since uw is determined in part by r we must consider the two together. Clearly u should be made as large (#:) dE ,/# b ‘gee ie eer hs Moet ne Lp ) dk, E, ratio between the slope of the 7,, EL, curve and the slope of the %, FE, curve, or the relative sensitiveness of plate current to changes in grid potential as compared with its sensitiveness to plate potential. In other words it depends not only on the steepness of the 2, EH. curve but also on the lack of steepness of the 2,, EH, curve. Therefore making r small, while it serves to reduce the denominator of our expression, also reduces the value of » and therefore at the same time reduces the numerator. As already stated, the experimental evidence shows that the value of plate potential, H,, is not critical as regards amplification. Very nearly the same results are obtained with values all the way from 150 to 250 volts. Considering both convenience and efficiency, a value between 160 and 200 volts is to be recommended. With the value of r thus approximately established, » should be made as large as possible. This means we should seek the steepest part of the 7, #, curve. But this leads us into the region where E, is positive, and it has already been shown that unless EL, is made negative the grid circuit absorbs current, and in that way detracts from the amplification. We should then seek the steepest part of the curve where this will not occur. The results shown in the table appear to substantiate this view, in that the greatest amplification here is obtained with the smallest negative value of grid potential, —2.0 volts. On the other hand, in other experiments with the same tube under otherwise closely similar conditions we have obtained just as great amplification with a grid potential of —3.0 volts; and in some cases almost as great amplification with —4.0 volts grid potential. It is therefore evident that with this tube the variations in » due to varying grid potential from —1.5 to —4.0 are not large, and do not make much difference in the resulting amplification. These figures must not be relied on for all tubes of this type. A new D-tube, already referred to in footnote 4, was found to absorb cur- rent with a grid potential of —1.5 volts; it was necessary to use as much as 3 volts negative bias to render the excursions independent of as possible. But pu is defined as equal to 452 ALEXANDER FORBES AND CATHARINE THACHER input resistance, R,, and to bring the tube to its point of maximum amplification. Each tube should be tested in this respect before it is used. In practice the most convenient arrangement is the insertion of either one or two dry cells (according to the requirements of the individual tube) in the input circuit with the negative terminal con- nected to the grid. With regard to by-pass resistance, FR, it is seen that when allow- ance is made for the effect of string displacement due to condenser leakage, an increase In amplification is obtained by increasing its value, provided the total battery voltage is increased at the same time enough to keep the actual plate potential the same. However, the increase in amplification after the by-pass resistance has attained a value of 60 per cent of the tube resistance is very small, and a prac- tical limit is placed on its further increase by the undesirability of installing a battery of much more than 300 volts for so slight a gain as is to be had in this way. The correct value may be readily determined after the battery voltage, E, is established and the tube resistance, R,, has been measured at a plate voltage of about 200 with the negative grid potential found necessary to prevent the grid circuit from taking . ( = 4 . current. The formula is R = Sree Assuming that the tube Ey is to be operated at approximately’200 volts, /,, the necessary data are provided. The value of S is of some importance in determining the amplifica- tion; larger excursions are obtained for a given impressed e. m. f. with a low resistance string than with a high resistance string. On the other hand, on account of large resistances necessarily involved in this system this advantage is less than it is in the case of direct re- ‘ording, except when the tissue is one of very high resistance. Further- more, in the choice of a string the main consideration—lightness and correspondingly small inertia, which enable a string to respond rapidly to currents of brief duration—is just as important when the string is used with the electron tube as without. The string will therefore be selected with reference to its availability for nerve work in general, and may be regarded as a constant in this system, and not as a variable to be selected. In some experiments the effect. of varying the cathode temperature was tested. Although most of the records were made with a filament current of 1.20 amperes, some were made with 1.25 and some with 1.10. Those made with 1.25 amperes showed little difference in amplification AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 453 from those with 1.20. Reducing the filament current to 1.10 amperes in the tube first used caused a slight but appreciable reduction in amplification. For example, in one experiment with the following values: filament current, 1.10; E., —1.5; FE, 270; R, 50,000 ohms; F,, 185 volts; and a resistance ratio of 0.46, the amplification on the basis of 10,000 ohms as the value of 2,,, was 10.3;—the greatest amplification on this basis obtained with any combination (not correcting for string displacement) was 11.5. But with the last tube used reduction of fila- ment current below 1.15 amperes caused a marked loss in amplification. In general it is advisable for the sake of the resulting increase in the length of life of the bulb, to use it at the lowest filament temperature that will give satisfactory amplification. In choosing the battery voltage, H. to be used when the installation was made permanent, we decided on 135 cells as being exceedingly convenient and at the same time efficient. Forty-five lead storage cells is the maximum number which, connected in series, can be charged efficiently from 110-volt power mains; 45 cells can be conveniently arranged in trays by placing them in nine rows of five cells to a row. Three such trays can be charged in parallel, and when connected in series will give at least 275 volts. As indicated above this voltage is sufficient to operate the tube at very near its maximum efficiency. In practice we have found that our battery when freshly charged has 300 volts on open circuit. In an earlier section it was pointed out that inasmuch as the elec- tron tube under proper conditions operates as an electrometer, meas- uring the difference of potential arising in the tissue irrespective of the resistance, the amplification of galvanometric excursion obtained by means of it, depends on the resistance of the tissue, since the excur- sion obtained with the galvanometer alone depends on tissue resistance as well as on electromotive force. In other words, since a large tissue resistance will reduce the response of the unaided galvanometer to a given e. m. f. but will not reduce the response of the galvanometer when working with an electron tube, a comparison between the two arrangements will show a greater gain in the case of a large resistance than in that of a small one. This fact is shown quantitatively in figure 21. Here the amplifi- cation is plotted for each of the three strings against the value of RF, expressed in thousands of ohms. The curve for string @ (resistance 20,000 ohms) is plotted on the basis of an amplification of 11.4 (cor- rected) at 10,000 ohms tissue resistance, this being a value obtained 454 ALEXANDER FORBES AND CATHARINE THACHER in a number of experiments and very nearly the greatest value obtained at all,—in one experiment Chaffee’s formula agreed exactly with obser- vation in giving this value. The curves for strings B and H, with resistances of 910 and 12,000 ohms, are based on the calculated ampli- fication under those experimental conditions in which this precise agreement was found; that is, the proper quantity was substituted for 20,000 as the value of S in the formula. It is notable that as mentioned earlier, amplification is replaced by reduction with the low resistance string when the tissue resistance PL | aE fee [eel fame: Oy NIP | Ca i | Seay 1) | DS deci bas en ae ee a - = aca 1 [ Plas 4 f sgt? 30; | } eee i T | 20 rayne ] ] 1 ; hisee , Pa ie eed esl: 10 20 30 40 50 60 70 80 40 100 Fig. 21. Amplification (ordinates) plotted against tissue or input resistance, Rn, (abscissae) in thousands of ohms, for the three strings used in experiments, and for the particular case in which » = 40, r = 71,000 and A = 100,000. becomes very small. With string B amplification ceases when R,, is reduced to about 1000 ohms. On the other hand with a tissue resist- ance of 42,000 ohms the amplification is identical with all three strings, and with greater tissue resistance the string of lowest resistance ampli- fies the most. With a tissue resistance of 100,000 ohms we obtain with string B a 55-fold amplification, whereas with the 20,000-ohm- string we obtain 45-fold amplification. These curves are only valid for the particular values of u, r and FR selected in this case. We have found that with a different choice of these values different curves are AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 455 obtained, similar in character and again all crossing at a point, but at a different point denoting a different value of R,. Below this point the amplification is always greatest with the string of highest resist- ance; above it the inverse relation always holds. The curves are of necessity straight lines. It should be understood in estimating from these curves the relative merits of low and high resistance strings that the gain of amplification is not an absolute quantity, but only rela- tive; it is merely the gain in excursion over what would be obtained in recording under similar conditions without the electron tube. COMPENSATION FOR CONDENSER LEAKAGE In an earlier section it was stated that in most of our experiments we met the problem of string displacement due to condenser leakage, by simply moving the fiber case laterally and thus bringing the string rt R, 150,000N Fig. 22. Method of compensating for condenser leakage, proposed by Doctor Williams. back into the center of the field. At the suggestion of Dr. H. B. Wil- liams, an electrical method of compensation was tried and found satis- factory. A dry cell was connected with a resistance of 700 ohms. The negative pole of the dry cell was connected with the grounded end of the string. From a point in the resistance separated from this terminal by 200 ohms, connection was made with the other end of the string through the resistance of a graphite line drawn on ground glass, having a resistance of about 150,000 ohms. The arrangement is shown in figure 22. In this way enough current is tapped off from the dry cell circuit to neutralize the current in the string due to condenser leakage, and yet the resistance of the shunt circuit around the string is so much 456 ALEXANDER FORBES AND CATHARINE THACHER higher than that of the string itself that it does not act to any consider- able extent as a short-circuit. The combination described approxi- mately balanced the condenser leakage, so that when the protective shunt was opened very little displacement occurred. The amplifica- tion with this arrangement was measured and compared immediately afterward with the amplification obtained with exactly the same values of the usual constants and with the ordinary method of bringing the string into the field by moving the fibre case. The excursion ob- tained with electrical compensation was slightly the larger of the two. This shows that if there is any loss of efficiency in this method it is less than the apparent loss of efficiency caused by the increased ten- sion due to the usual displacement. If a standard value of each of the variables is adopted as a matter of permanent installation it is a simple matter also to keep permanently installed a dry cell with the necessary resistances, and to effect electrical compensation by the simple throwing of a single switch. By tapping from the compensating cir- cuit through a still larger resistance, say a megohm, there would be practically no loss of efficiency. PROPORTIONALITY OF EXCURSION TO INPUT VOLTAGE For certain purposes it is desirable to know whether our system in amplifying gives a faithful record of the relative values of the disturb- ances amplified. That is, we should ascertain whether an e. m. f. of 2 millivolts applied to the grid will produce just twice as great an ex- cursion of the string as 1 millivolt. In some experiments this point was put to test. In some cases the excursions resulting from 1 milli- volt and 1.5 millivolts were compared in quick succession; in others comparisons were made between 1 and 2 millivolts. The propor- tionality regularly was good; in most observations it was accurate to within 2 per cent, and in no case was the value out of proportion by more than 4 per cent. This is within the limits of experimental and observational error. APPLICATION TO PHYSIOLOGY In order to put the electron tube method to the practical test of recording action currents, experiments were made with the sciatic nerves of both cats and frogs, and with the muscles of the human forearm led off through the skin by the usual method employed in making electromyograms (9). AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 457 The experiments on nerves will be described first. The resistance of a given length of a frog’s sciatic nerve is very much greater than the resistance of the same length of a cat’s sciatic nerve. Therefore the action current in the frog’s nerve should be susceptible of far greater amplification than that in the cat’s nerve, since the resistance in the tissue serves to reduce the current obtainable in the galvano- meter when directly connected with the nerve, whereas we have seen that this resistance causes no reduction in the excursion of the galvano- meter when used with the electron tube. A preliminary experiment was made with a cat’s nerve, using the heavy string (string B). This was done before the best values of the variables for amplification had been found. The result showed amplification, but not as great as was obtained under the more favor- able conditions developed later. After the experiments on maximum amplification had been completed and the permanent installation had been arranged as described above, an experiment was made to illus- trate the amplification of the monophasic action current in a frog’s nerve; in this experiment the new string, H, with a resistance of 12,000 ohms, was used. The sciatic nerve of a good-sized frog was placed in a moist chamber with stimulating electrodes on the central end, and the distal end laid across non-polarizable boot electrodes 15 or 20 mm. apart. The nerve was crushed at a point midway between the boot electrodes to render the response monophasic. The resistance of the section of nerve and the electrodes in circuit with the string was measured by substitution, and found to be 85,000 ohms at the be- ginning of the experiment, rising as a result of drying to over 100,000 ohms at the end. Such a high resistance as this should prove favor- able to great amplification, as indicated above. Specifically, with the combination of tube constants employed, which had previously shown 10.4-fold amplification with 10,000 ohms for R, and a string resistance of 20,000 ohms, there should be with a nerve of 85,000 ohms resistance and a string of 12,000 ohms, a 40-fold amplification. In practice it was found that a maximal stimulus sent the shadow of the string off the film in every observation with the electron tube. ° The lack of constancy in the condition of the nerve made impossible any accurate quantitative comparisons in the case of submaximal stimuli, but as nearly as such comparison could be made it indicated about 45-fold amplification. Records made in this experiment with and without the electron tube are shown in figure 23. In the first pair, (submaximal responses THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 458 ALEXANDER FORBES AND CATHARINE THACHER produced by make shocks) the stimuli were of the same strength. In the second pair, since the maximal stimuli with the electron tube caused the excursion to exceed the width of the film, the response to a slightly submaximal break shock with the tube is compared with a maximal response with the unaided galvanometer. The speed of film is the same in all, but since the tuning fork causes interference with the electron tube arrangement, time is only shown on the records taken without it. It is noteworthy that the duration of the electrical dis- turbance in the nerve is much greater than one would suppose from the records made with the unaided galvanometer, amounting to as % oa Fig. 23. Action currents in frog sciatic nerve (see text). String H. Tension, 1 em. = 4.16 X 10~‘ amp. A, electron tube arrangement; B, standard arrange- ment. Speed of film the same in all. Al:—I,, 1.10; E, 300; Ep, 189; R, 60,000; H., — 1.5. Stimulus submaximal make shock. A2:—I,, 1.10; E, 300; Hy, 207; R, 50,000; H. — 1.5. Stimulus, break shock, 8.0 Z units (Martin scale). B1, Stimulus. make shock of same strength as Al. B2, Break shock, 34.5 Z units. much as a tenth of a second. The observations were made at ordi- nary room temperature, 1.e., about 20°C. In this experiment and in another under similar conditions, every action current recorded, whether produced by a weak or strong stimulus, showed this surpris- ingly long duration. These experiments were performed in December and January with one of the original D-tubes mounted rigidly in the AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 459 open air. Three experiments performed on frog nerves in late Feb- ruary and early March failed to show any such prolonged duration (see fig. 26 A). The experimental conditions were identical except that the new D-tube already mentioned, having slightly different char- acteristics, had been substituted for that used in the earlier experi- ments (this being broken), and the tube was now mounted in the sound- proof box. Tests with an artificial source of current fail to reveal anything in the new tube or its mounting which could account for this great change in the duration of the recorded action current. We are therefore led to the view that the difference in duration is physio- logical, possibly due to some seasonal change in the condition of the nerve. It should also be noted in figure 23 that the direct action of the induction shock (“‘escape of current’’) appears in each record, pre- ceding the action current (8, pp. 186-198). This was the case in all experiments with frog nerves even when the stimuli were subminima!. From its absence in the case of a nerve of low resistance (see fig. 26 B) we may infer that the prominence of this effect depends on a high resistance in the grid circuit. In figure 24 are shown several electromyograms with and without the electron tube, some made with string B, a string of small resist- ance and large inertia, and some with string H having large resist- ance and small inertia. In the later of the two experiments when the high resistance string was used, care was taken to see that the con- tractions in each corresponding pair of observations were of the same strength, as indicated by the readings on the dynamometer which the subject gripped. EXTRANEOUS OSCILLATIONS In figures 23 and 24 it will be seen that those records which were madewith the electron tube and string H manifest fine oscillations at a frequency of about 305 a second. These are superimposed upon the coarser oscillations due to physiological activity. Similar oscillations were found in some of the earlier experiments with string G (also one of small inertia) as well as with string H, when used with the electron tube and an artificial source of current, but they were largest in the experiments with the frog’s nerve. They were absent in all experi- ments made with the heavy string, B. On the other hand it is evi- dent that they are not mere mechanical vibrations of the lighter strings, for they are only evident when the galvanometer is used in connection with the electron tube. B ol VEUVEYVTETUVET TTYSUTYUV ETT TVTOVYTYTYTOV VET TVT SY TTY YCTYTTTTV ETT San mare a VoL. Sal MFI WM r f \ he an Py NE sicrmetacngereacatassonsanaoustnpenttastbrtnssncn sila Pectatedinstguilentsivsitaesaiaedasgtiarieasiciiatciigiipeansaiagiae LSRAAAAASAAAALAALASAARARAAAAAASARARAALAAAABASAAASAAAAASAARABAAAAAAAAAAAARASALAAA CN GES, OO Ge CRP AI ad gage hirg J Rg SP tery OL fd ed a LPR nm, Rynrreenm /\ pa AA VARA ny ANA 0 N a ARAAAAALAAARSAAASAAADABAAAAAAAAAAAAAAAAAAAAALAAAAAAAAAAAAALAAAAAAAAAAZAAAAAARL een wh , . ah f ‘ } ff’ w % ie a ald AF nt wy Nei MARS wars \ ribs \ WW wf Nog” AA AF | ph . ze a) A tae Po | aki, | hy 4 a ies Yea YA ait ™ J UM \ oy uN We WE ie ae ay 868 @ eae " @eeseeegcseeae 3 a6 ss) Fig. 24 AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 461 A careful search was made for the cause of this disturbance. Since the electron tube arrangement had proved so sensitive to induction from the 110-volt motor used to drive the film in the older camera we were led to suspect that induction from other motors in the build- ing might be the cause. Records were taken with all other motors in the building running and with all shut down, and the oscillations persisted under both conditions. The 12-volt motor which moved the film in the new camera was suspected. Records were therefore taken in which this motor was shut off while the clutch was in and the shutter open, so that the film continued to move as the motor slowed down. Still the oscillations persisted. The only significant fact noted in these experiments was that the oscillations were larger when a large resistance was included in the grid circuit, as was of course the case when the action current of the frog’s nerve was being recorded. Finally, it was suggested to us by Dr. H. B. Williams that the cause of these oscillations might be modifications of the elec- tron stream due to vibrations of one or more of the elements in the tube; filament, grid and plate. Such vibration is apt to result in these tubes from mechanical jarring or sound waves. This point was tested by tapping the tube gently while the film was being exposed; also by the sound of the human voice singing on the pitch indicated by the oscillations in the record; that is, about 305 a second. For comparison the sound of the voice singing on a pitch of 260 vibrations a second was tried. The effects of these three procedures are shown in figure 25. Evidently the tube has just such a natural period of vibrations and is extremely sensitive to mechanical jarring. It is also rather highly resonant to sound waves of that frequency. In one experiment these vibrations were reduced to insignificant dimensions by greatly increasing the tension of the string and apply- ing Kinthoven’s method of damping with a condenser across the string terminals (6). This remedy would answer fairly well for certain pur- poses, but it involves decreasing the excursion resulting from a given disturbance. Also it fails to give complete abolition of the oscillations. Fig. 24. Electromyograms of human forearm flexors in voluntary contraction, subject squeezing dynamometer. A, B and C, string B; tension, 1 em. = 5.25 xX 10-7 amp. All others string H; tension, 1 cm. = 4.16 X 10-7 amp. A, strong contraction (strength not recorded), standard arrangement. B and C electron tube; Z,, 1.25; FE, 238; R, 90,000; E., — 3.0. B, weak contraction (10 kilos); C, strong contraction (28 kilos). D, E and F standard. G, H and I, electron tube:—I,, 1.10; E, 300; Ep, 207; R, 50,000; E,, — 1.5. D and G, dynamometer reading, 6 kilos; E, 10 kilos; H, 9.5 kilos; F and I, 20 kilos. 462 ALEXANDER FORBES AND CATHARINE THACHER On the advice of Dr. H. B. Williams, a sound-proof box was con- structed to contain the electron tube, and was suspended by the method of Julius (7). The box was made large enough to contain the tube with an air space of an inch or two between it and the lining. It was made of ¢ inch wood, lined with } inch of lead and an inch of acoustic felt inside of that, except on the bottom where the socket was secured directly to the lead lining. Holes were drilled in the sides of the box to admit leads of flexible stranded bell-wire, which were soldered to the terminals of the tube socket. Outside of the box these leads were soldered to insulated and lead-sheathed wires by which the tube Fig. 25. Oscillations due to vibration of electron tube elements. String H. Tension 1 cm. = 4.16 X 10-7 amp. Condenser method. J,, 1.10; EH, 300; R, 50,000; Fe, — 1.5. A shows effect of gently tapping tube. B, human voice singing on pitch of 305 a second, about a foot from tube. C, singing 260 vibra- tions a second, about the same intensity as in B. ‘Time indicated as usual below. was connected with the rest of the system, still in accordance with the wiring plan shown in figure 17, except that the miliammeter was elim- inated as being unnecessary in the permanent installation. The lead lining of the box and the lead sheaths of the wires were grounded. The box was suspended by means of four steel springs such as are used for sereen doors, attached to screw-eyes in the sides of the box at approxi- mately the level of the center of gravity. The springs were hung from a bench secured solidly to the outside wall of the building, just under the galvanometer. AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 463 When the tube had been mounted in this box thus suspended, the same procedures that had caused the large oscillations shown in figure 25 were repeated. Loud sound waves from the human voice at a fre- quency of about 300 a second failed to produce any perceptible oscil- lations in the light string, H, which had vibrated so vigorously before. Screening from sound was evidently complete. The top of the bench directly over the suspension of the box was rapped vigorously, and this produced small vibrations whose amplitude was only a small fraction of that of the vibrations produced previously by rapping much more gently the table on which the tube had originally been mounted. When neither of these procedures was tried, the string ap- peared wholly free from the characteristic oscillations, whereas these had previously persisted, probably as a result of building vibrations, throughout the greatest quiet obtainable. The new mode of protecting against vibrations was thus found to be quite satisfactory. More experiments were then carried out with both frog and cat nerves. In the case of frog nerves, having in cireuit resistances of over 100,000 ohms, oscillations were still found, but they were much smaller than before the tube was protected, and they were irregular, having frequencies between 190 and 300 per second, with finer oseil- lations of about 600 per second superimposed. In the ease of the sciatic nerve of a cat, the resistance of the part in circuit being about 10,000 ohms, the oscillations are almost if not wholly eliminated. Figure 26 shows maximal monophasic action currents in both frog and cat nerves recorded with the electron tube in the sound-proof box. In each ease there is shown for comparison the response of the same nerve to the same stimulus recorded with the unaided galvano- meter. Calibration curves show the excursion of the string in each case with the same tension and with the same resistance in series when a constant voltage is applied. The irregular oscillations in the experi- ment with the frog’s nerve may be seen in the figure. Apparently the remaining oscillations are dependent on a large resistance in the grid (input) circuit. In seeking further light on their nature it was found that fairly regular oscillations of about 220 per second appeared whenever a large resistance, FR, (e.g., 40,000 ohms) was introduced into the grid circuit as in figure 11. This is the same frequency with which the brushes shift contact on the commutator segments in the power house generator which supplies 110-volt direct current to the power mains throughout the building. Induction from this source is the probable cause of these oscillations. They are much 464 ALEXANDER FORBES AND CATHARINE THACHER smaller than those which were traced to vibration in the tube. If the same resistance was introduced in the filament side of the input cir- cuit, and if the greater part of A; lay on the opposite side of the key from the point at which the input circuit is led off, there was a brief excursion of the string in the opposite direction from that usually seen, following the closure of the key and preceding the usual excur- sion such as is shown in figure 14. This initial excursion is due to a capacity effect which need not be analyzed here. It has no bearing on the recording of action currents. But it is mentioned because of the confusion which may result if the resistance used to determine whether the negative grid potential is sufficient to prevent the absorp- seaeees ‘Needs eineeenieennenel : : A 3 : \AAAAAALA AAAABASA 1 2 Fig. 26. Action currents recorded with tube in sound-proof box. String H. A, electron tube; B, standard; C, calibration curves. 1, frog sciatic nerve, resistance 115,000 ohms. Tension, 1 cm. = 8.3 X 107 amp. Stimulus in each case, 14.6 Z units. 2, cat sciatic nerve, resistance about 10,000 ohms. Tension, 1 em. = 8.3 x 1077 amp. Stimulus in each case, 46.2 Z units. Al, I,, 1.17; E, 300; R, 70,000; E,, — 3.0. A2, I,, 1.06; E, 300; R, 70,000; E,, — 3.0. , 60 = 30 S + 10,000 my ’ S + 100,000 Amplification in 2 is somewhat reduced by low value of J. AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 465 tion of current, is introduced in this side of the grid circuit, and if the primary resistance, R;, is arranged in two parts with the key between them as was the case in most of our experiments. The oscillations persisting in the records of frog nerves are probably in part due to the induction from the generator just mentioned. Their irregularity may indicate some further disturbance of physiological origin, or may be due to some extraneous cause not yet identified. At. all events, the oscillations persisting in spite of the sound-proof box are insignificant as compared with a maximal action current in a nerve trunk; and with a resistance as low as is common in mammalian experi- ments they are imperceptible. PRECAUTIONS Several precautions in the use of the electron tube should be enum- erated and emphasized. The most important of these concerns the avoidance of charging or discharging the condenser through the string. The rule is not to open or close any switch in the filament, grid or plate circuit, in short, anywhere in the electron-tube system, nor to make any adjustment in any part of the system while it is connected with the string galvanometer. All adjustments of current and voltage connected with the tube, and of by-pass resistance, should be made while the switch leading to the string is open and while the protective shunt (or equivalent short-circuit) is closed to enable the condenser to be charged from the B-battery. After all adjustments are made the switch leading to the string may be closed, and finally the protective shunt opened; but it should not be opened until at least 5 seconds after the closure of the switch which permits the charging of the condenser, for on account of the large resistance through which the charging cur- rent must pass, this time is required for the current to subside to a value which is safe to pass through the string. The string should be disconnected again before any switches in the rest of the system are opened. The least neglect of this rule will probably result in the loss of the string. There is one exception; the mere closing of the protec- tive shunt, without disconnecting the string, is sufficient protection during minor adjustments of the filament current if made with suffi- cient caution. If there is any possibility that one of the electrodes may be dislodged from its contact with the tissue under observation during an experi- ment, a shunt should be provided to prevent the grid circuit from THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 466 ALEXANDER FORBES AND CATHARINE THACHER being broken. If a grid “bias’’ is used, the change in tube resistance which would result from breaking the grid circuit, would cause a current in the string so great as. probably to destroy it. As a safeguard a graphite pencil line of about a megohm resistance, drawn on paper, has been connected across the leads to the non-polarizable electrodes, and has been found by experiment not to reduce measurably the ex- cursion resulting from the action current of a nerve. The filament current should be shut off when the tube is not in use. With the tube out of sight in a sound-proof box, one is apt to overlook this at the close of an experiment. The life of a tube is usually long, but not unlimited. In one case a tube was permanently impaired by accidently leaving this current on until the battery was discharged. Other precautions have been touched on in connection with the electrically induced oscillations caused by the Sandstrom kymograph motor and by the electrically driven tuning fork, and in connection with the oscillations traceable to vibration of the tube. Two dis- tinct classes of disturbance are met with, those electrically induced and those of a mechanical origin. Two ways of dealing with the electrical disturbances are open,—prevention of the cause, and inter- vention between the cause and effect by shielding. It is desirable to avoid the use of electric motors in the vicinity of the system, and to minimize sparking from any source. To drive the film a low volt- age motor completely encased in grounded metal is satisfactory. To record time a device which is free from the sparking of an electro- magnetic tuning fork, should be used. It has been found that rapid motion of the body close to the grid circuit, or shuffling of the feet produces marked excursions of the string. These are to be explained by the electrostatic induction in the grid circuit by the more or less charged body. On this account the experimenter should stand as still as possible during an observation. This disturbance is somewhat reduced by grounding the body of the experimenter. This has been effected by keeping the hand on the clutch which sets the film in motion, and which is grounded. If, when a stimulating inductorium is set up near the electron tube ‘system, the primary circuit of the inductorium is touched with the hand, an excursion of the string will result when the circuit is made or broken, even if there is no connection between the inductorium and the recording system (see p. 428). It is essential, therefore that if the stimulating key be operated by hand it should be done through an insulated handle. AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 467 It is important to see that all connections are good; they should be soldered as far as possible. A loose connection will introduce serious disturbance and may result in the loss of a string. It is also very important that all batteries, especially the B-battery, should be in good condition. Dry cells which are nearly exhausted or storage cells in bad condition or bubbling as a result of too recent charging, may cause troublesome excursions of the string. The danger of mistaking static effects or other extraneous disturb- ances for physiological effects is always present in the study of action currents with the string galvanometer. Amplification with the elec- tron tube enormously increases this danger; one cannot be too careful to guard against it both by eliminating all possible sources of dis- turbance and by scrutinizing all results obtained with reference to latent period, dependence on vitality in the tissue studied, and any other possible criteria which may serve to verify the physiological origin of the current observed. SUMMARY 1. The electron tube is a device wherein a stream of electrons emitted from a hot cathode in a vacuum and drawn to an anode (plate) by a high voltage battery, is modified by small changes of potential applied to a third electrode known as the grid, placed in the intervening space. In this way it serves as an amplifying relay and is commonly used as such in radio telegraphy. 2. This device offers great possibilities of amplifying such action currents in the nervous system as are too small to be recorded satis- factorily with the string galvanometer alone. When the string gal- vanometer is used to record the current amplified by this device it cannot be placed directly in the path of the high voltage current, as is done with the telephone, without destruction of the string. This must be protected from the direct current, and yet so connected with the system as to record changes in the electron stream. 3. Three methods of connecting the string with the tube in such a way as to afford the needed protection, were tried. These are desig- nated the transformer method, the bridge method and the condenser method. They are illustrated diagrammatically in figures 3, 4 and 5. The condenser method was the only one which proved wholly satis- factory. It consists in placing a very large condenser in series with the string, and shunting the plate current by both through a resistance 468 ALEXANDER FORBES AND CATHARINE THACHER of the same order of magnitude as the plate-to-filament resistance of the tube. By using a condenser with a capacity of as much as 15 microfarads the distortion due to: gradual discharge of the condenser when the system is unbalanced becomes negligibly small in the case of action currents of brief duration. A permanent wiring arrangement was installed by means of which the electron tube system can be made available for the amplification of action currents in a tissue under observation, by merely throwing a few switches and making two or three adjustments, a process requiring less than two minutes. ‘ D-tubes” loaned by the Western Electric Company, have been used in these experiments. 