■^^t\ :te aw^j; sm ■mll^' i lii ^"H. ADOLPH CO PHYSIOLOGICAL REGULATIONS By EDWARD F. ADOLPH Associate Professor of Physiology in the University of Rochester THE JAQUES CATTELL PRESS LANCASTER, PENNSYLVANIA 1943 Copyright, 1943, by THE JAQUES CATTELL PRESS PRINTED IN U. S. A. THE SCIENCE PRESS PRINTING COMPANY LANCASTER, PENNSYLVANIA To the memories of JOHN SCOTT HALDANE 1865-1936 and LAWRENCE JOSEPH HENDERSON 1878-1942 Bold explorers and teachers of a larger physiology Investigators of physiological regulations "Life is an example of the way in which an energy-system in its give and take with the energy-system around it can con- tinue to maintain itself for a period as a self -centered, so to say, self-balanced unity. Perhaps the most striking feature of it is that it acts as though it * desired ' to maintain itself. ' ' — C. Sherrington : Man on His Nature, p. 84; Cambridge, 1941. PREFACE ''His body temperature is always the same; if it were not he would be sick." That empirical rule is rediscovered daily. To physiologists the rule is a challenge to seek further relations among bodily functions. How are the specific properties of an organism maintained? What sorts of processes are provided so that the temperature is not upset in a body subjected to twenty-fold in- creases in heat production and to winter's icy blasts? Physiological regulations are patterns of processes; the out- come of all those operating characteristics that assure the con- stancy of a property such as body temperature. The creature's endowments are not merely material ; activity is provided continu- ously. Of all the possible combinations of chemical velocities, physical responses, and organismal behaviors that are possible, those very ones prevail that fit into a scheme of self -perpetuation. This scheme is the object of investigation. The invisible pattern according to which in a given physiological state, whether of action or of rest, these particular velocities are operating and others not, constitutes regulation. Physiology seems to me more than a science of individual work- ing parts. I found it fun to compare the manners in which a kid- ney excretes various substances. But then I wondered whether there was some rhyme in the heights of the several thresholds and the hypertonicities of various urines. It next became obvious that measured ingestion saved excretion many a day's work. Evi- dently excretion is subservient to organism, a kidney to a body. The scale of renal clearances seems utterly arbitrary until one thinks of the millions of body cells that have to tolerate and even transform what the kidneys do not excrete. At one time I supposed that the pattern of physiological re- search was fixed once for all. The newcomer merely went into more detail with modified apparatus. Now, I believe that fingers and levers are to be supplemented by concepts. There is no limit to the patterns of physiological investigation, for every concept adds a pattern of search. Physiology is more than a technology, more than information. It develops new aspects at every turn; as long as it lives it will include the unorthodox. The unorthodox of today becomes the standard of tomorrow; let no one make a VI PREFACE religion of an emphasis. Let not wisdom scoff at strange notions or isolated facts. Let them be explored. For the strange notion is a new vision, and the isolated fact a new clay, possible founda- tions of tomorrow's science. Lest anyone shy at the title before he has sampled the book, I may state that it describes interrelations among a limited number of rather familiar facts. It is not a derivation of the imagination, nor an exposition of a hypothesis, nor a development of a nebulous idea. Whatever ideas went into the work were, I think, either sub- stantiated or killed. The presentation that results therefrom takes on an almost mechanical plan consisting of numerical data and of generalization from them. Often the facts seem exciting enough so that no imagery or creed is required. This book is not a survey nor a review; it is a report of an investigation that became too extensive for presentation in a journal. It would be gratuitous in such a work to note what raw materials were omitted, or even what other lessons might be drawn from the same sludge. Let it be supposed, as for Basil Montagu, that ''He puts the facts before us in the full confidence that they will produce on our minds the effect which they have produced on his own" (T. B. Macaulay: Essay on Lord Bacon). The best evidence that the study of regulations actually helps one to understand physiology, lies in predictions about organisms. Here is some substance J always present in muscle. That fact per se implies that there is a source and a sink for J. But further, the fact suggests that acquirement of J is probably faster whenever its content tends to diminish. So metabolism is viewed as an intricate pattern of interrelated processes which persists by virtue of the very multiplicity of its quantitative steps. Once the steps have been enumerated, their correlations become self-evident. Truly each relation is new fruit, which would be discovered to be ''neces- sary" to some future brilliant hypothesis, if it were not firmly indicated by present facts. Those who prefer to take their facts on faith may avoid wading through the presentation of details by reading the concluding sec- tion of each chapter. In this way the progress of the inductive development may be partially followed, leading at the close to certain generalizations concerning the functional constitution of organisms. PREFACE Vll I wish to acknowledge the generous encouragement of the late Professor L. J. Henderson, who furnished the environment for the inception of this specific investigation. He took the trouble to read the manuscript shortly before his untimely death. Possibly years ago he and the late Professor J. S. Haldane unwittingly implanted a curiosity about the regulative aspects of physiology. I am grateful for the comradeship of my students who have shared in the laboratory studies concerning water and heat metabolism. Material aid was accorded the prosecution of this work through various administrative officers of the University of Rochester. The publication is made possible by the School of Medicine and Dentistry of the University. E. F. Adolph October, 1942 ABSTRACT This is a monograph recording an investigation in quantitative physiology. Regulations in organisms are maintenances of rela- tive constancies. They are described from suitable data by selected relations of the following sorts : (1) Variations, especially in the successive values found for one sort of measurement in an individual. These represent the results of all those events that antagonize the occurrence of wider variations. (2) Changes and exchanges as correlated with excesses over, or deficits under, the control values (contents). These are rates of processes, and express what is done to recover the usual physio- logical state. (3) Behaviors that exhibit preferences for environments which either promote or prevent exchanges of diverse components. Water is chosen as the prototype of component that tends to be in quantitatively regular amounts in living units. Data for the dog show in how far water contents (body weights), intakes, and outputs vary from hour to hour and from day to day ; and how the rates of intakes and outputs are modified with each experimentally provided content. The latter relation is investigated both in sta- tionary states of unusual content, and in recoveries from them. Preservation of content may be regarded as a pattern of particular relations among the rates provided in diverse paths of exchange. Man, frog, and many other species of animals, including several of invertebrates, manifest similar relations in water exchange. Organs, tissues, and cells are also recognized as having the pattern. From these materials the general features common to many living units are formulated. Quantitative differences among species are also ascertained. A variety of other measurements, particularly upon blood and other parts, is correlated with water content, to characterize the physiological states of water excess and deficit. Thus an intensive and comprehensive account of water relations of animals and their tissues results. Analogous data are set forth concerning other quantities than water. Some of these are heat, glucose, oxygen, carbon dioxide, lactate, frequency of heart beat, and blood pressure. For each a similar pattern of equilibration is found. ABSTEACT The -uniformities and the contrasts among them are pointed out. The general forms of variation in content, of time relations in recovery, and of rates of exchange in relation to excesses and deficits (equilibration) are induced. Simultaneous equilibrations of two or more such quantities are studied. The interplay of quantities indicates the organism's choices in handling at one time the various items that call for recovery. It is inferred that the organism is a compound of rela- tions among components in mutual adjustment. Descriptive pro- cedures are illustrated that exhibit the multiple relations in com- prehensible manners. These selected relations among data are manifestations of the processes commonly meant by the term physiological regulations. They provide a quantitative means of visualizing what organisms do to maintain constancy not only of composition, but of energies, forces, structures, and functioning. CONTENTS ^^ ** ^ PAGE Preface v Abstract ix Chapter I. Introduction 1 $ 1. The project 1 2. How regulations may be studied 4 3. Extensions of the study 7 4. Mode of treatment 8 5. Outline of the investigation 11 PAET A. WATER RELATIONS OF ANIMALS Chapter II. Water Exchanges of Dog 17 $ 6. Introduction 17 7. Sudden excesses of water 18 8. Deficits of water 27 9. Equilibration 32 10. Velocity quotient 37 11. The water-time system 40 12. Stationary states of excess 41 13. Stationary states of deficit 47 14. Summary 49 Chapter III. Other Types of Water Increment (Dog) 50 § 15. Introduction 50 16. Excesses of water 50 17. Deficits of water 59 18. Distinctions among water increments 62 19. Modifications of water content at balance 68 20. Summary 71 Chapter IV. Variabilities of Water Relations (Dog) 73 $ 21, Introduction 73 22. Variations of water content 73 23. Variations of turnovers 78 24. Variations in rates at diverse loads 84 25. Summary 86 Chapter V. Water Relations of Man 88 § 26. Introduction 88 27. Single ingestions by mouth 88 28. Repeated ingestions by mouth 93 29. Water privation 95 30. Equilibration and recovery 100 31. Comparison with dog 102 32. Variations 103 33. Characterizations and tests 105 34. Diverse types of water load 107 35. Summary 108 Chapter VI. Water Relations of Frog 110 § 36. Introduction 110 37. Water exchanges 110 38. Variations 116 xi iriC' 56bt> Xll CONTENTS PAGE 39. Some other types of water load 118 40. Summary 120 Chapter VII. Water Eelations of Other Species 122 § 41. Introduction 122 42. Eabbit 123 43. Eat 126 44. Garter snake. Eeptiles 132 45. Limax 134 46. Helix 135 47. Insects 137 48. Phascolosoma 138 49. Lumbricus : 142 50. Bipalium 144 51. Arbacia egg. Echinoderm eggs 144 52. Freshwater Zoothamnium 146 53. Marine Zoothamnium 147 54. Ameba 148 55. Note on plants 148 56. Summary 149 Chapter VIII. Equilibrations in Parts of Organisms 151 $ 57. Introduction 151 58. ■ Volumes in situ 151 59. Blood and plasma 153 60. Arm 158 61. Other organs 158 62. Cells and nuclei 160 63. Summary of parts in situ 160 64. Isolated parts 161 65. Muscles 161 66. Cells '. 164 67. Summary 166 Chapter IX. General Features of "Water Exchanges 168 $ 68. Introduction 168 69. Some limitations 168 70. Some parameters comparing exchanges 170 71. Time relations 178 72. Equilibration diagrams 188 73. Beha:vior 195 74. Water contents and turnovers 195 75. Tolerated loads 201 76. Summary 202 Chapter X. Some Other Correlatives of Water Content (Dog) 206 §77. Orientation 206 78. Volumes of parts 207 79. Water contents (dilutions) of parts 210 80. Blood and plasma 213 81. Concentrations of urine 224 82. Concentrations of other body fluids 225 83. Others compositions 225 84. Correlated metabolisms and behaviors 227 CONTENTS Xlll PAGE 85. Quantitative characterizations of water loads 233 86. Summary 236 Chapter XI. Some Other Correlatives of Water Content (in Other Species) 238 § 87. Introduction 238 88. Man 238 89. Frog 246 90. Rabbit 248 91. Eat 250 92. Comparisons 251 Chapter XII. Further Correlatives of Water Content and Exchanges 255 § 93. Introduction 255 94. Possible forces 255 95. Permeability 258 96. Body size, age 263 97. Temperature 275 98. Races .: : 276 99. Summary 277 Chapter XIII. Water Balances and Exchanges, Recapitulations 279 $ 100, Introduction 279 101. Classification of variables 279 102. Scopes of the classes of variables 280 103. Interrelations of the variables 284 104. Procedures or steps used 285 105. Outline of water relations 286 106. Uniformities ; 290 107. Diversities 292 108. Agents and types of load 294 109. General theories of water constancy 296 110. Summary 297 PART B. REGULATIONS OF SEVERAL COMPONENTS AND IN GENERAL Chapter XIV. Heat 301 § 111. Introduction 301 112. Maintenance in man 302 113. Heat loads (man) 304 114. Recoveries (man) 307 115. Heat exchanges of rabbit 312 116. Comparisons among species 318 117. Summary 321 Chapter XV. Diverse Components 323 § 118. Total substance 323 119, Glucose in dog 332 120. Carbon dioxide in man 340 121. Oxygen in man 343 122. Lactate in man 346 123, Summary 348 124. Heart frequency 350 125, Arterial blood pressure 352 126, Volume flow of blood 353 XIV CONTENTS PAGE 127. Eegeneration of tissue 354 128. Excitability of isolated nerve [ 357 129. Other data 353 130. Summary 359 Chapter XVI. Uniformities and Comparisons among Components 361 $ 131. Variabilities 361 132. Behavior and maintenance 364 133. Sequences in time 366 134. Loads _ 37O 135. Eates of exchange 373 136. Velocity quotients 377 137. Paths of exchange 381 138. Types of loading 335 139. Comparison of kinetic parameters 386 140. Changes in tissues, other metabolisms, etc 389 141. Species 390 142. Equilibration diagrams 394 143. Summary of variables studied 396 144. Types of relations 398 Chapter XVII. Interrelations among Components 400 § 145. Introduction 400 146. Heat load and water load in man 400 147. Several components during physical exercise (man) 405 148. Some other components 409 149. General features 413 150. Meanings of interrelations 419 Chapter XVIII. Choosing Physiological Variables 422 § 151. Introduction 422 152. Present procedures 422 153. General features 424 154. Description 427 155. Other modes of procedure 431 156. Summary 436 Chapter XIX. Physiological Eegulations 437 $ 157. Meanings of regulation 437 158. Classification 440 159. Measurement 443 160. Kinetic equilibrium 444 161. Multiple regulations 446 162. Signs and tests 448 163. Summary 451 Chapter XX. Some Speculations Concerning Eegulations 453 $ 164. Introduction 453 165. Distinctions between generalizations and theories 454 166. Maxima and minima 455 167. Origins of equilibrations 458 168. Forces in biological equilibria 461 169. Spencer's views 463 170. Survival and death 465 171. Integrations 466 CONTENTS XV PAGE 172. List of theories 457 173. Summary ^gg Chapter XXI. Conclusions 47I $ 174. Introduction 47]^ 175. Abstract of the investigation 47I 176. Equilibrations 474 177. Alternative studies of the same material 475 178. Contributions made 476 179. Conclusions concerning physiological constancies 477 180. Retrospect 473 References 4gO Index 494 LOCATIONS OF TABLES TABLE 1 ... 2 ... 3 ... 4 ... 5 ... 6 ... 7 ... 10 11 12 13 14 15 16 17 18 19 20 21 22 23 PAGE 30 51 58 66 152 156 157 166 171 172 172 175 176 178 182 187 189 196 197 198 199 201 208 ABLE 24 PAGE 216 25 230 26 246 27 248 28 249 29 251 30 269 31 277 32 289 33 303 34 . 305 35 311 36 313 37 314 38 336 39 361 40 372 41 378 42 380 43 ... 381 44 440 45 443 46 „ 449 Chapter I INTRODUCTION <§. 1. A living organism is much the same day after day, not only in form but in function. If I count my heart beats for a minute each morning before getting out of bed, I find the frequency of beats only less constant than the length of my foot. The foot looks like a permanent mold in which matter is cast. Actually it swells and shrinks with each pulse of blood coming to it, each change of posture, and each meal. Its substance changes from month to month. The beats of the heart appear to be accurately meted out without a visible pendulum. During a day many an acceleration of frequency occurs ; but when rest or other uniform state of body prevails, the frequency quickly returns to its characteristic value. That constancy represents a physiological regulation. Every biologist, indeed every owner of a pet, expects his animal to be nearly the same every day ; to show the same movements, the same responses to signals, the same posture, and the same food con- sumption. If not, he decides there is ' ' something wrong. ' ' The certainty that he feels about the animal's health is similar to that which he feels about the performance of his automobile, his kitchen, or his piano. He has come, by experience, to expect uniformities of particular sorts. Failure of the car's motor to start is for him as uncommon as failure of the car's fender to keep its shape. Accumulated experiences seem to make a man more and more confident of always finding particular features that he can count on in his machine or instrument. In much the same way, the biologist becomes intimate with the uniform characteristics of each animal dealt with, and the phy- sician notices small departures from regularity in his old patient. All this is a tribute to constancy of physiological function, a recog- nition that each living individual tends in many ways to be like itself at each observation. Regulations mean, then, in physiological science, those self- managements of organisms which result in constancy of function. TV^at do animals do to maintain their physiological constitutions and activities ? Functional arrangements evidently work hand-in- hand with structural, chemical, and other kinds of arrangements ; 1 2 PHYSIOLOGICAL EEGULATIONS it would be quite arbitrary to consider traditionally functional aspects of organisms by themselves, and the term physiological merely indicates the aspect to be stressed here above others. And now the questions are : how may this recognition of con- stancy be removed from the limbo of vagueness, and subjected to quantitative study? What shall be measured to assure me of the concrete existence of regulations 1 How may a physiologist record what an organism is doing to maintain its functions so constant? Why does anyone investigate constancies anyway? Various excuses have been set forth by others for this perennial interest. Any property seems worth knowing about, just in proportion to its prevalence. Constancy of function is a feature of whatever constituents and manifestations have been observed in all organ- isms. By the degree of constancy I can characterize diverse mani- festations in one species, can compare similar manifestations in diverse species, can ascertain what each animal does to insure con- tinuance of function, and can observe how the animal manages many properties simultaneously. If I had not guessed at ways to investigate how constant various components are, and what hap- pens when a component is disturbed, I might have been content with lip service to the existence of constancies. But now that some of my guesses have found substantiation in data and correlations, I am not satisfied with anything less than a quantitative treatment of the elements concerned. There is nothing in physiology or any other science to indicate that understanding is promoted by not intimately prying into the relations involved in a class of phe- nomena. I propose first to outline briefly the subject of maintenance of physiological constancies, giving the methods, results, and conclu- sions that I have found appropriate. Thereafter, a more detailed and ordered treatment of the material will be given, leading to each of the generalizations to be mentioned. Since my generaliza- tions have arisen in the study of water relations of animals, it is most direct to use water content as the property whose constancy shall be visualized (Part A). Whatever further features can be found with respect to content or manifestations of other com- ponents of organisms can be noted later (Part B). In an early stage of physiological science the statement was often made that an organism's water content is kept constant by virtue of kidneys, or integuments, or some other structures. Later INTRODUCTION 6 it was asked how those structures "knew" when the content was the same as usual, and ''knew" what to do when it differed. Further it was observed that the regularity of water content is related to systematic modifications in the exchanges through those or other structures, in accordance with the actual contents prevail- ing. To certain contents the body is indifferent or acquiescent ; in the presence of unusual amounts of the same components there is violent action that usually succeeds in changing them. Moreover, however satisfying it may be to know that kidneys or alimentary tracts are pathways that care for the exchanges of water, it is soon realized that many organisms exist without those organs, and yet probably every organism has means of correcting unusual water contents. Further still, an organ to remove excesses and another organ to make up deficits serve the body only so long as each grades its activities to whatever disturbance of content prevails, and only so long as the two organs are related to one another, in agreement as to what water content will be accepted by both. Is an organism actually exposed to unlimited loss, or to unlimited gain ? This topic concerns structural, mechanical, chemi- cal and other features that somehow protect the body from suffer- ing rapid changes, in the mode of existence peculiar to its species. But is water content entirely a matter of faster and slower exchanges? Does its constancy not depend on availability of water in the environment? Water content is usually part of a stationary state that requires a source and a sink; these are en- vironmental features toward which the behavior of the organism is frequently found to be appropriate. Biologists in all ages have visualized one or another of these questions. Taken together, their views (§ 157) form a valuable background to an intensive treatment of self-maintenances. As long as 2400 years ago, Alcmaeon stated that "The preservation of health consists in an accurate adjustment of forces." This is the oldest known generalization of physiology. It represents the vague realization that the normal state of the organism is main- tained by specific processes which resist change. And today it persists as a fruitful generalization that can be sharply delineated. By methods that were scarcely tried by my predecessors I can hope to fit together observations that help in comprehending the manner in which constancies of many properties are maintained. I seek to describe the operating characteristics of these maintenances. 4 PHYSIOLOGICAL REGULATIONS The peculiar method of this description is to use detailed and quantitative materials to lead up to a picture of certain types of regulation. Then the maintenance of a specific constancy in an organism may appear no longer as a conjecture or a pronounce- ment, but as a conclusion from many observations. Regulation takes on a patterned, and at times a geometrical, concreteness. Heretofore the abstraction of regulation has been an accepted part of physiology, though often regarded more as a frill than as an inevitable aspect of everyday phenomena. Usually it is conceived in the form in which Claude Bernard (1878, p. 144) presented it. He postulated that "in animals having free existence there neces- sarily exists a group of arrangements controlling the losses and the gains in a manner that maintains the quantity of necessary water in the internal medium" of the body. A quantitative de- scription of any considerable group of such arrangements has been given only for blood (Henderson, '28) ; and blood has commonly been referred to as though it were a special sort of system, as though to it alone did that scheme for representing interrelations apply. By sifting and putting together selected data I ascertain what occurs when some quantity (content) of the organism gets too large or too small for continuance. In this way a circumscribed func- tioning of the organism is quantitatively described, which later (chapter XIX) proves to be a part of a larger functioning. § 2. How REGULATIONS MAY BE STUDIED The approximate definitions of regulations, indicated above, at once suggest means of investigating them. What procedures are available for measuring constancies in living bodies, and what features are present by virtue of which the constancies are per- petuated? I think that there are broad categories of measure- ments which characterize the maintenances of all sorts of proper- ties, and that in the end it will be quite unnecessary to restrict specifications to water content or to heat content or to any other one. The following phenomena appear to be implicated. (1) Exchanges. In a stationary state which is the organism, constancy of content means a rate of gain equal to a rate of loss. But, of numerous components, as water, animals do not character- istically have equal inflow and outflow in every minute of life. Both undergo endless fluctuations. They are compatible with con- INTRODUCTION O stancy in so far as the content is juggled by compensations. First gain exceeds loss, then loss exceeds gain, and in diverse degrees. The mean rate of gain is found by measurement to exceed the rate of loss when the content of substance is lower than usual. In other words, the modifications of exchange that are seen when content is disturbed, whereby gain is faster when content is low and loss is faster when content is high, indicate the way in which exchanges are modified so that in the average instance the usual content tends to be restored. This relation of exchanges to contents describes how corrections of content occur. It specifies the precise connection between proc- esses of gain and processes of loss, and the actual gradation of each to the content. Corrections of unusual contents may be thought of roughly, though not strictly, as internal regulations. Though faster or slower exchanges stand ready to compensate when the organism has gotten into trouble, provisions for keeping out of trouble, for preventing departures from constancy, might save many a need for cure. Those processes of the body and its parts that discourage excesses and deficit of some property J in the routine of life are therefore observable means of preserving constancies. One method of preservation is complete isolation ; a dead animal in a museum might remain constant in that way. Since complete isolation is incompatible with much of living, partial isolation appears more often. To judge when it is present is feasible in two ways. It may be shown that a dog, or some por- tion of its surface, exchanges J more slowly than does a sponge soaked with saline solution and of equal size, shape, temperature, and so forth. Or, it may be shown that modifying or removing the surface hastens the exchanges. In general, the constitution of the organism is such that the environment exchanges with the body more slowly than it would with another constitution believed to be less differentiated. (2) Behavior in selecting among environments. Given the physicochemical constitution of the organism, its peculiar gearing to rates of exchange, and its sensitivity to changes of content, still a factor in its preservation is the keeping of itself in environments that are not too hard on it. By selection of surroundings, the organism may obtain the cheapest maintenance if it find a place in which to be temporarily isolated, or in which water or other sub- stance is optimally available. In general the organism seeks some conditions, and avoids others. b PHYSIOLOGICAL REGULATIONS To accomplish this, the organism uses sensory and locomotory (sensorimotor) equipment. There is thus a distinction between the behaviors of alimentation and absorption, which are exchanges ; and the behaviors of making water available (hydrotropism) or absent. The latter is a frequenting of certain environments, a group of external regulations. The study of selection among environments is carried out by well-recognized methods. A rat (§43) is compelled to pass certain obstructions to obtain water. Or, a rat is put in an apartment having both moist atmospheres and dry ones ; when poor in body water it shows preference for the moist ones, thereby minimizing its further water loss, while when rich in body water it is more indifferent about moisture. In other words, there is a correlation between the animal's water content and the water content of the environment that it frequents. (3) Variability of content. This answers the question, how constant is the content of component J? And, how far does the content change before something is done about it? In a first ap- proximation an investigator might choose a hundred individuals under specified conditions and ascertain the content of J in each. But sooner or later it might be found that the individuals chosen did not belong to a homogeneous population. It seems to me preferable to make, if possible, a hundred determinations of con- tent upon one individual at successive equal intervals of time. All the while the individual studied will be maintaining itself under conditions that the observer believes are adequate, or under those particular conditions that the animal itself frequents. The outcome is a succession of values of content. There are various ways of summarizing this succession. If it be shown that the points are random in time, then a distribution of frequencies of content, or its standard deviation, might characterize the vari- ability. Without that demonstration I believe the physiological variations are best accounted by ascertaining the differences be- tween successive points. These differences are what the organism allows itself in the way of fluctuations of body substance. Con- tents outside the range observed simply do not occur spontane- ously ; this organism in this environment sees to it that they do not occur. Such a variability can be measured in respect to most kinds of content, and represents the net outcome of all regulatory processes. INTRODUCTION ( By the three aspects: modifications in exchanges, behavior to- ward environments, and variability of contents, regulations are to be described. Information that falls in these categories will be the materials examined. § 3. Extensions of the study Each of the three ways of studying self -maintenances of con- stancies is applicable, so far as I can now ascertain, to all organ- isms and portions of them (chapters XIV to XVI). All three can be further studied (to limited extents) by interfering with what the organism is doing, i.e., by surgical, pathological, or pharmacologi- cal means, especially in breaking connections among its functional units. One means of gaining insight into regulations is to observe to what extents various animals are provided with them. Provisions for correcting the contents of some components are absent or in- complete to diverse degrees in immature animals. Or, comparisons of species show regulations of a few components (as, heat) in some but not in others. In some, regulations are faster than in others ; in some they are lacking over certain ranges of bodily content. Instances are found in which compensations become more rapid with practice, and even in which non-existent ones may be called into being. All these special instances may help in grasping how self -maintenance is carried on. Any compensation