4. An effort was made to determine those values of filament-heating current, plate-battery voltage, grid potential, and by-pass resistance which would afford the greatest possible amplification. In general, amplification depends on a large proportional change in tube resistance for a given change in grid potential, together with other factors whose interaction is rather complex. 5. Complete mathematical analysis of the problem, using the em- pirically established formulae for tube characteristics, leads to insol- uble expressions. Dr. E. L. Chaffee devised a method of obtaining an approximate formula for expressing in terms of two easily measur- able quantities in the tube characteristics and of the resistance of the string and the by-pass resistance, amplification in any given iar uR dE. (R+S)r+RS rent in the string; E, is grid potential; R is the by-pass resistance; S is the resistance of the string; r is the virtual resistance of the tube as determined by the reciprocal of the slope of the curve correlating plate current with plate potential; « expresses the sensitiveness of plate current to grid potential as compared with its sensitiveness to plate potential, in other words, the ratio between the slopes of the curves correlating plate current with grid potential and with plate potential respectively. 6. A large number of experiments in which the amplification was measured with a large variety of combinations of the quantities above enumerated, verified Doctor Chaffee’s formula. Unless the grid is given an initial negative potential the grid circuit will absorb current and thereby detract considerably from the amplification. It was found that best results are obtained with a filament current of 1.1 to 1.2 am- peres, with a negative grid potential of 1.5 to 3.0 volts (the larger value ease. The formula is In this 7, is the cur- AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 469 being necessary with some tubes), with a plate voltage of approximately 200, and with a plate battery voltage as large as is convenient to install in the laboratory, the difference between plate and plate battery voltage being consumed in the by-pass resistance. This resistance should be at least half as large as the resistance of the tube, and preferably more than that. In practice the combination should depend on preliminary tests of the individual tube. Convenient and efficient arrangements with the D-tubes used have been within the following limits—fila- ment current 1.1 to 1.2; grid potential —1.5 or —3.0 (one or two dry cells); plate battery voltage 300; by-pass resistance 50,000 to 80,000 ohms. 7. The use of paper condensers, which are most convenient when a very large capacity is desired, introduces sufficient leakage of cur- rent to cause a considerable displacement of the string. With the Cambridge string galvanometer the string may be simply brought back into the field by lateral adjustment of the fiber case. A somewhat better method is to tap off from a dry cell through a high resistance an equal and opposite difference of potential. 8. As nearly as we could measure it, the amplified current in the string galvanometer was strictly proportional to the e. m. f. applied to the grid circuit. 9. If the grid potential is kept negative with respect to the filament by a sufficient amount to prevent the flow of current in the grid cir- cuit, the excursions of the string depend wholly on the e. m. f. im- pressed on the grid, and are independent of the resistance of the input circuit, i.e., with values of resistance at least up to 150,000 ohms. In this way the system operates as an electrometer rather than a galvanom- eter. The amplification, or ratio between excursions of the string with and without the tube depends both on the resistance of the input or tissue circuit and on the resistance of the string. The electron tube necessarily introduces high resistance into this system; therefore if the current to be recorded arises in a tissue of small resistance, and is recorded with a string of small resistance, the use of the electron tube will afford no amplification. But if either string or tissue has a large resistance there will be amplification. Given the resistance of the string and the amplification obtained with any given tissue resistance, the amplification obtainable with any other tissue resistance may be simply calculated. In the case of a frog’s sciatic nerve having a re- sistance of 100,000 ohms between the leading-off electrodes, the ampli- fication may be more than 50-fold. 470 ALEXANDER FORBES AND CATHARINE THACHER 10. Records have been made with the electron tube of the action currents in frog’s nerve, and in human muscles during voluntary con- traction, both with a high resistance string and with a low resistance string. These have been compared with action currents recorded di- rectly with the unaided galvanometer under otherwise similar condi- tions. In the case of the frog sciatic nerve an amplification of more than 40-fold has been found. 11. When a motor operated on the 110-volt power mains was used to drive the film for photographic recording, oscillations were induced in the galvanometer through the electron tube system, which we were unable to prevent by screening. It proved necessary to record with a camera designed to work with a 12-volt motor completely encased in metal and run from a storage battery within the room. This camera has certain adyantages over those previously used, and a brief descrip- tion is included. Oscillations were also induced by an electrically driven tuning fork. These oscillations were probably traceable to high frequency currents arising in the electric spark and rectified by the tube. It was necessary to dispense with this method of recording time. 12. Other oscillations occurred in the galvanometer when connected with the electron tube system, which were traceable to vibrations in the elements in the tube, causing fluctuations in the electron stream. This difficulty was eliminated by installing the electron tube in a sound-proof box, suspended by the method of Julius. Smaller oscilla- tions persisted in experiments with nerves of very high resistance, traceable probably to induction from the power mains. They are so small as to be insignificant, and were absent in experiments with tissues of medium resistance. 13. In addition to the precautions indicated in the last two sections, it is most important that the string be protected from the high volt- ages used in this system by careful avoidance of adjustments in any part of the system while the string is connected with it, and by allow- ing the condenser to become charged through a protective shunt or short-circuit before any connection with the string is made. Other confusing effects due to static electricity or bad connections or batteries in poor condition, must be carefully guarded against, if satisfactory results are to be obtained. We wish to express our thanks to Dr. H. B. Williams, Dr. H. B. Arnold, Mr. E. H. Colpitts, Mr. 8S. W. Dean, Mr. Sewall Cabot and AMPLIFICATION OF ACTION CURRENT WITH ELECTRON TUBE 471 Prof. G. W. Pierce for suggestions and advice, and especially to Dr. E. L. Chaffee for the mathematical analysis, and to Miss M. D. Ring for help in preparing the work for publication. BIBLIOGRAPHY (1) VAN DER BiJL: Phys. Rev., N. S., 1918, xii, 171. (2) DeForest: U. S. Patents no. 841, 387, 1907; no. 879, 532, 1908. (3) WHITTEMORE: Phys. Rev., N. S., 1917, ix, 434. (4) Forsres anp Grece: This Journal, 1915, xxxvii, 118. (5) Lanemurr: Proc. Inst. Radio Engineers, 1915, iii, 261. (6) ErntHoven: Annal. d. Physik, 4th Folge, 1905, xvi, 20. (7) Juuius: Annal. d. Physik, N. S., 1895, lvi, 151. (8) ForsBEs AND GreEGeG: This Journal, 1915, xxxix. 172. (9) ForBEs aND RappLeye: This Journal, 1917, xlii, 228. VARIATIONS IN THE RESPIRATORY DEAD AIR SPACE DUE TO CHANGES IN THE DEPTH OF BREATHING R. G. PEARCE anp D. H. HOOVER From the Cardio-Respiratory Laboratory, Medical Service, Lakeside Hospital Received for publication April 8, 1920 The variation which occurs in the capacity of the dead air space of the lungs because of changes in the tone of the bronchial muscles or in the amount of air inspired at each breath, is unsettled. In a previous paper of this series (1) it was shown that the dead air space has a fairly constant volume at ordinary depths of inspiration, and the great variation in the capacity of the dead air space found by Haldane (2) and by Henderson (3) and co-workers was not sustained, while the findings of Krogh and Lindhard (4) were in the main substantiated. These latter observers found evidence of a small increase in the capacity of the dead air space when maximal breaths were taken. While we were unable to obtain absolute evidence of even this small increase, we believe that our early results suggested such a possibility. Our experiments were free from many errors present in earlier work, but owing to the manner of combining our figures we obtained averages of the dead air space for all depths of inspiration. Later work by us indicated that, at moderate volumes of inspiration, the dead air space was a relatively fixed volume. Unfortunately, in the method employed we did not and could not use extreme volumes of inspiration. Moreover the spirometers we used had a dead air space of their own which, although we tried to eliminate it by filling with expired air before the observation, must have contributed toward making our results too high. In order to put this subject to further test, we have devised a method which we believe will, if properly carried out, give a fairly accurate measure of the capacity of the air passages at all depths of inspiration. In this method, as in the former method, we used the mathematical formula given in our first paper (5). By it the carbon dioxide contents of a large and a small expiration following a measured inspiration are compared with the volume of air in a large and a small expiration in 472 VARIATIONS IN RESPIRATORY DEAD SPACE 473 a manner so that the air from the dead space may be determined. . In the application of this formula to estimating the capacity of the dead air space at different lung volumes, we assume that the amount of carbon dioxide which is picked up by the definite volume of air inspired and expired quickly, following a forced expiration to residual air, is fairly constant in amount for a given state of bodily activity. That this is the case we have satisfied ourselves. We also assume that the carbon dioxide in the alveolar air collected for analysis has a uniform concentration. We admit that such is not the case actually, but for practical purposes of the experiment, providing the expirations com- pared do not differ by more than 400 cc., the assumption may be allowed, separ ead rr ete apeloe EN iin a MER Fie. 1 It is also assumed that, given the same volume of inspiration, the capacity of the air passages will be the same in the different observations necessary to collect the required data. By progressively increasing the amount of air inspired, any variation of the dead air space due to stretching of the bronchial tree can be detected if such variations be greater than the experimental error of the method. The essential feature of this method is that a sharp, quick expiration to residual air be followed by a quick inspiration of a measured volume of air, and this followed by a quick expiration, the first portion of which is collected for measurement and analysis. The amount of air collected must be varied between 400 and 800 to 1000 cc. for best results. Briefly the experimental procedure is as follows. Reference is to figure 1. A7A R. G. PEARCE AND D. H. HOOVER Bag A is a flat rubber bag capable of holding 5 liters of air. It connects with a three-way stopcock, G, which in turn connects with stopcock H, which has a mouthpiece, J. Bag B is a flat rubber bag of 1200 cc. capacity attached to cock G. C is an automobile grease gun, delivering 340 cc. of air at a stroke, and D a three-way cock for taking air into the gun and delivering itinto bag A. Hisa100 cc. Luer syringe, and F a three-way cock attached to bag B for measuring the volume of airin bag B. K isa tube with cock for obtaining samples for analysis. The bags are flat and do not hold an appreciable amount of air when evacuated. - Bag A is filled with a measured amount of air by C. The cocks are then as shown in position 1. The subject is seated, and after five minutes of rest makes a quick forced maximal expiration. This must have followed a normal inspiration. Cock H is then turned to position 2, and a quick inspiration empties bag A. The breath is held while cock G is turned to position 3 and a forced expiration is made into bag _ B. In this case, however, cock G is quickly turned to position 4 during the expiration. By this means it is possible to limit the amount of air collected in bag B. The air in bag B is then analyzed accurately for its carbon dioxide content, and its volume measured by means of syringe E. For each inspiratory volume it is necessary to make at least four observations, in which two small volumes and two somewhat larger volumes of the first portion of the expired air are collected in bag B. The amounts of air collected in bag B and their carbon dioxide con- tent are combined in the formula for the estimation of the dead air space as given in our previous papers. The formula reads: Volume of large multiplied by the volume of the small portions of expired air, multiplied by the difference in the percentage of carbon dioxide in the large and small portions of expired air, divided by the difference in the total volume of carbon dioxide found in the large and small portions of expired air, equals the volume of the dead air space. The validity of using this method for obtaining the volume of the dead air space depends on certain factors: a, For each individual series of observations in which a given volume of air is inspired, the amount of carbon dioxide produced per unit of time must be constant during the experiment. For this reason it is absolutely necessary that the subject rest 3 to 5 minutes before each observation. 6, The inspiration of the measured amount of air in the large bag must follow the deepest possible expiration, which in turn must have followed a normal inspira- VARIATIONS IN RESPIRATORY DEAD SPACE 475 tion. This must obtain, since otherwise the amount of carbon dioxide in the lung air and the amount which is delivered by the blood to the lung air during the moment of observation would vary. c, The small bag, which collects the first part of the inspiration, must not contain any air except that which has been received from the first part of the forced expiration. d, The analysis of the carbon dioxide content of the air in the small bag must be very accurate, and the volume of the air collected in the small bag must be measured very accurately. e, The periods of time taken for the observations must be about equal, in order that the carbon dioxide eliminated by the blood be the same for each observation. The results obtained on the subjects are given in tables 1, 2, 3 and 4, and are plotted as curves on cross-section paper, figure 2, the abscissae of which represent the volumes of air inspired and the ordinates the volumes of dead air space found with the respective inspiratory volume. It will be noted that when ordinary volumes of inspiration—1500 to 2500 cc. of air—are taken after a forced expiration (which amount is practically that which is ordinarily present in the lungs at the end of the normal inspiration), the variations of dead air space are very little but progressively increase; while with larger volumes of inspiration the dead air space increases from 100 to 115 per cent maximum. In other words, the increase in dead air space with the increasing volumes of inspiration is not a linear function to the depth of inspiration but a powered function. The explanation of this phenomenon no doubt rests in the fact that the small bronchioles contribute much to the volume of dead air space and are expanded somewhat during the in- spiratory effort. The greater the inspiratory effort, the greater must be the negative pressure of the intra-thoracic pressure. Since the bronchi and bronchioles are made up of more or less elastic tissue, it is reasonable to suppose that their walls are dilated more or less pro- portionately to the degree of intra-thoracic pressure. Since the area of a circle varies as the square of its radius, it is apparent that a slight progressive increase in the diameter of the bronchioles will cause a relatively disproportionate and larger increase in the area of the cross section of the bronchioles. The curves obtained very strongly suggest that this condition obtains to a greater or less degree during deep breathing. It is interesting to note that the maximum variation of dead air space, obtained when the deepest possible inspiration was made, is about 115 cc., correspond- ing closely to the figures obtained by Krogh and Lindhart (4). It 476 R. G. PEARCE AND D. H. HOOVER TABLE 1 Subject R. G. Pearce, height 5 feet 9 inches, weight 176 pounds. Vital capacity: about 4000 ce. ER MENE AIR ANE FROM | AIR pea ey INTO CO: SAGs Mapebiee CO2 IN ec. cc. per cent 1 1700 660 3.50 23 .00 2 1700 690 3.55 24.50 3 1700 390 3.10 12.10 4 1700 300 3.00 10.50 1 2380 685 2.90 19.80 2 2380 590 2.80 16.50 3 2380 303 2.25 6.80 1 3060 585 2.60 15.25 2 3060 383 2.15 8.20 1 3740 720 2.40 21.70 2 3740 655 2.85 18.60 3 3740 665 2.80 18.50 4 3740 395 2.10 7.80 5 3740 365 2.10 7.30 When 1700 cc. of air are inspired: By combining 1 and 3 we have 94.5 ce. dead air space By combining 1 and 4 we have 92.0 cc. dead air space By combining 2 and 3 we have 97.5 ce. dead air space By combining 2 and 4 we have 95.0 cc. dead air space Average dead air space when 1700 ce. are inspired is 94.8 ce. When 2380 cc. of air are inspired: By combining 1 and 3 we find 104 ce. of dead air space By combining 1 and 4 we find 101 ce. of dead air space Average dead air space when 2380 ce. are inspired is 102.5 ee. When 3060 cc. of air are inspired: By combining 1 and 2 we have 143 cc. of dead air space When 3740 cc. of air are inspired: By combining 2 and 3 we find 201 ec. dead air space By combining 3 and 5 we find 182 ec. dead air space By combining 1 and 4 we find 184 ce. dead air space By combining | and 5 we find 165 cc. dead air space By combining 3 and 4 we find 196 ce. dead air space By combining 3 and 5 we find 172 ec. dead air space Average dead air space when 3740 ce. are inspired is 183 ce. VARIATIONS IN RESPIRATORY DEAD SPACE 1700 ce. 2380 ce. 3060 ce. 3740 cc. Summary inspired gives dead space of 94.5 cc. inspired gives dead space of 102.5 ce. inspired gives dead space of 143.0 cc. inspired gives dead space of 183.0 cc. TABLE 2 477 The following results were obtained on D. H. H. between 4:00 and 6:00 p.m., June 3, 1918, the subject being seated in a chair at rest. as above, save that the expired air was collected in a small spirometer. The experiment was the same The dead space of the apparatus when these data were collected was probably 20 cc. greater than with the improved method using the rubber bag B. His vital capacity was about 4000 cc. EXPERIMENT NUMBER INSPIRED EXPIRED SPIROMETER| CQO2SPIROMETER B B CO: sPIROMETER B or wn om © bt ee BROW oo Li) bo S S S me whe & S Oo cc. per cent 560 2.75 525 2.70 930 3.05 1035 3.10 735 2.70 670 2.60 1070 3.05 1085 3.10 1130 2.50 1060 2.47 500 1.80 525 2.00 525 1.85 1035 2.20 1125 2.32 1105 2.33 620 1.75 630 1.75 680 1.85 675 1.75 1125 2.25 1085 2.15 478 R. G. PEARCE AND D. H. HOOVER When 1500 cc. of air are inspired: - combining 1 and 4 we find 120 ec. dead air space - combining 2 and 4 we find 121 ec. dead air space - combining 1 and 5 we find 130 ce. dead air space 7 combining 2 and 4 we find 121 ce. dead air space Average dead air space when 1500 ce. are inspired is 128 ce. By By By By When 2100 cc. of air are inspired: - combining | and 3 we find 120 ce. dead air space By combining 2 and 3 we find 129 ce. dead air space - combining 1| and 4 we find 140 ce. dead air space By combining 2 and 4 we find 139 ce. dead air space Average dead air space when 2100 ce. are inspired is 132 ec. By By When 2900 cc. of air are inspired: By By By By By By combining 1 and 3 we find 208 cc. dead air space combining 2 and 3 we find 206 ce. dead air space combining 1 and 4 we find 156 ce. dead air space combining 1 and 5 we find 208 ec. dead air space combining 2 and 5 we find 209 cc. dead air space combining 2 and 4 we find 166 cc. dead air space Average dead air space when 2900 ce. are inspired is 192 ce. When 3600 cc. of air are inspired: By combining 1 and 4 we find 238 cc. dead air space By combining 1 and 5 we find 245 ce. dead air space By combining 2 and 4 we find 261 cc. dead air space By combining 3 and 4 we find 245 cc. dead air space Average dead air space when 3600 cc. of air are inspired is 247 ce. When 4300 cc. of air are inspired: By combining 1 and 3 we have 241 ec. dead air space By combining 2 and 3 we have 280 cc. dead air space By combining 1 and 4 we have 201 ce. dead air space By combining 2 and 4 we have 253 ec. dead air space Average dead air space when 4300 cc. of air are inspired is 243 cc. case 3800 cc. was about the vital capacity. 1500 ce. 2100 ee. 2900 ec. 3600 ce. 3800 ec Summary inspired gives dead space of 123 ec. inspired gives dead space of 132 ce. inspired gives dead space of 192 ce. inspired gives dead space of 247 cc. . (maximum) gives dead space of 243 ec. In this This is given in form of curve in which 20 ce. are deducted from each figure because of the dead space present in the spirometer. This figure is an approxi- mate one only for the actual dead space of the Krogh spirometer is difficult to measu re accurately. VARIATIONS IN RESPIRATORY DEAD SPACE 479 TABLE 3 Subject R. G. Pearce Re ee ee | commas ee ae cc. cc. per cent 1 | 1700 830 3.05 25.40 2 | 1700 925 3.00 27.75 3 1700 560 2.80 15.70 7 1700 455 2.70 12.25 1 | 2680 720 2.30 16.55 2 2680 520 2.10 10.90 3 2680 470 2.00 9.40 a 2680 340 1.75 5.95 1 3740 750 2.30 17.40 2 3740 663 2.25 14.90 3 3740 430 1.75 7.55 of 3740 425 1.75 7.45 When 1700 cc. of air are inspired: By combining 1 and 3 we find 120 ce. dead air space By combining 1 and 4 we find 101 cc. dead air space By combining 2 and 3 we find 85 cc. dead air space By combining 2 and 4 we find 82 cc. dead air space Average dead air space when 1700 cc. are inspired is 97 ce. When 2680 cc. of air are inspired: By combining 1 and 3 we find 142 ec. dead air space By combining 1 and 4 we find 128 ec. dead air space By combining 2 and 4 we find 125 ce. dead air space Average dead air space when 2680 cc. of air are inspired is 132 cc. When 3746 cc. of air are inspired: By combining 1 and 3 we find 176 ce. of dead air space By combining 2 and 3 we find 176 ce. of dead air space By combining 1 and 4 we find 191 ce. of dead air space By combining 2 and 4 we find 189 cc. of dead air space Average dead air space when 3740 cc. of air are inspired is 183 ec. Summary 1700 cc. inspired gives a dead air space of 97 cc. 2680 cc. inspired gives a dead air space of 132 cc. 3740 ec. inspired gives a dead air space of 183 cc. 480 R. G. PEARCE AND D. H. HOOVER TABLE 4 Table 4 gives the data collected in the case of Dr. M. D., who is 5 feet 93 inches in height and weighs 180 pounds. His tidal air amounts to about 500 cc., and his complemental air to 1200 cc. When breathing in 2200 cc. following a forced expi- ration, his lungs would contain about 500 cc. more than at the end of a normal inspiration, or when breathing 3600 cc. as above, 1900 cc. more than is present at the end of anormal inspiration. His vital capacity was about 3600 cc. EXPERIMENT | AIR INSPIRED FROM | AIR EXPIRED INTO | CO2INSPIROMETER | AMOUNT CO2IN NUMBER BAG A SPIROMETER B B SPIROMETER B (7 ce. per cent 1 2200 1000 3.55 35.5 2 2200 1000 3.50 35.0 3 2200 790 3.40 26.8 4 2200 700 3.30 23.1 5 2200 650 3.25 21-2 6 2200 400 2415 11.0 1 3600 1500 3.00 40.5 2 3600 1050 2.90 30.4 3 3600 800 Ate: 22.0 4 3600 575 2.40 13.8 When 2200 cc. of air are inspired: By combining 1 and 6 we find 130 cc. dead air space By combining 1 and 5 we find 132 ec. dead air space By combining 1 and 4 we find 140 cc. dead air space By combining 6 and 3 we find 121 cc. dead air space By combining 6 and 4 we find 128 ce. dead air space Average dead air space when 2200 ce. are inspired is 131 ce. When 3600 cc. of air are inspired: By combining 1 and 4 we find 185 cc. dead air space By combining 1 and 3 we find 182 cc. dead air space By combining 1 and 2 we find 195 ce. dead air space By combining 2 and 4 we find 182 ee. dead air space By combining 3 and 4 we find 196 cc. dead air space Average dead air space when 3600 cc. are inspired is 184 ec. Summary 2200 ce. inspired gives a dead air space of 131 ce. 3600 cc. inspired gives a dead air space of 184 cc. VARIATIONS IN RESPIRATORY DEAD SPACE 481 is most important to note that variations in the dead air space in the ordinary volumes of respiration are slight, and this point makes it possible to determine the percentage composition of the alveolar air with reasonable accuracy by the examination of the gaseous content. of the expired air, correction being made for the dead air space. It. is also interesting to note that the volume of the respiratory dead air space thus determined is somewhat smaller than has hitherto been thought. c. Cc. DEAO AIR 230 igo 150 Cc, INSPIRED i500 20co 2500 3000 3500 4000 Fig. 2 It was hoped that evidence might be obtained of normal variations in the respiratory dead air space at different times of the day and under varying conditions. Such variation has been noted previously by us when the dead air space was determined by the method described in previous articles of this series. The experimental error of the method is probably too great to allow this to be done. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 482 R. G. PEARCE AND D. H. HOOVER SUMMARY 1. A new method of determining the capacity of the dead air space is described. 2. The volume of the respiratory dead air space is relatively constant at all volumes of respiration. 3. The changes which occur in the volume of the dead air space when the variations in depth of breathing are such as ordinarily occur during moderate effort, are not of as great magnitude as when forced respiratory efforts are made. 4. The percentage composition of the alveolar air can be determined with approximate accuracy from analysis of the expired air and an assumed volume of the dead air space, if the depth of breathing is not excessive. This method is as reliable as the best direct methods now used for the determination of the percentage composition of alveolar air, providing the respiratory volume is large enough to completely wash out the dead air space. BIBLIOGRAPHY (1) Pearce AND Hoover: This Journal, 1917, xliv, 391. (2) Haupane: This Journal, 1915, xxxviil, 20. (3) HENDERSON, CHILLINGWORTH AND WHITNEY: This Journal, 1915, xxxviui, 1. (4) Krocu anp LinpHARD: Journ. Physiol., 1917, li, 59. (5) Prarce: This Journal, 1917, xliii, 73. THE SUB-ARACHNOID AND INTRA-ARTERIAL ADMINIS- TRATION OF SODIUM BICARBONATE AND OTHER ELECTROLYTES J. B. COLLIP From the Department of Bio-chemistry and Physiology of the University of Alberta, Edmonton, Canada Received for publication April 8, 1920 INTRODUCTION It has been shown that extensive voluntary hyperpnoea may produce mild tetany in the normal human subject (1). The causation of this effect was interpreted as a mild alkalosis due to the “washing out”’ of CO, from the blood at the lung surface. A certain amount of evi- dence in support of this view is given by the fact that on two occasions mild tetany was produced in dogs by large injections of sodium bi- carbonate. It would therefore appear that such tetany is due either to a decrease in the C,, of the blood or else to a disturbance in the sodium, potassium, calcium equilibrium within the nervous tissue. In order to throw further light upon this subject a number of experiments were carried out on dogs, rabbits, frogs, and turtles, in which the ef- fects of sub-arachnoid and intra-arterial injections of various electro- lytes were studied. It has been shown by Weed and Wegeforth (2) that irrigation of the spinal and cerebral sub-arachnoid spaces with bicarbonate-free Ringer’s solution can be carried out without serious effect, whereas irrigation with isotonic solutions of sodium chloride results in the manifestation of severe toxic effects. They found also, when the irrigation fluid was free from KCl that no noticeable toxic symptoms developed, while irrigation with an isotonic Ca-free solution resulted in the same toxic manifestations as when 0.7 per cent to 1 per cent NaCl in distilled water was used. The importance of the Ca ion in relation to physiological activity has frequently been emphasized (3). Locke (4) showed that the Ca ion is necessary for the transference of the excitory process from nerve to muscle and Overton (5) demon- strated that it is equally necessary for the transmission of the excitory 483 484 J. B. COLLIP state through the synapse. It is very probable that a variety of effects which may be elicited in various ways, such as will be subsequently de- scribed, are in a large measure due to a disturbance in the Ca balance within the nerve cell. METHODS The effects of sub-arachnoid injections of various electrolytes were studied in both normal and anesthetised animals. Dogs were used for the most part but experiments were also carried out on rabbits, frogs, and turtles. The animal, in practically all experiments where graphic records were taken, was first placed under morphine-ether an- esthesia, a cannula was then inserted in the left carotid artery and con- nected to the mercury manometer, tracheotomy was performed and anesthesia maintained by the ether bottle method described by Jack- son (6). The respiratory movements were recorded by means of a Marey tambour which was connected by rubber tubing to a glass can- nula which was inserted into the rubber tubing connecting the tracheal cannula with the valved top of the ether bottle. The valves in the ether bottle were so adjusted that uniform anesthesia was maintained, the tracing of the Marey tambour lever then being a relative index of the rate and depth of the respiratory movements. Two needles were, as a rule, placed in the spinal canal, one in the lumbar, the other in the subcerebellar region. Injections into the sub-arachnoid space were made very slowly and cautiously and the volume of fluid injected at any one time was small as the records indicate. As a rule 2 or 3 cc. of spinal fluid (in large dogs) was drawn off before an injection was made or else the obturator was removed from the needle for two or three minutes prior to making the injection, and this latter was not replaced until a few minutes after the injection had been made. These precautions were taken in order to exclude the possibility of any effects elicited by spinal injection being due to increased pressure of the cerebro-spinal fluid. Intra-cerebral injections were made through a needle placed in the third ventricle, a smal] opening having first been made in the parietal bone just lateral to the median plane. Intra-venous and intra-arterial injections were also made. The ex- ternal jugular vein was chosen for the former while the vertebral or carotid artery was selected for the latter. Intra-spinal injections were made on rabbits after lumbar puncture had been performed under local anesthesia, while intra-spinal injections SUB-ARACHNOID INJECTION OF SODIUM BICARBONATE 485 were made in frogs and turtles by means of a small hypodermic needle which was placed through the occipito-atlantoid ligament. When dogs were not placed under morphine-ether anesthesia, a small amount of morphine was given by subcutaneous injection and then lumbar punc- ture was performed under local anesthesia. RESULTS Sub-arachnoid and intra-arterial injections of NaHCO;. It was found that the injection of small amounts of sodium bicarbonate solution into the lumbar sub-arachnoid space resulted almost immediately in the manifestation of marked tetany of the musculature of practically the _ whole body, the muscles of the posterior half of the body being, how- ever, more affected than those of the anterior region (protocols 1, 2 and 3). Experiments have been carried out on twenty dogs and six rabbits and in every instance where NaHCO; has been injected into the spinal canal in appreciable amounts, tetany has resulted. Tetanic convulsions were also elicited in frogs and turtles by intra-spinal in- jections of mere traces of NaHCO 3. The effect can be produced in both anesthetised and unanesthetised animals. The strength of the bicarbonate solutions used varied from 1 per cent to 10 per cent. Intra- spinal injections of NaHCO; in addition to producing tetany resulted in intensive and sometimes very much prolonged stimulation of the chief medullary centers which was manifested by a great increase in lung ventilation, rise in blood pressure and various degrees of cardiac vagus activity (fig. 1). If the NaHCO; were injected in sufficient amount to cause most intensive tetanic spasms, respirations were in- hibited and artificial respiration had to be resorted to until the spasm had modulated to a sufficient degree to make spontaneous respiration possible. It was found that tetany could be antagonized to a con- siderable extent by the intra-spinal injection of CaCl (fig. 1), while subcutaneous injection antagonized only to aslight degree. The medul- lary stimulation was, however, not counteracted to the same degree by the intra-spinal injection of CaCl. When injections of NaHCO; were made into the sub-cerebellar cistern after occipital-atlantoid puncture, the medullary stimulation was much more marked, while tetany was less readily obtained. If but a small quantity of NaHCO; was injected in this region there was no tetany but only violent hyperpnoea, increased blood pressure and evidence of cardiac inhibition (fig. 2). If the dose was slightly increased, in addition to the above symptoms mild 486 J. B. COLLIP tetany of the muscles of the face, neck, shoulders, and fore limbs was produced, while a still larger injection caused the tetany to become gen- eral. Intra-cerebral injection of NaHCO; through a needle placed in the third ventricle produced somewhat similar effects as injection into the subcerebellar cistern (fig. 3, e). While intra-venous injections of comparatively large quantities of NaHCO; were without marked effect it could be shown that the medullary stimulation following the sub- arachnoid injection of NaHCO; was due in part at least to the direct | 3CC if Gch, (CC 10/FaHCO, BCE, . . ICS.10/FHEO Taba thie | i ! - Fig. 1. Tetanie convulsions following intra-spinal injection of NaHCO 3. Ef- fect antagonized by CaCl». Time, 3-second intervals. action of the latter upon the nerve cells of the medulla. This was dem- onstrated in the following manner. «4 : t oe i] I Ih Hf 490 J; Bs) COLLIE Sub-arachnoid injections of KCl. Hypertonic solutions of KCl in- jected into the sub-arachnoid space or into the carotid or vertebral artery were found to have the same effect as hypertonic solutions of NaCl, but to a much greater degree (fig. 3, g). CaCl. antagonized the action of KCl in much the same manner as it did NaCl. Sub-arachnoid injections of (NH,)2CO3, MgSOs, NasHPO, and NaH2PO;. (NH4)eCO; was found to have a mild stimulating effect on Fig. 4. Effect of injection into carotid artery of X, NaOH, “ neutral phos- 10 9 216 phate, distilled water and “ acid phosphate. Time, 10-second intervals. the medulla (fig. 2), while MgSO,, as has been well established, had a depressant action. Both basic and acid phosphates injected into the subcerebellar cistern caused increased lung ventilation while neutral phosphate had a slight stimulating action. Intra-arterial injection of both acid, basic and neutral phosphate solutions caused increased re- spiratory activity (figs. 3 a, and 4).- Intra-spinal injection of basie phosphate produced tetany in the rabbit (protocol 6). SUB-ARACHNOID INJECTION OF SODIUM BICARBONATE 491 Sub-arachnoid injections of NaOH. One cubic centimeter of 5}, NaOH injected into the lumbar sub-arachnoid space of a rabbit caused mild tetany in all muscles of the body which was followed by a tempo- rary paralysis of the muscles of the lower extremities (protocol 7). Three cubic centimeters N. sodium hydroxide injected into the lumbar region of the spinal canal of an 8-kilo dog under the influence of mor- phine caused increased tonus of all muscles. Two cubic centimeters of 10 per cent NaHCO; injected into the same region after an interval of twenty minutes caused intensive tetanic spasms of the whole mus- culature. Three cubic centimeters of <=; NaOH injected into the lumbar region of a dog under morphine-ether anesthesia produced very little effect while 1 cc. of ;4, NaOH injected into the sub-cerebellar re- gion of another animal caused a decrease in respiratory movement and a slight fall in blood pressure. One-half cubic centimeter of =~ NaOH placed in the sub-cerebellar cistern caused in another instance decreased lung ventilation and stimulation of the cardio-inhibitory and vaso- motor centers. Injection of 10 cc. of =; NaOH into the carotid artery of one animal caused definite stimulation of the respiratory center (fig. 4). Irrigation of the spinal canal. Irrigation of the spinal canal with isotonic Ringer-Locke’s produced little effect, while irrigation of the sub-arachnoid spaces with Ringer-Locke’s containing 0.3 per cent NaHCO; produced intense hyperpnoea. It was found when the cere- brospinal fluid was washed out from the sub-arachnoid spaces and re- placed by isotonic saline that, after a period of half an hour had elapsed, it was practically impossible to secure fluid from the sub-cerebellar cistern. This was attributed to the inability of the practically protein- free fluid to resist absorption from the sub-arachnoid area. The protocols which follow indicate in greater detail the manner in which some of the experiments referred to above have been carried out. PROTOCOLS 1. Rabbit &, weight 2 kilos 3:55 p.m. Cocaine hydrochloride 0.5 grain subcutaneous and intra-muscular over the fourth lumbar vertebra. 