appears as a correlation between content of J and some feature of exchange. Of all the possible exchanges, organisms are patterned to modify the exchange of J itself, which alone restores the content of this particular component. To limited extents, however, other contents and exchanges are modified (chap- ters X to XII), usually to smaller extents. A study of the numer- ous correlations among measurable properties yields a notion of the number of strains suffered by the organism whenever it has departed from balance of J. Further, each of the properties that is disturbed is itself under- going adjustment toward restoration. That gives a picture of the interrelations existing among components whereby simultaneous processes act to bring the organism to its most persistent state (chapter XVII). Only with violence to the organism does the investigator study its components one at a time. Nevertheless, after examination of what organisms do when single components 8 PHYSIOLOGICAL REGULATIONS are varied in content, uniformities of pattern are evident. These general features form quantitative pictures of the manner of regu- lations. Instead of having a different basic scheme for handling each component, the organism apparently assigns to some portion of itself the fulfilling of the specific operations, consistent with the general pattern that it has for other components. While the spe- cific physical machines for liandling excesses of water, heat, oxy- gen, and glucose might be infinite in variety, the kinetics of restoring content is practically singular. That general fact calls for intensive study. In brief, diverse sorts of living units and of a variety of proper- ties, may be examined in the light of relations of compensation, preference, and variability. Studies of stages in development of the individual, comparisons among species and tissues, and exami- nations of simultaneous disturbances of many properties, are indi- cated. The task is to discover those features that are common to numerous kinds of regulated properties. •^ 4. Mode of treatment Every investigation appears to start from vague notions about the relations among observable phenomena. Having supposed that certain quantities might inform me about regulations, and tested the suppositions, I find scarcely worth mentioning what the work- ing hypotheses were and what vicissitudes they have undergone. Then how shall I proceed to record the results, contributing order to the concrete observations'? (1) Description. There are often said to be two kinds of scien- tific ordering, causal and mathematical. In physiology the causal has frequently been essayed, and has not, I believe, revealed all that can be discovered with respect to regulations. I propose to try the other procedure, indicating descriptive relations among processes that seem concerned in maintenances of organisms. I will ask how, and not the unanswerable why. It is often supposed that mathe- matical descriptions do not satisfy the desires for understanding in the way that causal orderings do. That seems to depend on the individual scientific appetite. For I find that mathematical rela- tions in physiology, even as in modern physics, may prove just as explanatory as any others that have been proposed. Of that no one can judge who has not attempted to use them. Readers may feel the omission of the usual excuses, arguments, INTRODUCTION y and sophistries that ordinarily accompany statements of fact de- rived by inspection, much as they might miss their own clothing. The practice of combining facts with non-descriptive inferences is so prevalent that facts seem not to sit comfortably in some minds without that kind of reasoning. The urge toward rationalization is as though natural phenomena did not subsist in their own strength but needed bolstering before they become acceptable. Statements are said to be tedious and dry; they are supposed to need flavoring with imagination and hypothesis before they are presentable. I propose to depend heavily on correlations for flavors and significances. If a reader were to overlook this explicit procedure (§ 154) he might make the mistake of supposing that some data presented have no connotations, as though they had been set down in useless and thoughtless disorder. Recognition of this purpose rather than of some other purpose is a means of grasping quickly the plan and ends of the investigation. (2) Quantitative data are obtained and utilized at every point possible. They make precise whatever phenomena are studied. They allow conclusions to be set down in concrete numbers and dimensions. They avoid certain of the intrigues of words and statements whose generality would require further investigation. The data selected in what might seem at first to be a prolix pro- fusion lead by induction to conclusions concerning maintenances and regulations. "Illustrations" have been suggested heretofore in the pursuit of these same general phenomena. To me the factual material would not be sufficient for the present inductions so long as it was qualitative. Even in what is currently termed quantita- tive biology, only qualitative conclusions are regularly drawn. But numerical relations seem to me to be precise parts of conclusions ; phenomena are not just large or small, more or less, yes or no, proved or disproved. Equations and graphs may express more forms of relations than words do (§ 152). (3) Generalisations result from the quest for uniformities among the quantitative descriptions of particular kinds of physio- logical phenomena. Instead of inferring that a relation observed will turn out to be a constant one, it seems preferable to count how frequently, with what variability, and under what varieties of con- ditions each uniformity appears. The reiteration of certain quan- tities, of combinations, of correlations, becomes the basis of under- 10 PHYSIOLOGICAL REGULATIONS standing the apparent vagaries and diversities of living units (§153). (4) Interrelations among measured phenomena become ex- tended and multiple. A considerable part of scientific activity (sometimes believed to be the whole of science) is the placing of facts in relation to one another. It might be said that any datum may be viewed in an almost infinite number of ways. Ordinarily the first observer of fact A places it in relation with facts B, C, etc., arbitrarily according to his own knowledge, interests, and interpre- tations. This prejudices the future relations in which the fact A will be thought of, considered, and used ; especially this limits the relations of which the discoverer will think. Another person is, so far as I can discern, freely entitled to reisolate the fact A and put it in relation with facts D, E, etc. Often the second scientist will put the fact A to as fruitful use as the first did. Nevertheless, the ignoring of those relations among facts that have previously been promulgated, and the emphasizing of other relations, often arouses resentment. Divorcing facts from conno- tations is perhaps as important as, and is often more difficult than, finding connotations for them in the first place. The plan of re- cording an investigation is a responsibility ; but to enable another to separate the observed fact from the scheme of representation is also a responsibility. I know of no means of predicting beforehand which, of the semi- infinite number of correlations that are possible, will prove to be more interesting than the average, in the eyes of many generations of scholars. After correlations are established, the positive and negative ones look happier than the zero ones ; and those with small perturbations have more statistical significance than others. But it seems as though any correlation that is made, takes on signifi- cance to the investigator as he observes it and thinks upon it (§ 144 and §150). Hypotheses are gradually formulated about it, and an interrelation that to one physiologist looks arid will to another be pregnant with meaning. Hence, the procedures are such as to emphasize : description of relations, quantitative data and comparisons, generalization from similarity of relations and multiple interrelations. No doubt the conclusions to be found depend upon the particular data utilized; insofar as they have been tested, the conclusions that are reached appear to be of considerable generality. inteoduction 11 § 5. Outline of the investigation" The plan of this investigation is the equivalent of an inter- weaving of at least five studies. Each of them could be under- taken separately, but their interplay multiplies the inductions that can be formulated from the data presented. (1) Water exchanges in animals. (2) Rates (kinetics) of certain classes of physiological proc- esses ; time sequences. (3) Quantitative comparison of like functions in diverse spe- cies and individuals. (4) Organ and tissue exchanges; the study of specializations, localizations, paths, and routes. (5) Components (constituents, endowments, and properties) ; similarities and contrasts in their metabolisms and econo- mies. Emphasis throughout is upon interrelations among simultane- ously occurring activities. Some features into which this study delves, incidentally to its chief purpose of describing physiological regulations, are : Water metabolism of man (chapter V) Tolerance curves (§ 71 and § 133) Stationary states in organisms (§ 138) Variabilities of organisms (chapter IV, and § 131) Recovery processes (§ 71 and '^ 133) Signs of disturbance and disease (§ 162) Interacting maintenances (chapter XVII) Temperature regulations (chapter XIV) Selection of environments (§ 43 and ■§ 132) Storage and depots (§79) Ontogeny of regulatory processes (§ 96 and § 115) Comparative physiology (§ 107 and § 141) Cells and tissues as regulated units (chapter VIII) Volumes of distribution (§ 58 and § 80) Blood volumes (§ 59 and § 80) In designing the investigation, it is useful to distinguish certain categories of measurement with which the data will deal. The variables initially selected for study fall into six classes, none of which excludes from membership in other classes : 12 PHYSIOLOGICAL REGULATIONS (1) Rate of exchange (R), e.g., Total gain of energy, of water (2) Species or Living Unit (U), e.g., Dog, erythrocyte (3) Path of exchange (p), e.g.. Urinary, radiative (4) Tissue (portion) studied (s), e.g., Blood, whole body, cell (5) Component (J), e.g.. Water, heat, pressure (5a) Type of displacement (f), e.g.. Privation, injection (5b) Quantity of one component (C), e.g., Excess, deficit (6) Temporal parameter (t), e.g.. Time, duration The number of kinds of correlation among 6 variables is 57, or taken two at a time is (6 X 5)/(l X 2) or 15. The number of spe- cific data available in each class to be correlated is semi-infinite. The procedure chosen is, at first, to keep the component (J) constant, then the species or individual or living unit (U) con- stant. Within the study of one component, the type or means of disturbance (f) influencing the component is recognized; later the tissue or other portion measured (s) is differentiated. This re- duces any one set of data to three variables (J, p and t) to be dealt with; p often can also be selected and held constant through one series of data. The other three are entered in separate correla- tions, until summaries calling for regrouping are required. A more complete view of the course pursued may be gained after part of the treatment has been covered (§ 101 and § 152). Not all the possible combinations of variables are presented. To do so would be a completion but also a tedium. Those omitted are not less important to the organism, so far as anyone can judge ; but are of three classes: those for which suitable data were not obtained, those unfamiliar to me, and those in which the correla- tion did not seem to me to yield illuminating relations. Limita- tions are imposed also by the number of variables that can be con- veniently and profitably handled at one time. Variables, in fact, may be numerous and unknown, but the ones recognized are chosen for their reproducibility, statistical significance, and apparent in- terrelations. Only by continual classification and limitation can the investigation be kept, at each point in its progress, within comprehensible limits. Some of the studies here presented may appear to be exhaustive. But I am the last one to consider them so, for hosts of lacunae and possible extensions have come into view. It is conceivable that physiology of the future will be still more quantitative, more detailed, and more interested in interrela- tions and variabilities. INTEODUCTION 13 Of course, most of the objectives were not apparent, or even implicit, in the initial conscious formulations of this investigation. As in all studies, possibilities became evident as the methods and results accumulated. Initially the research appeared only as a means of relating certain quantities from studies of water in organ- isms. Thereafter the generality of the relations and their role in regulations gradually unfolded. It is also fair to say that I have long hoped to find general methods of understanding regulations in organisms, and it is almost usual that methods turn up to the conditioned (prepared) mind. The inquiry is divided into two chief parts : Part A deals with selected data and aspects of ivater exchanges and water content. Generalizations among measured quantities concerned with main- tenance of water metabolism lead, in Paet B, to comparison and extension of the same types of description to other components and exchanges. In this way certain general methods of study and some general properties of physiological processes are examined. Part A WATER RELATIONS OF ANIMALS Chapter II WATER EXCHANGES OF DOG § 6. Of the several approaches to the study of regulations, the compensations manifested when an animal has unusual amounts of water in the body call for intensive consideration. The object is to find how the water content is restored after it is disturbed from complacency, after water is forcibly added to or subtracted from what is usually there. Among the relations of water content, it seems to me desirable to study features available in all organisms and their parts and aggregates. The amount of water present in the living unit, and the rate at which it is gained or is lost are, in that respect, suitable quantities. Each has distinct dimensions. The amount of water present is measured either in physical units, or else in physiologi- cal units that are obtained by comparing the state of the organism having much or little water witl^ its control state as it exists before or after or omitting the treatment or condition. The physical unit is the liter or the gram ; the physiological unit is a relative measure, such as a relative increment of weight. The initial investigation of water exchanges by living units is limited, while dealing with one individual animal (g) in one set of observations (f) and by one path of exchange (p), to variables of four dimensional sorts: Water increment (AW), water exchange (R, or SW/At), time elapsed (t) after establishment of a new water content, and velocity quotient (k or 1/At). Rate of exchange, R, is thus a ratio of a quantity of component to an interval of time ; and velocity quotient, obtained by dividing a rate of exchange by a content, is a reciprocal of an interval of time. Intensive treatment of these four quantities (in chapters II to IX) concerns the one bodily component water. With regard to the component water, excesses are first imposed in the form of water by stomach and deficits in the form of priva- tion of water. Two types of water excess or deficit are succes- sively studied : the temporary state following a single administra- tion or deprivation of water (§ 7 to § 10) and the stationary state in which excess or deficit is maintained approximately constant for some period of time ('§12 and §13). Total exchanges are mea- 17 18 PHYSIOLOGICAL REGULATIONS sured, followed by separation into paths: Urinary water, evapora- tive water, fecal water, water formed in metabolism (oxidative), and water ingested as such (ingestive). One species (dog) is initially studied, and the whole body is a unit, it being of no concern for the present where within the body the water is. Confusion may be avoided by further defining a few quantities to be used, particularly in connection with the graphical represen- tation of data. Any departure of water content (W) from the control content (Wo) is designated as water load, ± AW. Quanti- tatively the unit of load is defined as an amount of water, weight, or other equivalent equal to one-hundredth of the control live weight of the body (Bo) or AW = 100 (W — Wo) /Bo. Each type of modification of the water content has a slightly different definition within the general class ; each may be measured by a partially dif- ferent procedure, or under different conditions ; and some will be in milliliters per gram, others in grams per gram. Similarly, interval of time (to — ti) is designated as At. Usually the unit of time is the hour, but the hour selected may be various : the initial hour, the hour of maximal load, an hour's duration of an instantaneous rate, hour computed from half -hours, quarter-hours, minutes, or days, and many others. It seems to me that clarifica- tion is obtained by multiplying distinctions just so far as actual or probable differences exist in the phenomena described. While the dogs are endowed with the diverse increments of water content, their rates of total water exchanges (SW/At) are measured, during the passage of time (t) after the increment is established. Rates are expressed in per cent of body weight per hour, or aW/At = 100(Wo — Wi)/BoAt. Thus I have elected an experimental situation and a set of vari- ables to be measured. Their definition constitutes the design of the initial investigation. § 7. Sudden excesses of water Water contents are conveniently modified by administering known quantities of water to dogs that are initially in control state. How long do excesses remain? Knowledge that water was the only liquid or solid given the dog saves the great trouble of evaluating the water contents of dogs sacrificed and analyzed at diverse inter- vals in each test. The quantities of water present and the ex- changes are both measured by body weight, since the amounts of WATER EXCHANGES OF DOG 19 other substances lost and gained are small in comparison to water and are known to be almost independent of water content. In each actual test a dog unfed for 16 hours but allowed water ad libitum is given water by stomach tube (at zero time). Urine is collected, and body weight is measured, every 0.25 hour. In plan the test requires little apparatus, for all measurements are made with a graduated cylinder, a sensitive balance, and a clock. In 0 12 3 Hours Fig. 1. Course of total water load (per cent of the body weight, Bo, at zero time) after diverse single quantities of water are given by stomach tube. Dog, four individuals. The number of tests indicated is also the number of measurements averaged for each point. In series A the initial water load was 6.14 per cent of Bq. New data of Kingsley and Adolph. practice each dog needs to be trained to stand still during several hours, to take a stomach tube without struggling, and to submit to continuous observation. Previously a bladder fistula or an exteri- orization of the urinary bladder is surgically produced. Ulti- mately many computations are required to place the results in coordinated and comparable form. 20 PHYSIOLOGICAL REGULATIONS In the course of time (fig*. 1), water is lost until the initial weight is approximately reattained. The rate at which water is lost in- creases as more is administered, and the whole process of recovery lasts but little longer. For each curve, all tests in which similar amounts of water were initially given contribute values of relative body weight ; all those values occurring in the same serial interval of time are averaged arithmetically. Rates of elimination are also averaged arithmetically, though with equal suitability harmonic means might have been used. The rates of elimination (fig. 2) vary with time as well as with the 0 12 3 Hour5 Fig. 2. Eate of total water output ( % of Bo/hour) in relation to time after diverse quantities of water (shown in figure 1) are given by stomach tube. Each rate is ob- tained from the change of body weight during 0.25 hour, and the average is plotted at the middle of the period in which it prevailed. quantity of water put into the stomach. Initially there is a lag period in which loss is unmodified ; thereafter rates increase quickly to maximal, followed by more gradual fall. With regain of control water content, the rate of water output approaches the initial rate. Very similar relations are shown (figs. 3 and 4) by measuring the exchanges through urinary channels alone. Exact comparisons and differences between total losses and urinary losses are not computed, inasmuch as figures 3 and 4 contain added tests that are not represented in figures 1 and 2. In this general account of mean results, no account is taken of WATER EXCHANGES OF DOG 21 0 12 3 Hours Fig. 3. Course of sensible water load (% of Bo) after water is given by stomach in a single dose. Sensible load is total ingesta minus urine collected. The same tests are represented as in figure 1, plus a few additional one on the same four individuals. 0 12 3 Hours Fig. 4. Eate of urinary water output (% of BoAour) in relation to time after a single dose of water is given by stomach tube. Same tests as in figure 3, plotted in manner of figure 2. 22 PHYSIOLOGICAL EEGULATIONS the variation among tests upon one individual. Also, recognizable differences exist among individuals, as though the capacities for metabolizing water were of diverse orders; these too are disre- garded. Further, there is significant acclimatization in an indi- vidual that has been repeatedly given water in large doses, whereby the speed of disposal is increased. Such matters may await later treatment. The course of water exchanges is partially different if, instead of the dog as weighed, one considers water content of the dog minus the alimentarv tract. Then the water in stomach and intestine is +2 - % I — r 1 1 1 dV A' c a// \ Dog - r V ^^ 1 1 1 1 ■~-— Hours Fig. 5. Course of water load ( % of Bo) after water is given by stomach tube. Loads represent the means in 9 individuals upon each of which several tests of urinary water output were completed. Absorbed volumes are the means in those 9 individuals when killed at various intervals within 0.6 hour after water administration, in order to find how much water remained in the alimentary tract. Data of Klisiecki et al. ('33a). Eedrawn by permission of the Eoyal Society of London. counted as being outside the body, and the rate of absorption modi- fies the increment or load of water present at any instant. Absorp- tion has been directly measured at only one load administered; with considerable variation among individuals, the amount ab- sorbed was found to be proportional to time elapsed up to complete absorption in 0.6 hour. The load characterizing the dog minus alimentary tract is indicated in figure 5. This diagram serves partially to locate the water at each instant during the test. Actu- ally most of the water given is still unexcreted when absorption is apparently complete. WATEE EXCHANGES OF DOG 23 Water is gained during these periods of water excess in only small amounts. When water is offered, the dog consistently re- fuses to drink it. Small quantities are being continually formed in the body by oxidation of organic compounds containing hydro- gen, and possibly by other processes, which do not vary signifi- cantly with water content or load (Bidder and Schmidt, 1852; Rubner, '02, p. 62; Heilner, '07; Lusk, '12). In general, losses of water are enormously increased, while gains of water are somewhat decreased, for some hours after water is given by stomach. The greatest modification is in rate of uri- nary output. The full amount of water ingested is only rarely realized or returned in the urine before the rate of urine formation comes back to approximately that of the control state. After administration of more than 2 per cent of the body weight, about 76 per cent is returned as urine in 3 hours (fig. 3). In total output the corre- sponding return is 82 per cent of the volume ingested (fig. 1). If corrections are made for the basal (control) rates of water loss, either urinary or total, the return is still less. Apparently the body does not treat quite all the water administered as excess. Some- what greater returns than those exhibited occur only after special preparation for the experiment, consisting in a previous adminis- tration of an excess of water on the same day (Klisiecki et al., '33a ; Kingsley and Adolph). Not only can the returns be ascertained after diverse periods of time have elapsed, but from figures 1 to 5 may be read the times required for initiation of diuresis, for maximal rates of excretion, for half return or half -life (or any other fraction) of the adminis- tered load, and for cessation of diuresis. All of these intervals of time except that for initiation are longer as the load is greater. Figures 1, 3 and 5 represent loads in relation to time. This relation in certain components of organisms is commonly termed a tolerance curve; the designation may be applied therefore to all curves relating load to time. I suppose the "logic" of the word tolerance is that the curve indicates how much added component the organism tolerates by removing it. High tolerance for water means fast disposal of an excess or a deficit of it, the opposite of indifference toward the increment. Rates of exchange in relation to time (figs. 2 and 4) may be designated as exchange curves. Water content in control conditions is approximately propor- 24 PHYSIOLOGICAL REGULATIONS tional to body weight, being about 62 per cent of it, in the intact dog (see table 18). A 1.0 per cent load or increment (± AW) as based on body weight, is therefore a 1.6 per cent increment in body water. Loads of water may be classified (fig. 5) into: (1) Total Load = Excess body weight, CDD' (2) Sensible Load = Excess administered, minus urine ex- creted, CEE' (3) Administered Load = Excess administered, C (4) Absorbed Load = Excess administered and not in the ali- mentary tract, AA' (5) Absorbed Total Load = Excess absorbed, minus amount lost judged by body weight, BD' (6) Absorbed Sensible Load = Excess absorbed, minus urine. excreted, BE' It might be considered poor technique to employ a large number of types of water load. In reality the number studied is the small- est number suitable for the purposes in hand. In the course of study it is poignantly observed that several confused variables are easily and usually thrown into a single category by virtue of having the same dimensions or names. The distinction of varie- ties of water contents may be a step toward clarity and precision in the study of the physiology of water. The tolerance curves shown above are approximately confirmed in unanesthetized dogs by less complete data of other investigators (Falck, 1872; Molitor, '26a; Rioch, '30; Abe, '31a; Hatafuku, '33a; Theobald, '34; Pickford, '36; Rydin and Verney, '38). Now I have described what happens to loads and rates of ex- change in the course of time. By correlating the ordinates of figure 2 with the simultaneous ordinates of figure 1, I compare rates of water elimination at diverse contents of water in the body. The course of each test or set of tests may be followed on the new coordinates (figs. 6 and 7). If corrected for absorption (fig. 8), only the early portion is modified. At any one time after the initial 0.5 hour (fig. 2) the rates of water loss increase with the excesses present. During a chosen interval of total elapsed time this is also true (fig. 9). After a while a maximal rate of excre- tion is attained, and thereafter at any one load (fig. 6) the rate of excretion is roughly independent of time. I believe it is signifi- WATER EXCHANGES OF DOG 25 cant, however, that at a given load slightly faster rates follow more recent maximal rates. The measurements of water exchanges after sudden establish- ment of water excesses lead to a relation of output to content that is rather uniform, for one is roughly proportional to the other. More exactly, if Rw is rate of water output at load + AW, a is rate of output at no load, and k' is a coefficient of proportion, then Rw-a = k'(AW). In figure 6, a = 0.15% of Bo hour, and k' = 0.7/hour. +> o /^\ -(^< o\ ^ -A - V ^f ' 1 Fig. 6. Eate of total water output (% of Bo/hr.) in relation to total water load (% of Bo). Each point in figure 2 is plotted against the mean load in figure 1, in the corresponding 0.25-hour interval. Under the conditions chosen, only urinary losses of water, and losses by no other paths, increase; evaporative losses from skin and from lungs do not, losses do not. Gains by ingestion of water are nil and metabolic production of it is independent of water increment. The net exchange of water (loss — gain) is on the average infinitely increased over that in control conditions, for in them loss equals gain over sufficient periods of time. In summary, excess of body water suddenly created in dogs by gastric administration of water leads to rates of water loss typi- 26 PHYSIOLOGICAL EEGULATIONS +2 Sensible Wa+er Load Fig. 7. Eate of urinary water output (% of Bo/hr.) in relation to sensible water load (% of Bo). Each point in figure 4 is plotted against the mean load (in figure 3) of the corresponding 0.25-hour interval. 3 O a a: Wa+er Load Fig. 8. Eate of water output (% of BoAr.) in relation to water load (% of Bq). Same data as in figure 5, each point representing a period of 0.25 hour. Data of Klisieeki et al. ( '33a). BE', absorbed sensible load; BD' absorbed total load; CEE', sensible load; ODD', total load. WATER EXCHANGES OF DOG 27 cally greater than usual. At each time, after a latency which is related to absorption of the water from the alimentary tract, the rate is proportional to the load. A family of curves of water tolerance is thus described. Recovery of water content ensues within a few hours ; it takes but little longer after large excesses than after small. The urinary route is the only path of output -H^^ 1 1 ■ I ■ 1 1 1 D09 • 3 O-h? - - u ^_^,.„— • „, "^ a ^^^^^^^"""""^^ • :^ +1 _ • % ^Jff^ — -^ • H- • • i O 5 Fig. 19. Total velocity quotient (1/hour) in relation to time after single administra- tion of water by stomach. These velocity quotients are obtained by dividing rate of total water output by total water load, using the data of figure 6. Times at which half the administered load had been disposed of are indicated (1) from figure 1. rates of excretion and at loads above + 1% of Bo serve to charac- terize more broadly the exchanges that prevail. Under limited conditions, then, the velocity quotient (rate/load) compares numerically the exchanges that occur at various loads, by various paths, at various times. Very often it is constant throughout considerable ranges of load or of time. It will be seen later (§71) that it is a parameter in an equation that describes the 40 PHYSIOLOGICAL, EEGULATIONS kinetics of recovery from water load. It allows numerical repre- sentation of the very exchanges that constitute adjustment of water content. § 11. The water-time system The data presented have been restricted to variables of four dimensions: AW (water load), SW/At (rate of water exchange), t (time), and 1/At (velocity quotient). Each variable appears in diverse conditions, and continued care is necessary to limit the data in any one comparison to those values obtained simultaneously or under stated conditions. Correlations among these four variables taken two at a time are of six types : AW vs. SW/At equilibration diagram (figs. 6, 11, 13) AW vs. t AW vs. 1/At SW/Atvs.t SW/Atvs.l/At 1/At vs. t tolerance diagram (figs. 1, 5) load-velocity diagram (fig. 17) exchange diagram (fig. 2) exchange-velocity diagram (not exemplified) velocity diagram (fig. 19) 0 +1 +2 ^3 To+al Water Load Fig. 20. Contour diagram interrelating rate of net water exchange (% of Bo/hour), total water load (% of Bo), time (hours), and velocity quotient (1/hour). Dog. Heavy lines represent sequelae of single ingestions of water by stomach, derived from figures 2 and 6 in positive loads, and from figure 10 in negative loads. Solid lines HHH' represent the dogs maintained at various steady water loads, derived from figure 29 of section 13. WATER EXCHANGES OF DOG 41 Three variables at a time can be represented in one diagram by one set of accessory contours. With a three-dimensional figure, four variables could be represented. For the present system of variables that is unnecessary, since one of the above four variables is a ratio between two others, and a two-dimensional figure is ade- quate (fig. 20). On this one diagram may be represented the data contained in all the foregoing figures. These four variables may be said to constitute the water-time system for this particular species under the prescribed types of load and the named conditions of environment. Such a system is delimited by the describer. Correlations among these variables characterize the responses to disturbances of water content with respect to time and to accuracy of recoveries. § 12. Stationary states of excess Thus far the responses have been described that follow a sudden addition of water or end a privation of water. Other physiologi- cal states have also been studied, in which the increment of water in dogs in maintained steadily. Do they also help in understand- ing equilibrations ? Positive increments (+ AW) are created by administering re- peated doses of water by stomach at equal intervals of time (figs. 21 and 22). The process of loading the body with water is most effective at first, for then nearly all the water put into the stomach stays in the body. With the passage of time and the increase of load, output becomes faster (figs. 23 and 24). Still, only at low rates of water administration (M and N) does the rate of output come to equal (at 1.5 hour) the rate of forced intake. Output also becomes stationary over a period of time, and might remain so for an indefinite period, if intake continued. When water administra- tion ceases (at 2.5 hours) excretion is well under way, and no lag occurs in the recovery ; in other words, the maintenance of a steady output and the beginning of a recovery output are one process. Throughout recovery the rate of water loss is closely proportional to water load, or, the falling curves (as well as the rising curves) of figures 21 and 22 are exponential with time. All this is shown in still another way by comparing simul- taneous rates of intake and of output (fig. 25). The sequence of exchanges is emphasized by this correlation. When the rates of water output are correlated with the loads 42 PHYSIOLOGICAL EEGULATIOISTS of water prevailing (figs. 26 and 27), it becomes explicit that the rates during loading are not equivalent to those during unloading. Hence even the procedure of repeated administration has its time factors, but not of a sign that would be concerned with lag in ab- sorption of administered water. During unloading, rates are nearly linearly proportional to the loads. 0 2 4 6 Hour^ Fig. 21. Course of total water load (% of Bo) when water is administered by stomach tube in ten equal portions (1.00 or 0.60 % of Bo each) during the first 2.5 hours. Dog, 2 individuals. The actual loads found by weighing are connected by dotted lines, while the mean loads of each 0.25 hour are marked by points. After 2.5 hours the points are actual body weights. New data of Kingsley and Adolph. Comparing figure 26 with figure 6, and figure 27 with figure 7, I find that the rates of output during recovery from repeated water administration are less than during recovery from single doses. I think the difference is significant in spite of the fact that diverse individuals, and the same individuals in dissimilar stages of accli- matization to water excesses, were used in some of the tests that are averaged. Hence it may be concluded that the rate of recovery depends not only upon the magnitude of the load but also upon the WATER EXCHANGES OF DOG 43 duration and course of the previous loading. Slow and prolonged administration yields slower excretion at any one load. Further evidence of difference is observed in other ways. The highest out- put rates in series K to N are found when intake rate exceeds output rate ; series A to F further exaggerates this factor. During intake the rate is higher than after it, as though incoming water is more promptly excreted than water already incorporated. Q 2. d 6 Hours Fig. 22. Course of sensible water load (% of Bo), i.e., ingested water minus urine excreted, with repeated water administration. In the first 2.5 hours, water is given by stomach tube in ten equal portions, and the loads are designated as in figure 21. Half of the maximal load is returned at 1. The total amounts of water given are: K 10.0% •of Bo, L 6.0%, M 2.14%, N 1.40%. Data of Kingsley and Adolph. Which of the data shown may be considered characteristic of the stationary state of load? A stationary state strictly exists only in those limited periods where neither contents nor rates of •exchange manifest a marked trend. A marked trend is (for the present) one in which the component may be shown to have changed, either consistently or statistically significantly, within three of the .successive periods of time in which its exchange is measured. 44 PHYSIOLOGICAL REGULATIONS 0 2 4 6 6 Hours Fig. 23. Eate of total water exchange (% of Bo/hr.) in relation to time. In the first 2.5 hours water is given by stomach in the equal portions at the rates of intake K' and L'. Eates of output are ascertained every 0.25 hour. The same tests are represented as in figure 21. Actually, the data show that both in the stationary state and in the recovery state following it {i.e., during diminishing loads), the rate bears the same proportionality to load. In other words, after the load ceases to increase, the rates of output in series K and L are regularly related to water load, and not to time. This conclusion could only be guessed until the consistent data were obtained based 0 2 4 6 8 Hours Fig. 24. Eate of urinary water output (% of BoAr.) in relation to time. The same tests, with repeated water intake by stomach, are represented in figure 22. WATER EXCHANGES OP DOG 45 0 12 3 4 Rate of Forced Water Intake Fig. 25. Eate of total water output (% of Ba/hr.) correlated with simultaneous rate of forced water intake (% of Bo/hr.). Water is given by stomach tube every 0.25 hour for 10 periods. Times are indicated by light dash lines; heavy lines follow the mean progress of any one rate of administration. The data are the same as in figures 23, M representing one test additional. upon many loads. Hence the rate of loss seems to be linear with load within the range of increments studied (fig. 26). More extreme positive loads of water were observed in dogs by Greene and Rowntree ('27) and by Harding and Harris ('30). Most of their data are incomplete for the present purpose, since either loads (body weights) were not reported or rates of loss were +1 +2 +3 +4 +5 +6 +7 Total Water Lood Fig. 26. Eate of total water output (% of Bo/hr.) in relation to total water load (% of Bo). Eepeated water intake during 2.5 hours, followed by recovery. The data are those represented in figures 21 and 23. In rising loads points are knobbed to the right, in f aUing loads to the left. 46 PHYSIOLOGICAL REGULATIONS ascertained only over periods lasting several hours. Many of their rates are higher than in figure 26, the maximum being 3.4% of Bo/hour. But whether they are still linearly proportional to water load cannot be ascertained; I infer they are not. Certain of the data given by Harding and Harris indicate that in some individuals low rates of water output accompany large loads, as though under stress of continued administrations of large amounts of water (10% of Bo/hour), water automatically accumulates (up to +20% of Bo) in individuals that excrete slowly. Descriptive physiology may proceed in the correlation of quan- tities without inquiring whether the hypophysis, or any other organ, regularly influences the excretion of water. Very often 0 +1 +2 +3 +4 +5 5en3ible Water Load +6 +7 Fig. 27. Eate of urinary water output (% of Bo/hr.) in relation to sensible water load (% of Bo). The data are from figures 24 and 22, water being given by stomach every 0.25 hours for ten periods up to maximal load. such an organ is regarded as a dictator of water exchanges. From the relations shown above it is evident that the dictator, if such there be, precisely grades the exchange to the load. Whether that fact robs the dictator of his title, is of little moment. To know in what anatomical part, if any, the grading of exchange to load is managed, is a project for research, but is not required for the present description of water maintenance. For, the quantitative relations shown in equilibration diagrams are explanations of water regulation to the same degree as relations of any other kind that have been discovered. What has been shown is that repeated gastric administrations of water at brief intervals ultimately produce a stationary rate of water loss. At diverse rates of administration, the resulting losses are proportional to water loads. The same proportionality con- WATER EXCHANGES OF DOG 47 tinues during recovery after administration ceases, but each rate of loss is somewhat less than when sudden excesses are given the dog to obtain an equal load. § 13. Stationary states of deficit Negative increments of water are readily maintained in dogs with fistula of the esophagus. Water taken into the mouth escapes into a measuring vessel, thus indicating the amount drunk by the dogs, but does not enter the remainder of the body to modify its water content. On constant diet these dogs may receive inadequate amounts of water by stomach, creating water (weight) deficits of Tig. Woter Load 28. Eate of sham-drinking of water (% of Bo/hr.) in relation to negative water load ( % of Bo) in two dogs with esophageal fistulas. Each measurement is the mean for a period of 24 hours. Twelve such measurements are averaged, in order of water or weight deficit, for each of the points plotted, giving the regression of rate on load. Eedrawn from Adolph ('39a). diverse magnitudes. Hour after hour the sham-drinking inter- mittently continues, at mean rates (fig. 28) proportional to the deficit prevailing. The intermittent drinking of the dog with esophageal fistula approaches in velocity only momentarily the sudden drinking of the dog that ingests (Adolph, '39a). But spread over an arbitrary interval of one hour or more the steady rate is found to surpass the initial rate. The intermittent character of drinking ordinarily allows time for absorption to follow ingestion. Evidently the ali- mentary tract meters the water taken, before any postabsorptive influence upon bodily composition occurs. The subsequent failure of the postabsorptive factors to confirm the earlier alimentary fac- 48 PHYSIOLOGICAL REGULATIONS tors as to whether water was obtained, releases more alimentary activity (Bellows, '39). Data are now available for a second equilibration diagram that concerns stationary states of water loads (fig. 29). Comparing in it the positive increments with equal negative increments, I note that all rates of net gain surpass rates of net loss. It is as though intake were on a larger scale (oversize) than output. The same statements are represented in another form by the velocity quo- tients (fig. 18, HH'). At any one load, simultaneous rates of gross gain and of gross loss of water may be compared numerically by taking their ratio. 6 ^ 1 1 1 — -I D09 « C o \ ■ >.4 3 \Goin o \ ■ -H Losi„^ — \ ^^^^.^-^"'^ n Loss ^--^"■"""^ Gen Total Woter Lood Pig. 29. Eates of water exchanges (% of BoAour) in relation to water load (% of Bo) in steady states of load. Equilibration diagram. In negative loads the data are from figures 28 and 12; in positive loads the losses are from figure 26, the gains from figure 13. This ratio (total gain/total loss) has been termed by Huber ('24) the water economy, or economy quotient. At balance the economy quotient is 1. The economy quotient thus measures the relative role of gain and of loss during attempts to recover. Thus, in selected conditions of nearly steady water load, uni- form rates of water exchange are observed. In them the equilibra- tion diagram measured is of wide generality, for it is independent of time. Whenever the water load is known, total and partitioned exchanges are also known. Whenever one rate of total exchange is known, the others and the water content may be predicted. water exchanges of dog 49 § 14. Summary To find how the dog adjusts the content of water in its body, unusual amounts of water are experimentally provided, and subse- quent movements of water into and out of the body are observed. It is found that excesses are removed chiefly through the kidneys, at rates that are nearly proportional to the excesses ; while deficits are made up with amazing promptness and exactitude by drinking. Both recoveries are by modifications in rates of exchanges that already are operating in turnover. Alternatively, excesses may be experimentally maintained by continual addition of more water, and deficits by not allowing the water that is drunk to be absorbed (esophageal fistula). Then the time elapsing since the water load was imposed ceases to be an important factor ; the rate of exchange is stationary. In all circumstances the relations between exchange and load serve to describe the processes concerned in recovery from water load. Such relations, represented in equilibration diagrams, indi- cate the events by which the usual water content is restored, and ordinarily is maintained. Partial representations are afforded by various numerical means : velocity quotient, ratio of modifica- tion, and economy quotient. These are several ways of comparing the modifications of water exchanges that occur in the presence of increments of water content. They concern only four sorts of variables which for convenience are said to constitute the water- time system of the dog. The relative effectivenesses of the separate paths of exchange (urinary, evaporative) are rated according to the speeds with which water flows through them at various increments. The net effects in compensating for unusual water contents are the alge- braic sums of these speeds. The fact that each rate of net water exchange is proportional to water load appears to be a condensed account of what the dog does to compensate for any disturbance of its water content. Chapter III OTHER TYPES OF WATER INCREMENT (DOG) § 15. It seems useful to examine further compensatory ex- changes of water, before the dog's other manifestation of water regulation are examined. In the investigation of water equilibra- tion already presented, only one general method of producing excess and only one of producing deficit are considered; positive loads are imposed through administration by stomach and negative loads through privation. These two types of positive and negative increment, and various sorts of stipulated conditions, were arbi- trarily chosen; scores of alternatives are possible. Data exist for water exchanges following some of the alterna- tive modifications of water content, and I now inquire what features of the physiological recoveries are similar, and what ones differ, among them. There is no way other than actual comparison of finding whether or not, for instance, what Keith ('24) called ''dehydration" is the physiological equivalent of what Gamble ('29) called ''dehydration." How does the dog indicate equiva- lence among possible deficits or excesses of water ? Does ' ' excess ' ' of water content always call forth polyuria, and one rate of urinary output 1 § 16. Excesses of water Types of water load may be provisionally grouped according to manners of their production. a. Does an anesthetized dog with water load differ from the same individual unanesthetized? It is widely recognized that it usually does ; the present object is to treat the differences as quanti- tative ones. In positive loads of water, the rates of elimination are diminished under the influence of ten out of eighteen narcotics tested by Bonsmann in a variety of concentrations. A few results are shown in table 2. The other narcotics, such as papaverine in the dose tested, do not diminish water diuresis. None of them augments the returns by significant amounts, indicating that, as usual, processes are not hastened by imposed agents. In negative loads the rates of water intake under anesthesia are zero. When stated thus, any alternative seems preposterous ; but that does not keep physiologists from doing experiments which 50 OTHEK TYPES OF WATER INCREMENT 51 TABLE 2 Urinary water losses of dogs, in the first 2.0 hours after giving water ty stomach plus administering narcotic or anesthetic Physiological state Num- ber of tests Water given, % of Bo Urinary water loss in 2 hrs., % of Bo Urinary water loss in 2 hrs., % of the water given Source of data Control 4 20 16 3 8 11 4 1.99 2.20 1.99 2.15 1.77 2.18 2.78 1.74 1.56 0.37 1.62 0.90 0.38 0.55 87 71 19 75 51 18 20 Fig. 3 Control Morphine (1 mg./kg.) Papaverine (6 mg.Ag-) ■•• Paraldehyde (1 ml./kg.) ... Luminal (50 mg./kg.) Ethyl ether Bonsmann ('30a) < < C ( C t C ( Bonsmann ( '30b) < ( I c presuppose that water balance is restored or maintained during anesthesia. Actually, no method has been devised of demonstrat- ing whether an animal that is in water balance before anesthesia continues to be so under anesthesia. Therefore all such tests rest on the supposition that anesthesia has not shifted the relation between water content and water balance. 