4:00 p.m. Injection of 2 cc. 10 per cent NaHCO; in distilled water into the sub- arachnoid space of the lumbar region. 4:01 p.m. Tetanic convulsion onset of which caused the animal to turn a com- plete somersault. 492 J. B. COLLIP 4:01 to 4:30 p.m. Gradual diminution of the tetany with hyperpnoea through- 4: 11 11 12 WAS 174E 148 is PE 12 PAs Ne 45 p.m. 255 alm. 758 a.m. :00 noon. 05 p.m. 06 p.m. 073 p.m. 10 to 12: 13 to 12: 745 p.m. 50 p.m. xt day. out. Animal able to run around. Animal died the following day. 2. Dog 9 6.5 kilos 1 gr. morphine subcutaneous. 0.5 gr. of cocaine subcutaneous and intra-muscular over fourth lumbar vertebra. 4 ec. of 5 per cent NaHCO; injected into the lumbar sub-arachnoid space. Tetanic spasm of the entire body but most marked in the posterior half. Tetany practically gone. 2 ce. of 1 per cent CaCl, injected into the lumbar sub-arachnoid space. 5 ec. of 5 per cent NaHCO; injected without effect. 8. Dog &, 40 kilos 4 gr. morphine subcutaneous. Ether anesthesia. Tracheotomy and cannula in left carotid connected for recording blood pressure. Lumbar puncture. Clear fluid from needle. 1 cc. of 10 per cent NaHCO; injected through needle in lumbar sub- arachnoid space. Tetany confined to posterior half of body developed at once, rate and depth of respirations increased, and increase in blood pressure from 120 mm. to 200 mm. Hg. 1 ec. of CaCl, by lumbar needle. 1 ce. of 10 per cent NaHCO; by lumbar needle. No tetany. Hyperp- noea and rise in blood pressure from 130 mm. to 165 mm. Hg. 4. Rabbit #7, 2 kilos 0.5 gr. cocaine hydrochloride sub-cutaneous and intra-muscular over fourth lumbar vertebra. Lumbar puncture. 2 ce. distilled water slowly injected into spinal canal. Intense tetanic spasm of all muscles. 13 p.m. Artificial respiration. 45 p.m. Tetany diminished gradually. Animal has control of fore-limbs, head and shoulders. Animal able to run about. Completely recovered. SUB-ARACHNOID INJECTION OF SODIUM BICARBONATE 493 5. Rabbit &, 1.5 kilo 4:00 p.m. 0.5 gr. cocaine hydrochloride sub-cutaneous and intra-muscular over fourth lumbar vertebra. 4:05 p.m. Lumbar puncture. 4:07 p.m. 2 cc. of 10 per cent NaCl intra-spinal. 4:08 p.m. Tetany of whole musculature. 4:13 p.m. Tetany considerably diminished. 4:30 p.m. Animal running about. Next day. Animal normal. 6. Rabbit 7, 2 kilos 4:00 p.m. 0.5 gr. cocaine hydrochloride sub-cutaneous and intra-muscular over fourth lumbar vertebra. M 4:05 p.m. 0.5 cc. 5 Na2HPO; intra-thecal. 4:05} p.m. Tetanic spasm most marked in posterior extremities. 4:10 p.m. Tetany practically gone. 4:15 p.m. Animal running about. 4:20 p.m. 0.5 ce. of NaH2POs intra-thecal; no effect. 7. Rabbit o, 2 kilos 4:00 p.m. 0.5 gr. cocaine hydrochloride sub-cutaneous and intra-muscular over the fourth lumbar vertebra. 4:02 p.m. Lumbar puncture. N 4:05 p.m. 0.5 ce. of i0 NaOH intraspinal. 4:053 p.m. Mild tetany of whole musculature. 4:15 p.m. Tetany practically gone but animal has no control of lower extremities. 5:00 p.m. Only partial control of hind legs. Next day: Slight paralysis of posterior extremities. Five days later: Complete recovery. DISCUSSION It was thought at first that the tetany following intra-spinal injection of NaHCO; was due in part to the change in the C, of the nerve cells as a result of increased concentration of bicarbonate and in part to the specific action of the HCO;ion. When, however, further work disclosed the fact that similar effects could be elicited by intra-spinal injections of such substances as distilled water, NaCl, KCl and NaOH, it became apparent that the factors of H and HCO; ion concentration were prob- ably of secondary importance. The change in concentration of Ca ion within the nerve cell either relatively or absolutely in relation to the 494 J. B. COLLIP concentration of Na and K ions would seem to be the direct cause of the symptoms manifested in these experiments. As the effect of NaHCO; is relatively greater than that of the other electrolytes studied, especially of NaOH, it is possible that the bicarbonate ion as such may have a specific action. This would concur with the opinion of Macleod (7) that the respiratory center is sensitive to the HCO, ion. It is as- sumed, therefore, that change in the relative as well as absolute concen- trations of the various kations is the chief underlying cause for the va- rious phenomena noted in the experiments which have been quoted. Disturbance of the concentration of the Ca ion is, however, the chief factor. As physiologically balanced saline solutions are practically without effect while distilled water and solutions of Ca-free electrolytes in various concentrations produce such marked symptoms, it is evident that in all these latter instances the concentration of the Ca within the nerve cell must have been reduced as a result of the diffusion of the same out of the cell due to the difference in concentration established as a result of the injection. The administration of weak alkaline solu- tions such as NaOH, NaHCO; and NasHPO, would cause as well an actual fixation of a certain amount of Ca within the tissue, the concen- tration of Ca ion thereby being appreciably diminished. The fact that sub-arachnoid injection of sodium bicarbonate caused effects which in magnitude are out of all proportion to the somewhat similar effects produced by NaOH would tend to show that the HCO; ion or the NaHCO; molecule has a specific stimulatory action. Large amounts of NaHCO; intra-spinal are definitely toxic, and result in par- alysis of the motor nerve cells. It is also true that paralysis of the respiratory center results after large injections of distilled water or almost any neutral electrolyte. It is possible that the phosphate ion may also have a specific stimulatory action on the medulla as both the acid and basic salts produce somewhat similar effects on the medullary centers. Intra-spinal injections into the lumbar region, however, dis- closed the fact that tetany is readily produced by the basic but not by the acid salt. This, taken in conjunction with the medullary stimu- lation by acid, basic and neutral phosphate solutions suggests a certain amount of specific action of the phosphate ion on the medulla. The tetany which is produced by lumbar sub-arachnoid injection of NaHCoO;, distilled water, hypertonic solutions of NaCl and other elec- trolytes, is due in the main to direct stimulation of the motor cells in the anterior horn. One cannot exclude, however, the possibility of © the stimulation of the anterior horn cells as a result of irritation of the SUB-ARACHNOID INJECTION OF SODIUM BICARBONATE 495 afferent fibers in the posterior roots or the intercalated neurones in the spinal reflex arcs. The stimulation of the respiratory and vasomotor center must under the above circumstances be largely reflex as the effect is manifested before the injected fluid can possibly reach the medullary centers. When, however, one makes the injection into the sub-cere- bellar cistern there is, no doubt, direct stimulation of the nerve cells in the medulla. The fact that cardiac inhibition through the vagi is ' produced much more readily by sub-cerebellar injection than by lum- bar injection is very significant in this respect. The cardio-inhibitory center is not stimulated reflexly to the same degree as are the vaso- motor and respiratory centers, while direct stimulation of the medullary centers results in a very definite cardiac inhibition, which is manifested by a marked fall in blood pressure. This effect then gives way and the blood pressure rises due to general vasoconstriction. The experiments which are quoted in this paper demonstrate very clearly that substances in solution administered by the sub-arachnoid channel act in a very short space of time directly on the nerve cells of the brain and cord. While the effects following sub-arachnoid injec- tions are, for obvious reasons, much more pronounced than those follow- ing injections into the carotid or vertebral artery, yet the similarity of the symptoms manifested after arterial and sub-cerebellar injec- tion of NaHCOs, for example, are very significant. The very definite stimulation of the respiratory and vasomotor centers following intra- arterial injection of a potentially strong alkali like 10 per cent NaHCO; is positive proof of the sensitivity of these centers to the HCOs ion irrespective of blood C,. The decrease in lung ventilation which fol- lows, as a rule, slow and continuous intra-venous injections of NaHCOs, is probably due to the lowering of the hydrogen ion concentration of the blood, without there being a concomitant increase of sufficient magni- tude in the bicarbonate ion to neutralize or to overcome the effect of the latter. Definite stimulation of the respiratory center can, however, be occasionally obtained during intra-venous injection of NaHCOs. It is of interest in this respect that such an effect is more likely to be manifested by an animal under ether rather than under morphine-ether anesthesia. When massive doses of NaHCO; are given comparatively rapidly by intra-venous injection, it is possible to obtain sudden collapse of the heart without paralysis of the respiratory center. In one instance spontaneous respiratory movements continued for two minutes after the heart had failed completely. 496 J. B. COLLIP While it is possible to obtain direct action of a substance on the nerve cells by sub-arachnoid injection, the passage of such substance into the general circulation may take place very slowly or in some instances it may not be removed from the spinal canal at all (8). The demonstration of Wilson (9) that parathyroid tetany is asso- ciated with an alkalosis of the blood, and the findings of MacCallum, Kellogg and Voegtlin (10) that symptoms like those of tetany can be induced by deficiency of calcium and symptoms following parathyroidectomy relieved by calcium administration would be amplified by the experiments herein recorded. CONCLUSIONS It would seem logical to conclude that the tonus of effector nerve cells and probably of all nerve cells is regulated not only by the C, of the blood and spinal fluid but also by the maintenance in them of a definite equilibrium between the various ions. Of these latter the Ca ion is preéminent. Decrease in Ca ion leads to increased tonus, and increase in the Ca ion to decreased tonus. Slight increase in the con- centration of either Na or K is marked by definite stimulation of the medullary centers while the stimulatory: effect of the bicarbonate ion is very definite. The importance of Na, K and Ca ions as regards reg- ulation of the respiratory center has been suggested by Howell (11). It would seem probable that one of the underlying causes of tetany, as observed clinically, is a disturbance in the kation equilibrium within the nerve cells of the brain and cord. Results, such as are herein reported, serve to emphasize the impor- tance of definite ions of electrolytes in definite concentration within the living cell, a principle which has never been lost sight of since the pio- neer experiments of Sidney Ringer (12) were published. In conclusion I wish to acknowledge the assistance of Dr. P. L. Backus in the experimental work reported in this paper. SUMMARY 1. Sub-arachnoid injection on the lumbar region of NaHCO:, NaCl and distilled water is followed by violent tetany and definite stimula- tion of the medullary centers. 2. CaCl; administered by intra-spinal injection antagonizes tetany so produced. SUB-ARACHNOID INJECTION OF SODIUM BICARBONATE 497 3. Injections of NaHCO;, NaCl and KCl into the sub-cerebellar cis- tern cause marked stimulation of the respiratory, vasomotor and cardio- inhibitory centers. CaCl, antagonizes this effect to a limited extent. 4. Injection into the carotid or vertebral artery of NaHCO;, NaCl or KCI caused definite stimulation of the respiratory and vasomotor centers. 5. Intra-venous injection of NaHCO; may cause increased lung ventilation. 6. It is held that the bicarbonate ion has a specific stimulatory effect upon the nerve cells of the medulla and the motor cells of the anterior horn. 7. Disturbance in the concentration of the various kations particu- larly of the Ca ion within. the nervous tissue produces most marked effects. 5 BIBLIOGRAPHY (1) Coturp anp Backus: This Journal, 1920, li, 568. (2) WreED aND WEGEFORTH: Journ. Pharm. Exper. Therap., 1919, xiii, 317. (3) Howe..: This Journal, 1898, ii, 47. (4) Locker: Zentralbl. Physiol., 1894, viii, 166. (5) Overton: Pfliiger’s Arch., 1904, ev, 346. (6) Jackson: Experimental Pharmacology, St. Louis, 1917, 38. (7) MactEop anp Knapp: This Journal, 1918, xlvii, 189. (8) Becut: This Journal, 1920, li, 1. (9) Witson, Stearns, THOMPSON AND THURLOW: Journ. Biol. Chem., 1915, xxlil, 89. (10) MacCauuvum: Journ. Exper. Med., 1909, xi, 118; Ibid., 1913, xvili, 646; Journ. Pharm. Exper. Therap., 1911, ii, 421. (11) Howe: Textbook of Physiology, Philadelphia, 1918, 705. (12) Rincer: Journ. Physiol., 1880, iii, 195; Journ. Physiol., 1884, iv, 29. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52. NO. 3 THE INFLUENCE OF INTERNAL SECRETIONS ON BLOOD PRESSURE AND THE FORMATION OF BILE ARDREY W. DOWNS From the Physiological Laboratory of McGill University, Montreal, Canada Received for publication April 16, 1920 In connection with work previously reported (1) on the influence of internal secretions on the formation of bile, a study of the blood pres- sure was made. The object of this was twofold: to observe the effect on blood pressure of the particular gland substance being studied, and to determine what relation, if any, existed between blood pressure changes and the amount of bile secreted after a gland substance had been administered intravenously. The following gland substances were employed: mammary, orchic, ovarian, pancreatic, splenic, thymic and thyroid. These were all obtained from Armour & Company. To this list was added solution of adrenalin chloride prepared by Parke, Davis & Company, and secretin prepared by a method described in connection with other experiments (2). These were all given by intra- venous injection. Nearly all of the experiments were performed on dogs but occasionally cats were used. A description of the technique will be found in the article to which reference has been made (1) and the numbers of experiments mentioned in this contribution correspond to the experiment numbers there given. All observations were made for arbitrary periods of twenty minutes each. The usual procedure was to record blood pressure and the number of drops of bile secreted during three twenty-minute periods, one immediately preceding the administration of the gland substance, a second commencing with the beginning of the injection, and a third immediately following the second. In a few cases the third record was not secured. The dose in every experiment, except those with adrenalin and secretin, was 10 mgm. of the gland substance per kilogram of body weight of the animal, dissolved in 100 ce. of physiological saline solution. This was warmed to 37°C. on a water bath and injected into the jugular vein by means of a burette. The dose of adrenalin was 0.1 cc. of a 1:1,000 solution of adrenalin chloride per kilogram of body weight, and the dose of secretin was 10 mgm. of a dried acid extract per kilogram. 498 EFFECT OF GLAND EXTRACTS ON BILE FORMATION 499 Adrenalin. Five determinations were made with adrenalin. There was invariably an immediate rise of pressure which averaged 81.25 mm. of mercury. In the case of adrenalin the injections were made much more slowly than with the other substances studied and the maximum to which the pressure was allowed to rise was 220 mm. This high blood pressure, with slight fluctuations, was maintained for 5 minutes as an average and at the end of the 20-minute period the pressure had always fallen to or slightly below the original level. The amount of bile secreted after the injection of adrenalin was always less than during the preceding 20 minutes and the striking feature was the marked decrease during the period of high blood pres- sure. For example, in experiment 13 the initial count was 140 drops distributed fairly uniformly throughout the period; during the first 5 minutes of the injection period 1 drop was secreted, in the next 5 minutes 3 drops, in the third 5 minutes 3 drops, and during the last 5 minutes 42 drops—a total of 49 drops for this 20-minute period. The diminution produced by adrenalin in the amount of bile secreted may be explained by a decrease of both the arterial and venous inflow. As has been shown by Burton-Opitz (3), (4) adrenalin injected into the hepatic artery or portal vein exerts a local constricting action and diminishes the inflow and it is not improbable that even when the adrenalin is introduced through a vessel more distant from the liver the same action takes place. Mammary. Mammary substance was used eight times and in all but one there was an initial rise in blood pressure followed by a fall and a slow return to normal. The average figures for these experi- ments are as follows: blood pressure at beginning of injection 109.2 mm.; Maximum pressure 117.6 mm., reached 26.4 seconds after injec- tion started; minimum pressure 95.4 mm., 71.4 seconds from beginning of injection; pressure 103.8 mm. at end of first 5 minutes, and 106.8 mm. at end of period. Experiment 46, which failed to show an initial rise in blood pressure, was similar in other respects to the experiments of the group. Ott and Scott (5) report a slight fall in blood pressure for a few seconds after intravenous injection of mammary substance but there is no record of the preliminary rise which occurred so uni- formly in our experiments. So far as bile production was concerned the effect of mammary gland substance was not uniform but it would seem to be a temporary decrease. In experiment 11 the falling off in the rate of secretion was most marked immediately after the injection. Twenty-three drops of 500 ARDREY W. DOWNS bile were secreted during the twenty-minute period following the injection as compared with an initial count of 46; 14 of the 23 drops came during the last half of the period whereas the original 46 drops had been quite uniformly distributed. Experiment 8 was similar; the initial count was 3 drops, but after the injection there was no secretion until the last 4 minutes of the period when 2 drops were recorded. On the other hand, experiments 16, 46 and 57 showed a progressive decrease in the formation of bile. The other experiments of the group showed no change or an increased production. From these records it appears that not only is mammary substance inconstant in its effect on the secretion of bile but that there is no rela- tion between changes in blood pressure and bile formation under its influence. Orchic. Seven experiments were made with orchic substance. Four showed an initial rise of pressure followed by a fall slightly below nor- mal, a gradual return to the normal level and the pressure at the end of the period approximately the same as at the beginning; two did not show an initial rise, but a slight fall and then a gradual rise above the original level persisting to the end of the period. In one, experiment 4, no blood pressure record was made. The average figures for the first group of experiments, viz., 18, 41, 42 and 55, are as follows: blood pressure at beginning of injection, 128 mm. of mercury; maximum pressure 147.2 mm., reached 62.5 seconds from beginning of injection; minimum pressure 124 mm. at 110.7 seconds after injection com- menced; pressure at end of first 5 minutes, 134.7 mm., and at end of period, 126.5 mm. For the second group of orchic experiments, num- bers 19 and 62, the initial pressure is 113 mm., followed by a minimum of 99 mm. in 20 seconds, a rise to 127 mm. in 125 seconds, 123 mm. at end of first 5 minutes and 132 mm. at the end of the period. The reported observations of the effect of the intravenous injection of testicular extract on blood pressure vary somewhat but agree in general that the pressure is lowered. In 1901 Dixon (6) described an immediate and considerable transient fall of blood pressure accom- panied by cardiac slowing following the injection of orchitic extract. Vincent and Sheen (7) and Miller and Miller (8) note the production of a fall in pressure after testicular extract. Vincent (9) and Ott and Scott (5) state that the fall in pressure is slight and lasts for a few seconds only. Wheelon (10) records a slight fall in blood pressure after castration. Bingel and Strauss (11) were unable to produce a change in pressure by the administration of extract of testis. EFFECT OF GLAND EXTRACTS ON BILE FORMATION 501 Orchic gland substance caused a decrease in the amount of bile formed in five of the seven experiments and in four of these the effect was progressive, a further decrease occurring in the second 20-minute period after the injection. One interesting feature presents itself in an analysis of these records: there was a primary decrease during the first 15 minutes following the injection, a rise during the next 15 min- utes, and a second decline beginning about 30 minutes after the gland substance was administered. This second depression of secretory activity was more marked than the first. A comparison of blood pressure changes with variations in the amount of bile secreted fails to show any uniformity. Of the four experiments showing a rise in blood pressure, then a fall and a gradual return to normal, three gave a decrease in bile formation and one, number 55, an increase of 108.33 per cent. In the two experiments where there was a fall of pressure followed by a rise to a point 18 mm. above the original height at the end of the period, the changes in pres- sure were accompanied by decreased production of bile in one, experi- ment 19, and increased production in the other, experiment 62. Ovary. Ovarian substance was used five times, in four of which blood pressure records were made. Two cases, experiments 20 and 59, showed a rise of pressure shortly after the injection began, then a fall below normal, a second rise above the original level, and a gradual return to normal. Averages: initial pressure, 130 mm. of mercury; first maximum, 148.5 mm., 19.5 seconds after injection began; minimum 119.5 mm., 44 seconds from beginning; second maximum 137.5 mm. at end of first 3 minutes, and final pressure 132.5 mm. Two other exper- iments, 43 and 56, with ovarian substance failed to show changes similar to the above, the pressure remaining almost constant. In experiment 5 no blood pressure record was secured. The effect of ovarian gland substance or extract of ovary on blood pressure has been reported by Ott and Scott (5), Vincent and Sheen (7), Miller and Miller (8), Vincent (9) and Gonalons (12). These investigators agree that the effect is a lowering of blood pressure which is usually slight and transient. Ovarian substance invariably produced a decrease in the amount of bile formed. The average decrease was 44.59 per cent in the first period and 59.36 per cent in the second period after the injection. Pancreas. This substance was employed in nine experiments and for six of these complete blood pressure records are available. The effects on blood pressure fall into two groups, one showing a prompt 502 ARDREY W. DOWNS and very marked fall in blood pressure, the other showing a prelim- inary slight rise followed by a fall. In all of these records there is a rather characteristic lowering of blood pressure which is still in evidence at the end of the period, but from which there is gradual recovery during the second 20-minute interval. Forty minutes after injection the blood pressure averaged 121 mm. as compared with 123.6 mm. immediately preceding the injection. The first group consists of four experiments, 22, 38, 48, 50, and shows the following: initial pressure 123.2 mm.; fall began in 5.5 seconds and the minimum pressure, 60 mm., was reached 26.7 seconds from the beginning of the injection. At the end of 5 minutes the pressure was 86.5 mm., and at the end of the period, 95.7 mm. The second group, experiments 14 and 52, gives these averages: initial pressure, 124 mm.; 17 seconds later pressure began to rise and reached a maximum of 131.5 mm. 19 seconds after the injection was started; then fell to 77.5 mm., 40 seconds from beginning; at end of first 5 minutes was 118.5 mm., and at end of period 107 mm. With the exception of Popielski (13), who reports a marked and pro- longed rise in blood pressure as the result of the injection of an acidu- lated watery extract of pancreas, investigators agree that injection of the substance of the pancreas or of saline extracts of pancreas causes a fall in blood pressure. Ott and Scott (5) state that the lowering of pressure is more marked than that obtained with ovary, testis, mam- mary, spleen and thymus. Biliary secretion was decreased in every case in which pancreatic substance was administered. The reduction averaged 45.45 per cent in the first period and 44.28 per cent in the second period following the injection. The striking feature in this connection was the great reduc- tion in secretory activity immediately after the injection. The counts made during the periods preceding the injections show the drops of bile falling at a fairly uniform rate. In the first 10 minutes subsequent to the introduction of the gland substance only 19 per cent of the total secretion for the period was obtained, the other 81 per cent occurring during the second half of the period. In the next 20-minute period, the second after the injection, the rate of secretion was more uniform but still only a little more than one-half the original. Secretin. Seven experiments were carried out with secretin. The blood ‘pressure records show an average pressure preceding the injec- tion of 114.1 mm. of mercury; 20.2 seconds after the injection was begun pressure commenced to fall and reached a low point of 56.8 mm. EFFECT OF GLAND EXTRACTS ON BILE FORMATION 503 40 seconds from the starting point; at the end of 5 minutes the pres- sure was 110.5 mm., and at the end of the period, 118.4 mm. In two experiments, 26 and 54 of this series, there was a slight rise of pressure immediately following the injection. This increase above the original pressure averaged 15 mm. of mercury and in experiment 26 was suc- _ eeeded by a drop similar to that which took place in the other experi- ments of the group; in no. 54 there was no abrupt fall in pressure but a gradual decline with the pressure at the end of the 20-minute period nearly the same as at the beginning. The amount of bile produced was greatly increased in every case in which secretin was employed. This increase averaged 241.52 per cent for the first 20 minutes after the injection and 413.78 per cent for the second 20 minutes. Bayliss and Starling (14) state, “‘ Acid extracts of the mucous mem- brane ( of the duodenum and jejunum) normally contain a body which causes a fall of blood pressure. This body is not secretin, and the latter may be prepared free from the depressor substance by acting on desquamated epithelial cells with acid.”’ This has been confirmed by v. Firth and Schwarz (15). Matsuo (16) also concludes that the depressor substance is separate from secretin, especially as acid injected into the duodenum, while producing copious pancreatic secretion, was | followed by no change in blood pressure. He was, however, not able to obtain a secretin preparation which did not produce some fall in blood pressure, but the degree of the fall and the activity of the various preparations were not at all proportionate. In our experiments it will be observed that while there was, as a rule, a fall in pressure immediately following the injection this had been recovered from in 5 minutes and from then on the pressure re- mained within a few millimeters of the original. At the same time the amount of bile was tremendously increased during the 40 minutes the experiments lasted. Spleen. Injection of substance of the spleen was practically without effect on blood pressure. Seven trials were made and the average pressure prior to the injection was 104.4 mm.; 20 minutes after the injection it was 107.4mm. Only two experiments showed any fluctua- tions that could be attributed to the injection. In experiment 21 the pressure rose from 128 mm. to 135 mm. in 20 seconds, then fell to 124 mm. 22 seconds later, rose to 136 mm. 63 seconds later, and declined gradually to 129 mm. at the end of the period. The initial pressure in experiment 61 was 130 mm. With the beginning of the injection 504 ARDREY W. DOWNS blood pressure commenced to fall and in 40 seconds had reached 124 mm.; from there it rose gradually to 140 mm. at the end of the period. Ott and Scott (5) report a slight fall in blood pressure for a few seconds after the intravenous administration of splenic substance. Vincent and Sheen (7) and Vincent (9) noted the production of a tran- sient fall in pressure by splenic extract. Oliver and Schafer (17) state that spleen produces a preliminary fall of pressure followed by a gradual rise above normal and then a gradual return to normal. Bingel and Strauss (11) found inconstantly a temporary rise of pressure followed by a sharp fall. Muller and Miller (8) state that in their hands saline extracts of spleen invariably caused a rise in blood pressure, which was usually but not always followed by a slight fall below normal. So far as the formation of bile after the introduction of splenic sub- stance by vein is concerned we found no constant effect. In the first period after the injection four experiments gave a decrease, two an increase and one no change. Alterations in the rate of secretion during the second period after injection were also without harmony. Thymus. Five experiments were carried out with thymic substance and no consistent effect on blood pressure noted. Experiment 39 showed a slight rise in pressure followed by a sharp fall, and a return to normal 6 minutes after the injection began. Where any effect on blood pressure was produced by the injection intravenously of thymic gland substance, watery extracts or saline extracts of the thymus, observers are almost unanimous in reporting it as a decrease. Reference is made to Ott and Scott (5), Miller and Miller (8), Schafer (18), Popper (19), Basch (20) and Lucien and Parisoot (21). Popielski (13) states that an acidulated watery extract of thymus caused a rise of blood pressure. ‘Reference to our previous paper (1) on this subject will show that we found the amount of bile secreted after the administration of thymic substance to be decreased. This effect was still in evidence in three out of four experiments in which a count was made for a second period of 20 minutes after the injection. Thyroid. Substance of the thyroid gland was injected into nine dogs. In two of the experiments, 6 and 33, no record of blood pressure was obtained. The other seven all show an initial rise in pressure, a sharp fall, a second gradual rise and a fall to normal. Pressure during the second 20-minute period following the injection was practically the same as in the period preceding the injection. Average figures for these experiments follow: initial pressure 105.8 mm.; first maximum EFFECT OF GLAND EXTRACTS ON BILE FORMATION 505 119.4 mm., 19 seconds from beginning; minimum 86.5 mm., 43.4 sec- onds from beginning; second maximum 117.7 mm., 191.4 seconds from beginning, and pressure at end of period, 110.2 mm. Other observers have usually found a fall in blood pressure to result from the administration of thyroid extract or thyroid gland substance. Oliver and Schafer (17), Haskovec (22), Georgiewsky (23), Guinard and Martin (24), Fenyvessy (25), v. Cyon and Oswald (26), v. Fiirth and Schwarz (27), all report a fallin pressure. Ott and Scott (5) used iodothyrin and obtained a marked fall with subsequent gradual rise above normal. Schafer (18) observed a considerable fall when using thyroid extract. Vincent (9) usually noted a fall but occasionally a rise of pressure. Levy (30) records no appreciable alteration in blood pressure after intravenous injection of Kendall’s crystalline thyroid iodine compound. A rise in pressure is reported by Popielski (13), Heinatz (28) and Livon (29). So far as bile formation is concerned thyroid substance was without constant effect. Experiments 6, 28, 40 and 51 show an increase aver- aging 82.29 per cent in the first period after the injection. Experi- ments 12, 17, 33, 37 and 58 show a decrease averaging 33.85 per cent for the same period. Nevertheless, in spite of the wide variations in amount of bile secreted, the blood pressure ran a very similar course in all of these experiments. DISCUSSION While the foregoing experiments are too few in number to permit definite conclusions to be drawn, one thing seems certain, viz., that there is no constant relation between blood pressure and the amount of bile secreted. Adrenalin, it is true, consistently raised blood pres- sure and lowered bile formation; secretin, on the other hand, where it caused a change in blood pressure, produced a lower pressure and a great increase in the flow of bile. It might be urged that thyroid gland substance owes any action it exerts upon blood pressure and bile for- mation to the intervention of the adrenals and this cannot be entirely controverted by our experiments. That thyroid substance increases the output of adrenalin has been shown by Biickner (31), Rudinger, Falta and Eppinger (32) and Gley and Quinquaud (33). In our experi- ments, however, there was no constant relationship between blood pressure and bile production after the administration of thyroid gland substance. We did not find bile production regularly decreased when blood pressure rose or vice versa. With others of the gland substances 506 ARDREY W. DOWNS employed the blood pressure might be lowered and bile production decreased at the same time. The most striking example of this is in the series with pancreatic substance. Intravenous administration of substance of the pancreas caused lowering of the blood pressure and lessening of the output of bile. To a lesser extent thymic substance acted in the same manner. If we place in contrast with this the effect of secretin it would seem that we are not justified in concluding that the effect on bile formation is due to the altcration in blood pressure. With the other gland substances employed, viz., mammary, orchiec, ovarian and splenic, the results were inconstant. Orchic substance, for instance, caused a rise of blood pressure in some of our experiments and a drop in others, while the production of bile was definitely lowered. After the administration of substance of the ovary the blood pressure showed oscillations, or waves of higher and lower pressure. The pressure at first rose above the original level, then fell below the initial, rose again and returned to the original. Biliary secretion was lowered in every case and an examination of the individual records of these experiments does not show any synchronism between changes in blood pressure and the rate at which the bile was secreted. CONCLUSIONS 1. As a result of the experiments set forth in this paper we feel inclined to believe that some at least of the endocrine organs exert a specific influence on the secretory activity of the hepatic cells leading to the production of bile. 2. The output of bile in the dog is increased by the administration of secretin. 3. The output of bile in the dog is decreased by the administration of adrenalin, and by mammary, orchic, ovarian, pancreatic and thymic gland substances. 4. The amount of bile secreted is not affected in a constant or defi- nite manner by the substance of the spleen and thyroid gland. 5. Blood pressure is raised by adrenalin. 6. Blood pressure is lowered by pancreatic substance and the secretin preparation employed. 7. A fall of blood pressure, ordinarily preceded by a slight rise, is caused by orchic and mammary gland substances. 8. Oscillations of blood pressure are caused by ovarian and thyroid gland substances. 9. Blood pressure is not usually affected by splenie and thymic gland substances. EFFECT OF GLAND EXTRACTS ON BILE FORMATION 507 BIBLIOGRAPHY (1) Downs AND Eppy: This Journal, 1919, xlvili, 192. (2) Downs anp Eppy: This Journal, 1917, xliii, 416. (3) Burtron-Opitz: Quart. Journ. Exper. Physiol., 1912, v, 309. (4) Burron-Oprrz: Quart. Journ. Exper. Physiol., 1912, v, 329. (5) Orr anv Scott: Amer. Med., 1914, xx, 249. (6) Drxon: Journ. Physiol., 1901, xxvi, 244. (7) VINCENT AND SHEEN: Journ. Physiol., 1903, xxix, 242. (8) MiLtuer AND Miter: Journ. Physiol., 1911, xlii, 242. (9) Vincent: Internal secretions and the ductless glands, 1912, 29. (10) Wuereton: This Journal, 1914, xxxv, 283. (11) Brneext anv Strauss: Deutsch. Arch. f. klin. Med., 1909, xevi, 476. (12) Gonatons: Surg., Gynecol. and Obstet., 1918, xxvi, 196. (13) Poprreusxt: Zentralbl. f. Physiol., 1909, xxiii, 137. (14) Bayxiss AND STaRLING: Journ. Physiol., 1902, xxviii, 325. (15) v. Firru anp Scuwarz: Pfliiger’s Arch., 1908, cxxiv, 427. (16) Martsvo: Journ. Physiol., 1913, xlv, 447. (17) OLiveR AND ScHAFER: Journ. Physiol., 1895, xviil, 277. (18) Scudrer: Brit. Med. Journ., 1895, ii, 343. (19) Popper: Sitzungsb. d. K. Akad. d. Wissensch. z. Wien, Math.