7 0 I E 3 4 5 6 Hours Fig. 30. Course of sensible water load (% of Bo), i.e., forced intake minus nrinarj output. Dog. Water given by rectum (A), mean of 5 tests on one individual; data of Falck (1873). Water given by vein (B, C, and D) during 0.2 hour; data of Faick (1872). At 1, half the load is returned in urine. 52 PHYSIOLOGICAL EEGULATIOlSrS Sensible Water Load Fig. 31. Eate of urinary water loss (% of Bo/hour) in relation to sensible water load ( % of Bo) . A, water administered by rectum ; B, C, and D, water administered by vein; E, maximal rate in the third hour after giving water by stomach to the same 3 individuals. Each load is the mean during 1.0 hour. Same data as in figure 30, of Falck (1872, 1873). b. Various routes of administering water have been studied. Falck (1872) gave 9% of Bo of water by vein (figs. 30 and 31) with results far different from those when the same individuals received + 10 r^ 1 1 I - r *05 . D09 \ 5^ n - 1 1 1 1 1 0.1 2 hours Fig. 32. Course of sensible water load after single equal volumes of water are introduced by five types of procedure. Dog. Each curve is the mean of 4 tests and indi- viduals with bladder fistulae. A, tapwater introduced quickly by stomach; B, tapwater subcutaneously in 0.2 hour after zero time ; C, tapwater by vein in 0.2 hour ; D, tapwater subcutaneously in 1.0 hour; E, tapwater by vein in 1.0 hour. Data of Hashimoto ('14). OTHER TYPES OF WATER IISrCREMEISrT 53 water by stomach. The diuresis that follows infusion is delayed, less intense, and greatly prolonged ; thus diuresis is not apparent for six hours, during which oliguria prevails (D). Even with smaller administrations (1.9 and 4.1% of Bo, B and C) seven hours are required to eliminate in urine the entire volumes injected. At any one load, the rates are all lower than when the water is given by stomach. But 15% (Chiray et al., '38) to 21% (Falck) may be tolerated when given by vein. ■^ Q75 -♦- 3 o 3 S. O £1 "fe Q25 O a. Q50- - Dog 1 1 u A-i A ' M 1 1 0 +0.5 +1.0 Water Load Fig. 33. Eate of urinary water output ( % of Bo/liour) in relation to sensible water load (% of Bo). Single equal volumes of water are administered by five types of pro- cedure. Same data of Hashimoto as in figure 32. Hashimoto ('14) injected water by vein in less amount (1% of Bo). When given during a period of time equivalent to Falck 's (0.1 to 0.2 hour) only oliguria results (C, figs. 32 and 33). But given during 1.0 to 1.2 hour (E) some diuresis occurs, yet only sufficient to return as urine during it one-fifth of the fluid adminis- tered. Water given subcutaneously in the amount of 1% of Bo yields no diuresis whether injected rapidly (B) or slowly (D, fig. 32). When water is given by rectum (3.6% of Bo, tests A in figs. 30 and 31) diuresis is prompt, but less in rate and less prolonged than after the same load is given by stomach. 54 PHYSIOLOGICAL KEGULATIONS The upshot is that water is not the same everywhere. The site of the water and the rate of administration of it are distinguishable factors. It may be inferred that the water content of the body is, after one of these administrations, eventually adjusted to its initial value ; meanwhile an excess is present, often for many hours. It was found above (§13) that sudden additions of water lead to faster elimination than gradual additions of equal amounts. Hence any decrease in rate of passage or absorption is likely to decrease the rate of recovery. By no known rule is the rate of output regu- larly limited by a capacity of the kidneys ; time relations through- out the body are factors. +4 -I O 2 hours Fig. 34. Course of sensible water load in dogs subjected to four diverse regimes. Single doses of water are given by stomach. Each point represents the mean of 6 or 7 tests on as many individuals, the same individuals being catheterized hourly in each set. A and A', standard state; B, fed thyroid substances; C, deprived of food for 8 days but allowed water ad libitum; D, deprived of food and water for 8 days. In D the sensible water load does not represent the total water load, since water balance did not prevail at zero time as it presumably did in the other tests. Data of Hatafuku ( '33a, '33b). c. Dogs subjected to various regimes are given water by stomach. If deprived of food and water for 8 previous days (D, fig. 34), they show no diuresis when 3% of Bq of water is given. If previously given no food but allowed water ad libitum (C) they have diuresis, yet only half of the water given is returned in urine within 4 hours ; while all of it is returned by the same individuals upon control days. It is quite arbitrary to define water balance under such regimes. OTHER TYPES OF WATER INCREMENT 55 If fed with thyroid substance, diuresis follows the introduction of water to the stomach just as promptly, but sometimes yields less complete returns (B, fig. 34). The same procedure (so far as stated) in another laboratory showed continued greater rates of water output and equally complete returns during thyroid feeding (B, fig. 35). Here is a regime, the only one known at present, that augments the exchange above the usual. Privation of food super- imposed on the procedures mentioned diminishes the response to that without thyroid administration (Hatafuku, '33b). 3 o 3 a: Sensible Water Load Fig. 35. Eate of urinary water output in relation to sensible water load during the elimination of single doses of water given by stomach. A, standard state; B, thyroxin- fed. Each point represents the mean for four dogs, in 18 and 20 tests altogether, the same four individuals with * ' extended ' ' ureters being used for both sets A and B. Data of Klisiecki et al. ( '33b, p. 534). If the dog is injected with pituitrin, the introduction of water to the stomach induces no diuresis for some hours (Molitor, '26a; Klisiecki et al., '33), after which the effects of pituitrin disappear. Altogether less than two-thirds of the water administered is returned in urine. Poisoning with phosphorus (C, fig. 36) appears to reduce the 56 PHYSIOLOGICAL EEGULATIONS rate of output of water in recovery from water excess. But ad- ministering the drug novasurol (D) seems to remove the effect of phosphorus treatment. Physical exercise of running prevents or reduces the usual response to water introduction by stomach (Rydin and Verney, '38). This inhibition of diuresis may intervene at any time after diuresis has begun, and may outlast the exercise by various lengths of time. +4 0 12 3 4, Hours Fig. 36. Course of sensible water load after single ingestions of water by stomach. Each of four individuals is tested once in each series. Data of Abe ( '31c). A, in control state; B, given novasurol intramuscularly; C, treated with phosphorus; D, treated with novasurol and phosphorus. The general conclusion is that many regimes give rise to physio- logical states that modify the recovery of water balance. d. Various solutions may be substituted for water to constitute positive loads. Given by stomach, such solutions induce almost any degree of polyuria and oliguria, according to the solute, its concentration and amount (Chanutin et al., '24; Rioch, '30; Mel- ville, '36 ; Kaunitz, '37). Given by vein the return is no more com- plete (D, fig. 37) than when water is substituted (fig. 30). Super- imposed upon diverse negative loads of water (fig. 37) almost no OTHEK TYPES OF WATER INCREMENT 57 diuresis miglit be observed. Given by peritoneum still other rela- tionships are expected (Darrow and Yannet, '35) ; for oliguria and aposia (not drinking) now prevail together. Enough instances have been cited to illustrate the variety of responses (table 3) that may be observed after administration of excesses of fluid to dogs. To consider the dogs under these various administrations, conditions, regimes, and solutions in a single cate- gory leads to confusion. . Some of their contrasts are indicated in the right half of figure 38. In diverse sets of measurements after water is given by stomach under supposedly identical conditions, +100r I £ 3 4 5 Hours Fig. 37. Course of total water load (relative to volume infused) in dogs in four states. About 8 % of Bo of 0.15 M sodium chloride solution is infused by vein at zero time. Data of Davis and Dragstedt ( '35). A, 2 tests after deprivation of drinking water for 12 days; B, 11 tests after continued total loss of pancreatic juice; C, 3 tests after total loss of gastric juice for 8 to 11 days ; D, 3 tests in control state. the velocity quotient varies between 2.4 and 0.3/hr. The highest rates occur with thyroid administration (velocity quotient 3.5/hr.) ; the lowest with pituitrin and with intravenous water (velocity quotient 0.02/hr.). Yet all these varieties of water excess (and others too) are commonly referred to as states of ''hydration," ''positive water balance," "hydremia," and the like. In reviews {e.g., Adolph, '33, p. 348) observations from one or several of them are quoted in the same sentence as being coordinate facts, as occur- ring or predicted for all forms of water excess. Once it is recognized that the sequelae, of any one means or of 58 PHYSIOLOGICAL EEGULATIONS 1-^ S so »H to '^ 2>e 5- CO 'i^s v-,\ 05 1^ ?€<^ § s 72 1 Tjt CO CO H ^ - O o S^ bJD be bD tototo'«'«« is CO CO CO - " ^ rJH (m"- CO CO g o3 CM 00 ^ 73 03 •- Mcrt'- ^ CO ^^in t:^ . ^^^ 0) « « C<1 ^^ -^ -^ CO OQ CO O ••v "^-M ni Qj, (n C3 03 03 ^-tf • ■i± ^ u ■3.2-^ Ci CO ■<*< 00 <0 (M 1— 1 OT-Ht-tOt~tO-*(MTtlTt< O -tfrH 50 lO ■<*< I— 1 1— 1 1-H lO ostot^coo5ininoin(M O O O O O O .H (MOiHOOOOtHOO ^1^ • Ci °5 g o CQ ca o rH -"+1 o COCOMtOCOt^tOM .2-2^ «0 ?0 lO -rj^ t-; OJ CO .-Hi-HOT-iooi-<**in tSS^ O r-l O rH O O O ■ OOiHi-lrHOOO bt^ ^ '^ ■,: Oi t^ 0 -* 05 lOtOtOCOtMOdCOaSrHtO rH iH O O O O iH OaOOOOCIr-lTHiHiHO P ^ f^ 03 03 +3 fH t>- CO as o od th 00 icocqinrtitototot- o o CO OS in oj CO rHt^-*IO(M-^-^0 1-5 oa iH CO rH W CCl CO O , OOOiHiHOtHCQ Load at Ihr. t^ OS -^ to to O t- OiOOSrHCvl-^Ot-tOOO pH i-H Oi i-H CO O C<1 ooi— icooscotMinooi— ito o 72 OTHER TYPES OF WATER INCREMENT 59 any one agent that produces water excess, are not uniform, quanti- tative relationships can supersede categorical statements in de- scribing the enhancement and inhibition of water diuresis. At one load and time the rates of recovery of water content are diverse but are numerically comparable. 1 r m second and later hours -2 0 +2 Total Wo+er Load Fig. 38. Comparison of net exchanges of water in dogs subjected to several types of water load. Ordinates, % of Bo/hour; abscissae, % of Bo. The curves selected are each derived from 4 or more tests, as indicated in figures 39 (J), 16 (M), 29 (L, T), 35 (N, P), 34 (R, V), 6 (K), 31 (S, U) ; and treated with pituitary extracts studied by Molitor, '26a (X). Gains in negative loads are compared in the first 1.0 hour of recovery, except for the steady ingestion (L) by the dog with esophageal fistula; losses in positive loads are com- pared at times after 1.0 hour, being the maximal rates of net exchange found. § 17. Deficits of water Water-drinking ordinarily follows water deficit. No dog un- assisted gains water by route other than the mouth; dogs without food drink less (Kleitman, '27) ; dogs after physical exercise drink more (Gerhartz, '10) ; dogs ingest other amounts of any solution offered them in place of water (Wettendorff, '01). Both excretion and ingestion are quantitatively graded activities; so also are mobilization and absorption. An example of recovery from another type of water deficit is the sequel of sucrose infusion (Keith and Whelan, '26). After two hours of intravenous administration of 1.46 M solution, four hours are allowed for further loss of water through diuresis accompany- ing excretion of the sugar. Then water is offered during one to five hours, with the recoveries shown in figure 39. The quantities ingested are less at every deficit than those of figure 11, as figure 40 y 60 PHYSIOLOGICAL KEGULATIONS shows. Some of the differences might be due to reliance upon body weight alone as a measure of water deficit in both situations. The recovery from "dehydration" by sucrose infusion may differ from the recovery from the first type of water deficit cited (privation of dietary water, fig. 11) less than many other types do ; there are no suitable measurements of water ingestion after other types. In addition to the types listed in table 3, some that have been termed ' ' dehydrations ' ' in dogs are : privation of food as well as of water (Mayer, '01), hemorrhage (Wettendorff, '01), catharsis by magnesium sulfate (Tobler, '10), pyloric obstruction (Gamble and Ross, '25), intestinal obstruction (Haden and Orr, '23), re- moval of pancreatic juice (Gamble and Mclver, '28), superficial burns (Butler et al., '31), gas poisoning (Underbill, '19), histamine dosage (Underbill and Kapsinow, '22), insulin administration (Drabkin and Shilkret, '27), adrenal insufficiency (Loeb et al., '33), and parathyroid treatment (Shelling et al., '38). While certain features such as decreased body weight or increased concentration of hemoglobin in whole blood may be common to all these states, the danger of having a single term for all the states lies in the prevalent assumption that other characteristics such as water load and water exchange will be uniform. On the contrary, not only qualitative, but particularly quantitative, diversities may be pecu- liar to each type of modification. The statement is repeatedly made that ' * dehydration is accom- panied by reduction in the urine flow . . . and by sensation of thirst" Adolph, '33, p. 349; Gregersen, '38, p. 917). A search makes it evident that investigators who produced states of "dehy- dration" in dogs did not heretofore report the rates of urine flow, and no one has yet directly measured the sensation of thirst in dogs. Perhaps the statement is correct sometimes ; but no attempt has been made to limit the term "dehydration" to instances where the reduction and the sensation have been demonstrated. The investigators had, of course, other criteria, either in the blood or in the previous loss of fluid, of the fact that the body had less water than before. The term "dehydration" has many meanings, there- fore, often being not equivalent to "negative water load." No term in previous usage appears to be sufficiently exact to charac- terize an experimental state of water content and exchange. A second measure of recovery in negative loads of various types is the retention of injected solutions (fig. 37). Dogs in water OTHER TYPES OF WATER INCREMENT 61 10 - c o 0) -I- o cr 1 1 o 1 1 1 - o - o o\^ Dog o \. o ° \ o o o\ o\^ 1 1 i 1 1 > -\5 -10 Water Load Fig. 39. Water intake in relation to water load after sucrose infusion by vein. A 50 per cent (1.46 M) sucrose solution is injected by vein during 2 hours until 16 gm. sucroseAg. of Bo have been given. Then 5 hours more elapse before water is offered in limited amounts (for about 1 hour). Data of Keith ('24) and Keith and Whelan ('26). \ ' ' ' ' 1 r \ \ r \ Dog \ \ \ \ \ \ \ \ - \ \ . \ ^L \ \ \m \ y • y^ "-\\ y ^^^=-^=-^^^"M Loss "X ■ IP 1 1 — Gain 1 1 -4. -£ O ^2 +4 Totol Water Load Fig. 40, Eate of total water exchange in relation to total water load. Quantitative comparison of equilibration diagrams in dog. J, gain in the first 1.0 hour of recovery after sucrose injections (from fig. 39) ; K', loss in periods aper the first 1.0 hour of recovery after single administration of water by stomach (from fig. 6) ; L, L', gain and loss in stationary states of defiicit and excess (from fig. 29) ; M, M', gain and loss in the first 1.0 hour of recovery from deficit and excess of water (from fig, 13). 62 PHYSIOLOGICAL REGULATIONS balance rapidly eliminate most of the sodium chloride solution infused; those previously depleted do not. Comparable results follow the infusion of 0.28 M glucose solution in each of the types of depletion. The present study is being strictly limited to what occurs after the positive or negative increment of water content has been estab- lished. The additional paths, such as loss by rectal diarrhea, drainage of saliva or nasal secretions or lymph, or loss of blood or sweat, that might be involved in securing the water losses (load- ing), are not concerned in recovery except as the types of load brought about are distinctive of the particular manner of initial loss. "With the recognition of many qualitative types of water deficit, the expedient indicated is to designate each type by the procedure used in loading. Before the types are some day quantitatively characterized, the rates of exchange that prevail during recoveries from diverse increments may be measured. § 18. Distinctions among water increments Meanwhile how shall I judge when a water deficit exists? In one type of water privation, constant food is being added to the body. In ''dehydration" by sucrose a fluid is added to the blood stream such that more fluid than given is eventually lost from the body. In vomiting, dissolved substances, especially electrolytes, are lost in diverse proportions. In hemorrhage all substances are lost in the proportions present in the blood, but this is not the pro- portion present in the body as a whole. Only arbitrary judgments of what constitutes a deficit are available ; but once a criterion is chosen its consequences are provisionally accepted for purposes of classification and comparison. When 0.15 M solution of sodium chloride is infused, the distur- bance of water content is termed an excess (positive load). Is 0.30 M also an excess? The answer seems to depend in part on whether the urine subsequently formed is more concentrated or less concentrated than the fluid administered. A systematic study of continuous intravenous infusions (Wolf and Adolph) compares the gains with the outputs of water, after 7 hours during which steady rates of output are gradually attained. With a particular rate of inflow (0.9% of Bo/hour) the outputs become equal to the intakes if the salt concentration infused is either 0.11 M or 0.29 M. OTHER TYPES OP WATER INCREMENT 63 Between these limits water is slowly added to the body; below 0.11 M the water is mostly excreted and draws some salt from bodily reserves with it ; at concentrations higher than 0.29 M the water leaves the body faster than it enters, but some sodium chlo- ride stays behind. It is clear that the mere infusion of fluid is no guarantee that a positive load of water is established. The precise relations to solute and to previous depletion determine what regu- latory activities toward water come into play ; and the relations are already sufficiently complex that no one is likely to predict the responses to these types of water load from a knowledge of some other type. If 3.4 M solution of sodium chloride is infused, the dog promptly drinks water (Gilman, '37; Bellows, '39). The drinking itself might be regarded as a criterion of water deficit, even though the body weight meanwhile is increased. If 6.7 M solution of urea is infused in equal volume, less drinking results ; is the deficit less? One distinction that is useful is between an absolute deficit and a relative deficit, the latter representing a change in the proportion of water to at least one, several, or all other constituents of the body. In addition, an increment that is initially a relative deficit may become during the processes of adjustment and metabolism a relative excess, and vice versa. In both, however, recovery with respect to water consists in net loss of water in excess and net gain of water in deficit, for no other events constitute a restoration of water content. The rule of procedure which emerges is that every state of water content requires quantitative characterization by rates of water exchanges at least. Other characteristics may be studied to great advantage; such will be considered later (chap- ter X). Whereas in the water exchanges considered in chapter II only five chief paths are distinguished: urinary (sensible), evaporative (insensible), fecal, ingestive, metabolic (oxidative) ; in the water increments of other magnitudes and other types additional paths may be involved. Water in loads above +10% of Bo arouses in- tense salivation and actual large losses of water thereby (Weir, Larson and Rowntree, '22), especially when pituitrin is also ad- ministered (Theobald, '34). Vomiting is a response to rapid water administrations, but only when the water is put into the alimentary tract (Rowntree, '22). The partition of water exchanges among paths thus shows large contrasts among the several types that have been investigated. 64 PHYSIOLOGICAL KEGULATIONS It is apparent that an increment of water content {± AW) may be obtained by means that may not ordinarily be thought of as dis- turbing the component water, nor as displacing it in a recognized direction. There is no certain means of foretelling which condi- tions affect water load and which not. Indeed, it becomes probable that there are relatively few states in which the organism can be found that do not involve water loads and water exchanges. It is only arbitrarily and for present purposes that I concentrate atten- tion on increments of water and omit other modifications that accompany them. A limited method by which water exchange can be known surely to involve the same type of water load over a period of a few hours is to use a single kind of analysis or measurement for both rate (SW/At) and load (AW), provided water and oxygen alone are available from without. Thus, the weight changes of the dog give both data (exchange and increment) from one difference of weights taken at two times. In summary, a few factors may be specified that affect water exchanges under diverse types of water loads. An increment of water {± AW) is not often just a change in water content, even though no other chemical constituent of the organism be known to have changed. (1) Time (since ingestion or privation of water) makes a dif- ference in the rates of water exchange. (2) Means of addition of water matter. Thus, intravenous in- jection of water was found to produce highly variable results ; the administration being sometimes termed " unphysiological. " (3) Means of subtraction of water matter. Thus, the sequelae of catharsis by rectum may not a priori be confused with those of water privation. (4) Any accompaniment of addition or subtraction may be of consequence. There probably is no "pure" change of water balance. (5) Kation or regime upon which the water load is superim- posed may matter. Thus, whether the alimentary tract is empty or full while recovery is proceeding may be important. In general, each type of water load concerns water and a variety of circumstantial factors such as time, locality, responses incidental to introducing or subtracting the water. The bodily system studied is never a homogeneous one, as though a solution were diluted or OTHER TYPES OF WATER INCREMENT 65 concentrated with prompt mixing; but a highly diversified one, such that only one type of carefully specified procedure may give a set of reproducible results. What at first appears to be an incre- ment of a single chemical entity is in reality an arbitrarily chosen complex; indeed, no increment without a "complex" seems physi- ologically possible. The term "hydration," like "dehydration" and many other terms that might be cited, may have been first used (in connection with animals) to designate a particular change in the organism. Later supposedly similar changes were assigned the same name. The early attempt to group like states has now probably passed its usefulness, and a need prevails to separate these states according to their dissimilarities. A dozen types of water excess are here compared (table 3), and at least thirty more have been experimen- tally observed in part. Additional procedures and responses may at any time be distinguished and the present types be subdivided accordingly. Finally, I recognize that many investigators are more inter- ested in what goes on within the dog than in the overall responses to water load. How does the body recognize the presence of load? What tissues are excited, what ones transmit messages in accord- ance with the load present ? Implicit in the fact that exchanges are correlated with load and with one another, is the existence of co- ordination and its machinery. To small degrees their locations may be made out by methods of isolation and interference, making use of physical, chemical, pathological, and surgical procedures of various sorts. All those tools are also, however, specifications of the diverse types of water load; they are conditions of recovery. Here the emphasis is not upon the parts played by each anatomical or chemical bit of the organism ; yet the same facts are included in the account exhibited above. Those facts seem to me to furnish help in the study of regulations in this one respect, namely, how differently do dogs get along when their compensations are abolished? If all are abolished for long, dogs do not survive. But over limited periods of time each path of exchange, each means of communication and distribution, and each excitable tissue may be out of commission. Specific physiological factors for recovery of water balance in the dog are far from intimately known. For polyuria to follow, water may be administered by many routes; the most prompt 66 PHYSIOLOGICAL KEGULATIONS F^ 3 h-1 W => < Eh 5J) !» s t-. ^_^ CO 03 •-H OO "c3 13 Ph 1-1 00 03 fl £? o. O GO o ITS 1-1 p iH (M 00 ^ ^ « O oT to vi ft fl c3 • r^ -^ CO CO Pm 00 rH o CD o fl *« 00 -(J a < ft < i I-H -H ft ^« bi Veloci quotie of recove 1-1 1—1 05 00 i-H iH r-i d r-i Water ingested in first 1 hour of recovery, % of Bo O 00 r-J 05 _= Mean peak load, %of B cc C<5 !>. «q 00 CO 00 (O 05 Tt ai '^ i6 r-i l> t- IC o O. 1—1 1-H 1—1 1 1 1 1 1 1 1 1 1 «M tC Net rate o oadin %of Bo/hr CT o fl- 5 00 00 c c O c o o o 00 Co' r- co oc t-- ^ 2 fi Sq Tt to « 3 00 ^ eg IN 5C to S o 5 > 05 tc > -* 00 1 =4-1 _ Num ber 0 tests i iH -<1 H 00 1—1 t- (M u: y-{ nd ^ ^ C o o rf ^ ^ TS r^ rO « T3 « 2 '^ "S -s s c c ^-S 111 fi ■a a £ ra -g =« i =« 0 Q 1- PC 1 fi P 1 o O S" .B E _c in 03 a "S 'E nS 13 O 0 .2 > > fcH -^ -fl «H d 03 g^ CD =e ^ O fl -1-^ 5D O 03 o c ft .g be .-. ft m O o ^ 1 Do 'S Eh -rH 4^ ••:3 53 "^ CO 00 to '^ rH ""* P- P^ PL PM Pin tH ©q I- (M 1 OTHER TYPES OF WATER INCREMENT 67 elimination of the excess follows introduction by stomach. A whole circulatory system and at least one intact kidney are re- quired. The kidney need not be innervated, nor need any adrenal be present. For polyposia to follow, deficit is created in the pres- ence of intact swallowing machinery, but no one group of cranial nerves need be intact. The esophagus may be disconnected from the remainder of the alimentary tract. Some day all that informa- tion may possibly be arranged so as to explain water regulations to students of physiology. At the present time it seems to form little but a special paragraph in the anatomical story of the dog. The diversities of water exchanges sho\\Ti among the types of load-production that have been cited (tables 3 and 4) are : (1) Rates of both intake and output sometimes are increased during recov- eries (sucrose, sodium chloride). (2) Rates of both may be de- creased, thus preventing recovery (luminal, ethyl ether). (5) Water excesses may be temporarily retained, with less rates of out- put than are shown after introduction of water by stomach to control individuals. (4) Polyuria may appear only after a delay of some hours, or (5) polyuria may disappear before the positive load of water has been completely returned. The exact evaluation of the statistical significances of the diver- sities awaits further data. Physiological classification of the types of load might be based on the rates of exchange or the velocity quotients that result, thus minimizing the emphasis upon agents and conditions that prevail. What are the uniformities found among all the types of water load that have been mentioned? (1) Some change occurs either of absolute water content (ml. per 100 gm. wet weight) or of relative water content (ml. per gm. dry weight, or weight of some com- ponent). But the relative content may differ in sign as well as in magnitude from the absolute content. (2) Rates of water ex- change are modified with load. (3) Other uniformities are con- cerned with quantities outside of water exchanges and water con- tents of the whole body. (4) The fact that some features such as increased rates of urinary loss are common to several types of water increment is no guarantee that other features such as rates of salivation will also appear common. (5) Many possible rela- tions of water in the body do not occur in water loads of any sort. Some of these are : increased rates by paths other than ingestive, urinary or salivary; increased rates of ingestion in relative water excess ; augmentation in rate of loss without any lag. 68 physiological eegulations § 19. Modifications of water content at balance While positive and negative increments of water have been examined, little has been said of states in which water balance pre- vails. These states are both kinetic and stationary. When balance is defined as equality of intake and output, further qualifications are still needed, for in the dog intake is usually intermittent, and output is ordinarily continuous. Moreover, diverse unusual bal- ances can be maintained by repeated ingestions {i.e., in stationary states of load as fig. 24, N), for then within limits the average rate of output equals the rate of intake. Hence it is desirable to specify that the term usual water bal- ance applies to those states in which neither intake nor output is forced, where neither privation nor manipulative procedures inter- fere, and where sufficiently long periods (usually 24 hours) elapse, so that rhythms of feeding and sleeping shall be minimized. In particular cases control dogs may be put under more rigidly or less rigidly uniform restrictions of diet, movement, temperature, and the like; it seems quite impossible to define water