-Naturwis- sensch. Klasse, 1905, exiv, 539. (20) Bascu: Jahrb. f. Kinderheilkunde, 1906, Ixiv, 70. (21) Lucien aNnp ParisooT: Compt. rend. d. 1. Soc. d. Biol., 1909, Ixvui, 377. (22) Haskovec: Revue Neurol., 1896, iv, 193. (23) Grorairwsky: Zeitschr. f. klin. Med., 1897, xxxiii, 153. (24) Guinarp AND Martin: Compt. rend. d. |. Soc. d. Biol., 1899, xi, 161. (25) Fenyvessy: Wien. klin. Wochenschr., 1900, xii, 6. (26) v. Cron anp Oswa.p: Pfliiger’s Arch., 1901, Ixxxiii, 199. (27) v. Fixrs anp Scuwanz: Pfliiger’s Arch., 1908, exxiv, 361. (28) Heinarz: Quoted by Georgiewsky, Zeitschr. f. klin. Med., 1897, xxxili, 153. (29) Livon: Compt. rend. d. 1. Soc. d. Biol., 1898, x, 98. (30) Levy: This Journal, 1916, xli, 492. (31) Bicxner: Compt. rend. d. 1. Soe. d. Biol., 1908, Ixiv, 1123. (32) Ruprnemr, Fauta anp Eppincer: Zeitschr. f. klin. Med., 1908, Ixiv, 1. (33) GLEY AND QuiInquaup: Compt. rend. Acad. d. Sci., 1913, elvi, 2013. GASTRIN STUDIES V. CHEMICAL STUDIES ON GASTRIN BODIES F. C. KOCH, A. B. LUCKHARDT anp R. W. KEETON From the Hull Laboratories of Physiological Chemistry and Physiology of the University of Chicago and the Laboratory of Pharmacology of the University of Illinois, College of Medicine Received for publication April 16, 1920 As a result of numerous physiological studies with extracts from duo- denal and gastric mucous membranes and from many other tissues investigators, at present, are divided into two schools. One school considers pancreatic secretin on the one hand and gastrin or gastric secretin on the other hand to be different chemical substances and to act as distinctly specific secretagogues. The other school considers the substances to be identical, that they are widely distributed ‘vaso- dilatins” and that they do not act as specific secretagogues. The for- mer view is generally accepted in our modern textbooks and this is especially true as topancreatic secretin. In a series of previous papers (1) we referred to the literature on gastric secretin and we there reported more particularly on the physiological phase of the work. Although much of the evidence there presented together with the investigations published by others leads us to believe that very likely secretin and gastrin are different substances; nevertheless, we have not been able to convince ourselves as to the correctness of either of the two views now held. In this paper we report observations which bear more par- ticularly on the chemical phase of the work. It is evident that the iso- lation of these secretagogues in pure form will do more toward solving these important questions than biological studies with variously puri- fied extracts. METHODS Animals. The studies reported here were controlled on dogs with Pawlow and chronic fistula stomachs. The same general procedure was followed as previously reported, that is, a collection of a one- or 508 CHEMICAL STUDIES ON GASTRIN BODIES 509 two-hour spontaneous secretion before the injection of the solution and a one- or two-hour collection after the same. The volumes were noted and the free and total acidities titrated in the usual way. All the solutions used were slightly acid and were well sterilized before in- jecting intramuscularly. Extraction. It may be recalled that Bayliss and Starling (2) at- tempted to meet Popielski’s criticism as to the presence of “vasodil- atins”’ in secretin solutions by extracing the dried material with abso- lute aleohol and discarding the absolute alcohol-soluble portion. In our earlier studies we also employed this method, but later we arrived at the conclusion that we were by no means warranted to assume that the absolute alcohol did not also remove some of the gastrin activity. The studies reported here are on material which in most cases has not been thus treated with absolute alcohol and which was prepared from filtered artificial gastric juice as prepared for the manufacture of pepsin. The filtered material as obtained from the factory was coagulated by heat, filtered, and the filtrate concentrated under diminished pressure in a bath of boiling water. The thin syrup thus obtained was neutral- ized to faint acidity and precipitated with five to six volumes of redis- tilled 95 per cent alcohol. After standing for several days the solution was filtered and the filtrate again concentrated under diminished pres- sure with complete removal of the alcohol. The solution left, if not distinctly acid, was made-so_and stored under toluol in glass stoppered cylinders. It retained its activity perfectly. This concentrated solu- tion was used as the material for various attempts at further purifica- tion and for various studies as to stability. PRECIPITATION STUDIES Action of basic lead acetate. The addition of basic lead acetate in excess to this partly purified solution of gastrin does not precipitate the activity as is shown by a number of experiments in which both the precipitate and filtrate were freed from lead and tested physiologically. In each case very little or no activity was found in the precipitate and most of the original activity was recovered in the filtrate. As an illus- tration one of three equally reliable experiments is given below. 510 F. C. KOCH, A. B. LUCKHARDT AND R. W. KEETON Dog IIT. Pawlow accessory stomach, January 14, 1916 HYDROCHLORIC ACID TIME JUICE Free Total ce: per cent per cent 10:10 a.m. to 11:10 a.m. el 02095 0.12 Mied Ov arnasteen) ange Injected 2 ec. Sb4B* UBIO Wow cayenne ele ge 6.9 0.36 0.39 Obs ore eee 1.4 0.36 0.40 1:10 pum. as oes eee en ectedy2 ices Sb4AS 2 AO Meme eee 0.6 0.26 0.31 2520 pumbeiae else lelnjecteds2scesnob4 be Ee rem: 2.20 0.27 0.32 AEC) oN a0 ieee ae || 1.8 0.36 0.41 * A crude extract, Sb4, was precipitated with an excess of basic lead acetate, filtered and washed. The precipitate and filtrate were decomposed to remove the lead and concentrated under diminished pressure. Both solutions were then made up to the same volume so that 2.5 ec. were equivalent to 1.5 ce. of the original Sb4. Sb4B representsthe filtrate and Sb4A the lead acetate precipitate. This step in purification was frequently used in many of the other studies referred to below. Precipitation by picric and picrolonic acids. Many attempts were made to precipitate the activity by picric acid in various acidities and concentrations, but in every case the activity was recovered in the filtrates. Picrolonic acid also yielded inactive precipitates. It is of particular interest to note, however, that when histamine hydrochloride is added to the same gastrin solution one obtains an active precipitate which undoubtedly is the histamine picrolonate. In these experi- ments the precipitate obtained was in each case suspended in dilute hydrochloric acid and shaken with benzol or amyl alcohol to com- pletely remove the picric or picrolonic acid. The aqueous hydrochloric acid solution was then concentrated under diminished pressure and rediluted to the original volume. The filtrates from the precipitates were decomposed in the same way and also made up to the original volume after concentrating under diminished pressure. The results below are characteristic and were duplicated on other animals. CHEMICAL STUDIES ON GASTRIN BODIES Sit Dog III. Pawlow accessory stomach, January 22, 1916 HYDROCHLORIC ACID TIME JUICE REMARKS Free Total ce. per cent per cent fet nour... .. 0.7 0 : Trace Slime Injected 1 ce. SIA (an alcoholic extract of dried crude gastrin) 1st hour after.. Le 0.36 0.46 2nd hour after. . ail 0.46 0.50 ws Same dog on January 27, 1916 12:30-1:30 p.m.. 1.0 | 0 | 0.05 Mucin 1:30 p.m.......| Injected picrate precipitate from 1 ec. SIA 1:30-2:30 p.m.. ial 0 | Trace 26) 00) eee Injected picrate filtrate from 1 cc. SIA 2:30-3:00 p.m.. 5.4 0.24 0.26 3:00-3:15 p.m.. 1.8 0.43 0.45 3:15-3:30 p.m.. 0.8 0.46 0.49 3:30-4:00 p.m.. 0.7 0.44 0.47 If to 10 cc. of the same SIA solution we added 10 mgms. of histamine hydrochloride and then proceeded with the precipitation by picric acid in the same way and examined the precipitate and filtrate for ac- tivity, the following results were obtained on the same dog. HYDROCHLORIC ACID TIME JUICE REMARKS Free Total ce. per cent per cent Spe OUN -........<. 0.6 0 Trace Mucin Injected 1 cc. of the concentrated filtrate+1 ce. SIA 1st hour after... 7.8 0.38 0.41 2nd hour after.. 5.9 0.50 0.54 3rd hour after. . Aas 0.46 0.51 Injected 1 cc. from precipitate+1 cc. S1A plus 1 mgm. histamine hydrochloride 0.28 0.12 Oni 1.8 0.27 0.09 1st hour after... 2nd hour after.. The results indicate a very poor precipitation of the added histamine by the picric acid. §12 F. C. KOCH, A. B. LUCKHARDT AND R. W. KEETON With picrolonic acid we have as stated before, not been able to ob- tain an active picrolonate precipitate from various gastrin solutions, but when we add histamine to the same solution and then precipitate in the same way with picrolonic acid we not only obtain the histamine in the precipitate but also recover the gastrin activity otherwise, just as before, in the filtrate. In conducting this experiment we first showed that we could quantitatively recover 0.010 gram histamine hydrochlor- ide from 10 cc. water by precipitating with 50 ce. saturated aqueous picrolonic acid solution. This was shown chemically by the diazo test. Next the same amount of histamine hydrochloride was added to 10 ce. of SIA gastrin solution and precipitated in the same way by picrolonic acid. The precipitate and filtrate were decomposed in the usual way and after concentration made up to the original volume. The physio- logical response was as indicated in the protocol below. Dog III. Pawlow accessory stomach, February 9, 1916 HYDROCHLORIC ACID TIME JUICE REMARKS Free Total Ces per cent per cent 1ehours eee OFS 0.00 Trace J ,HOUTASSeee eee 0.2 0.00 Trace Injected of the picrolonate filtrate=1 ee. SIA 1st hour after... 5.8 0.37 0.43 2nd hour after. . 1.5 0.47 0.52 | Injected of the picrolonate precipitate=1 cc. S1A plus 1 mgm. histamine hydrochloride Ist hour after... 5.5 0.37 0.41 2nd hour after. . 0.3 Present On comparing the results here with those obtained with 1 ec. of the original SIA as given in the table for January 22, we find that the re- sponse is of the same order for the picrolonate filtrate as for the original SIA. This together with the inability to precipitate gastrin by picro- lonic acid shows that the added histamine was quite completely precip- itated by the precipitant. This was also verified by the diazo test applied on the precipitate and filtrate. These studies certainly indicate that gastrin and histamine are not identical chemically although they both act as good secretagogues. CHEMICAL STUDIES ON GASTRIN BODIES 513 Precipitation by mercuric chloride in acid solution. After the removal of considerable foreign material by basic lead acetate and removal of the lead from the filtrate, the acid filtrate when treated with a fair ex- cess of cold saturated aqueous mercuric chloride yields a precipitate possessing considerable activity. A second mercury precipitate could be obtained by adding an alcoholic saturated mercuric chloride solu- tion, but both this precipitate as well as the filtrate therefrom possessed very little or no activity. There appears to be a distinct loss of ac- tivity in this mercury treatment. The protocol below illustrates the character of the findings. Dog IIT. Pawlow accessory stomach, January 29, 1915 HYDROCHLORIC ACID TIME JUICE REMARKS Free Total cc. per cent per cent Letehour. «2. 0.1 0 0 Mucus 10:50 a.m....| Injected 3 cc. Sb5D3* ESO lam 0.1 Trace Mucus 11:50 a.m....| Injected 1.5 cc. Sb5D2* 12550 pam... .|': 0.1 Trace Mucus 12:50 p.m....| Injected 1.5 cc. Sb5D1* 50 p.m... . 3.5 0.38 0.41 735,199 ee 0.9 0.38 0.47 * The material started with Sb5B had been treated with basic lead acetate. The first mercury precipitate, Sb5D1, after decomposition and removal of the mercury was concentrated under diminished pressure and made up to the same volume as the original Sb5B started with. The second mercury precipitate, Sb5D2, was likewise freed from mercury and after concentration made up to the same volume as the original Sb5B started with. The filtrate from the second mercury precipitate, Sb5D3, similarly treated, was made up to twice the volume of the original Sb5B. It is to be noted that nine days previously an injection of 1 cc. Sb5B in the same dog caused a secretion of 4.0 and 1.5 cc. gastric juice during the first and second hours following the injection with acidities of 0.46 and 0.50 per cent re- spectively. In other words, a dose of two-thirds the size caused a better secre- tion than any of the fractions above. Similar experiments on other dogs and repetitions of the same experiment on other extracts confirmed these findings. Purine base precipitation. To 10 ec. of a solution which had been earried through the treatment with basic lead acetate and through the mercuric chloride precipitation in acid solution, we added sodium hydrate until the solution remained only very slightly acid. We now THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 514 F. C. KOCH, A. B. LUCKHARDT AND R. W. KEETON added 2 grams sodium acetate, 2 cc. 35 per cent sodium bisulphite solution, 4 cc. 10 per cent copper sulphate solution and boiled for 3 minutes, filtered at once and washed with boiling water. Both the precipitate and filtrate were freed from copper by hydrogen sulphide, concentrated under diminished pressure and made up to definite vol- ume. The precipitate contained no activity but the filtrate showed good secretagogue action. Other precipitants. In part of the work, as shown below, it was dis- covered that the gastrin activity can be extracted from an alkaline solu- tion by amy] alcohol and that it can be recovered therefrom by shaking with dilute hydrochloric acid in water. A solution was thus prepared from material which had been treated as above; that is, with lead ace- tate and mercuric chloride. The hydrochloride solution thus obtained was evaporated to dryness under diminished pressure, then redissolved in water and diluted to one-fourth the original volume. The original solution referred to was already a very active one. Of this solution 1 ec. portions were taken and treated on watch glasses as indicated below. The precipitates and filtrates were separated and freed from precipi- tating reagent as well as possible and were then tested for physiological activity and for color reaction by the diazo benzene sulphonic acid reagent. REAGENT AND AMOUNT IMMEDIATE RESULT RESULT AFTER 20 HOURS Saturated picrolonic Flocculent precipitate Same acid, 2 ce. Alcoholic zine chloride, | No change No change IL oe Alcoholic cadmium chlo- ride, 1 ce. Iodine in potassium 1o- dide, 3 cc. Potassium mercuric 10- dide, saturated, 5 drops 5 per cent platinie chlo- ride, 1 ce. 5 per cent gold chloride, il Ge Saturated silver sulphate and Ba(OH)s. until al- kaline No change Heavy brown to black precipitate Slight precipitate No change Heavy orange precipi- tate Precipitate white and brown Slight precipitate No change No change Reddish yellow precipi- tate No change Black precipitate CHEMICAL STUDIES ON GASTRIN BODIES 515 The results of the physiological and diazo tests are recorded below. PRECIPITATE FILTRATE REAGENT —_—_——_— Diazotest | Physiological Diazo Physiological Ficrolonic acid. ....../...... Negative | Negative Positive | Positive Cadmium chloride.........| Negative | Negative Positive | Positive IC anil 1S 5 eee eae en Negative | Negative Positive | Negative Platmie chloride... .:...2.<. Negative | Negative Positive | Positive (?) Gold chloride: ::. 2%. .).:..:.. Negative | Negative Positive | Positive AgSOx and Ba(OH):........ Positive Positive Positive | Positive The results again confirm the former findings as to picrolonic acid and also again suggest a similarity to imidazol derivatives in that the pre- cipitation by silver in an alkaline solution is positive and in that the diazo reaction usually parallels the activity. This precipitation by silver sulphate and barium hydrate was re- peated later on larger quantities of material. It was found that not only was the activity left to a large extent in the filtrate but that there seems to be a loss of activity by this treatment. STUDIES WITH IMMISCIBLE SOLVENTS An active solution obtained by the mercuric chloride precipitation process was rendered alkaline to litmus by sodium hydroxide and then shaken with various immiscible solvents. These solvents were later shaken with dilute aqueous hydrochloric acid and the latter concen- trated under diminished pressure, diluted to definite volume and tested physiologically. In every case a precipitate formed in the alkaline aqueous solution; this was filtered off and found physiologically in- active. The results on the immiscible solvents show that the gastrin is not soluble in carbon disulphide, chloroform, benzene, ether and ethyl acetate. The only immiscible solvent by means of which we were able to extract the activity from an alkaline solution is amyl alcohol. The amyl alcohol extraction was slow but clear-cut. We used Kahlbaum’s purest grade amyl alcohol for the first trials and found that careful controls thereon showed no activity. Later we used the same alcohol as such or after recovery by shaking with alkaline solutions, removing the water by anhydrous potassium carbonate and fractionating by dis- tillation. The discovery of the solubility in amyl alcohol was next applied on pure histamine solutions and there also we found that the free base can be removed from alkaline aqueous solutions by means of 516 F. C. KOCH, A. B. LUCKHARDT AND R. W. KEETON this solvent.!. This was shown by means of the diazo-benzene sulphonic acid reaction. An attempt was made to carry on a continuous Soxhlet extraction of a dry alkaline mixture of crude gastrin, but the amyl alcohol extract became very much discolored, gave indications of resin formation, and contained very little or no activity. Very likely the activity was lost in the prolonged extraction at the boiling temperature of amyl alcohol. The similar solubilities of gastrin and histamine in amy] alcohol are suggestive, and indicate that the substances are of similar character. They do not appear to be identical because, as shown above, the gastrin solution obtained from an amyl] alcohol extract does not yield a physio- logically active precipitate with picrolonic acid as does histamine. STABILITY STUDIES Action of alkaline solutions. To 10 ec. of the same preparation as used for the immiscible solvent studies we added 5 ec. 10 per cent sodium hydroxide in excess beyond the neutral point and then evaporated to dryness on the water bath. The material was next redissolved in water, neutralized by dilute hydrochloric acid and diluted to definite volume. This solution was compared with one treated in the same way only the evaporation was canducted with the solution in a neutral condition. The protocol below gives the physiological responses from the material and shows the destructive action of alkali on gastrin. Dog III. Pawlow accessory stomach, May 12, 1915 HYDROCHLORIC ACID TIME JUICE REMARKS Free Total ce. per cent per cent 9:34-10:34 a.m. 23} 0.12 (Oe 7/ Some mucin 10:34-11:34 a.m. 2.0 Trace 0.09 Clear 11:34 a.m.- 12:34 p.m. 1.0 Trace 0.05 12:34 p.m. Injected 1 ec. of the solution evaporated in neutral condition 12:34-1:34 p.m. 11.0 0.41 0.45 Clear 1:34-2:34 p.m. 1,524 0.36 0.39 Clear 2:34 p.m. Injected an equivalent amount evaporated in alkaline solu- tions* 2:34-3:34 p.m. 5.7 0.35 0.36 Clear 3:34-4:34 p.m. 0.6 0.24 0.27 Clear * Slight loss in injection, not enough to account for the difference in secretion. 1 This property was later applied in the quantitative separation of histamine and histidine as recently published by Koessler and Hanke. 66.566, GPAll. Journ. Biol.Chem., CHEMICAL STUDIES ON GASTRIN BODIES 517 Action of acids. 'Toasolution of gastrin obtained from the decompo- sition of mercuric chloride precipitate, there was added an equal volume of concentrated hydrochloric acid (sp. gr. 1.19) and this was boiled for twenty-four hours under a reflux condenser protected from oxygen by a pyrogallol seal. The acid was removed by evaporating to dryness under diminished pressure, redissolving the residue in water and again evaporating in the same way. The solution finally was diluted to the original volume. In the protocol below Sf is the original solution and SfA the acid treated material. Dog III. Pawlow accessory stomach, March 25, 1915 HYDROCHLORIC ACID TIME JUICE = Se a a REMARKS Free Total cc. per cent per cent 9:18-10:18 a.m. 1.9 Negative Bare trace Mucus 10:22a.m. Injected 1 cc. Sf 10:22-11:22 a.m. 7.4 0.29 0.32 11:22-11:52 a.m. 0.8 | 0.27 0.33 11:52 a.m. Injected 1 cc. SfA=1 cc. Sf 11:52 a.m.- 12:52 p.m. 9.2 0.36 0.40 12:52-1:22 p.m. 1.5 0.35 0.36 In another experiment a similar preparation was boiled for forty- nine hours, without protection from oxygen, with 20 per cent hydro- chloric acid. The physiological results confirm the findings given above. This remarkable stability in acid solution is further confirmed by our experience with gastrin solutions kept for months in a distinctly acid solution under a layer of toluol. COLOR REACTIONS OF GASTRIN SOLUTIONS It is to be noted that in the precipitation studies reported above one finds the diazo benzene sulphonic acid reaction positive in very nearly every case in which we found activity. This was found to be generally true in all our separations and, in fact, this color reaction was frequently used as a rapid test for detecting complete washing of pre- cipitates to be discarded. The intensity of this color reaction always was found to parallel intensity of the physiological activity. It again suggests a similarity in the chemical character of histamine and gastrin. Other color reactions were carried out on variously purified gastrin 518 F. C. KOCH, A. B. LUCKHARDT AND R. W. KEETON solutions with the conclusions given below. Negative reactions were obtained by ninhydrin, di-methyl amido benzaldehyde, glyoxylic acid and furfural condensation. The Millon reaction was very faintly pos- itive with a heavy precipitate. Knoop’s bromine reaction was posi- tive as with histidine, and the biuret test was positive but developed slowly on a solution which had not been hydrolyzed with hydrochloric acid. After hydrolyzing for 49 hours the activity remained, but the biuret reaction now was negative while the bromine and diazo reactions remained positive. The xantho-proteic acid reaction was very faintly positive on the mercury precipitated gastrin preparation. PHYSIOLOGICAL STUDIES WITH SUBSTANCES OF KNOWN COMPOSITION It was thought that by following the secretion of gastric juice after the injection of various substances of possible biological importance, but of known composition, we might be able to throw some light on the probable chemical character of the active substance in the gastric mu- cous membrane. The results of these studies are given in the table below. TABLE 1 PHYSIOLOGICAL RESULTS s NUMBER-OF |2 2) 23 ae SUBSTANCE TOTAL DOSE EXPERIMENTS Posi- New Doubt: tive tive ful mgm Triethylamine HCl), W505 < 2.,-5.- 10.0 3 2 1 Tetraethylamine HCl]............ 10.0 1 1 CVSUN. Jae coe mins ee cee Ee 20.0 4 1 2 1 inyptophane's +. 2). Ss.2ec ee se 20.0 4 4 @reatine tor (occ t nee eee Wed 4 4 Creatinine =..55:0.65- ao ee Ud 5 5 C@holineschloride@ se eee Lee 10.0-20.0 16 5 10 1 EGS GGING 25 ec eee Ee 10.0 2 2 Methylimidazol HCl............. 10.0 il 1 Hydroxymethylimidazol HCl..... 10.0 1 1 Guanidine HCW... 5. eee 10.0 5 4 1 Glucosammes HCI? 2. syne 20.0 4 4 URtAMING $5522 ch Seas eee ee 0.25-1.0 | Numerous | All 0 0 Apparently the only substances which have caused a secretion are choline and histamine. However the response from choline is so un- certain and so slight as compared with gastrin solutions and other tissue extracts that the physiological action of these extracts cannot be attrib- CHEMICAL STUDIES ON GASTRIN BODIES 519 uted to the presence of choline as was at one time claimed by various investigators (3). It is of interest to note that we never were able to cause a secretion of gastric juice by feeding gastrin by mouth in two- to threefold doses, nor have we ever been able to detect gastrin activity in dog or human gastric juice even after concentrating under diminished pressure and injecting intramuscularly. GENERAL DISCUSSION OF THE RESULTS Our studies thus far indicate that the active substance in gastrin solu- tions obtained from hog gastric mucous membrane is a very stable sub- stance in acid solutions, that it is not easily oxidized in acid or neutral solutions, but that it is destroyed appreciably by heating with strong bases in water solution. The active substance does not lose its activity by prolonged acid hydrolysis and thus appears to be a distinctly basic substance as shown by its precipitation from acid solution by mercurie chloride and by its behavior toward immiscible solvents in the alkaline and acid conditions. It is not precipitated in the purine base precipi- tation method. The constant diazo reaction also suggests a basic sub- stance and is particularly interesting in that pilocarpine is also an imi- _dazol derivative. All of these properties suggest a similarity to pan- creatic secretin as reported by others, with the exception that pancreatic secretin is generally considered to be more labile, especially in an atmos- phere of oxygen (4). The properties enumerated above, as well as the physiological action, further suggest the identity of gastrin and hista- mine, but thus far we have not been able to precipitate the activity of a gastrin solution by means of picric or picrolonic acid. The work of Abel and Kubota (5) also suggests that histamine is the active constitu- ent in our extracts, but as their work was based mostly on physiological evidence the proof is by no means conclusive. Koessler and Hanke? have in fact not been able to confirm the work of Abel and Kubota in that they did not find histamine by careful chemical search in casein nor in the fresh ox hypophysis. The results of Abel and Kubota (5), Koessler and Hanke (6), Tomaszewski (7), Fawcett, Hackett and Rahe (8), as well as our own results indicate the very wide distribution of substances, probably of basic character and similar in physiological action, in many tissues and possibly in the proteins themselves. It is very interesting indeed that we find these substances in relatively high concentrations in the same organs in which ammonia is generally most 2 Unpublished verbal communication. 520 F. C. KOCH, A. B. LUCKHARDT AND R. W. KEETON , prominent immediately after death. It may also be recalled that these substances do not act physiologically when given by mouth and that they appear to be absent from gastric juice. The substance or substances in question may later be found to be identical or very closely related, possibly peptamides of very closely related structures. CONCLUSIONS 1. Our results suggest a striking similarity between histamine and gastrin. 2. Gastrin appears to be more stable than secretin is reported to be, although they are similar in many respects. 3. Gastrin has not been precipitated by picric or picrolonie acid. 4. Gastrin appears to be a basic imidazol derivative which can be extracted as the free base by amy] alcohol. 5. Thus far histamine and pilocarpine are the only known tmidazol derivatives which stimulate the gastric mechanism to secretion. 6. Choline acts with uncertainty and then only faintly as compared with histamine or gastrin. The authors wish to express their thanks to Drs. Karl Koessler and Milton Hanke for the imidazol derivatives used in this work. BIBLIOGRAPHY (1) Kerron anp Kocu: This Journal, 1915, xxxvii, 481. LuckHarptT, Keeton, Kocn anp La Mer: This Journal, 1920, 1, 527. Kerrton, Kocu anp Luckuarpt: This Journal, 1920, li, 453. Kerrton, LucKHARDT AND Kocu: This Journal, 1920, li, 468. (2) BAYLIss AND STARLING: Journ. Physiol., 1902, xxviii, 325. BAYLISS AND STARLING: Journ. Physiol., 1903, xxix, 174. (3) von Férte anp Scuwarz: Arch. f. d. gesammt. Physiol.. 1908, exxiv, 427. (4) Drxon'‘anp Hamiuu: Journ. Physiol., 1909, xxxv, 314. (5) ABEL AND Kusota: Journ. Pharm. Exper. Therap., 1919, xiii, 243. (6) KorssLER AND HaAnKE: Journ. Biol. Chem., 1919, xxxix, 521. (7) TomaszEwsk1: Arch. f. d. gesammt. Physiol., 1918, clxx, 260. (8) Fawcett, Rocers, Rane AND BEEBE: This Journal, 1915, xxxvii, 455. Fawcett, Rane, Hackett AND Rogers: This Journal, 1915, xxxix, 154. Rocers, Rane, Fawcerr AND Hackett: This Journal, 1916, xxxix, 3495. ESSENTIALS IN MEASURING EPINEPHRIN OUTPUT WITH FURTHER OBSERVATIONS ON ITS RELATION TO THE RATE OF THE DENERVATED HEART G. N. STEWART anv J. M. ROGOFF From the H. K. Cushing Laboratory of Experimental Medicine, Western Reserve University Received for publication April 17, 1920 The measurement of the rate at which epinephrin is given off from the adrenals under any given conditions necessarily involves the meas- urement of a mass and a time, since that rate is the quantity of epi- nephrin liberated per unit of time. It seems absurd to set this down in so many words. And the absurdity would strike every reader if our subject were the rate of output of blood from the left ventricle, the rate of output of carbon dioxide from muscles, the rate of hydrolysis of starch by amylase or the rate of excretion of urea by the kidney. But curiously enough, in the case of epinephrin some writers seem to assume that the mathematical conditions which govern the measure- ment of every velocity can be circumvented. And quantitative con- clusions as to the influence of this or that factor upon the rate of epi- nephrin output have been confidently deduced from reactions which are not even specific qualitative reactions for epinephrin (such as the paradoxical dilatation of the pupil or acceleration of the denervated heart) and from observations which did not comprise a single measure- ment of the quantity of epinephrin given off per unitof time. Indirect qualitative methods may have value in corroborating the results of direct quantitative methods, but in case of conflict the presumption must be in favor of the latter. We have recently (1) had occasion to eall attention to the bearing of this generally accepted principle upon certain discrepancies in the literature in regard to the influence of asphyxia upon the rate of epinephrin output. The straightforward way of measuring the mass of epinephrin given off per unit of time is to collect blood from the adrenal veins for a known time and to assay its concentration in epinephrin. Various methods of obtaining the blood may be adopted depending on the kind of animal o21 522 G. N. STEWART AND J. M. ROGOFF and other conditions of the experiments. Thus, in good-sized dogs a cannula may be inserted into the lumbo-adrenal vein on one side extraperitoneally, and blood collected from one adrenal. In cats the most convenient method is to form a cava pocket. This is a method which has been employed by a number of investigators, including Hoskins and McClure! (2) who, in addition to the cava and the veins entering it which Biedl (3) tied off, ligated the abdominal aorta. We also prefer to tie the abdominal aorta and renal arteries, in order to prevent too much blood from stagnating in the hind end of the animal and the kidneys when the cava and renal veins have been ligated. The arteries are tied a little time before the veins, so that as much blood as possible may drain out of the occluded parts. There is no obvious reason why the exclusion of the hind legs and kidneys should affect the output or the nervous mechanism governing it. Indeed, the better blood pressure maintained after ligation of these vessels ought to favor rather than to hinder any reflex nervous effects elicited, e.g., by stimu- lation of the central end of sensory nerves or any excitatory effects upon nerve centers exerted by such conditions as asphyxia, since the conductivity of the reflex ares and the excitability of the nerve centers must be better maintained when the blood flow is increased. The ordinary spinal reflexes are well obtained in the parts whose circulation has not been interfered with, and indeed in the hind end for some time after ligation of the vessels (see, e.g., protocol of cat 436). When Doctor Cannon (4) attributes to the peculiar method adopted by us for collecting the blood our failure to corroborate his results, he does not explain how our method is unfavorable to the eliciting of the reaction. He attempts to show that the output of epinephrin meas- ured by us in adrenal vein blood, collected from the cava and assayed on rabbit intestine (and uterus) segments, is an artificial phenomenon. He says ‘“‘the effect of opening the abdominal cavity, clamping off the inferior cava and repeatedly manipulating the abdominal contents, ‘The epinephrin output determined by these observers in 5 dogs by assay of adrenal blood, collected in the manner described, on rabbit intestine segments agrees well with our results. If the 1: 1,000,000 adrenalin solution, in terms of which the output is expressed, is assumed to contain about 75 per cent of base (it was probably not assayed) the average output would be 0.00019 mgm. per kgm. per minute. Our average for 17 dogs was 0.00022 mgm. per kgm. per minute. The average weight of our dogs was exactly half that of the dogs employed by Hoskins and McClure, and there is reason to believe that in the larger dogs the output of epinephrin per kilogram of body weight is somewhat less than in the smaller. MEASUREMENT OF EPINEPHRIN OUTPUT 523 either in pressing blood out of the inferior cava or withdrawing it by syringe, must be examined.”” We do not press blood out of the inferior cava nor do we withdraw it by syringe, if that is a more serious insult to the animal, nor do we repeatedly manipulate the abdominal contents. We open the abdominal cavity, prepare the cava pocket and clamp it off when blood is to be collected from the cannula inserted in the lower end of the pocket. We have examined the influence of several of the factors mentioned upon the output and have not found any effect detectable by our methods (5), (6). The alleged influence of opening the abdomen upon epinephrin secre- tion. Doctor Cannon does not state the result of his examination. All he says is that ‘fully twenty years ago Bayliss and Starling called attention to the profound effect which opening the abdominal cavity has on the intestines in causing them to become absolutely motionless. These well-established facts make an interesting commentary on the use of the cava pocket as a mode of obtaining evidence of normal or natural secretion.’”’ He proceeds, ‘‘There is no doubt that secretion from the adrenal medulla is subject to impulses delivered by the splanchnic nerves and there is no doubt that opening the abdominal cavity under anesthesia results in a discharge of impulses along these nerves. The adrenal glands, therefore, are continuously and ab- normally stimulated if the abdomen is opened.” If Bayliss and Starling had found that opening the abdomen caused the setting up of inhibitory impulses for the intestines, this would, of course, form a complication in the study of the intestinal movements under these conditions. De- spite this, Cannon accepts their results as, so far as we are aware, everybody else does. But because he imagines that opening the abdomen sets up inhibitory impulses for the intestinal movements which, however, did not prevent Bayliss and Starling from making trustworthy observations, he concludes that our results on a totally different object, the adrenals, are completely vitiated by impulses set up in an altogether different group of nerve fibers, the epinephrin- secreting fibers, although neither he nor anybody else has ever shown that opening the abdomen influences them in the least. The only com- mentary we need make upon this is that Bayliss and Starling (7) do not call attention at all to the effect of opening the abdomen in causing the intestines to become absolutely motionless. On the contrary, they state that when they opened the abdomen in the warm-saline bath they found the intestines collapsed and absolutely motionless if the splanchnic nerves were intact. They drew the conclusion, not that 5DA G. N. STEWART AND J. M. ROGOFF the opening of the abdomen caused the inhibitory impulses to the in- testines to be set up but that these impulses were already descending the splanchnies before the abdomen was opened. Singularly enough, Cannon quotes their statement verbatim: ‘‘These facts suggest that in the intact animal, at any rate under the conditions of our experiment, tonic or reflex influences are continuously descending the splanchnic nerves, inhibiting the activity of the intestines.”” To be logical, then, he ought to conclude that since the impulses which pass tothe intestines along the splanchnics are not called into existence by opening of the abdomen but are already present, impulses descending the splanchniecs to the adrenals and sustaining a normal epinephrin output must also be present in the intact animal. If opening the abdomen so com- pletely deranges the action of abdominal viscera, including the adrenals, a very large part of our supposed physiological knowledge must be wiped out. The truth is, of course, that we must neither assume in any particular case that opening of the abdomen is indifferent to the experiment nor that it absolutely contraindicates it, but must always test, as far as possible, whether and in what direction this operation, as well as the other experimental conditions, influence the result. We have pointed out elsewhere (22) that the relatively narrow range within which the output, (as measured by observers, including our- selves, who have employed methods correct in principle), varies with different anesthetics and different operations suggests strongly that what we term the normal or ordinary spontaneous output is not ini- tiated or sustained by the trauma or the anesthesia. To emphasize his criticism of our method of obtaining adrenal vein blood by opening the abdomen, Cannon states that the acceleration of the denervated heart, caused by stimulation of the central end of the sciatic, is hardly ever observed after opening the abdomen, although asphyxia still gives the reaction. He apparently considers this remark so important that he italicises the following sentence, ‘‘in the entire series of cases with opened abdomen there was only one in which sensory stimulation caused any effect ascribable to adrenal secretion.” Since we have shown (8) that this heart reaction has no significance as an indication of augmented epinephrin output, the statement that acceleration on sciatic stimulation is rarely obtained after the abdomen has been opened would, even if true, have no bearing upon the question at issue. It is easy to demonstrate, however, that the reaction is readily obtained not only soon after the abdomen has been opened but long thereafter, and when the abdominal aorta, renal vessels and cava or ~”) on MEASUREMENT OF EPINEPHRIN OUTPUT have been tied or clipped off in making the cava pocket. Indeed, there was some indication that the reaction might be greater after clipping the abdominal aorta, presumably owing to the better blood flow in the nervous centers (see, e.g., protocol of cat 437 in the previous paper (8). There is plenty of evidence in the protocols of that paper that opening of the abdomen does not interfere with the reaction in question, and this is further illustrated here in figures 1 to 4. Fig. 1. Parts of blood pressure tracing from cat 175. A, before and B, a por- tion commencing 24 seconds after beginning sciatic stimulation; (abdomen opened, left splanchnic cut nearly 2 hours). In all figures line of zero pressure corresponds with time trace; time in seconds; numbers above time trace represent heart rate per minute. The protocols of the experiments from which these tracingsare taken have already been published (8). In figure 1 are reproduced two por- tions of the blood pressure tracing (from cat 175), one just before, A, and the other during stimulation of the sciatic, B. The pulse rate was increased by nearly 30 beats per minute. The abdomen had not only been opened almost 2 hours before, and the left splanchnic nerve cut, but the intestines had been purposely manipulated. The heart re- action had been obtained many times after opening the abdomen, 526 G. N. STEWART AND J. M. ROGOFF both by excitation of the sciatic and by excitation of the left splanchnic before this tracing was taken, but an excellent rise of pressure was still caused when either of these nerves was stimulated. Stimulation of the sciatic before the abdomen was opened and with both splanchnics intact gave an acceleration of 22 beats in one observation and 25 beats in another. Figure 2 from another cat (177) shows an acceleration of almost 40 beats produced by sciatic stimulation after opening the abdomen and Ly VA AAA Fig. 2. Parts of blood pressure tracing from cat 177. A, before and B, a por- tion commencing 32 seconds after beginning of sciatic stimulation (abdomen open). Reduced one-nineteenth. section of one splanchnic. Before the abdomen was opened, stimulation of the sciatic caused a maximum acceleration of 48 beats, with both splanchnies intact. In figure 3 are reproduced portions of a curve (from cat 179) dem- onstrating a maximum acceleration of 34 beats, produced by sciatic stimulation before opening of the abdomen. One splanchnic had been previously cut extraperitoneally. After the abdomen was opened, a number of sciatic and splanchnic stimulations having been made in the meantime, stimulation of the sciatic caused an acceleration of 24 MEASUREMENT OF EPINEPHRIN OUTPUT 527 VA a\ ve ANNA ws dy oy Fig. 3. Parts of blood pressure tracing from cat 179. A, before and B, a por- tion commencing 32 seconds after beginning of sciatic stimulation (left splanchnic cut). A ny VAIN Cea ee at iy Fig. 4. Parts of blood pressure tracing from cat 179. A, before and B, a por- tion commencing 25 seconds after beginning of sciatic stimulation (left splanchnic cut and abdomen opened). C, before and D, a portion commencing 23 seconds after beginning of sciatic stimulation (both splanchnies cut). 528 G. N. STEWART AND J. M. ROGOFF beats (fig. 4, A and B). In this, as in all the other experiments, so long as sciatic stimulation gave a good rise of blood pressure, it gave a sub- stantial acceleration of the heart. Later on in the experiment the second splanchnic was cut. The blood pressure fell to 44 mm. of mer- cury, and stimulation of the sciatic now caused little or no acceleration of the heart and only an insignificant rise of blood pressure (fig. 4, C and D). If Cannon believes that the presence of the reaction indicates in- creased epinephrin output, it is incumbent upon him to explain why in adrenal blood collected at a time and under conditions when this reac- tion is well obtained, we are unable to detect, by a direct and sensitive method of assay (intestine segments), any sensible increase in the rate aa ¥ Wnin, yw VW . yy Winn, nie 5 ww Wynne Vay Fig. 5. Parts of blood pressure tracing from cat 436. A, before asphyxia for 60 seconds; B, 7 seconds before end and 16 seconds after end of asphyxia; C, a portion commencing 40 seconds after end of asphyxia (abdominal aorta and renal vessels tied). Reduced to four-fifths. of output. In the case of asphyxia it is still more necessary that he should explain why we cannot detect such great increases in the epi- nephrin output as he assumes to occur. For he admits that the acceler- ation of the denervated heart associated with asphyxia is not abolished by opening the abdomen. We can confirm this entirely, and can add that the reaction can also be obtained after ligation of the abdominal aorta, renal arteries and veins and other vessels, as done in the for- mation of a cava pocket for the collection of adrenal vein blood. This is illustrated in figure 5 (from cat 436). Condensed protocol. Cat 436; male; weight 2.26 kgm. Left superior cervical ganglion excised 28 days previously. Under urethane (6 grams) cut vago-sympathetics, excised stellate ganglia, prepared central end of left sciatic for stimulation. The anesthesia was very deep and considerable depression of respiration and circulation was present from the start. MEASUREMENT OF EPINEPHRIN OUTPUT 529 Rate Pressure 12:42 p.m. Before sciatic stimulation....... 5 an eaasiaithdpssoGeyt on tye OO 56 6 seconds after beginning o inteee Sars verchleyie ne LOO 62 20 seconds after beginning evimulenern? Srice skthacteeceoe a. aU 64 40 seconds after beginning stimulation............... 204 64 12:52) p.m. Before asphyxia for 45 seconds..... ... 4... ..25 5.62.20. 135 26 Just lafteribeginning asphyxia... «4% «ease Sianasioewiee 140 30 10 seconds after beginning asphyxia.................. 162 46 26 seconds after beginning asphyxia.................. 195 45 1:05 p.m. Opened abdomen, tied abdominal aorta, renal arteries and veins 1310 pim. Before asphyxia for 45 seconds. ; 2. .f22 220.6. 62. 0.2 ee) 184 49 Just, atter Decinning ASPDYRIB Leys) csaccras qyeeie 4 deers nterave 180 54 15 seconds after beginning asphyxia.................. 181 62 USP oaeLeE CNG. OF ASPAY X18... «5 ua ueus = susigacs «oe cee 198 60 1:15 p.m. Intravenous injection of 50 cc. Ringer’s solution Betore Ringer injection 1.0. oso. cet seeks ches abe 200 53 Miner shingersimjection.\:) 28621. O4f 4.RO.. bE Sa Se 204 80 iso p.m. ..Before.asphyxia for'60 seconds! 00°... ois.ce.m. wel Jem ads 194 66 dust after beginning asphyxia.:. 2. 2. <..40cccaemeent a 199 78 30 seconds after beginning asphyxia................4. 178 56 MUSscearver eng Of ASPMYXIRs cs. ce ccc wecis cee Se eee 62 Z20iseconas- after ‘end ‘of ‘asphyxia’..iA02.0/ 5 ib ates ane ee 53 40 seconds after end of asphyxia....................+- 220 54 1:40 p.m. Tied coeliac axis and superior Eeeebaienia see clipped cava; hind leg and tail reflexes still present 142 p-m. Before asphyxia for 60 seconds... ... :..-..0.0.eescdeene 200 63 During first 30 seconds of asphyxia.................. 201 64 Just. atter end of asphyxia. qs.) yaetiho! oe ee sate es asete ane 204 65 i5seconda after end; of:asphy xian. eine dss Seed sok 218 60 30 seconds after end of asphyxia..:.......-2::------ 220 59 1:48 p.m. Completed cava pocket iam. “before closure of pocket. .. . .. 5. 0. «mss ao sans & eek 208 56 Murine closure’ of POCKEUs. <27-<754 a ce ceases ies 190 42 Just after release of pocket blood.................... 189 44 15 seconds after release of pocket blood.............. 215 58 In any case how does Cannon conceive asphyxia to act upon the epinephrin output? He says a short asphyxia acts through the nervous system, a longer period of asphyxia has a direct action. Now, why should the stimulating action of asphyxia on the part of the central nervous system which presides over the adrenal epinephrin secretion be abolished by tying the abdominal aorta? The effect of its ligation is to raise the blood pressure and therefore to keep up a more efficient circulation in the central nervous system and adrenals. How does this prevent the stimulation of the nerve centers by asphyxia? It does THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 530 G. N. STEWART AND J. M. ROGOFF not do so, and the reason we do not find an increased output of epi- nephrin with asphyxia must be that when tested by a quantitative method, the alleged increase is non-existent or too small to be detected. The relation of the epinephrin output of the adrenals to the acceleration of the denervated heart associated with asphyxia. Cannon states that tying the lumbo-adrenal veins on both sides of the adrenals may not diminish in the least the acceleration of the heart produced by asphyxia of a given duration. This seems very good evidence that the reaction cannot be a quantitative reaction for epinephrin output. It is fully in accord with this conclusion that Cannon himself sometimes obtained a very large acceleration by asphyxia after section of both splanchnics. He endeavors to explain this by assuming that prolonged asphyxia stim- ulates the adrenal medullary cells directly. He brings forward no real evidence that this occurs. v. Anrep (9) states that after section of both splanchnics (in the dog) no reaction is elicited by asphyxia on the blood vessels of the denervated limb, and concludes that asphyxia has no direct action upon the adrenal medulla. Pearlman and Vincent (10), however, have not been able to obtain on the denervated limb any reaction in asphyxia which could be attributed to epinephrin. If asphyxia lasting for 90 seconds can stimulate the adrenal medulla directly so as to cause enough epinephrin to be given off to accelerate the heart by 68 beats a minute after section of the splanchnics, our failure to obtain evidence of this effect in adrenal vein blood directly assayed would be more puzzling than ever. For Cannon himself admits that the asphyxial acceleration with intact splanchnics is not interfered with by opening the abdomen, and how then should opening the abdomen abolish an action of asphyxia exerted directly upon the adrenal cells? Cannon’s conclusion, that asphyxia stimulates directly the adrenal medulla, is simply a misinterpretation of his result that sometimes after division of both splanchnics asphyxia may cause a quickening of the denervated heart. Even if the heart reaction were specific for epinephrin, this would only indicate that some portion of the nerve supply of the adrenals might have been spared, and would not of itself prove a direct action. But as soon as it is shown that the reaction is not specific for epinephrin, and canbe obtained after removal of the adrenals, the whole argument is seen to be baseless. Cannon cites Czubalski (11) as having ‘‘adduced evidence that asphyxia if sufficiently prolonged may have a direct stimulating action on the adrenal me- dulla.”” The paper referred to is a preliminary note in which it is stated that even after section of the cord and bulb and also after division of the splanchnies MEASUREMENT OF EPINEPHRIN OUTPUT 531 and vagi, a rise of blood pressure can be produced beginning usually at the end of the third or beginning of the fourth minute of asphyxia. The rise of blood pressure was very often accompanied by marked slowing of the heart. Czu- balski attributes the increase of pressure to adrenalin liberated by the direct action of asphyxia on the adrenal medulla. The heart, of course, was isolated as regards its extrinsic innervation, and, according to Cannon, the characteristic response of the heart thus isolated to increased epinephrin output is an accelera- tion. It is curious that he should find confirmation of his result on the direct action of asphyxia on the adrenals in an effect upon the heart precisely the oppo- site of that which he associates with epinephrin. Czuhalski’s assertion that the general blood of a dog after asphyxia contains so much epinephrin that its pres- ence is demonstrated by marked inhibition of a rabbit’s intestine segment, when the defibrinated blood is made to displace ‘‘a nutritive solution,” is of itself suffi- cient to show that his conclusions are not of any value. His statement that after removal of the adrenals asphyxia no longer causes a rise of blood pressure, is invalidated by the observations of Gley and Quinquaud (13). In one of Cannon’s experiments he obtained with an asphyxia lasting 90 sec- onds, which is what he considers a prolonged asphyxia, an acceleration of 28 beats before section of the splanchnics. After section of the splanchnies no ac- celeration was produced by an asphyxia of 60 seconds, but he does not state what acceleration, if any, was caused by an asphyxial period of the same duration as the control before splanchnotomy, so that there is no way of judging whether the section of the splanchnics in this case had anything to do with the negative result. If the asphyxial rise of pressure is a factor in the acceleration, it is easy to see that section of the splanchnics or injury to them in removal or ligation of the adrenals must interfere with the reaction, apart altogether from the interference with the epinephrin output. In the figure given by Cannon (fig. 2) there seems to have been no rise of pressure after removal of the adrenals till 45 seconds from the beginning of asphyxia, whereas before removal of the adrenals the curve shows a substantial rise of pressure before this, how much earlier it is impossible to say from the section of the curve reproduced. It is unnecessary, however, to go into details of this kind. The best proof that Cannon’s supposed demonstration of the augmenting effect of asphyxia upon the epinephrin output by means of the heart reaction is fallacious is that asphyxia can elicit marked acceleration in the ab- sence of the adrenals. Experiments showing that asphyxia can cause acceleration of the de- nervated heart in the absence of epinephrin from the adrenals. Several instances have been given in the previous paper (8). Thus in cat 438 the heart was beating at the rate of 175 a minute. The second adrenal had been removed 35 minutes previously by the abdominal route. Asphyxia was induced for 60 seconds and the heart rate rose to 225 beats a minute, the maximum acceleration being attained after resump- tion of respiration. The blood pressure was 64 mm. of mercury and pa V4 G. N. STEWART AND J. M. ROGOFF did not change much. Sciatic stimulation at this time caused little if any acceleration of the heart. although earlier in the experiment and after removal of the second adrenal it gave a good acceleration. In cat 440 after removal of the adrenals and exposure of the cervical cord, which was associated with a fall of blood pressure of 20 to 30 mm. of mercury, asphyxia for 45 seconds caused an acceleration of 13 beats during the asphyxial period and of 18 beats counting from the end of asphyxia, the blood pressure rising only slightly (from 84 to 88 mm. of mercury). Stimulation of the sciatic, which had caused good accelera- ations (as much as 42 beats a minute) after excision of the second adrenal, had at this time only a very slight effect upon the heart rate. With a subsequent asphyxia of 30 seconds, the heart rate diminished from 216 a minute before to 203 during the asphyxia, but recovered to the initial rate on stopping the asphyxia. In a number of experiments the preliminary dissection for excluding the adrenals from the circulation was made by the abdominal route at the beginning of the experiment, and ligatures were placed but not tied. The abdomen was then closed. After one or more periods of asphyxia with the adrenals discharging, the ligatures were tied so as to occlude the adrenal arteries, the adrenal veins and the lumbar veins just before they cross the glands. The abdomen was closed and the effect of asphyxia on the heart rate was again observed. Finally the adrenals were excised by cutting between them and the ligatures. There was no bleeding and no further ligatures were required, showing that the glands had been completely excluded from the circulation. The pro- tocol of cat 445 is given as an example. Protocol. Cat 455; female; weight, 2.27 kgm. Under ether cut vago-sym- pathetics, excised stellate ganglia, opened abdomen and placed ligatures in posi- tion to occlude adrenal vessels but did not tie them. Rate Pressure 12-52 p.m | Belorerasphy.xia, (GORSeconds) ease cree reer 284 130 Just after end of asphyxia. 5).2:. -.as. seo ese eee 118 15, seconds after endiiofiasphyxiaes. +24. ose eee 127 30 seconds after end of asphy-xai... sae ere eee 304 126 12:55 p.m. Before asphyxia (45,seconds): - 5.--+-45--heeeeee ees 125 During first.15 seconds asphyxia. .. 3... 20> eee eee 130 During next 15 seconds asphyxia. =. °....s.. oo eee 114 During next 15 seconds asphyxia. .!)...5.: .¢ 522 Sosa eee 100 Just after iendiot asphysdaner eee lode dene ice eee 312 120 20 seconds) after end of ‘asphysda,. sos eeecie oe eee 323 132 30 seconds after end) of asphyxia eee cere cise 316 138 1:05 p.m. 12pm: 1:14 p.m. 1:18 p.m. 1:23 p.m. 1:30 p.m. 1:40 p.m. 2:06 p.m. MEASUREMENT OF EPINEPHRIN OUTPUT Before asphyxia’ (G0! seconds): /4.4s42-6s hee eee ae During first 15 seconds asphyxia sicacees A) 2550 Sales During next 15 seconds asphyxia... . Just atter.end: of asphy xiaiise44.nbaesns dol eee eee 15 seconds after end of asphyxia. .................. 30 seconds after end of asphyxias).. 2... 0... 2sk ee scinoees 50 seconds after end of asphyxia. ...................0. Tied off adrenal vessels Before asphyxia (40 seconds). ..............0eceeceees During first 12 secondsvasphyxias ..-...650:00%0 00 +0 sess justrattercend of asphyxia. 2). $1). ROLLER Sa ee. l5vseconds- after ‘end, of asphyx1a..). a2 scree. --oee alisecondsratter endyof asphyxia... .2..¢%.4-. cass adter as 45 seconds after end. of asphyxia... ...... 002... 42-002 -0: DELO asphyxia (GU SECOHGS) codmia.cs «0 sa.s ns oe Sees eee During last 15 seconds asphyxia... . Justrativerrendvot asphyxia. asees ca eee ances eek Before aspliyxia(60,séconds)s: {a..js2 2% YA. heeees During. first, 15. secondsiasphyxia...< 2 2.» $a. sa- seeder wae During next 15 seconds asphyxia. . During next 15 seconds asphyxia..... Duringinext | 5'seconds aspltyxiay. 2.58 se ee ee eee ee Just-atter end of aspiy xia. ..% ace ios ae. Bae ee 15 seconds after end of asphyxia. ..........-.00.0scens 30 seconds after end of asphyxia. ............4........ Excised both adrenals; prepared central end of sciatic Before wsphyxal(s5!Seconds))\. eae eee nee ee ee oe During first 20 seconds asphyxia. ........ just, ater end: ofasphyxitad |.)2P.0 SI Nel (5iseconds after end of asphyxia. 3. 22 oe+se. ses ee ee B0lseconds) afbersendsot asphyxian seco ec e eee ee Ad seconds atter/end ofasphyxiay) see ge: sooo ote Beforeisciatic stimulations (Gems )a. a ees seeae cee eee Justiatterjendtorctimulationet.ces..- see eee eee 15 seconds after end of stimulation................... before asphyxia’ (60 seconds) 1:3 2 EVOL eee Dirine asphyxia’ Mews Je LA) 2 A Pa Se justatieriend ofasphyxia ices eo see eee emcee 20nseconds.-aftereng, Of aSpHNIS .9.22 22. da oo eee cp ee SOESECONGS ALLER ENC aOisasplivexdan co c1n cesar Cele erae es Before asphyxia (45,seconds) 20. 292) 0 thos eee During first:15 seconds asphyxia. V/.0.. kee Duringnext 15 seconds asphyxais: 6 c.ci te catiie. 4-4 JUSsianterenduotasp hi Xia). 4. )eete. oe eee eects lS ysecondsiartersend) of asphy- bile serie = 50 secondsiaiter endof asphyxia’)... 2.2. eee ne nes 533 Rate Pressure 308 102 534 G. N. STEWART AND J. M. ROGOFF Rate Pressure 2:14 p.m. Before sciatic stimulation...............--.---...-.. 238 49 During first 15 seconds stidation bones LAE Bae 239 63 During next 10 seconds stimulation.................. 242 64 During next 10 seconds stimulation.................. 246 65 Just after end .of stimulation.) 20, 202. 2/124 ese eee 242 64 2:32 p.m. Before asphyxia (45 seconds)..............--2+.0+0.5 24 62 During first 25 seconds asphyxia. ...................-- 245 54 Just after end of asphyxia. 2 G5 oe sees eee ee eee 242 45 20 seconds: after end of asphyxiaete se. ose see eee 240 49 It will be seen that ligation of the adrenals did not interfere with the acceleration of the denervated heart induced by asphyxia. The first asphyxia after ligation of the glands gave an acceleration of 38 beats, the second an acceleration of 43 beats per minute. The third asphyxia gave an acceleration of 16 beats, but the initial rate was already 300, which was 17 beats greater than before either of the two previous asphyxial periods. After actual removal of the adrenals in this experi- ment the maximum acceleration caused by asphyxia was slight (only 9 beats per minute). This, however, can have nothing to do with loss of epinephrin output since the glands had been entirely excluded from the circulation by the ligation. We have had abundant evidence, if evidence were needed, that repeated as- phyxiation is not an indifferent procedure for the heart. A heart which at first has responded by an acceleration during the asphyxial period, and usually a still greater acceleration immediately thereafter, may later on respond by some dimi- nution in the pulse rate, becoming more manifest toward the end of the asphyxia, and succeeded by an increase on stopping the asphyxia, which may carry the rate beyond the initial value or not. In figure 2 of Cannon’s paper (4) the heart rate toward the end of asphyxia (after the adrenals were tied off) was slower by 6 or7 beatsaminutethan before theasphyxia. It then increased tothe initial rate after respiration was resumed. As only a small portion of the curve is repro- duced, it does not show whether later on the rate went beyond the initial value or not. When successive periods of asphyxia are superimposed upon a steadily de- clining blood pressure in the course of an experiment the loss or diminution of the heart reaction is apt to be particularly evident. A good acceleration was still given when the sciatic was stimulated after excision of the adrenals, accompanied by a fair rise of pressure. It is scarcely necessary to point out that it isimpossible toreconcile these results with Cannon’s conclusion that the acceleration of the dener- vated heart is an index of the increased output of epinephrin from the adrenals caused by stimulation of the sciatic and by asphyxia. In cat 455 the heart rate after section of the vago-sympathetics and ex- cision of the stellate ganglia was exceptionally great. — MEASUREMENT OF EPINEPHRIN OUTPUT Bao In the next experiment to be cited the heart rate was unusually low, not much more than half the rate in cat 455 and the blood pressure was also only about half. Protocol. Cat 453; male; weight, 2.51 kgm. Under urethane (5 grams) cut vago-sympathetics, excised stellate ganglia, opened abdomen and placed liga- tures in position to occlude adrenal vessels but did not tie them. Rate Pressure 1b35)a.m: Before asphyxia (45 seconds)... ........6....0s cece 150 66 During first 20 seconds asphyxia. ................2.005 152 66 During next 20 seconds asphyxia. . 22.4.5. 0.2.00008.. 152 72 lnemaier cud of asphyxia. 3 soos ace eee ve wore e's 153 76 20 seconds after end of asphyxia. ..................00- 155 68 35 seconds after end of asphyxia. ................2.05. 156 71 ia ) erore aspliyxia, (60 seconds)... 2 .s.6.20 153 63 During first 15 seconds asphyxia. ..... wc a IE 6 151 63 During next'20 seconds asphyxia... :.....52-0052--.- LDL 69 During next 15 seconds asphyxia..................... 157 85 NMsinarier endvoet asphyxiae.¢ i. acco ak Lee LL 166 (/ 25 seconds after end of asphyxia. ..................22. 166 80 11:45 a.m. Tied off adrenal vessels Bea a i, setore asphyxia (60 seconds)... .......0. 5. .- 4). 0544 -cs eee 40 seconds after end of asphyxia. .... bs Cee Oe Prepared central end of sciatie for Sramilntion Before sciatic stimulation (6 cm.)..................... 240 During, seiatic stimulation... 202).42.0.5. . ss 5. oe Just after end of stimulation. i . 251 Before asphyxia (through 3 ft. abe) “(60 abeoada\es . 239 Durmp hrst 20isecondsjasphyxiap 4. eae eee eee oS During next 20"seconds asphysqa-)-4 oe eee eee Oe During next Z0lseconds asphyxiaa- 44s see eee es Just after end of asphyxia............. 241 20' seconds after end! of asphyxia. -225-- -5-- eee eee 45 seconds after end of asphyxia: ..............c..+s0- 247 Before asphyxia (105 seconds) interrupted by 5 seconds respiration each half minute: -.- 42. 5 soccer eee During first 20!secondsiasphyxdae -.55 94. e eee ee During next 20’ seconds asphyxia. :c. .. 2.222. During next 20/seconds asphy-xias 6) 4 94552 see eee During next 15 seconds asphyxiaio-.. 5 42) eee een eee During next 20 seconds ee s gia. daicln o bn oe eee Just after end of asphyxia. . : 246 Before sciatic stimulation (6 cm. a Pree es Ser o. 50; 242 During: first 15 seconds!stimulation. «24.45. eee 247 During next 20 seconds stimulation. ... 252 Just after end of stimulation.i.« 3. -64- see 2 22 115 12-14 pam.) Betore, asphyxia (30 seconds) ......... 22... 35. 225. e eens © 218 110 During first 15 seconds asphyxia. ...............-..... 223 118 During next 15 seconds asphyxia... (4:.4/22..22-00+e. 220 127 Jusizaiter end. of) asphyxia. 50.4 or<.3'< 20pas se repay eta PEeO 110 20 seconds after end of asphyxia... 05. .4 eee oes 220 120 35 seconds after end of asphyxia. . e223 120 12:18 to 12:35 p.m. Three more observations iit ne xia (60 ts 90 seconds) gave a small drop in the rate during the asphyxia; the blood pres- sure had fallen to 84 at the end of the experiment. 542 G. N. STEWART AND J. M. ROGOFF We do not think anybody who has studied our previous results will attribute the failure of the reaction in this experiment, after adrenal- ectomy, to elimination of the epinephrin output of the adrenals. But if by any chance a reader chooses to conclude, as apparently Doctor Cannon would do, that the drop of 36 beats a minute in the rate after removal of the glands is due to the absence of the epinephrin previously being given off from the adrenals (average 0.0002 mgm. per kgm. per minute in etherized cats) he will have difficulty in seeing why an in- crease in the output to many times this amount, with asphyxia, should be necessary to cause an acceleration of 42 beats a minute before the adrenalectomy, and why it should be impossible that a redistribution of the epinephrin without any increase in the rate of output might be responsible for the acceleration, even if asphyxia produces no change in the susceptibility of the heart to the action of epinephrin. The only possible conclusion, however, from our experiments is that whether the epinephrin takes any share in the acceleration associated with asphyxia or not, the reaction cannot be due solely to the adrenal epinephrin, and there is no evidence that epinephrin is concerned in it at all. As has been pointed out in the previous paper (8), our proof that the increase in the heart rate caused by stimulation of the sciatic can- not be an index of an increased epinephrin output, does not entail the obligation to explain the acceleration, which is probably a complicated reaction. Still less do we feel bound to explain the asphyxial accelera- tion, which is possibly more complicated still. It is for Doctor Can- non, who brings forward the reaction as a quantitative test for changes in the epinephrin output, to exclude the other possible factors. We have suggested, however, (12) one possibility which must be controlled in so far as epinephrin may be a factor at all, namely, that asphyxia might render the heart more sensitive to such amounts of epinephrin as were being given off before the asphyxia was induced. Since we have shown that asphyxia in the absence of the adrenals can cause accelera- tion, it seems obvious that the accelerating action of a given dose of epinephrin might be reinforced by asphyxia. In that paper (12) we stated that it is not permissible to make quantitative comparisons on the effect of such a condition as asphyxia, using a test object like the heart well supplied with oxygen in one observation but asphyxiated in the comparison observation, without controlling any possible effect pro- duced by the asphyxia upon the reactivity of the test object itself. There was every reason to point this out, as this factor is habitually MEASUREMENT OF EPINEPHRIN OUTPUT 543 neglected by certain writers, who have worked with test objects in situ, although it must be taken account of before their results can be used at all. Cannon now endeavors to control the possibility that asphyxia might alter the reactivity of the heart to one and the same dose of epinephrin, so that a reaction might be obtained simulating an increased output when the output was really unchanged. Although the question loses much of its interest, as a question in the technique of estimating epinephrin output, when it is known that the heart reaction relied upon by Cannon to demonstrate increased output is readily obtained in the absence of the adrenals, it may nevertheless. be noted how ‘Cannon takes account of the factor in question. He reproduces a curve taken while adrenalin, to the amount of 0.08 mgm. per minute, was being continuously injected into the veins ofa cat. This is, for an average cat under ether, 200 times the mean rate of output of epinephrin under the conditions of our experiments, which according to Cannon may artificially increase the output to an “‘unsur- passable limit.’”? The concentration of epinephrin in the blood passing through the coronary circulation must, therefore, have been much greater than could probably ever be sustained by the output from the adrenals, even if, accord- ing to Cannon, this were greatly increased by excitation of sensory nerves or asphyxia. The increase in the heart rate, however, was only 40 beats (from 132 to 172 per minute), an acceleration of the same order of magnitude as may be given by sciatic stimulation or asphyxia alone, without any artificial injection of adrenalin. Now, since the amount of adrenalin injected is, according to Cannon’s own data, far greater than that liberated from the adrenals by asphyxia, it must be assumed that the acceleration was maximal before asphyxia was allowed to act. How can a possible change in the reactivity of the heart to epinephrin under the influence of asphyxia be demonstrated if the heart is already making its maximum effort in response to an enormous dose of injected adrenalin? The slowing of the heart, with the drop of pressure as the asphyxia continued, does not indicate that asphyxia of moderate duration renders the heart less sensitive to epinephrin, but is a very common phenomenon in asphyxia when no adrenalin has been in- jected, and when none can be coming from the adrenals. Epinephrin liberated at the ordinary rate may possibly enable the heart to resist better, and for a longer time, the depressing influence of asphyxia, and to respond better to the factors responsible for the acceleration either during the asphyxial period, or when respi- ration is resumed. Another factor in the asphyxial acceleration, if epinephrin takes any share in it, may be, as already pointed out in the case of the acceleration caused by stimulation of sensory nerves (8), the redistribution of the blood associated with the vascular changes, which may result in a greater quantity of epinephrin per unit of time or a greater concen- tration of it being supplied to the coronary circulation, without any increase having occurred in the rate of output from the adrenals. 544 G. N. STEWART AND J. M. ROGOFF In another paper (12) we cited some evidence that certain reactions may be produced by epinephrin in this way, for instance, an experiment of v. Anrep (9) in which he shows that ‘‘if one splanchnic nerve is intact while the suprarenal on the other side is extirpated, stimulation of the splanchnic nerve on the side of the extirpated suprarenal may still cause constriction of a denervated limb. Only after the other splanchnic nerve is cut does the constriction disappear and the limb react passively to the change of blood pressure.’’ His interpretation of this result is that ‘‘this is due to the fact that stimulation even of the peripheral end of the splanchnic excites a certain number of afferent nerves, so that there may be a reflex excitation of the suprarenal of the other side through the intact splanchnic nerve.’”’ There is no evidence that stimulation of the peripheral end of the splanchnic nerve can affect reflexly the rate of epinephrin secretion from the other adrenal and excellent evidence against it. The true explanation, we be- lieve, is that so far as epinephrin is a factor in y. Anrep’s reaction, the constriction seen under the circumstances described is due to the shunting through the de- nervated limb of more of the epinephrin being given off at the ordinary rate from the other adrenal. Cannon has suggested that in the cat (efferent) adrenin secretory fibers may occasionally pass from one splanchnic to the oppositeadrenal. but he produces no proof of such a crossing. We have never seen evidence of any increase in the output of one adrenal when the opposite splanchnic was stimu- lated, or any evidence of diminished output when the opposite splanchnic was cut. Elliott’s results on the protection of the epinephrin store from depletion in the adrenal whose nerves are cut, as compared with its fellow, could hardly have been what they were if both glands were innervated from each splanchnic. In the dog, v. Anrep could scarcely have failed to obtain the limb reaction in the ex- periment mentioned, after section of the opposite splanchnic, if there was a cross- ing of efferent fibers from the stimulated splanchnic to the other gland. Redistribution of the blood owing to vasomotor changes is certainly not the only factor, if it is a factor, in the asphyxial acceleration, and in most of our experiments the rise of pressure produced by asphyxia after denervation of the heart was not great. Good accelerations were obtained with little or no increase of pressure, and the maximum acceleration was generally found when the pressure had fallen again, or, although rising after resumption of respiration, was still below the initial value. There was no obvious difference in this regard whether the adrenals had been excluded or not. Accordingly, such an experi- ment as Cannon illustrates in figure 7 of his paper (4) has no bearing on the question whether acceleration of the heart caused by asphyxia is due to increased epinephrin output. He denervated the heart, tied the limb and carotid arteries and severed the mesenteric nerves, and still got an increase in heart rate from 180 to 212 beats per minute after asphyxia, ‘‘with no previous noteworthy change in blood pressure.” The pressure rose about 25 mm. of mercury after stopping the asphyxia, MEASUREMENT OF EPINEPHRIN OUTPUT ; 545 and it was at this time that the maximum acceleration was counted. But how can this be a proof that asphyxia augments the output of epinephrin since equally large accelerations have been obtained by us in the absence of any epinephrin discharge? As regards Cannon’s method of demonstrating adrenal secretion by the rise of pressure caused by asphyxia after tying the carotid and limb arteries and ‘‘denervating the splanchnic area,” it may be noted that the curve reproduced in figure 3 of his paper shows a slight elevation (less than 10 mm. of mercury) after the first minute of asphyxia, and a very large rise (at least 110 mm. of mercury) after resumption of respi- ration at the end of two minutes of asphyxia. He attributes both rises to the action of an increased output of epinephrin, but unfortunately does not state how much epinephrin would have been required to raise the pressure by 110 mm. of mercury. If such an enormous outpouring of epinephrin as would seem necessary to cause an effect of this kind were actually induced by asphyxia, it is impossible to see why we always missed it in collecting adrenal vein blood. Our operation was assuredly no more “severe” than Cannon’s, the abdomen was opened by him also, and he states that the method yielded constant results so far as the belated influence of asphyxia was concerned, although ‘it was commonly disappointing as a means of demonstrating the early in- fluence of asphyxia.’’ If asphyxia can cause the liberation of so much epinephrin that a rise of pressure of this order of magnitude can be pro- duced, it is also unintelligible that the careful experiments of Gley and Quinquaud (13) should not have revealed a striking difference in the effects of asphyxia on the blood pressure curve before and after elimi- nating the epinephrin output. They found that the curves were practi- cally identical. Cannon’s statement that after tying off the adrenals the rises of pressure did not occur, may mean nothing more than that the con- siderably lower blood pressure (during asphyxia it fell to little over 25 mm. of mercury) had so injured the heart that it could not respond to the resumption of respiration by such an increased action as, in the restricted circulation, would cause a marked rise of blood pressure, whereas, with the better circulation before exclusion of the adrenals it could do so. These suggestions as to possible ways in which epinephrin may be a factor in the asphyxial acceleration of the denervated heart, without any increase in its rate of output having occurred, are not intended to imply that there is evidence that epinephrin takes any share in the acceleration. Our experiments show conclusively that it cannot be THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 546 G. N. STEWART AND J. M. ROGOFF the sole factor, if it is a factor at all, since marked accelerations can be obtained after exclusion or removal of the adrenals. As to the mech- anism of the acceleration, we do not propose to enter further into a dis- cussion of this reaction, in which direct effects of asphyxia upon the heart and indirect effects produced through the changed quality and quantity of the blood supplied to it may be intermingled. Our purpose was accomplished as soon as the statement that the reaction can be used to demonstrate an increased output of epinephrin during asphyxia was shown to be without foundation. Still less than in the case of the acceleration caused by sciatic stimulation can we imagine how such a reaction could be employed to measure quantitatively the rate at which epinephrin is given off by the adrenals. Knowlton and Starling, working with the heart-lung preparation, have shown that the rate of the isolated mammalian heart is very sensitive to changes of temperature, although not altered by marked variations of arterial or venous pressure. But we have not included change of temperature of the blood coming to the heart during as- phyxia or sensory stimulation, or change of temperature of the heart directly produced in asphyxia by stoppage of the heat loss through the respiratory tract, among the factors possibly concerned in the reaction because there is no evidence that under our experimental conditions this could have had any appreciable influence. Cannon states that “the completely denervated heart can be used as an indicator of adrenal secretion in testing the influence of emotional excitement quite as well as in testing the influence of sensory stimula- tion and asphyxia.’”’ We have shown that it cannot be used at all in the case of sensory stimulation or asphyxia, and as, according to Cannon, it has the same value as a test for the influence of emotional excitement, we do not judge it necessary to discuss the matter here at any length. Experiments on emotional excitement are necessarily less satisfactory than on sciatic stimulation or asphyxia because the animal must not be anesthetized. As a matter of fact, the evidence adduced by Cannon is extremely scanty and, in our opinion, will not bear examination. In one case, for instance, after removal of the left adrenal gland and section of the right splanchnic nerve in the thorax on the previous day an in- crease of 42 beats a minute was observed under excitement. In another case after a similar operation an increase of approximately 28 beats a minute was seen during excitement. These accelerations are of the same order of magnitude as those seen with both adrenals and their MEASUREMENT OF EPINEPHRIN OUTPUT 547 nervous supply intact. How is it possible to consider a reaction as a quantitative test of the rate of epinephrin output which is given as well with one adrenal removed and the innervation of the other crippled as when both are normally discharging epinephrin? These animals had had many hours to recover from the effects of the anesthetic and oper- ation and, therefore, gave a good heart reaction with excitement, a reaction which probably had little if anything to do with the assuredly diminished output of epinephrin from the remaining adrenal. That after removal of the remaining adrenal the reaction was not obtained when the animal was allowed to recover from etherization, is no proof that the failure of the reaction was due to the suppression of the residual epinephrin output of the gland, which was already largely denervated. The negative result is far more likely to have been due to the general condition of the animal just recovering from a second etherization and a second operation. The only other experiment mentioned by Cannon is one in which excitement increased the heart rate from 217 to 255 beats per minute with the adrenals intact. Two days after this observation the adrenals were removed, of course under anesthesia, and about 5 hours thereafter excitement caused the rate to increase from 217 to 221 beats a minute. The conclusion is drawn that the difference in the result before and after the adrenalectomy is due to the loss of the epinephrin secretion. The possibility that the general condition of the animal after the adrenalectomy might be less favorable to eliciting a large acceleration is not con- sidered. It must again be pointed out, as in the case of sciatic stimulation and asphyxia, that even if it had been proved that the failure of the reaction after removal of the adrenals was due to the absence of the adrenal epinephrin output and to this alone, the conclusion would not be justified that emotional excite- ment increases the rate at which epinephrin is liberated into the blood, until it was shown that the changes in the amount and concentration of the epinephrin passing through the coronary system, occasioned by the vascular effects of the excitement, were insufficient to account for the observed accelerations. We should have considered it a mere accident that the heart rate was the same (217 beats a minute) after removal of the adrenals as two days before. Cannon, how- ever, believes that this “‘is an indication, and the first reliable one, that under quiet, peaceful conditions there is no adrenal secretion or a secretion so slight as not to affect the denervated heart, an extremely sensitive indicator.’’ He argues that if there was a secretion of the magnitude observed by us there ought to have been a drop in the pulse rate when this secretion was eliminated. Doctor Cannon has not mentioned what the pulse rate was just before the adrenalectomy, so we do not know whether there was a decline in the rate or not after removal of the glands. We have, however, presented evidence both in the preceding paper (8) and in this, that other factors than the loss of epinephrin are concerned in the diminution of the pulse rate which usually follows removal of the adrenals and that sometimes that operation, or an equivalent interference with the epinephrin 548 G. N. STEWART AND J. M. ROGOFF output may not be followed by any diminution in the rate, while after exclusion of the adrenals a marked diminution may be caused by operative and other procedures associated with a fall of blood pressure. The catheter method. Cannon sums up his work with the heart reac- tion by the declaration that “the results obtained with the isolated heart used as an indicator of adrenal secretion confirm in every respect the results obtained eight years ago by the catheter method.” We have shown what kind of confirmation the isolated heart used as an indicator of adrenal secretion can lend to any other method, it could only render it suspect. Since, however, despite their confirmation by the isolated heart reaction, the results obtained by Cannon with the catheter method might be correct, we shall take this occasion to examine them on their merits in somewhat greater detail than has hitherto been done. This series of papers by Cannon and his co-workers deals with the influence of emotional excitement, asphyxia, stimulation of sensory nerves and some drugs upon the rate of epinephrin output. He collected blood from the inferior vena cava above the level of the adrenal veins by means of a catheter pushed up through the femoral vein. According to him, blood obtained from this region in cats in the absence of excite- ment, asphyxia, etc., does not cause inhibition of cat’s intestine strips or rabbit’s intestine segments, whereas blood obtained during or after the action of these factors causes marked inhibition, due to the out- pouring of epinephrim. After emotional excitement lasting for 10 minutes or more, so much epinephrin has been or is being liberated that it can be detected in the general venous blood (femoral vein), after passing not only through the lungs but through the systemic capillaries. It may be remarked that if this is correct, it ought not infrequently to be possible to detect epinephrin in blood collected from patients or animals by puncture of a vein. For much excitement may attend even so small an operation. But no satisfactory evidence has ever been brought forward that detectable concentrations of epinephrin exist in the general venous blood. To account for the extraordinary output of epinephrin which would be necessary in order that it should be clearly detected in femoral vein blood, Cannon suggests that adrenalin liber- ated into the blood stimulates the adrenals to further increase the out- put. ‘‘Thus the more marked effect as time passes (see fig. 3) may be due not only to further excitement, but in part to an autogenous con- tinuance of adrenal ‘secretion. Thus also the persistence of the emo- tional state after the exciting object has disappeared can be explained. Indeed it was the lasting effect of excitement on digestive processes MEASUREMENT OF EPINEPHRIN OUTPUT 549 which suggested this investigation.” In support of this idea he appeals to a statement of Elliott’s (14) that adrenalin itself causes depletion of the epinephrin store of the adrenals. Elliott himself later withdrew this statement (15), which in any case would not prove that adrenalin increased the output of epinephrin. Cannon complains that ‘‘authors have written as if Cannon had been attempt- ing to support the idea that emotional experiences were dependent upon circu- lating adrenin. Thus Stewart and Rogoff report, as if the matter had been questioned, that all signs of fright can be elicited by administering morphine to a cat with one adrenal removed and the other denervated.’’ This subject is not referred to at all in the paper quoted by Cannon. In another paper (26), we are discussing the depletion of the epinephrin store produced by morphine, which Elliott (15) suggests is directly due to the fright caused by the drug. In this connection we say ‘‘That signs which might be interpreted as those of fright are present in cats under morphine is, of course, not doubtful. Whether this inter- pretation is correct might be difficult to decide, and does not concern us here. It is, however, of interest to note that epinephrin seems to have nothing to do with those signs. The signs of morphine fright can all be elicited by administer- ing morphine to a cat in which one adrenal has been removed and the splanchnic supply of the other cut and in which accordingly no liberation of epinephrin through the splanchnics takes place.’? How can this be spoken of as ‘“‘writing ‘as if Cannon had been attempting,” etc.? As a matter of fact, in that paper we were not discussing Cannon’s psychological theories at all. We do not see that there was any impropriety in publishing observations showing that signs, which might be interpreted as those of fright, can be elicited by morphine in cats whose epinephrin output has been abolished or markedly diminished, even if we had been aware, which of course we could not be, that this detail in the pharma- cology of the drug would seem self-evident to Cannon. Cannon complains further that ‘“‘Rogoff points out that the secretion of suf- ficient adrenin to produce symptoms of fright would be impossible, again as if any claim had been made that these symptoms were due to secreted adrenin.”’ What Rogoff did say was: ‘‘In some of the cats (after removal of one adrenal and denervation of the other) it was found that the liberation of epinephrin from the adrenals was so interfered with by this operation that there could not have been one-thousandth of the normal liberation, if any epinephrin was being given off from the glands. These animals, nevertheless, responded readily to fright and other emotional disturbances with the usual symptoms of sympathetic excitation in the same manner as normal cats. Certainly, an outburst, through nervous influence, of epinephrin in such quantities as would be necessary to produce these symptoms would be impossible in these animals.’”’ Again, we fail to see the impropriety of quoting such a result, so long as it is true. We are glad that Doctor Cannon finds this result so obvious as not to be worth while mentioning. For it must then be obvious to him also that in these animals “the persistence of the emotional state after the exciting object has disappeared”’ cannot be explained as due to ‘“‘continuance of adrenal secretion,’’ whether ‘‘autogenous”’ or not. 550 G. N. STEWART AND J. M. ROGOFF Cannon seems to criticise our use of the term ‘‘denervated eye’’ for the eye after removal of the corresponding superior cervical ganglion, and says that he and de la Paz ‘‘tried the denervated eye method of testing for adrenal secretion but could not persuade themselves that an eye still innervated by the third cranial nerve was really denervated.’’ It is surely unnecessary to state that when we use the term ‘‘denervated eye’’ in this sense we do so for convenience, following the practice of other writers, for instance, Elliott (15). We frequently speak of ‘‘the (denervated) eye reactions of Meltzer,’’ or simply ‘“‘the eye reactions,” or of ‘‘the iris sensitized by previous removal of the superior cervical ganglion.”’ Cannon himself uses the terms ‘‘excited’’ blood and ‘‘quiet’’ blood for conven- ience, and quite properly, although he will not maintain that they are literally correct. We did not imagine that any reader would suppose that we were not aware that the oculomotor nerve supplies the eye. We agree with Cannon that “the prompt dilatation of the pupil in a paroxysm of rage,’’ in cats whose epineph- rin output has been interfered with by removal of one adrenal and section of the nerves of the other has nothing to do with epinephrin, but is due to inhibition of the pupillo-constrictor mechanism. We have attributed the dilatation caused by stimulation of sensory nerves, including the paradoxical reaction, which can be obtained after removal of the adrenals (8), to the same mechanism. Long ago, it was stated by one of us (27) that sexual excitation caused very marked dilatation of the pupils, in dogs of both sexes, after division of both vago-sym- pathetic nerves (without removal of the superior cervical ganglion) by central inhibition of the constrictor pupillae through the oculomotor. Apparently it is necessary to point out again, as has been done repeatedly before (1), (12), that when we employ the eye reactions to test for epinephrin, we eliminate factors which can affect the eye through the nervous system, by collecting the adrenal blood in a cava pocket and only releasing it, as far as pos- sible, after these factors have ceased to act. Cannon, on the basis of his experiments with the catheter method, concluded that emotional excitement, asphyxia and sensory stimulation increase the rate of output of epinephrin. Nicotine does the same, whereas urethane has no effect. Our own experiments with asphyxia and sensory stimulation, made by methods which permitted the rate of output of epinephrin to be estimated, did not support Cannon’s con- clusion, yielding no evidence of sensibly increased output. Such nega- tive results must always yield to positive results obtained by a better method, and this has been emphasized by us. Cannon’s catheter method, however, far from being a better method than those employed by us, is not a method by which the rate of ouput of epinephrin can be estimated at all. As a matter of fact, his papers do not contain a single estimate of the output of epinephrin before, during or after the action of the factors which he is studying. And it is clearly impossible that they should contain such data. The most that the method could yield would be the epinephrin concentration in the cava blood above the MEASUREMENT OF EPINEPHRIN OUTPUT 551 adrenal level. Without knowing the amount of blood with this epi- nephrin concentration passing the point of collection per unit of time, the output of epinephrin could not be calculated. But even the epi- nephrin concentration was not determined by Cannon. He gives not one estimate of the concentration, or any information as to the minimum concentration which his strips or segments could have detected. In the absence of information of this kind it is impossible to deduce any conclusion from the few curves reproduced. What, for instance, was the concentration of epinephrin in the femoral vein blood after emo- tional excitement? If we knew whether it was 1:1,000,000 or 1:10,000,- 000 we might be able to check the probability of the statement by calculating the concentration which must have existed in the blood of the adrenal veins, and seeing whether anyone has ever observed so high a concentration. If the minimum concentration clearly detectable by the segments and the concentration actually found in the cava blood during asphyxia, etc., had been determined, there would be some possi- bility of deciding whether the observed changes in concentration could be due to alterations in the rate of blood flow in the cava without any change having occurred in the rate of output of epinephrin. We have suggested that vasoconstriction, especially in the splanchnic area, which may be assumed to accompany the rise of blood pressure associated with stimulation of the central end of the sciatic or with asphyxia, will cause slowing of the blood flow in the inferior cava. Doctor Cannon seeks to invalidate this suggestion by the statement that an increased arterial blood pressure causes an increased blood flow through the adrenals. Had this argument not been repeated and emphasized in his paper we should have considered it a slip. Certainly the adrenal flow increases when the arterial pressure rises. But as the blood from the adrenals is only a small fraction (perhaps 745 to s}y) of the blood passing along the inferior cava this, of course, has no sensible influence upon the rate of cava flow. Cannon further argues that when the arterial pressure is raised by splanchnic vasoconstriction the blood flow in the inferior cava may be in- creased. We had supposed it was universally admitted that vasoconstriction renders the passage of the blood through a vascular area more difficult, and that the arterial blood pressure rises for this very reason when an important area, like the splanchnic area, is constricted. Weare not concerned here with compen- satory reactions which may occur in other areas, leading, for example, as Edwards (16) has shown, to increased flow in the superior cava. It is precisely because less blood is passing through the vasoconstricted areas that more blood can pass through areas which are not constricted. Vasomotor effects are not the only factors which may influence the rate of the blood flow in the inferior vena cava under the conditions of Cannon’s experi- ments. The respiratory movements and, therefore, the intrathoracic pressure are affected by sensory stimulation, emotional disturbance and very grossly by aoe G. N. STEWART AND J. M. ROGOFF asphyxia. It is not possible to produce such changes without markedly affecting the pressure in the great veins. The region of the cava from which a sample of blood is assumed to be drawn off through the catheter is right up against the diaphragm. How can the control sample be obtained under the same conditions, except for the hypothetical stimulation of the adrenal secretory mechanism, as a sample taken when the animal is gasping in asphyxia? In asphyxia the right heart and great veins become engorged withblood. The blood in the cava must, therefore, be dammed back and if the adrenal goes on steadily secreting epi- nephrin the concentration in the upper segment of the cava must increase without any change having occurred in the rate of output. If the animal, in struggling or gasping, markedly increases the intraabdominal pressure the cava might be more or less obstructed by flattening of its walls upon the catheter, especially above its upper end. There are still other ways in which the blood flow may be altered through effects produced upon the heart by the factors studied by Cannon. Yet in none of his papers has he indicated that there is any necessity to take into account possible changes in the rate of blood flow. It is the same with the question whether massage of the glands, particularly of the right adrenal, may not be produced by the catheter during the voluntary movements executed by the ani- mal when frightened or the movements of the diaphragm in asphyxia. The right adrenal lies close against the cava, separated from the catheter by little more than the thin wall of the vein, and it is well known that massage causes liberation of epinephrin from the adrenals. All these possibilities ought to be controlled. In referring to our negative results with the catheter method Cannon states, what could not be concluded from anything:in his previous papers, that ‘‘the method is difficult and exacting, and that not until after some experience with it did it begin to yield us positive results.’’ He rebuts our suggestion that some -of his positive results might be due to a fortunate location of the eye of the cathe- ter with reference to the orifices of the adrenal veins by saying ‘‘that it seems to — “have been made with disregard for the care exercised in making control obser- -vations under precisely the same conditions before and after stimulation.’”’ This is beside the point. There is no particular difficulty about imserting a fine catheter iuto the cava, so that in successive observations a string tied on the catheter is at the same level of the femoral vein, while the eye of the catheter is judged to be above the level of the adrenal veins. It is easy to turn the eatheter so that a mark on its circumference always occupies the same position. But the exact relation of the eye of the catheter to the orifices of the adrenal -veins or to the orifice of one of the veins, e.g., the right, since they do not enter the cava at the same level, can never be known till the abdomen is opened, if then, nor can it be assured that this relation will always remain precisely the same during such movements as the animal may make, even when securely tied down, during excitement caused by a barking dog or during stimulation of sensory nerves or asphyxia. The few curves published by Cannon, are not at all convincing. In all the work with the catheter method he has contented himself with one or two com- parative tests of the bloods. We have laid stress on the necessity of repeated comparison. His method of emptying the cylinder by sucking out one liquid MEASUREMENT OF EPINEPHRIN OUTPUT 55a before the application of the next is also liable to introduce error, as we have previously pointed out (18) and illustrated by curves. Not only may the tracing be deformed at the critical moment, but the strip or segment is suddenly exposed and then suddenly immersed in liquid. He refers toa paper by three young medi- cal men (17) as having confirmed his results on emotional excitement. But this paper contains not a single tracing or protocol, not one estimation of the epinephrin concentration in the blood or of the concentration which it would have been possible for the segments to detect. There is internal evidence that any reaction obtained could not have been due to epinephrin. For example, they state that when blood collected from the cava above the level of the adrenals, during the action of factors which increase epinephrin output, was allowed to stand for 25 minutes, it no longer caused inhibition of the rabbit’s intestine seg- ment, ‘‘and epinephrin is the only oxidizable substance in the body that produces inhibition of the intestine which can be oxidized in 25 minutes.’”’ Now, a sample of blood, known to contain epinephrin in amount sufficient to give a decided inhibition of the intestine, does not become ineffective merely by standing for 25 minutes. At one place it is stated that morphia prevents the epinephrin reaction from being given by the cava blood under the influence of Witte’s pep- tone, and in another place that morphia distinctly diminishes the normal output of epinephrin, as tested in blood taken directly from the adrenal vein. In still another place the statement occurs that “‘substances with which we have experi- mented consisted of the toxins of gonococci, of streptococci, of staphylococci, of colon bacilli, of tetanus bacilli, of diphtheria, foreign proteids, indol and skatol, leucin, creatin, feces extract, strychnin, morphia, carbolic acid, Witte’s peptone, sheep serum, ox serum, kitten serum. All of these have seemed to cause adrenal activation, and from their use 61 out of 66 consistent experiments can be reported Narcotics and anesthetics caused no increased output of adrenalin 4% The various agents which caused increased epinephrin output, or which diminished it, are either neutral as to motion and fever or diminish them. There was no epinephrin reaction in any animal (subjected to intense emotion, rage and fear) in which bilateral division of the splanchnic nerves existed . . . . Of great significance is the fact that allagents that produced ‘an increase of adrenalin output cause also a hyperchromatism of the brain fol- lowed by exhaustion. It mattered not whether this stimulating agent was a physical exertion in running, in fighting or in convulsions; or whether it was the emotion of fear, or of anger, or a reaction to anaphylaxis, to toxins, to indol, skatol, foreign proteids, or strychnin, in every case hyperchromatism and in- crease of adrenalin output went hand in hand.’’ All this without tracings, with- out any sample of the great mass of experimental details which would be necessary to establish so many conclusions. If it is to be a matter of taking anybody’s word for a physiological result, with all due respect to these writers, whom we know to be very competent men in their own subjects, we should much sooner take the word of an accomplished physiologist like Doctor Cannon himself. In Cannon’s paper on emotional stimulation the first figure shows that adrenal vein blood, obtained under ether, when substituted for inactive blood caused relaxation of the intestine strips, whereas blood from the renal vein did not give this effect. The most which could be deduced from this is that adrenal vein blood has a demonstrable content of epinephrin, a proposition which is generally 554 G. N. STEWART AND J. M. ROGOFF accepted. The second figure shows that when so-called excited blood (i.e., blood removed through the catheter from above the adrenal level during excitement) was substituted for Ringer’s solution, the tone of the intestine strip was markedly increased. It remained increased for about 43 minutes, and then the strip relaxed and the beats disappeared. Cannon interprets this extremely belated relaxation as an epinephrin reaction. This is contrary to all our experience. An epinephrin relaxation is produced immediately or not at all. A relaxation occurring 4 or 5 minutes after the application of a blood cannot be accepted as a positive reaction for epinephrin. A second application of the excited blood, this time replacing so-called ‘‘quiet’’ blood caused a prompt relaxation. The only conclusion which can be drawn from this figure is that two inconsistent observations with a given sample of blood on such an object as an intestine strip or segment cannot be accepted as evidence that the blood from the inferior cava, drawn during excite- ment, gave a positive epinephrin reaction. If an undoubted positive reaction had been obtained, this, as already pointed out, would yield no information as to whether the rate of output of epinephrin had been increased by the excitement or not. The third figure professes to demonstrate that with the prolongation of excite- ment the effect on the epinephrin secretion goes on increasing. This is done by comparing a record of the strip in defibrinated blood drawn from the cava after 11 minutes excitement with a record in the serwm of blood obtained after 15 minutes of excitement. In the defibrinated blood the strip retains the increase of tone caused by substituting the blood for Ringer’s solution, for about2 minutes, and then begins to relax. In the serum the strip relaxes more promptly. If the two samples of blood had the same concentration of epinephrin, the serum would necessarily give a stronger reaction than the blood since practically all the epinephrin is in the serum (19). After the initial increase of tone has lasted 2 or 3 minutes some relaxation occurs also with ‘‘quiet’”’ serum, and the beats become slower. No data are given by which it would be possible to judge how great the difference in epinephrin concentration in the different samples was. The fourth figure is intended to show that after removal of the adrenal glands, blood from the cava of an excited animal above the adrenal level gave no inhibi- tion of the intestine strip when substituted for Ringer’s solution. As there is no question that epinephrin exists in the upper cava blood, so long as the adrenals are intact, and must disappear when the adrenals are excised, nobody will dispute that a test object, if sensitive enough, will give a positive reaction in the first case and not in the second. In the legend of this figure, it is stated that ‘‘the strip later proved sensitive to adrenalin in blood in the ratio 1: 1,000,000,’ and in the fifth figure tracings are given showing the inhibition produced by blood with 1: 1,000,000, 1: 2,000,000 and 1:3,000,000 adrenalin when substituted for ‘quiet’? blood. The 1: 3,000,000 adrenalin blood gives a reaction much smaller, and the 1:2,000,000 adrenalin blood a reaction probably smaller than those figured by Cannon as given by “‘excited’”’ blood. He does not state specifically that these curves were obtained from strips used to test any of the ‘‘excited”’ bloods whose records are reproduced, and, as has been pointed out, it is unfortu- nate that he did not assay the concentrations of epinephrin which he believed these bloods to contain. But there would seem to be no point in reproducing tracings showing the inhibition of intestine strips given by such concentrations MEASUREMENT OF EPINEPHRIN OUTPUT 555 of added adrenalin, unless they had a bearing upon the magnitude of the reactions given by the cava bloods. Redfield (20) who worked with Cannon’s method, and partly under his direction, has stated that he succeeded in modifying the method so as to render it capable of detecting 1: 10,000,000 adrenalin. Redfield obtained such variable results in experiments, in which he endeavored to detect epinephrin in the blood (of the horned toad) during emotional excitement by means of the intestine reaction, that he says ‘‘no weight can be placed upon the experiments.”’ It seems pretty clear that in Cannon’s hands the method was not very sensi- tive, and that such inhibitions as he figures, if due to epinephrin, must have corresponded to epinephrin concentrations in the ‘‘excited’’ cava blood of the order of magnitude of 1: 1,000,000 or 1: 2,000,000. Now, these concentrations are double the average concentrations found in adrenal vein blood, with the average normal arterial pressure under the conditions of our experiments, and would represent 100 to 200 times our average normal output. Cannon in his last paper (4) has stated that with sciatic stimulation he can increase the output to 5 to 25 times our “normal” output. But that apparently is nothing to what can be obtained when a cat is frightened by a dog. For us, such concen- trations in the cava blood (corresponding perhaps to 1: 10,000 in the adrenal vein blood) are so improbable that we are unable to accept them on such evidence as Cannon has furnished. If anything like such concentrations of epinephrin could exist in the cava blood with ordinary rates of blood flow, it would be impos- sible for us to miss them with the method of assay which we employ. In the sixth figure a tracing is given, showing that ‘‘active’” blood applied to a strip caused relaxation after 2 minutes. The same blood, after oxygen was bubbled through it for 3 hours, when applied to a fresh strip is stated to have failed to cause relaxation. If the reactivity of the new strip was the same as that of the other (which was not controlled, but which must always be controlled if reliable results are to be obtained with such test objects) this would indicate that an oxidizable substance (epinephrin) was responsible for the original inhibition. Cannon cites Redfield (20) as having “‘reported that in th- horned toad nervous excitement causes a contraction of the melanophores in the denervated skin, a reaction which does not occur after the removal of the adrenal glands.’’ He adduces this as a proof that in this animal emotional excitement increases the epinephrin output. Asa matter of fact, Redfield found that in the great majority of his experiments removal of the adrenals ‘‘does not check the contraction of the melanophores.’’ Out of a large number of animals he only succeeded in finding two in which after adrenalectomy (involving the opening of the body cavity, removal of the gonads and a portion of the genital ducts and tying off the posterior cava) contraction of the melanophore pigment could not be ob- tained “‘when the mouth was stimulated by a weak faradie current.’’ There is no proof that in these two exceptional cases it was the loss of the epinephrin output of the adrenals and not some other factors, such as the general deteriora- tion of the animal caused by the operation, which was responsible for the negative result. However, his experiments with denervated skin and temporary ligation of the blood supply of portions of skin suggest strongly that epinephrin given off from the adrenals is a factor in contraction of the pigment. But we do not 556 G. N. STEWART AND J. M. ROGOFF find any proof that the effect is due to an increased epinephrin output. An experiment is quoted to show that blood from an excited horned toad causes a contraction of the melanophore pigment of an unexcited animal in the neighbor- hood of the point of injection, whereas blood obtained from an animal after destruction of the portion of the thoracic cord, whose integrity is essential to the production of contraction of the melanophore pigment through noxious stimuli, has no such effect. All that this experiment could indicate would be that the effect of the blood from the animal excited with intact cord was due to epinephrin. The experiment throws no light upon the question whether excitement can in- crease the epinephrin output. For if the injury to the thoracic cord prevents the contraction of the melanophore pigment associated with stimulation, by inter- fering with the ordinary output of epinephrin, the blood of the excited animal would necessarily contain more epinephrin than that of the control, without any increase in the rate of output having occurred in the former. We have shown (21), (22) that in mammals the upper part of the thoracic cord is intimately re- lated to epinephrin secretion. The possibility must be considered that in this animal, on account of slower circulation, less active metabolism, and perhaps lower temperature of the blood and tissues, a greater concentration of epineph- rin may exist in the general blood than in mammals. The great concentration found by Abel and Macht (23) in the secretion of the so-called parotid gland of a tropical toad suggests that care may be necessary in applying results obtained in such animals to mammals. The above remarks are in no way intended as a criticism of Redfield’s paper, which seems to us a most interesting and sugges- tive piece of work. In Cannon’s paper on the effects of asphyxia, hyperpnoea and sensory stimu- lation on adrenal secretion there are three figures. Rabbit’s intestine segments were employed forthetests. The first figure shows a marked effect of disturbance of the curve by sucking out the liquid. If any weight can be laid on a curve of this kind, the ‘“‘normal’’ vena cava blood (i.e., from above the adrenal level in the absence of asphyxia) probably caused some inhibition and the cava blood from the same level after asphyxia, a greater inhibition. But how is it possible to determine whether the difference in epinephrin concentration between the two cava specimens is great or small? The second figure purports to demonstrate the failure of hyperpnoea to in- crease adrenal secretion. ‘‘The chest was opened and the chest walls held apart while the lungs were inflated in rapid repetition by means of bellows. The air forced into the lungs was permitted to escape quickly through an opening in the trachea.’’? The curve shows that when vena cava blood taken from above the adrenals with ordinary breathing is substituted for Ringer’s solution, there is some increase of tone, but the beats continue. When vena cava blood drawn from above the adrenals after the application of hyperpnoea is substituted for the previous sample of cava blood no notable change occurs in the tracing. The conclusion is “‘that hyperpnoea, to a degree resulting in acapnia is not attended by increased adrenal secretion.’’ This seems to us extremely probable, but what the tracing indicates is simply that no noteworthy change has occurred in the concentration of epinephrin in the cava blood, within the limits of sensi- tiveness of the segment. It is obvious that the conditions which in the asphyxia experiments may lead to slowing of the blood flow in the cava are absent here. * MEASUREMENT OF EPINEPHRIN OUTPUT 557 The chest being widely opened, changes in the intrathoracic pressure and in the movements of the diaphragm are eliminated. The situation as regards vaso- motor and cardiac changes, especially splanchnic vasoconstriction, is likewise quite different. Since Cannon has here performed a considerable operation in an anesthetized animal, and since, according to him, the trauma and anesthesia in our experiments increase the ‘“‘ordinary”’ output of epinephrin so much that it may be impossible for asphyxia or sensory stimulation to cause any further increase, the question would seem pertinent why he obtains in this experiment a totally negative reaction for epinephrin in the cava blood taken above the adrenal level. Our explanation is that the small adrenal output is too much diluted by indifferent blood in the cava to be easily detected. The third figure in this paper purports to show that sensory stimulation in- creases adrenal secretion. What it does show is that when cava blood from above the adrenal level, obtained during sciatic stimulation, is substituted for cava blood obtained without sciatic stimulation, the writing point descends abruptly. A large part of the descent appears to be due simply to the fact that when the latter blood was sucked out of the cylinder the weight of the segment, coming on the lever, raised the point considerably. When the cylinder was again filled up with blood, the weight of the segment was nearly neutralized and the writing point necessarily descended. The true effect of the inhibition of the segment is, therefore, considerably exaggerated. All that can possibly be deduced from the tracing is that the epinephrin concentration in the sample collected during sciatic stimulation was greater than in the other. No assay of the concentration was made and, therefore, it is impossible to know what the difference was. It might have been a very moderate difference or a very great one. Inspection of such a curve can yield no information as to this. Also, it is not stated how many times the result figured was confirmed on the same samples of blood. In his paper on the effect of nicotine on adrenal secretion Cannon has published three figures. The first shows that cava blood from above the adrenal level, secured before nicotine injection, when substituted for Ringer’s solution caused an increase of tone of an intestine strip. When cava blood from the same level, obtained after injection of nicotine, was substituted for the other cava blood, there was a prompt and large diminution of tone. The conclusion is drawn that the adrenal secretion was increased by the nicotine. The most, however, which can be deduced from the tracing is that the concentration of epinephrin in the second specimen is greater than that in the first. If the blood flow is slowed, as would be the case after the doses of nicotine employed, an increased epinephrin concentration would be consistent, not only with an unchanged, but even with a diminished output. Wehavealready pointed out (24) how completely Cannon, owing to the faulty method employed, has missed the details of the nicotine action on the epinephrin output, the most outstanding and most durable effect being a depression of the output. It is stated in the text that the several specimens of blood were tested for their content of epinephrin. But not one tracing illustrating the assay is reproduced, nor is a single concentration given, although the statement is made that ‘‘the characteristic inhibition of the rhythmic contractions of the muscle, even when it has been for some time removed from the body, occurs at a dilution of adrenalin 1: 2,000,000 in defibrinated blood.’’ If it is implied that this is the order of 558 G. N. STEWART AND J. M. ROGOFF magnitude of the concentrations which cause inhibitions like that displayed in the figure under discussion, we can only repeat that it is in the highest degree improbable that such concentrations can be obtained in a fair sample of the cava blood above the adrenal level, unless there is an extreme slowing of the cava flow. Our own work on nicotine shows that at the time Cannon collected his samples the initial brief stimulating action of nicotine on the epinephrin output must have long since passed into the stage of depression. The effect of the first cava sample when substituted for Ringer’s solution cannot, of course, be compared with that of the second cava sample when substituted for the first. To assay the two specimens they should be caused separately to displace the same sample of indifferent blood free from epinephrin. The displacement of one sample of epinephrin-containing blood by another sample of epinephrin-contain- ing blood is not usually satisfactory for purposes of assay. For a, the first sample has already produced an epinephrin effect upon the test object which may render the inhibition, caused by the second sample when it displaces the first, different from what it would otherwise be. 6, The epinephrin content of the first sample in contact with the test object, especially if left for a considerable time in contact with it, is diminished, so that the content of the second sample when now caused to displace the first may be exaggerated. The second figure in Cannon’s nicotine paper shows that the quantity of nicotine employed could not have accounted for the inhibition of the intestine strip produced by the cava blood after nicotine injection. This is unquestionably correct. The third figure demonstrates that cava blood obtained from an animal after removal of the adrenals and after administration of nicotine, does not cause an inhibition of the intestine strip. This is no doubt perfectly true, but it simply indicates that in the absence of the adrenals no epinephrin detectable by the method employed exists in the cava blood. It would be much more to the purpose to state what were the concentrations of epinephrin in the cava blood and the rates of cava blood flow before and after injection of nicotine. But the reader searches in vain for this essential information. That the intestine strip was inhibited by the addition of one drop of adrenalin 1: 1000 to a cylinder con- taining 2.5 ee. of blood (which would make a concentration of about 1: 35,000 or 1: 40,000) is no doubt correct. But what possible relation has a huge con- centration like this to anything which ean occur in the cava blood? SUMMARY 1. The acceleration of the heart induced by asphyxia after division of the vagi and excision of the stellate ganglia cannot be taken as an index of increased epinephrin output from the adrenals, as assumed by Cannon. Forit may be obtained and may not be diminished in amount after the adrenal veins have been ligated, after the adrenals have been completely isolated from. the circulation, and after the adrenals have been excised. 2. Toward the end of an experiment, after repeated periods of asphyxia and when the general condition of the animal, including the MEASUREMENT OF EPINEPHRIN OUTPUT 559 circulation, has deteriorated the reaction is apt to become much less marked or to fail altogether. An operation like adrenalectomy, when it accelerates the deterioration of the animal and of the heart by lower- ing the blood pressure decidedly, may sometimes seem to be responsible for the failure of the asphyxial acceleration. But there is evidence that this is not due to any specific effect of the adrenalectomy (loss of epinephrin output) but to the general effect of the operation. When asphyxia fails to cause a decided acceleration it is often seen that some slowing has occurred during the asphyxial period, succeeded by a quick- ening of the beat on resumption of respiration. It is clear that it will depend upon the state of the heart whether this subsequent acceleration will bring the heart back to, or nearly to, its original rate or cause it to pass considerably beyond the original rate. Whether an acceleration of some magnitude, as compared with the rate before asphyxia, can actually be counted on the trace is then more or less of an accident, depending upon whether the deteriorated heart comes through the period of asphyxia in such a condition as to permit it to attain a rate considerably in excess of the initial rate. 3. There is no good evidence that asphyxia causes an augmentation of the rate of epinephrin output by a direct action upon the cells of the adrenal medulla. 4. The acceleration caused by asphyxia, like the acceleration caused by stimulation of the central end of the sciatic, is not interfered with by opening the abdomen. There is no evidence that the operation necessary for collecting adrenal vein blood for direct assay of its epine- phrin content would render it impossible to detect an increase in the epinephrin output, if this were caused by asphyxia or stimulation of sensory nerves. 5. No real evidence has been adduced that the epinephrin output, measured by those observers who have adopted the fundamentally correct method of collecting the blood and assaying it on suitable test objects, is an output artificially increased by anesthesia and trauma. On the contrary, the remarkably narrow range of the output in different animals, under different anesthetics and with the different operations strongly suggests that it is a physiological output already going on, not initiated and probably not much modified, but merely unveiled by the experimental procedures necessary for its measurement. 6. As a method of estimating changes in the rate of output of epine- _ phrin from the adrenals, the catheter method is defective in principle. For at best all that could be measured by it would be changes in the 560 G. N. STEWART AND J. M. ROGOFF epinephrin concentration in the blood of the inferior cava above the level of the adrenals. Changes in the rate of flow of the blood are not taken account of. The quantity of epinephrin passing along the inferior cava to the heart per unit of time cannot be estimated, nor the changes, if any, produced in this quantity by the conditions studied. We have pointed out that in the cases in which Cannon claims to have obtained evidence of an increased output of epinephrin, all he can possibly have observed is an increased concentration in the cava blood, and that a slowing of the cava flow would cause an increase in the concentration, if no change whatever had occurred in the rate of output. We have suggested certain factors, associated with all the conditions studied by him, which might cause such changes in the cava flow as would tend to increase the epinephrin concentration, even in the absence of an in- creased rate of output. However, the complete lack of assays of the concentrations of epinephrin supposed to have been present, and of estimates of the concentration which could have been detected by each segment or strip employed, renders it impossible to determine to what extent positive reactions, when obtained, were actually due to epine- phrin. It cannot be assumed that by sucking blood from a catheter opening into the cava immediately below the diaphragm, a region where the blood flow is necessarily less steady than anywhere else, owing to the influence of changes in the intrathoracic pressure, the same aliquot part of adrenal blood will be drawn off in successive samples without regard to the gross effects upon the respiration produced by such con- ditions, e.g., as asphyxia. Such mechanical changes may alter the pro- portion of adrenal blood to the much greater quantity of blood from other sources drawn off through the catheter. Our experience with the catheter method fully justifies the criticism of Richards and Wood (25) that the method is “highly faulty in that the blood from the supra- renals is diluted with that from all of the structures whose veins enter the cava below the suprarenals.”’ BIBLIOGRAPHY (1) Stewart AND Rogorr: This Journal, 1920, li, 366. (2) Hoskins aNnpD McCuvre: Arch. Int. Med., 1912, x, 348. (8) Brepu: Pfliiger’s Arch., 1897, Ixvii, 448. (4) Cannon: This Journal, 1919, 1, 399. (5) Srewart AND Rocorr: This Journal, 1920, li, 375. (6) Stewart AND Rogorr: This Journal, 1919, xlviii, 22. (7) BayLiss AND Sraruina: Journ. Physiol., 1899, xxiv, 122. MEASUREMENT OF EPINEPHRIN OUTPUT 561 (8) Srewart AnD Rocorr: This Journal, 1920, lii, 304. (9) v. ANREP: Journ. Physiol., 1912, xlv, 318. (10) PEARLMAN AND VINCENT: Endocrinology, 1919, ili, 121. (11) Czupausxt1: Zentralbl. f. Physiol., 1913, xxvii, 580. (12) Stewart anpD Rocorr: This Journal, 1918, xlvi, 90. (13) GueyY AND QuINquaupD: Journ. d. Physiol. et Path. gén., 1918, xvii, 807. (14) Exxiorr: Journ. Physiol., 1905, xxxii, 401. (15) Exuiorr: Journ. Physiol., 1912, xliv, 374. (16) Epwarps: This Journal, 1914, xxxv, 15. (17) Hircuines, SLOAN AND AustIN: Cleveland Med. Journ., 1918, xii, 686. (18) Stewart AND Rocorr: ‘This Journal, 1917, xliv, 543. (19) SrewarT AND Rocorr: Journ. Pharm. Exper. Therap., 1917, ix, 393. (20) RepFieLp: Journ. Exper. Zodl., 1918, xxvi, 275. (21) Stewart AND Rocorr: Journ. Exper. Med., 1917, xxvi, 613. (22) Stewart AND Rocorr: This Journal, 1920, li, 484. (23) ABEL AND Macur: Journ. Pharm. Exper. Therap., 1912, iii, 319. (24) Srewart AND Rocorr: Journ. Pharm. Exper. Therap., 1919, xiii, 183. (25) RicHarps AND Woop: Journ. Pharm. Exper. Therap., 1915, vi, 283. (26) Stewart AND Rocorr: Journ. Exper. Med., 1916, xxiv, 709. (27) Srewart: Centralbl. f. Physiol., 1902, xv, 617. THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, NO. 3 THE EFFECT OF SPLENECTOMY UPON GROWTH IN THE YOUNG SAMUEL CHESTER HENN From the Hull Laboratory of Physiology of the University of Chicago Received for publication April 19, 1920 INTRODUCTION With but few exceptions the removal of the spleen has been done on adult animals. Might it be that splenectomy in the young would exhibit some change not yet observed? It is known that removal of thyroid glands in young rabbits is followed by much more striking changes (experimental cretinism) than are manifested by removal of these glands in the adult animal. LITERATURE Wolferth (1) reports blood changes in rats splenectomized at thirty days of age, comparable to those changes in the blood of adult sple- nectomized rats. There is a slight to moderate anemia persisting for three or four weeks with recovery in a few months. Accompanying this is a slight leucocytosis and an increased resistance to hypotonic sodium chloride solutions and hemolytic agents. Tachigara and Takayi (2) find that in general the effects of sple- nectomy in newborn dogs are the same as in adult animals. The most marked result is an increase in the number of red corpuscles. The younger the dog at the time of operation, the greater is this increase and the earlier does it appear. Their protocols likewise show that the splenectomized puppies operated from one day to two months make gains in weight that are nearly parallel to normal animals through periods of six to eight months. They also observed that the spleen of young animals weighs more in proportion to the body weight than it does in older animals. This is a factor that has been reported clini- cally by Gross (8) in a statistical study of the spleen. On two male dogs from the same litter, one splenectomized and one a laparotomy control, operated by Doctor Luckhardt at four weeks of 562 EFFECT OF SPLENECTOMY ON GROWTH 563 age, the weight curves were found to differ. The splenectomized animal gained in weight but not in the same proportion as did the con- trol. These dogs suggested a difference in metabolism as judged from the weight curves. These dogs were very kindly given to me by Doctor Luckhardt! at four weeks after operation for further observa- tions. (Vide infra). Removal of the spleen in adult animals presents results that differ, yet suggest that the spleen may have an influence in metabolism. Richet (4) reports in his metabolism experiments that splenectomized dogs must eat a much greater quantity to maintain the same nutrition as normal dogs. Spitta and Mayo (5) report an increase in weight following splenectomy. Paton (6) concluded that the spleen had no influence upon metabolism. Goldschmidt and Pearce (7) although differing from Paton on his blood studies, concluded that there was no effect upon metabolism, but observed that splenectomized dogs had a tendency to become obese. From observations in this laboratory upon adult splenectomized dogs it has been noticed that for a time following splenectomized dogs were emaciated but later became quite fat. The possibility that the spleen assumed a greater functional impor- tance in the young than in the adult animal as suggested by the pro- portionate differences in the weight of the organ at these two periods of life, as well as the contradictory results in metabolism of splenec- tomized dogs, led to this question: What effect has the removal of the spleen upon the metabolism of young animals as judged from the growth and weight curves? METHODS Operative procedure. In this work rats and rabbits have been the principal animals upon which the work has been done. Observations were made upon the dogs already mentioned and upon two series of kittens. Unfortunately distemper and accidents prevented further data on kittens and puppies. The rats and rabbits were operated at fourteen days of age under ether anesthesia. Of the cats recorded, one series was operated at fourteen days and the other at seven weeks. It was found inadvisable to operate earlier than two weeks, on account of the extreme delicacy and softness of the young tissues. 1The work reported in this paper was done upon the suggestion of Dr. A. B. Luckhardt whose kindly interest and constant help has been most gratefully appreciated by me. 564 SAMUEL CHESTER HENN Removal of the spleen was made through an incision in the left rectus. Curved forceps were used in removing the spleen so that the incision might be small. In rats no hgatures were placed upon the splenic vessels but sight compression by forceps prevented any hemor- rhage. In the rabbits, kittens and puppies silk ligatures were placed upon the splenic vessels. A frog board makes an excellent operating table for the rats. That the operative procedure might be nearly comparable in the splenectomized and control animals, the controls were kept under ether the same length of time and the viscera were manipulated for a short time. In rats a few animals were observed which were unoper- ated controls. This was done as a check upon the effects of the operative procedure. Diet. The diet of rats consisted of oats, bread, milk, carrots and, when obtainable, lettuce or cabbage. At intervals of two or three weeks raw meat was fed. The diet of rabbits was that of the rats save that bread, milk and meat were not fed. The diet of the kittens was table scraps, bread and milk, and meat. The dogs were fed a diet similar to cats without the addition of milk. Weighing. The animals were weighed just before operation and at weekly intervals prior to the time of feeding. RESULTS That individual variations might not lead to prejudiced or erroneous conclusions, composite weight curves were made of the rats and the rabbits. Rats. In figure 1 is given the composite weight growth curve of five male operated controls and eight male splenectomized rats over a period of twenty-three weeks. Of these rats there is in table 1 a record of one typical set. In this table there is also a record of one unoperated control. This record is typical of all unoperated animals. The com- posite curves show that there is very little difference between the splenectomized and operated animal controls, the difference never exceeding 10 grams and this in favor of the splenectomized animals. Figure 2 represents the composite weight curves of nine female sple- nectomized and eight female operated control rats’ over a period of twenty-three weeks. A typical set from this series is presented in table 2. As in figure 1, it may be seen that there is no appreciable difference in the weight curves. Here again the splenectomized ani- EFFECT OF SPLENECTOMY ON GROWTH 565 fs ee Pas aa O©r23%5 6789 Wl 12 13 4% 1516 17 [819 20 27 22.23 WEEKS. Fig. 1. Composite weight curve of male rats. Five operated controls——-—; eight splenectomized........ Operated at 0 when fourteen days of age and carried through a period of twenty-three weeks. TABLE 1 Male rats Three rats: no. 12, an unoperated control; no. 13, a splenectomized animal; no. 14, an operated control. The latter two animals were operated at fourteen days of age, and have their weights included in those of the composite curves of figure 1. The weights of the unoperated control were begun at fourteen days of age, and with those weights of the other two animals were carried through a period of twenty-four weeks. DATE OF OPERATION 2/4/19) 2/12 | 2/18 | 2/25 | 3/4 | 3/11 fen. Unoperated’ control ....2.25..........), 17 24 | 30 37 | 44] 49 Pees ESDIEDECTOMIZED «6. i... ee ww a as 17 24 | 31 45 | 58] 68 io an@perated,control. ..o:... 0... 3-2. 18 25 | 34.5) 50| 59] 66 3/18 | 3/25 4/1 4/8 | 4/15 | 4/22 |. 4/29 5/6 | 5/13 eee se oe | 62 | 70, | $2.5) 98.08 117 | 120 | 127 Usp lo rae ot sehr ce cee oe ee Det 74 | 85 | 92 {112 | 137 |144 | 147 | 154 | 160 eee re S272 1 83. 92 110 130 |137 145 | 153 | 162 ee a, Stee ~ fences | 129) | 284 | 140-1137 oI Si Se” TO.) 145.145 ert. cova oben dc +e) LOD 1 162 | 1081169) | Te Weg. | TS, 1s. | 18s ee ee . . c| 162 | 168: | 165 NTA. | Vee VOD 8 16s All weights are in grams. 566 SAMUEL CHESTER HENN Pop eae A EATERS = ae Feveey oo a PoP ee eee Or 242434 & © T B F 10 4% 13 14 15 16 14 16 19 20 21 22 23 WeEEexs Fig. 2. Composite weight curve of female rats. Eight controls , nine splenectomized........ Operated at 0 when fourteen days of age and carried through a period of twenty-three weeks. TABLE 2 Female rats Two rats: no. 27, an operated control; no. 30, a splenectomized animal. These animals were operated at fourteen days of age, have their weights included in the composite weight curves of figure 2, and were carried through a period of twenty-four weeks. DATE OF OPERATION 2/8/19] 2/15 | 2/22 | 3/1 | 3/8 | 3/15 Rat 27... Operated control. 7...22..- see seek cee 11 | 18.5] 22.5] 26.5) 29 33 Rat 30. , .Splenectominediet : 3.2 ......2.-2.:--- 13°| 21. | 30, | sa42 46 3/22 | 3/29 | 4/5 | 4/12 | 4/19 | 4/26 | 5/3 | 5/10 | 5/17 RB 2T so. cies che lee le since sa] O98) S40nl” S10) 58) 160 el tosmnae 74 | 72 Rat 30 o...:5 2. <6 «05's warm sees elt MOO S| aAOTE |, ON asO4S| (Syenal as ae ee 97 | 99 5/24 | 5/381 | 6/7 | 6/14 | 6/21 | 6/28 | 7/4 | 7/11 | 7/18 ROU Ql acces 6c. 50 esc ecu see eele 20 Fol | @FOU) SOU Se aameeeee eres 98 | 104 PURO een Ponti aio a's stole teens 109 | 115 | 109 | 109 109 {114 {112 | 111 | 112 All weights are in grams. © EFFECT OF SPLENECTOMY ON GROWTH 567 mals have an average that exceeds the operated controls during the period of growth prior to maturity. It may be seen that the female rats gain in weight at a rate equivalent to the gains made by the male rats until they are about fifty days of age, at which time the male rats gain at a slightly faster rate. This is the normal manifestation shown by normal unoperated rats as observed by Donaldson, Dunn and Watson (8). In two litters of rats weight records of one female and three male unoperated rats were made as a control upon operative procedure. rt Hac la Fa mle ei 2 SaaS eres a eb | ah eo PE aed a il Na aa Ms a ae he aa ke , aa fe LE Ls ra | 927234 567 &9 100 % 13 HIS 1617 18 319.20, 21 22.23 Fig. 3. Composite weight curve of all rats. Five male and eight female oper- ated controls ; eight male and nine female splenectomized........ Oper- ated at 0 when fourteen days of age and carried through a period of twenty- three weeks. The weights were below those of the operated animals, significant of the negligible influence that operation had upon growth of the young animals. In figure 3 there is a composite weight growth curve of all splenec- tomized and operated control rats (male and female). In this graph it is evident that the weight curves are almost identical during the period of growth prior to maturity. 568 SAMUEL CHESTER HENN Rabbits. Figure 4 contains the composite weight growth curves of four splenectomized and four operated control male rabbits. In table 3 the weights of a typical set of animals from this group and also the weights of an unoperated control are recorded. In figure 5 there are the composite weight growth curves of four splenectomized and three gperated control female rabbits. In table 4 a typical set from these series is recorded. [Ea ee eS Kvleata Sees bea ised ie ae Pa ea aa loo 012345 67 8 9 10 1I2Weeus. Fig. 4. Male rabbits. Composite curve of four operated controls and four splenectomized rabbits........ Operated at 0 when fourteen days of age and carried through a period of twelve weeks. It may be noticed that the general average of the male operated con- trols is above that of the splenectomized male rabbits and that the average of the female splenectomized rabbits is above that of the female operated controls. The differences are too small, considering the difference in their weights at the beginning of the records, to attempt an explanation through any effect the spleen may have upon metabolism. EFFECT OF SPLENECTOMY ON GROWTH 569 TABLE 3 Male rabbits Three rabbits: no. 5, a splenectomized animal; no. 6, an operated control, and no. 7, an unoperated control. The operated animals were fourteen days of age at time of operation and have their weights included in the composite weight curves of figure 4. The unoperated control was weighed at fourteen days of age, and thereafter weights were taken each week as were those of the operated animals. DATE OF OPERATION 1/24-19| 1/28 2/6 2/12 2/19 2/22 3/5 Rabbit 5. Splenectomized....... 144 |} 156] 227} 300] 290; 340] 360 Rabbit 6. Operated control..... 174 | 195 | 284] 310] 357 | 352] 390 Rabbit 7. Unoperated control...| 145 | 202 | 292] 365 | 359] 342] 375 All weights are in grams. ISS ee Se re ects |e |e G2 BEES Pe ie ral a) VA A i a led z 4 ea ae fe La ail ins Fig. 5. Female rabbits. Composite curves of three operated controls ——— and four splenectomized........ Operated at 0 when fourteen days of age and carried through a period of twelve weeks. 570 SAMUEL CHESTER HENN In figure 6 are given the composite weight growth curves of all the male and female rabbits. Here it is evident that there is practically no difference in the growth curves of the splenectomized and operated control animals, especially when it is taken into consideration that the average of the operated controls was 20 grams above the weight of the splenectomized animals at the beginning of the records. Kittens. In figure 7 there is the record of the weights, over a period of nine weeks, of kittens operated at fourteen days of age. Their rec- ord although irregular has the same general tendency as with the rats TABLE 4 Female rabbits Two rabbits: no. 11, a splenectomized animal; no. 9, an operated control. These animals were operated at fourteen days of age, have their weights included in the composite weight curves of figure 5, and were weighed through a period of thirteen weeks. DATE OF OPERATION 3/9/19 | 3/15 3/22 3/29 4/5 4/12 Rabbit 11. Splenectomized............| 155 | 182 | 239 | 319} 382) 512 Rabbit 9. Operated control.......... 136 | 163 | 224] 247 |} 299] 330 Rabbit iets nae pete oe nee Se 506 | 602 | 614] 634] 507 | 672) 725 Rabbit, 9... 2228% ssoc.cbexines aor e 307 | 346 | 425 | 420} 443 | 465] 470 All weights are in grams. and the rabbits, that there is a similar gain manifested by splenec- tomized animals as with the operated control animals. In other kittens operated at seven weeks of age the same relative curves were observed in the splenectomized animals as in the operated control kittens.” ? Of the first litter of two kittens the splenectomized kitten did not gain in weight as did its control mate. It died about the tenth week weighing 260 grams. Its mate at that time weighed 800 grams. At the time of operation each weighed about 150 grams. Necropsy revealed congenital (?) absence of the thymus of the splenectomized animal. It is worthy of note that the splenectomized animal possessed adrenal glands which were twice as large and weighed almost twice as much as the adrenal of the control animal which was sacrificed for examination on the day of death of the other.—A. B. L. EFFECT OF SPLENECTOMY ON GROWTH aval! Sees aS 6 TH9 Ween Fig. 6. Composite curves of all rabbits. Four male and three female con- trols ———; four male and four female splenectomized Boe Reet Delo oe © 12345 678 9 tO 4b 1213 Weexs. Fig. 7. Weight curve of kittens. One male control ; one male splen- ectomy ; one female splenectomy........ Operated at 0 when fourteen days of age and carried through a period of thirteen weeks. Operated at 0 when twenty- COMIZeCs .. nase . 4 ———; male splenec Male operated control Fig. 8. Weight curve of dogs. eight days of age and carried through a period of seventy weeks. EFFECT OF SPLENECTOMY ON GROWTH 573 Puppies. Figure 8 of the dogs over a period of seventy weeks after operation at four weeks of age shows a marked difference in weight, but failure of others to withstand distemper after operation, and in light of former work on puppies (Tachigara and Takayi) there is insuffi- cient data upon which any conclusions might be drawn in connection with growth. During the month following the operation these dogs _ had severe attacks of distemper. In addition to distemper the sple- nectomized animal had severe gastro-intestinal disturbances as was evidenced by vomiting and diarrhea. These factors may have had an inhibitory effect upon growth. OTHER OBSERVED EFFECTS OF SPLENECTOMY IN THE YOUNG Blood changes Dogs. As this problem did not primarily deal with the blood pic- ture, complete records have not been made. Blood counts were made on the dogs seven months after splenectomy. The splenectomized animal showed a general leucocytosis of 23,000 and moderate anemia (6,000,000 erythrocytes) compared to the counts made on the blood of control dog, which averaged 11,000 leucocytes and 8,600,000 erythro- cytes. The same pipettes were used in making the counts on the two animals. The differential counts of blood cells from both animals gave percentages of cell types nearly equal. a. Resistance to hypotonic sodium chloride solutions. Dogs. The blood corpuscles of splenectomized animals were found to be more resistant to hypotonic sodium chloride solutions than were the cor- puscles of normal dogs. Following the method of Karsner and Pearce (10) 0.1 cc. of cor- puscular mass was placed in each of a series of tubes containing 3.0 ce. of sodium chloride solutions varying from 0.1 per cent to 0.5 per cent by differences of 0.025 per cent. Readings were made at the time the corpuscles were mixed with the salt solutions and eighteen hours later, having been kept in a refrigerator near ice. Corpuscles from two male dogs, ten weeks after being splenectomized at four months of age, gave 100 per cent hemolysis at 0.4 per cent and all solutions of lower percentage, with only a trace at 0.5 per cent. From a female dog, nine months after splenectomy at ten weeks of age, and from splenec- tomized male of figure 8, there was practically no hemolysis in solutions above 0.25 per cent, with no trace of hemolysis above 0.4 per cent. In four animals unoperated and varying in age from ten months to eigh- 574 SAMUEL CHESTER HENN teen months of age, there was complete hemolysis at 0.35 per cent and below, with no trace of hemolysis above 0.45 per cent. These results are in keeping with the observations made by Karsner and Pearce. They report similar results as having been observed by Bottazzi, Banti, Pugliese and Luzzati, and others. TABLE 5 In this table there is a record of the coagulation time of the blood from three dogs. No. 1 was a male operated control, operated at four weeks of age and seventy-five weeks of age at this time. No. 2 was a male dog, splenectomized at four weeks of age, now seventy-five weeks of age. No. 3 was a female dog, splen- ectomized at ten weeks of age and now fifty weeks of age. FIRST EVIDENCE TIME OF SAMPLES ,. BLOOD DRAWN mie noe FIRM CLOT COAGULATION bie minutes ( 1 9:38 9:42 9:45 7.0 | 2 9:38 9:42 9:45.5 135 | 3 9:38 9:42.25 9:46 8.0 Drow aitis: 654% { 4 9:38 9:43 9:46 8.0 | 5 9:38 9:43 9:46 8.0 Average timMe...<..ckeehies tae See Cee ee Uff 1 10:09 As 10:10 1.0 2 10:09 * 10:10 1.0 3) 10:09 “8 10:10 1.0 Dog 2 ee 4 10:09 8 10:10 1.0 5 10.09 33 10:10 1.0 Average timeless than i’). 3. 2. (32.5. ese rg 1 2205 2:56 2357.5 2.5 2 Doe 2:56 2:57 2.0 3 2755 2:56 2:58 3.0 DO rte sar 4 2500 2:56 2:58 3.0 5 PARI) 2:56 2:58 3.0 Asyerage times). xcs tonw he eee ee Pers Rabbits. Corpuscles from a female rabbit, ten months after sple- nectomy at two weeks of age, were found to be no more resistant to hypotonic sodium chloride solutions than were the corpuscles of an operated control male rabbit of the same age. EFFECT OF SPLENECTOMY ON GROWTH 575 b. Coagulation time. Dogs. In drawing blood from the ear and saphenous veins, the blood from splenectomized animals clotted much quicker than that from operated controls and normal animals. In table 5 the decided difference may be seen between the time of coagula- tion of blood in an animal splenectomized early, one splenectomized later, and the blood from a normal operated control animal. Upon centrifuged blood of dogs of figure 8, it was found that cor- puscles were decreased and serum increased in the splenectomized animal. The percentages were: corpuscles, 31.12 per cent; serum, 68.87 per cent; while in the control there was 58.66 per cent corpuscles and 41.33 per cent serum. Rats. In seven male controls, seven male splenectomized, three female controls and four female splenectomized rats, there was no detectable difference in the time of the coagulation of their blood. Two samples of 1 cc. each were taken from the tail and the heart. Coagulation of the blood occurred in one minute. Rabbits. In rabbits the coagulation time was determined of the same two as are recorded under “resistance to hypotonic solutions.” The time of coagulation in five tests was found to be the same for each animal. These two animals lived longer than those that are recorded in the graphs. Appearance and temperament Rats. No difference in appearance and temperament could be observed between the splenectomized animals, the operated controls and the unoperated controls. Other people in being allowed to examine the rats and questioned as to what difference they might observe, could not differentiate between splenectomized and control animals. Kittens. In the kittens a difference in temperament but not in appearance was observed. The splenectomized animals were more active and pugnacious, but as cats are very emotional animals no particular importance was attached to this difference. Dogs. In appearance the splenectomized animal was emaciated, although fed an equal amount of food as the control. In temperament -and activity both dogs responded alike. The splenectomized animal contracted mange at two different times while the operated control was resistant, while living under the same conditions. Or ~I [op] SAMUEL CHESTER HENN . Fertility Rats. Two female splenectomized rats gave birth to litters of rats. The male in each instance had been completely splenectomized. The mothers cared for their offspring in a manner comparable to normal rat mothers. Rabbits. A female splenectomized rabbit gave birth to a litter of rabbits. The male was an operated control. The young rabbits were well cared for. Dogs. The male splenectomized dog of figure 8 was known to have copulated with a female animal which had been splenectomized at ten weeks of age. Two weeks later the mammary glands of the female became noticeably enlarged simultaneously with which there was a fal] in the leucocytic count from 25,000 to 18,000. Two weeks later the sw elling of the mammary glands disappeared and the leucocytic count rose to 29,000. No pups were ever delivered. Histological examina- tion of the male’s testes showed a very active spermatogenesis. Macroscopical and microscopical examination of tissues Rats. Macroscopic examination of the rats revealed no changes of especial note save that in most of the splenectomized animals there was red bone marrow in the femur. There were several of the control animals that showed traces of red in the femur, so that this finding was not limited to the femur of the splenectomized animals alone. In three of the splenectomized animals small bits of splenic tissue were found. It was interesting to note that these animals did not grow as rapidly as did the animals that had complete removal of the spleen or were only control animals. Microscopically the only difference that was detected was in the bone marrow. The bone marrow of the splenectomized rat was very rich in cellular elements composed principally of myelocytes, erthro- cytes, mononuclear cells and a few multinucleated cells. The fat content composed about one-third of the marrow. The bone marrow from the operated control was very cellular, composed of cells very similar to those of the splenectomized animal. Two-thirds of the marrow was composed of fat. Rabbits. Macroscopic examination revealed a liver in the splenec- tomized female that weighed more than in the male control although the control weighed more than the splenectomized animal. The EFFECT OF SPLENECTOMY ON GROWTH 577 difference in body weight was the normal difference between a male and a female. The marrow of the femur of the splenectomized animal was streaked with red, while it was distinctly yellow in the control. Microscopically the liver of the splenectomized animal was marked with many Kupffer cells containing erythrocytes in different stages of disintegration. The bone marrow was composed of one-third fat. The remaining portion consists of many multinucleated lymphoid cells, myeloblasts, mononuclear lymphoid cells, nucleated erythrocytes, erythrocytes and a few endothelial and giant cells. The mononu- cleated lymphoid cells appear very much like splenic cells. The liver of the operated control was moderately cirrhosed (probably explaining difference in weight), and contained a few Kupffer cells in which there were erythrocytes. The bone marrow was composed of nearly all fat, with a few areas near the periphery containing a few lymphoid cells and myelocytes. Kittens. Macroscopic examination revealed no observable changes. The bone marrow was distinctly yellow in the femur of the splenec- tomized animal as it was also in the control. No microscopic exami- nation was made. Dogs. Splenectomized animal: Macroscopically the most marked changes appeared in the lymph glands, bone marrow and in the weight of the liver which was 30.9 per cent of the body weight. The mesen- teric and retroperitoneal lymph glands were distinctly red in color. The inguinal, axillary and popliteal glands were enlarged and white and firm in consistency. The intestine gave evidence of only one extremely atrophied Peyer’s patch. The bone marrow was distinctly red in color and fairly firm in consistency. Microscopically the liver contained many Kupffer cells in which there were many erythrocytes and a few leucocytes. The atrophied Peyer’s patch was the seat of very little lymphoid tissue in the submucous layer. The mesenteric lymph glands in the periphery contained regularly arranged whorls of lymphoid tissue. In the medullary portion there were smaller and less regular whorls of the same tissue. There were marked sinuses in this portion containing few lymphoid cells, and many endothelial cells packed with erythrocytes, and a few leucocytes.: There were also many multinucleated lymphoid cells present in these sinuses. The inguinal lymph glands were composed of more irregular lymphoid tissue than that in the mesenteric glands. There were sev- eral endothelial cells containing erythrocytes and there were many more multinuclear lymphoid cells present than were in the mesenteric THE AMERICAN JOURNAL OF PHYSIOLOGY, VOL. 52, No. 3 578 SAMUEL CHESTER HENN glands. The bone marrow was densely packed with cellular material interposed with small amounts of fat. The cellular element was composed of erythroblasts, erythrocytes, myeloblasts, mono- and multi-nucleated lymphoid cells and a few giant cells. The testes showed a very active spermatogenesis. Operated control: The lymph glands were not enlarged, the bone marrow was distinctly yellow in color and the liver composed 23.2 per cent of the body weight. . Microscopically the lymph glands were rather compact and con- sisted of mononucleated lymphoid cells. The liver contained Kupffer cells but not to the marked degree that they were found in the liver of the splenectomized animal. The bone marrow was composed of a great amount of fat with very few mononuclear cells and myeloblasts interposed. The testes appeared very similar to those of the splenec- tomized animal both in structure and in showing a very active spermat- ogenesis. DISCUSSION In this work there are results from rats and rabbits in numbers large enough to permit of conclusion of the effect of splenectomy upon growth. It would not be justifiable in drawing conclusions upon the two series of kittens, although their weight curves are similar to the curves of the rats and the rabbits. In the dogs there were early influ- ences of distemper and gastro-intestinal disturbances that present a factor that was not ruled out by other animals. However from the carefully prepared protocols of Tachigara and Takayi it is suggested that puppies would present weight curves similar to the curves of the rats and rabbits. Mann, in speaking of rabbits, dogs and kids, says: “Our results show that some animals from which the spleen has been removed quite early in life develop in the same fashion as the unop- erated controls; furthermore, no definite specific change in the de- velopment due to the loss of the spleen was observed in any of the animals” (10). The resistance shown by the corpuscles against hypotonic sodium chloride solutions is found to be true of animals operated while young as well as adult splenectomized animals. This phenomenon suggests that the spleen may have the function of preparing the erythrocytes for destruction. The quickened coagulation time of the blood from splenectomized dogs may be a protective means against loss of blood by hemorrhage EFFECT OF SPLENECTOMY ON GROWTH 579 during the time of anemia. The results would suggest that the earlier an animal is splenectomized, the more rapid is the time of coagulation. It may be due to the general leucocytosis or to an increased amount of fibrinogen which may account for the large amount of fibrin secured when the blood from splenectomized dogs is defibrinated. Increased fibrinogen may be due to a greater production by the liver in the absence of the spleen. No other mention of coagulation time of blood from splenectomized animals was found in a review of the literature save by T. G. Orr (11) working on rabbits. He reports that in twelve splenectomized rabbits there was no change in coagulation time as compared to normal rabbits. Coagulation time of the blood from rats and rabbits was found to yield no important differences in our own series. In activity and temperament no difference was observed in rats, rabbits or dogs, but a slight difference in kittens was observed. This may be due to some individual variation rather than due to any Son following removal of the spleen. Loss of fertility was suggested by the female splenectomized dow’ s failure in bearing any pups; but the active spermatogenesis of the male’s testes and the swelling of the mammary glands at an early stage after copulation indicates that some other factor during early preg- nancy may have had an inhibitory influence upon the development of any embryos. In rabbits it was clear that the female was not sterile. In rats it was evident that neither male nor female was sterile. Histological examination indicates that the function of the spleen is in some way connected with the blood. This is evidenced by large hemolymph glands especially in the mesenteric region, which suggests that if it is due to a compensatory hyperplasia following splenectomy, it is important that such a function should be produced in the portal circulation. The bone marrow changes were rather striking. It may be that in disturbing a portion of the blood apparatus there is a com- pensatory functioning of other portions. The bone marrow through the different types of cells present, indicates that there is a process of erythrogenesis and leucogenesis taking place, while an increase of the Kupffer cells and endothelial cells of lymph glands indicates a process of blood destruction. Changes in lymph glands and in the bone marrow similar to the above have been described by Pearce and Pepper (12) and by A. S. Warthin (13). 580 SAMUEL CHESTER HENN CONCLUSIONS 1. The spleen is not essential to the life of young rats, rabbits, kit- tens and puppies. 2. Splenectomy has a negligible influence upon growth of rats and rabbits. It is suggested from the weight curves of the kittens that the same is true of these animals. Results of Tachigara and Takayi and Mann would indicate the same as being true of dogs. 3. Operative procedure has no debilitating effect upon the growth of young rats. 4. In dogs splenectomized while young there is a hastening of the coagulation time of blood. The younger the animals have their spleens removed, the more rapid is the coagulation time of the blood. 5. Corpuscles of young splenectomized dogs have the same property of more strongly resisting hypotonic sodium chloride solutions, than control and operated control animals, as do animals splenectomized after having reached maturity. 6. Accompanying the removal of the spleen there is an increase of Kupffer cells in the liver and changes in lymph glands and bone mar- row that indicate a compensatory function on the part of these tissues in supplying a missing factor in blood formation and destruction. 7. Male dogs, male and female rats, and female rabbits are not rendered sterile through the removal of the spleen while these animals are young. BIBLIOGRAPHY (1) WourertaH: Arch. Int. Med., 1917, xix, 105. (2) TacHIGARA AND Takayt: Mitt. a. d. med. Fak. d. Univ. Tokyo 1917, xvii, 563. (Protocols containing weights, 570-577). (3) Gross: Journ. Med. Res., 1919, xxxix, 311. (4) Ricaet: Journ. d. Phys. et d. Path. gen. 1912, xiv, 689. (5) Sprrra anp Mayo: Not available. Quoted from ‘‘The spleen and anemia,’’ Krumbhaar, Pearce and Frazier, 1918. (6) Patron: Journ. Exper. Med., 1915, xxii. (7) GoLtpscHMIpT AND PEarce: Journ. Exper. Med., 1915, xxii, 319. (8) Donatpson, DuNN anp Watson: Memoirs, Wistar Inst. Anat. and Biol., no. 6. (9) Karsner anD Pearce: Journ. Exper. Med., 1912, xvi, 769. (10) Mann: Endocrinology, 1919, iii, 303. (11) Orr: Journ. Lab. Clin. Med., 1917, ii, 895. (12) Pearce AND Prepper: Journ. Exper. Med., 1914, xx, 19. (13) Warruin: Journ. Med. Res., N. S., ii, 453. THE. IMPORTANCE OF VAGAL AND SPLANCHNIC AFFERENT IMPULSES ON THE ONSET AND COURSE OF TETANIA PARATHYROPRIVA WALTER LINCOLN PALMER From the Hull Laboratory of Physiology, The University of Chicago Received for publication April 19, 1920 INTRODUCTION The object of this work was to determine, if possible, whether the afferent impulses which reach the central nervous system through the vagi and the splanchnic nerves have any influence on the onset and course of parathyroid tetany. That these might be of some import- ance was suggested to us by the great frequency with which tetany and gastro-intestinal disorders are associated clinically. Thus Osler (1) in quoting Frankl-Hochwart on the etiology of tetany in adults de- scribes the “‘tetany of gastric and intestinal disorders, as dyspepsia, gastrectasis, diarrhoea, and helminthiasis,’’ and again says that in children ‘‘tetany bears a definite relation to gastro-intestinal disorders.” Paton, Findlay and Watson (2) mention this relationship in their clas- sic on the parathyroids but do not seem to attach any significance to it. McCann (3) attributes the tetany to the alkalosis resulting from the stagnation of the acid secretion in the stomach but Osler (op. cit.) says that “‘the type associated with dilatation of the stomach is rare, only thirty instances being reported” and hence this would not explain the great majority of cases. Ivy (4),from work done in this laboratory, reports that in eight dogs with duodenal fistulas which allowed every- thing the dogs drank together with the gastric juice to drain to the out- side, only one developed tetany. This animal was advanced in pregnancy and the tetany he attributed to eclampsia of pregnancy. However Rodman (5), Bene (6) and others consider gastric ulcer and dilatation to be more closely associated with tetany. Oppenheim (7) in discussing tetany says: Adults who suffer from gastro-intestinal affections are sometimes affected. . It has not been definitely decided whether the toxic agent is the cause of the spasm or whether the spasm has a reflex origin. The fact that it has occa- 581 582 WALTER LINCOLN PALMER sionally occurred directly after the use of a stomach pump or the passing of a tube or after profuse vomiting seems to point to a reflex origin, and the observation— which cannot be implicitly relied upon—that the spasm has disappeared after the evacuation of entozoa confirms this theory. It is remarkable that Gerhardt was able in a case of tetany with dilatation of the stomach to bring on the spasm by slight percussion of the stomach. In this laboratory it has been frequently observed that an attack of tetany may be precipitated in parathyroidectomized animals by the passage of a stomach tube, other sensory stimuli also having the same effect. Carlson and Jacobson (8) and later Friedman (9) also recorded the great frequency with which hyperemia, hemorrhages and ulcers of the stomach and intestinal mucosa are found in the post-mortem examina- tion of parathyroidectomized animals and were reminded of the rela- tionship between gastro-intestinal disorders and tetany in man, stating further that ‘‘it is not known to what extent this tetany is due to toxic products absorbed into the blood or to irritation of afferent nerve end- ings, but in late years numerous attempts have been made to bring all types of clinical tetany within the category of parathyroid tetany.” A priori we would reason that if irritation of the afferent nerve endings in the gastro-intestinal tract is ever of importance in tetany, it should be in that of parathyroid origin for in this condition we have not only these alimentary disorders but also a hyper-irritable condition of both the central nervous system and the peripheral nerves, as has been shown most completely by Paton, Findlay and Watson (op. cit.). Before undertaking this work it was thought that examination of the literature upon the effect of section of the cord at different levels might throw some light upon the subject but the work of Bied] (11), Falta and Rudinger (11), Paton, Findlay and Watson (2), Mustard (12), Carlson and Jac- obson (8), and of Luckhardt, Sherman and Serbin (13), is at such vari- ance that no conclusions on this point can be drawn. EXPERIMENTAL METHODS Two methods of procedure have been followed in these experiments. The first consisted in the determination of the effects of double vago- tomy and splanchnectomy on the onset and course of parathyroid tetany and the second in the determination of the effect of gastro-intes- tinal irritation by means of a drastic, i.e., croton oil, upon the onset and course of parathyroid tetany. It was recognized of course that this artificial irritation is not strictly analogous to the normal irritative condition. Dogs were used exclusively in the work. RELATIONSHIP OF NERVOUS IRRITATION TO TETANY 583 The technique of the first method was divided into three stages. In the first operation the thorax was entered on the left side between the sixth and the seventh ribs. Here the thoracic sympathetic trunk was picked up and all the rami communicantes torn loose in the lower part of the thorax, care being taken to destroy all threads of nerve tis- sue running from the rami communicantes to the left splanchnic. This nerve trunk was then grasped with the fingers and jerked out from the diaphragm. After thus getting rid of the left splanchnic, the eso- phagus was picked up and the vagi, which are found lying on each side at this level, were sectioned. Care was taken here to make sure that all of the vagus fibers were cut. The closing of the thorax completed the operation. Ether anesthesia alone was used, no morphine being employed because of its depressing effect upon the respiratory center. Artificial respiration was unnecessary unless the thorax was allowed to remain open for too long a time, in which case asphyxiation resulted. After a week or so when this wound was found to be well healed and the animal in good condition, the second operation was performed un- der morphine-ether anesthesia. This consisted in entering the abdomi- nal cavity through a right lumbar incision and picking up the right splanchnic just after it pierces the diaphragm. After dissecting this down to the point where it branches out to enter the coeliac plexus, the nerve and ganglion were removed and the abdomen closed. This op- eration is not so depressing as has been supposed, the dogs appearing lively and in good condition within a few hours after the operation and usually live in the best of health, apparently suffering no ill effects at all. The success of the two operations was determined at the subse~ quent post-mortem examination. After a week or ten days had again been allowed to elapse, a complete thyro-parathyroidectomy was performed, the glands being removed grossly with as much of the surrounding loose tissue as possible in order to avoid leaving behind any accessory parathyroid tissue in that region. The animals were kept under the same conditions as before the opera- tion and on the same diet (bread and meat). The electrical excitability of the femoral nerve was determined before the final operation and the change afterwards followed although it was expected from the work of Paton, Findlay and Watson that double vagotomy and splanchnec- tomy would have no effect upon the change in the electrical excitability of the peripheral nerves even though it did have an effect upon the tetany. The technique of the second method consisted also in complete thyro-parathyroidectomy in the manner just described, followed by the 584 WALTER LINCOLN PALMER daily administration of croton oil in varying amounts. The U.S. P. dose of this drug is 2 minims and hence the 2 to 4 minims given daily to the dogs produced very severe irritation. The animals were kept under the same conditions as those of the first series and in most cases the electrical excitability was followed here also. RESULTS The animals developed distemper so frequently that this proved to be quite a troublesome factor, especially in the first series. In a large pro- portion of the cases autopsy showed a terminal infection with pneu- monia. It is quite conceivable that this depressing condition might alter the course of the tetany somewhat but its effect was probably similar in the two series. Effect of double vagotomy and splanchnectomy. Table 1 gives a sum- mary of the results obtained in the first series of experiments. Nine of the sixteen animals developed tetany, five showed depression only and two are still living. The failure of these last two animals, dogs 48 and 52, toshow any symptom is probably to be attributed to the presence of accessory parathyroid tissue. Excluding these from consideration, nine out of fourteen animals, or 64 per cent, developed tetany and five showed depression only. In the series of thyro-parathyroidectomies on dogs reported by Paton and Findlay, twenty-six out of forty-two, or 61 per cent, showed tetany. The severity of the tetanic attacks seemed to be as great in these dogs as in normal parathyroidectomized dogs and even greater than in the dogs of series 2. The protocol of the dog showing the most pronounced tetany in this series is given in table 3. It will be noticed that this dog vomited frequently, which point is also of interest since It points to direct central stimulation as the cause of the vomiting. Effect of artificial gastro-tntestinal irritation. Table 2 gives a sum- mary of the results obtained in the second series of experiments. Here eleven out of fourteen dogs, or 78 per cent, showed tetany and three depression only. Although this is a somewhat higher percentage than that of series 1, it is scarcely of significance, especially in view of the fact that the severity of the tetany in these animals did not seem to us to be quite so great nor the attacks so prolonged as in those of the first series. The protocol of the animal which showed the greatest tetany in this group is given in table 4. Although this animal showed pneu- monia at autopsy while that of the first protocol did not, it lived longer in spite of the pneumonia and hence the comparison between the two is a rather fair one and illustrates the similarity between the two series. 585 RELATIONSHIP OF NERVOUS IRRITATION TO TETANY ‘ougAos+ ++ 00144 {pyr++omy {yy 81[s + ouo formyoue;ds— ds ‘youssq—aq ‘ eiume1adhy—sz9d Ay fesey11omey— wey { worywinp—anp { ApITIGVqyIOXxe [worIyooTo— "TTY { UoIssordap— dap { AyIaAVS—"IOAVS (SYOV}}V—"44R 2970 esoonul d11jSVs Ul ‘Wey ‘ajye;duos ‘1edo Yyog aur} -sojur Ul “Mey feje;dulod ‘1edo yo, qoeyur vult *[ds “41 !vrmuouMeUg BsOonul d11ySvs “10d -Ay foyotduioo ‘1ado yyoq SeruounoUug unueponp ul ‘way {oj0[du109 ‘19do yyoq ‘eluoUNeUg ayetduioo ‘1edo YyJOq ‘eruOoUMeUg SOUT}SO}UL UL “Wey ‘qyoejur [ds 44 JO pvoryy !evrmouneug qoequr vult *[ds “4y7 soul} -sojul Ul “Wey feyetdui0o ‘1edo yWOgG eke) (6 hoo Coys) ‘ado yyoq ‘esoonur o1myses ‘zadApy SOUTJSOYUL UT “WIOY :yoeyur ‘[ds qjoyl “1q ‘eruoumnoeud 4yyYysI[9 yowyur *[ds -4j[ “aq feruounoug ayo[durioo ‘19do yyoq feruoumnoug qowyur “[ds “44 “Iq ‘evIuOuINneUg SUIAYT Ud4JO SJIULO A. SUIAVT iodui04 -stp ‘yots Alo A radu1e4 *STUIO A AyUO S1OUI91} YYSITS Aouvusoi1d poouvaApy -sIp visoyd -ous {ATsnon -UIJUOD SUBOLY) 4sd0Lov SMUVNAA OU Wee celle cies lcip ates leet 8 Sciag |lamame |ierst «fuses aes PL L = + 0 0 0 0 0 0 ) 0 g ap tae) 0 0 oD Saisie |S Spo absinthe 09 | eg lo | O Gin liaiecie latencies alec Olee ligt 2 Bact apa gen ae ll |) Pas were dipsee a) cemna £Er 1G ee ie eee w lh (as 9% FO Joes ete ateite testa) ME al ae Weta ta Or lee tl = Ae 0 0 0 9 ye ay Scie Pera Gol shop Sunoy ara aa “dad “UCAS ee LASNoO fiunja} prowhiyyosnd uo fimojsouyounjds pup fiuojoboa ajqnop fo yall ay see eun eoecceees ISBIOAY or/9 | 2/9| 92¢/¢] 2 o1/9 | 29] 9¢/G] TS z1/9 | #/9]| 92/¢ | Og F1/9 | 22/9 | F/G | 6F 91/9 | 18/¢ | F2/9 | SP o1/9 | 22/S | O*/S | LF o1/9 | g2/¢ | 61/S | OF F1/9 | ge/S | 21/9 | or/9 | #/91 SI/S | OF 8/9 | #2/G | O1/G | 6E 01/9 | S1/¢ | OT/S | se or/9 | 6/S| s/s] 98 TIO. Roel Tige |, ee TH Sore | Tie | Se 11/9 | 18/¢ | O&/% | O8 ez/F | 6I/F | SP] 2 “ugddoO “ugddo "uqddo GUuIHL |GNOOasS | LSU pod PALMER - pt WALTER LINCOL 586 *a19A0S- ---+ 9014} ! pprur-+--+ OMy !4ysI[s-+ ouo fodvys0oMoy—'moy {yuoTIIedxe JO TOTyeinp —'dxo ‘inp :AqpIqeyioxe [RoLyoeja—'W “op Suorssordap—'dop {Ay1WoAOS—"OAOS {syou,ye—"4yR {]10 UO0JOID—TIO "DO 29/07 "SOOT SUIRIZ DAJOMT, , c tT /olpp ays ve ses OBIDAY, OUT}So}UL UL ‘UOY feIMOUMOU ol sear lseararlarararl! ~G tS 6 c27/9 | ST OUIJSO4JUT UL “Woy YYSI[s {vruoWNoUg 792 | ++ /++1/ 0 0 0 6I | 92/9] FI BSOONUL OII4SvVS Ul “Woy {vIUOUIMOUg 8 ++ }/4+4+-4+) ++ T SP OT $2/9 EI BSOONU OLI4SeB Ul “Woy {eIMUOUMOUg } = ee Sear if LE v $2/9 | OL BSOONU OLI4SBS UL “UI8Y ! vIUOUINOUg 76 | ++ |+++] 0 0 0) Sie fe lohon la tn BSOONU O1I4SVS UI “Uloy /vruOUNeUg POO] SqtUIO A 9 ++ )++ ;/ ++ Zz GG 9 81/9 | OT BSOONU OLIYSVS UL “UIOY {RruOUINOUg £e + ++ ais “6 LY OL | SI/9] 6 (i) Aus SULPSOPUL UL “UOFT | -30} Url YPeOd Ga leeteale =e 0 0 0 P 4/9 | 8 BSOONU OIIYSVS UL “Woy SBruOUINUg S}TUIO A. ¢ + + + T O¢ f PON | 972 oUuT}SOJUL UL “Woy {vrUOUMDUg £¢ ae aie z TG 8 c/o} 9 (4) Aue BSOONUL O1LYSVS UL “WUOFT | -309 Ul Yyeoq | 4z + ++ T SP 8 Cale i g oUuTJSOJUL UL “Woy ! vIuOuIMOUg SYTUIO A. p + + I t6 02 | O2/G | F oUulySoyUT UL ‘Woy !eruowMoUd 4yYsITS 1904S “pig 47, + + T SP ZL | O¢/S | § S}LUIO A 8 BSOONU OLIJSVS UL “ULOFT ffooys “pr <6) ++ )+++ VA 6c, lack |OUGh ZS sfiipp sinoy ASdOLAV SUUVNAY Gest bic aac ‘aaa |-ugads| “2Y | ragno rant aN 0G i “und See: ‘ 5 ‘ON “AHL funja) proshiyyound fo hiqisaaas ay) WO UOYD]LLL 1DWYSAIUI-O.SVB yoroyuyip fo walla ¢ A TAVL RELATIONSHIP OF NERVOUS IRRITATION TO TETANY 587 TABLE 3 Protocol of animal 51, series 1 ELECTRICAL REACTION IN MILLIAMPERES DATE HOUR PROCEDURE REMARES KCC | ACC | AOC 6 | 3:00 p.m 1203;|\ 4264: 7 | 11:15 a.m 1.8] 4.8 9] 8:30 p.m. PS -| 278 16 | 11:00 a.m. | 3rd op. 5:00 p.m O98} 26 17 | 10:00 a.m 0:8: [eae 5:00 p.m 0.6). ‘£0 Vomits frequently 18 | 9:00 a.m 0.4‘) 0.6 Vomits frequently 5:00 p.m 0.4} 0.6] 0.6 19 | 10:00 a.m 0.4} 0.5 Vomits frequently 5:00 p.m Slight tremors 8:00 p.m Tremor all evening but not severe 11:15 p.m Mild spastic tetany, particu- larly in forelimbs 20)| 3230 a.m. Mild spastic and clonic tetany all limbs 10:00 a.m Tetany continues, occasion- ally severe 2:00 p.m Mild tetany continues 3:30 p.m Tetany gone, tremors present 5:00 p.m Mild tetany again 9:00 p.m Mild tetany, occasionally se- vere 21 | 9:00 a.m. Tremors 10:00 a.m Mild tetany, at times severe 1:00 p.m Mild tetany continues 5:00 p.m Mild tetany has continued all afternoon frequently being severe 9:00 p.m Depressed 22 | 10:00 a.m Depressed 8:00 p.m Found dead _._ ee Ve ee ee ee eee ee ee ee ee eee Autopsy: 1. No pneumonia 2. Primary and secondary operations had been complete 3. Slight petechial hemorrhages in gastric mucosa 4. Other organs normal 588 WALTER LINCOLN TABLE 4 PALMER Protocol of animal 15, series 2 | | ELECTRICAL REACTION IN MILLIAMPERES REMARKS DATE | HOUR PROCEDURE KCC |} ACC | AOC 6 /21 5:00 p.m. 0.5 ie? 23, | 10:00 a.m. 0.8 1.0 25 9:00 a.m. 0.4 12, 0.8 11:00 a.m. | Thy-par 5:00 p.m. | 2d C. oill: 8:00 p.m. 0.4; 0.8] 0.8 26 | 9:00 a.m. | 2d C. oil Shght tremors 12:00 m. Mild tremors 27 | 9:00 a.m. | 2d C. oil 5:00 p.m. Severe tetany 10:00 p.m. Quiet 28 | 9:00 a.m. | 3d C. oil Severe tetany 9:00 p.m. Tetany 29 | 9:00 a.m. Severe tremors 8:00 p.m. Severe tetany 30 | 1:00 a.m. Mild tetany 10:00 p.m. Alternating severe tremors and mild tetany 4:00 p.m. Severe tremors. Depressed 9:00 p.m. Slight tetany 7/1 | 9:00 a.m. Depressed 12:30 p.m. Mild tetany 10:30 p.m. Mild tetany with intervals of depression 28:30) aeme Depressed. Occasional tre- mors 3 | 9:00 a.m. Very depressed 1:00 p.m. Found dead Autopsy: 1. Extensive pneumonia involving both lungs 2. Hyperemic gastric mucosa with petechial hemorrhages throughout small intestine. 3. Other organs normal DISCUSSION In the first table it will be seen that there is no relationship between the dogs which showed tetany and those in which the primary and sec- ondary operations were incomplete. Thus in dogs 40 and 49 the opera- RELATIONSHIP OF NERVOUS IRRITATION TO TETANY 589 tions were incomplete and the dogs showed no tetany, while in dogs 36, 38, 50 and 51 the operations were complete and yet the dogs showed tetany. Similarly in the second series there seems to be no relationship between the amount of irritation produced and the onset, severity or number of attacks of tetany. Comparison of the two series shows further that although the average length of time before the onset of tetany was somewhat shorter in the second series and the average length of life a little less, the difference is not enough to be significant especially in view of the fact that the severity of the tetany was some- what greater in the first series. The vomiting recorded in the protocol in table 3 and observed also in one other dog in the first series is rather interesting. Carlson and Jacobson observed that in parathyroid tetany “‘the vomiting is fre- quently severe and persistent, even in subjects taking no food.” Since it occurs in animals with both the vagi and the splanchnics sectioned, one must conclude that possibly this vomiting is of central origin. Subcutaneous morphine-atropine imjections preparatory to anesthetiza- tion for the tertiary operation was observed to have a similar effect in these animals. Carlson and Jacobson alse called attention to the groaning with evi- dences of abdominal tenderness so frequently seen in parathyroid tetany dogs. This was shown in our first series by no. 33 especially which was observed to groan continually and although autopsy showed a very small twig of the left splanchnic to be intact, it hardly seems probable that this groaning is due to stimulation of splanchnic or vagus afferent endings. The results obtained in this work, then, have been entirely nega- tive. The vagi and splanchnics are of no importance in parathyroid tetany in dogs. To reason from this that these nerves are of no im- portance in tetanies associated with gastro-intestinal disorders is, of course, unsound. However, the present finding does not support the theory of a nervous relationship between the two conditions, but rather lends favor to the view of a chemical relationship. CONCLUSIONS 1. Double vagotomy and splanchnectomy have no influence upon the onset and course of tetania parathyropriva. 2. Artificial gastro-intestinal irritation has no influence upon the onset and course of tetania parathyropriva. 590 WALTER LINCOLN PALMER The above research was done at the suggestion of Dr. A. B. Luck- hardt and it is a pleasue here to record my gratitude for his constant interest and helpfulness. The technique was developed by Mr. Ray- mond Householder in a series of preliminary experiments, for which I am greatly indebted. BIBLIOGRAPHY (1) Oster: Principles and practice of medicine, 8th ed., 881. (2) Paton, FinpLAY AND Watson: Quart. Jour. Exper. Physiol., 1916, x, 233, 315. (3) McCann: Journ. Biol. Chem., 1918, xxv, 553. (4) Ivy: Personal communication. (5) Ropman: Journ. Amer. Med. Assoc., 1914, xii, 590. (6) Bene: Journ. Amer. Med. Assoc., 1917, Ixviii, 1315. (7) OppENHEIM: Textbook of nervous diseases (trans. by Alexander Bruce), 11, 1274. (8) CaRLsoN AND Jacopson: This Journal, 1911, xxvii, 133. (9) Frrepman: Journ. Med. Res., 1918, xxxviil, 69. (10) Brepu: Innere Sekretion, 11 Auflage, 1916. (11) Fatra anp RupinGeR: Quoted from Falta and Kahn, Zeitschr. f. klin. Med., Ixxiv. 171. (12) Mustrarp: This Journal, 1912, xxix, 311. 3) LucKHARDT, SHERMAN AND SERBIN: This Journal, 1920, h, 187. INDEX TO VOLUME LII ADENI NE nucleotide, 193. BALDWIN, W. M. The artificial production of monsters demon- strating localized defects as the re- sult of injury from X-rays, 296. Barott, H.G. See Lanecworruy and Barortt, 400. Bett, A. E., H. P. Smrrs and G. H. Wuippte. III. Factors concerned in the perfusion of living organs and tissues. Artificial solutions substi- tuted for blood serum and the re- sulting injury to parenchyma cells, 101. 54. See Smitu, Bett and WHIPPLE, See Wurppie, SmitH and Be tt, 72. BerGemM, O. See MiLuer, BERGEIM, RexHFuss and Hawk, 1, 28. See Minter, Fowter, GEIM, REHFUSS and Hawk, 248. Bile formation, effect of gland extracts on, 498. Blood plasma, regeneration of proteins of, 54. -—— pressure, capillary, in man, 136. , effect of gland extracts on, BeEr- 498. proteins, relation of, to shock, 72. Brain volume, effect of stammering on, 168. Burce, W. E. The effect of acids, alkalies and salts on catalase produc- tion, 364. CAPILLARY blood pressure in man, 136. Carbohydrate tolerance in Eck fistula and hypophysectomized animals, 233. Catalase production, effect of acids, alkalies and salts on, 364. Coup, J. B. The sub-arachnoid and intra-arterial administration of so- dium bicarbonate and other elec- trolytes, 483. JPDANZER, C.S. and D. R. Hooker. Determination of the capillary blood pressure in man with the micro-capillary tonometer, 136. Digestion, influence of tea, coffee and cocoa on, 28. Downs, A. W. The influence of inter- nal secretions on blood pressure and the formation of bile, 498. J{PMUNDS, C. W. The character of the sympathetic innervation to the retractor penis muscle of the dog, 395. Epwarps, D.J. Segmental activity in the heart of the Limulus, 276. The arterial pressure curve as influenced by the occlusion of cer- tain vascular areas and by hista- mine, 284. Electron tube amplification with the string gaivanometer, 409. ‘ Energy expenditure in household tasks, 400 Epinephrin output and rate of dener- vated heart, 304. , ——, measurement of, 521. Ergographic response, effect of to- bacco-smoking on, 132. FORBES, A. and C.Tuacner. Am- plification of action currents with the electron tube in recording with the string galvanometer, 409. Fow.uer, H.L. See Minuer, Fow cer, BeRGEIM, ReEHFuss and Hawk, 248. 591 592 ({ASTRIC juice, psychic secretion Ort, il —— response to foods, 1, 28, 248. Gastrin bodies, chemical studies on, 508. Gland extracts, effect of, on bile for- mation, 498. GoLpMAN, A. MAN, 209. Grant, S. B. and A. Gotpman. A study of forced respiration: experi- mental production of tetany, 209. Growth, effect of splenectomy on, 562. —— of children, seasonal variation in, 121. See GRANT and GoLp- H4WkK, P. B. See Miter, Brr- GEIM, Reuruss and Haws, 1, 28. Hawk, P. B. See Mrtuer, Fowter, BrEeRGEIM, RenFuss and Hawk, 248. Heart, denervated, epinephrin output and, 304, 521. of Limulus, segmental activity in, 276. . Henn, S. C. The effect of splenec- tomy upon growth in the young, 562. Histamine, influence of, on pulse curve, 284. Hooker, D. R. See Danzer and Hooker, 136. Hoover, D. H. See Pearce and Hoover, 472. [NNERVATION, sympathetic, of re- tractor penis muscle, 395. Intra-arterial injection of sodium bi- carbonate, 483. JACOBSON, C. A study of the car- bohydrate tolerance in Eck fistula and hypophysectomized animals (posterior lobe removal). Liver function in the metabolism of sugars, 233. Jones, W. The action of boiled pan- creas extract on yeast nucleic acid, 203. INDEX Jones, W. The chemical constitution of adenine nucleotide and of yeast nucleic acid, 193. JKEETON, R. W. See Kocu, Lucx- HARDT and KEETON, 508. Kocu, F. C., A. B. Lucxsarpt and R. W. Keeton. Gastrin studies. V. Chemical studies on gastrin bodies, 508. ] ANGWORTHY, C. FH. Jand] HaaG: Barotr. Energy expenditure in household tasks, 400. LuckHarpT, A. B. See Kocu, LucK- HARDT AND KEETON, 508. M! CRO-CAPILLARY tonometer, 136. Miturer, R. J., O. Berecem, M. E. ReuFuss and P. B. Hawk. Gastric response to foods: X. The psychic secretion of gastric juice in normal men, 1. XI. The influence of tea, coffee and cocoa upon digestion, 28. ——, H. L. Fowuer, O. Berceim, M. E. Resruss and P. B. Hawk. The gastric response to foods. XII. The response of the human stomach to pies, cakes and puddings, 248. Monsters, production of, by X-ray, 296. Muscle, retractor penis, sympathetic innervation of, 395. PALMER, W.L. The importance of vagal and splanchnic afferent im- pulses on the onset and course of } tetania parathyropriva, 581. Pearce, R. G. and D. H. Hoovmr. Variations in the respiratory dead air space due to changes in the depth of breathing, 472. Perfusion of living organs, factors con- cerned in, 101. Permeability of placenta for adrenalin, 377. INDEX Placenta, permeability of, for adre- nalin, 377. Plasma proteins, stabilizing value of, We Porter, W. T. The seasonal varia- tion in the growth of Boston school children, 121. } Proteins of blood plasma, regeneration of, 54. Pulse curve after histamine and vas- cular occlusion, 284. RRAPHAEL, T. Certain changes noted in ergographic response as a result of tobacco-smoking, 132. Reuruss, M. E. See Mitier, Ber- GEIM, Reuruss and Hawk, 1, 28. See Miniter, Fow.uer, Ber- GEIM, ReHruss and Hawk, 248. Respiration, forced, tetany following, 209. Respiratory dead space, variations in, 472. Rossins, 8. D. A plethysmographic study of shock and stammering in a trephined stammerer, 168. Rocorr, J. M. See Stewart and Rocorr, 304, 521. SEASONAL variation in growth of children, 121. Segmental activity in heart of Limu- lus, 276. Suimipzu, Y. On the permeability of the placenta for adrenalin in the pregnant rabbit and albino rat, 377. Shock, relation of blood proteins to, 72. Smita, H. P., A. E. Bett and G. H. WuippeLe. I. Rapid blood plasma protein depletion and the curve of regeneration, 54. See Breit, Smita and WHIPPLE, See WHIPPLE, SMITH and BE tt, 593 Sodium bicarbonate and other elec- trolytes, injection of, 483. Splenectomy, effect of, on growth, 562. Stammering, effect of, on brain vol- ume, 168. Stewart, G. N. and J. M. Rocorr. Essentials in measuring epinephrin output with further observations on its relation to the rate of the dener- vated heart, 521. The relation of the epi- nephrin output of the adrenals to changes in the rate of the denervated heart, 304. Sub-arachnoid injection of bicarbonate, 483. Sugars, liver function in metabolism of, 233. sodium TETANY following forced respira- tion, 209. , relationship of nervous irritation to, 581. THACHER, C. THACHER, 409. Tobacco-smoking, effect of, on ergo- graphic response, 132. See Forses and VASCULAR occlusion, pulse curve after, 284. WHIPPLE, G. H., H. P. Siva and A. E. Bett. II. Shock asa manifestation of tissue injury follow- ing rapid plasma protein depletion. The stabilizing value of plasma pro- teins, 72. See Beit, Smira and Wuippte, 101. ——. See Smits, Beir and Wurepte, 54. X-RAY, production of monsters by, 296. YEAST nucleic acid, 193, 203. 4 b a Mid er hy fe) hy 7 i” UN , +p ae i , i ¢ H ; a 4 ar aP ' J hi dot i . ~ * ’ ay wah ’ . J a 7 ‘ 7 Ay, i shy of a7 on i : - yi bay ah ’ * M POT uy O48 if “abi ON bash ¥ is a 0 Aa tt ar, ) oudbilae hh rg he 1s Awe 2 Ol Ree, Oy ras vaRG ; hacia) Peas a ye ear givtinus j bls SIS id iM fe HTS a Re vo hast ob ie Chair J) AQ ad ae ’ dict ive ¢ ts ‘ oe Ge BREED Re , TAS Dae Mat ; leas Ja anes) bay 1s. deh, Figen watt %42 Nie gi if ABMS thy att. ihe at FA 2 ee ay ap 4 : arth linak ie th ee VT ‘ hale ' riewat } tof * yy y! hj j q 1] AM Lae J ye ei ad X y A ey a? , . 7 =e i A" ; : Bae = re i " sit 4 i ’ : f = ar ala ~ ah tee a oi = ia i } | A A: | - =e ba apace rege DOSES " . : 7 ein La *¢ i + ~* “th ’ Sone oe \ tr alah: Ma ane cy | : ; ae od ae Nate i Miata BINDING SECT. MAR4 1966 American Journal of Physiology Biological & Medical Serials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET | UNIVERSITY OF TORONTO LIBRARY STORAGE. tHe ‘ " ¥) ay aan eo) ie ile Beat ait ler Ca 1 7 et. va y we << eS pees SSres ose pee it et ee: wth = Se ; igethy pukantars a oe Se Ses ee tet Eon Se et ARR tt ! FSS