balance with great generality and yet with rigor. Often a steady balanced state is approached, when the environ- ment or the body is changed, that differs from the state that would be recovered in the original environment or organic state. The modifications of water content and the rates of these shifts (total, net; gain, loss) may sometimes be measured also during the trans- ition. I know of no data that compare accurately an equilibration of Wo that has been shifted by a known content with an unshifted equilibration diagram in the dog. What is here discussed are, therefore, consequences of partial data as generalized in the light of relationships so far outlined. Some instances of modified turnovers are as follows. If dogs are deprived of anatomical connections between hypophysis and brain, the intake and the output of water (turnover) increase enormously and in two cycles (Bellows and Van Wagenen, '38). Administration of desoxycorticosterone induces persistently high turnover (Ragan et al., '40). Surgical imposition of Eck fistula increases the exchanges of water (fig. 41), with gradual recovery toward usual rates. Chloroform-poisoning gives a smaller in- crease of turnover; but ligation of bile-duct and certain other surgical procedures do not. Where on the scale of water contents OTHER TYPES OF WATER INCREMENT 69 S 500 §400 CO -^3001- -P :5 4- O 200 (U '% 100 DC > 0 Dog o .. Bile-duct ligated (2) -200 0 200 400 600 Hours after Operation Fig. 41. Eate of water ingestion (relative to last control period), in relation to time before and after surgical operation. Each period of time lasts 1 week ; the number of individuals averaged is indicated for each series. Data of Crandall and Eoberts ( '36). the operated individuals fall, compared with themselves before operation, is unknown. Rates of turnover are the only items that are known to be augmented in all those instances. Any content of water at which gain of water equals loss of water evidently represents a state of water balance. A positive content at balance is upon this definition one in which the new Wo exceeds the control Wo, whether or not rates of exchanges are modified (fig. 42). Criteria for measuring a positive or a negative content at balance may be of diverse kinds, and a shift in Wo perhaps as judged from body weight may be positive, while in W© perhaps as estimated from analyses of tissues is negative. A growing dog is ..,.., ,. ■ I"" 1 \ \ \ \ I \ -l— 1 009 \a \ \ \b \c \p \ ^ "■^'^— — \J^ '^ Ci^-^ Total Water Load -aw lotal Water Load • +^iw Fig. 42. Equilibration diagrams representing possible relations of the dog's rates of total water exchange to control water content (W„) and to one another. 70 PHYSIOLOGICAL REGULATIONS in just this situation ; the total water in the body increases with age but the proportion of water to dry matter decreases (Thomas, '11). Privation of food may accompHsh the opposite (Bothlingk, 1897; Witsch, '26), decreasing the total water and increasing the propor- tion of it in the body. A drink of 0.08 M solution of sodium chlo- ride increases the water content both '" absolutely" and relatively, but one of 3.4 M only absolutely. No sharp criterion (other than recovery itself) is alone suffi- cient to distinguish (a) shifts of content at balance (Wo) from (b) changes of water content (AW) without shifts of the Wo to which it will return when allowed. Possible criteria would be the demon- +4 .T+2 - o CQ c a E -2 - 1 1 1 Dog B' 1 /^ -6 /AC^ 1 }f' ^fO> 1 1 1 1 1 0 10 20 30 Days Fig. 43. Sequence of body weights upon 27 successive days. The dog received constant food once daily; water was continuously available. New data. stration that rates of exchange at some three or four diverse water ■contents have changed (fig. 42). Biochemical analyses that could determine the water content are at present much less accurate than the measurements of physiological recoveries that ensue. Over periods of time exceeding a few days, or in inconstant conditions 'of food intake, it is customary to credit changes of body weight {i.e., of water content) to shifts of the zero water load {e.g., see fig. 43). In general, wherever prompt adjustment of water content to zero load does not occur when conditions are judged proper for recov- ery, it is tentatively supposed that Wo has suffered change. One further step is suggested. When it becomes possible to select some criterion by which the water content of the dog is mea- sured independently of the water exchanges, then diverse modifi- OTHER TYPES OF WATER INCREMENT 71 cations of exchanges may be represented in a series of equilibration diagrams (fig. 42). In them the shapes of the curves for intake and for output have not changed, it is supposed, but only their water contents at balance have shifted in diverse degrees. In practice it is hardly possible to distinguish case BB from case CC, since both show the same pattern of equilibration, and the displace- ment of Wo in absolute or relative value might be quite small. Inhibition of intake and output together {e.g., by anesthesia or pituitrin injections) may not change the analyzable content but does temporarily modify exchanges. In such cases (§ 16, a, c) there are two control or ''normal" rates of exchange, one for the individual treated, the other for modal individuals or the same individual untreated. Case DB', which exhibits both polyuria and polyposia, might be judged to be in some state of diabetes insipidus, or of Eck fistula, or other unusual character. Cases like DE', which have small turnovers of water, exhibit oliguria plus oligoposia. If either the curve for water intake alone (BB' to AB') or the curve for water output alone (BB' to BE') shifts, then the value of Wo also changes. This contrasts with shifts of Wo which may occur without any modification of the relative curves (BB' to EE') for water exchanges. Changes in water content at balance (W©) undoubtedly are as worthy of study as are those that excite equilibrations. The diffi- culties in their accurate examination are greater, for Wo shifts only in conjunction with contents of other components. Hence the cri- teria of body weight and sensible water content are usually unsuit- able, and two values of Wo can be compared only by {!) full accounting for all water exchanges between the two, or (5) chemi- cal analysis of paired individuals. Comparison of this kind re- veals, however, the definite nature of the physiological state of the dog with respect to water. § 20. Summary Excesses and deficits of water may be produced in the dog in various ways and under diverse conditions. Each type of load requires distinct denomination, and accurate measurement of the responses to it. Several states of narcosis, routes of water ad- ministration, modifications of regime, and kinds of solutes added to the water administered are here compared. Most rapid negative loadings are obtained by intravenous infusion of solutions of 72 PHYSIOLOGICAL EEGULATIONS sugars (table 4, column 5). By various means water exchanges in the diverse types of positive and negative load may be abolished altogether, or may be tripled in rate (tables 3 and 4, colmnns 7), as compared with those prevailing in the two types of water loads of § 7 and § 8. A few criteria are suggested for classifying difficult types into positive loads and negative loads. The relative rates of various water exchanges during recovery themselves usually serve as cri- teria of the direction and amount of disturbance in water content. Sometimes unusual paths of water exchange aid in adjusting the contents. The content of water at balance is also modified by many agents. The rates of exchanges (turnover) that prevail at balance may or may not suffer modification at the same time. Each state of the dog with respect to water therefore calls for characterization both in regard to water content and in regard to modifications in rates of water exchange, as greater and smaller loads of both signs are present. In the story of regulations of water content, there could be a chapter describing the behavior of the dog toward water in the environment. What acumen does the dog show in finding water and avoiding water? Does it hasten its water gain and minimize its water loss by choosing appropriate surroundings when it is in water deficit? Unfortunately for the present enterprise, no quan- titative information is available in answer to those questions. This particular method of regulation has been studied especially in the rat (§43) and in insects (§47). Qualitatively there is much evidence that dogs, like most other animals, seek out and stay in environments that favor their main- tenance of water content. They avoid desert areas, they system- atically search for sources of water supply. In other words, they use sensorimotor abilities to evade serious difficulties in supplying themselves with water, and in surrounding themselves with a para- dise of water. By such choices they forestall the frequent use of compensations on any large scale; this is a prevention that pre- cludes the need for cure. Behavior can be thought of as a sepa- rate line of defense against water loads, modifications of water exchange as an insurance when behavior fails. More strictly both are coordinate and specific means of maintaining and recovering water content. Chapter IV VARIABILITIES OF WATER RELATIONS (DOG) § 21. Physiological regulation concerns the preservation of reg- ularity in some property that might otherwise show larger changes. It seems to me that a way to measure how much preservation occurs is to ascertain how much regularity prevails. That task is accom- plished by finding the natural distribution of physiological states, or, more specifically, of diverse water contents. Any maintenance of content or exchange may be regarded as the systematic preven- tion of unusual states or contents. It may be realized that such a measure of regulation fails to distinguish between what the organism does to preserve itself and what the environment contributes. This realization is a part of the discovery that the organism and its environment are insepa- rable. The presence of water instead of liquid ammonia in the animal body is a tacit recognition of the fact that the environment lavishly aids in supplying water and does not abound in ammonia. It seems to me quite inadequate to consider the anatomical bound- aries of the organism as the boundaries of a physiological system ; for the initiation of isolation is the end of the stationary state. Even an excretory organ works in continuous reference to atmos- phere and hydrosphere, whether or not the correlation between them be one that is explicit in the reports of experiment. Suitable methods of characterizing whatever irregularities oc- cur in the dog's water content are first required. Preferably the distribution of contents is observed in a single individual at suc- cessive equal intervals of time. Is the distribution random? What parameters lend themselves to expressing it? <§> 22. Variations of water content The fluctuations of water content in the individual dog may be analyzed like any other characteristic. Most methods of study ask the blunt question, are the fluctuations random? And in what respects are the fluctuations non-random? According to one definition, regulation is that portion of the change in content of water that is non-random. It will shortly be evident, however, that in many series of data on water content no 73 74 PHYSIOLOGICAL EEGULATIONS such kind of regulation prevails. Instead, the physiological states of water content turn out to be random according to most criteria of randomness, but confined within the restricted range permitted to them. Two problems therefore arise, to see in how many re- spects the time-series of water contents is random, and to derive parameters by means of which the variability among water con- tents may be compared with the variabilities of other components. Data consist in the mere sequence of body weights (fig. 43) in a dog consuming the same kind and amount of food each day, living under conditions arbitrarily fixed. The important aspect is that measurements are made at equal intervals of time in an individual upon uniform regime. (1) There is a marked trend in the series. Hence to analyze the fluctuations as deviations from a mean (C.V. ± 2.09) is of no significance for the study of regulation of water content as such. (2) The trend may be found (B = 17.335 kg. + 0.0460t), and the fluctuations may be regarded as deviations from it. The root mean square of this deviation amounts to only ±: 0.62% of Bm- (3) More simply, first differences between successive values may be taken, their root mean square obtained, and divided by V2 to correct for the fact that each value enters twice in the series. This parameter I designate standard difference, and relative to the mean ordinate, coefiicient of difference (CA). Here CA is ±: 0.67% of Bm. Only on one-third of days is the shift of weight greater than this. The latter two parameters, which are in any random series (as here) identical, serve to characterize fluctuations. They do not, however, utilize all the information about temporal sequence. In that lies more grist for the algebraic mill. (4) For instance, the frequency of inversions of body weight may be counted. In a random series they occur in 50 per cent of the first differences. Here they occur 73 ± 10.2 per cent of the times, which is not very significantly different. After the ± sign the standard error is shown. (5) Frequency of movements toward the line of trend may be counted. Randomly they occur in 75 per cent of the first differ- ences ; here they occur in 64 ± 10.0 per cent. (6) Numbers of points succeeding in one direction may be ob- served. Such a succession is termed a run, and when first and last point are both included in each run, it randomly consists of 2.5 VAEIABILITIES OF WATER RELATIONS 75 points (Kermack and McKendrick, '37). Here the mean size of run is 2.30 ± 0.18. (7) Numbers of points succeeding from one maximum to the next may be observed. This succession is termed a gap, and usu- ally consists of 4 points (Kermack and McKendrick). Here the mean gap is 3.67 ± 0.37. Hence, of many possible tests, the five (3 to 7) that have been selected for their wide applicability all indicate randomness in the dog's daily body weights. Regulation, therefore, consists not in steering the body weight from one day to the next, but in prevent- ing wider fluctuations. From a somewhat different aspect this means that one day is too long an interval for the steering to be visible. Accordingly, fluctuations at shorter intervals of time may be examined. At one-hour intervals, in which either no food is fur- nished but water is allowed, or both food and water are allowed ad libitum, the fluctuations are no longer random. If no attention were paid to their succession, the non-randomness would not be evident; only the parameter CA = 0.098% would be known. In- stead it may be noted (figs. 45 and 46) that within 4 hours there are no inversions, and the runs are the full length of the series. No longer is the body weight fluctuating ; it is diminishing throughout. That fact leaves two possibilities ; either regulation is absent for several hours on end, or body weight is no longer a measure of water content. Evidence to be presented below indicates that both are factors. When observations are extended over a sufficient number of hours, runs succeed one another. The fact that they are greater than 4 hours in length (actually 4.5) implies that inversions of body weight (when the dog drinks) are at such intervals. When the dog is weighed every 0.25 hour, the same result is exaggerated ; the mean run has 18 points. Such an event as drinking is, hence, anything but random in occurrence; instead it comes at rather regular intervals, and by it the dog obtains amounts of water that carry the body through several hours. The fact of periodicity is the clear expression of non-randomness, and indicates a charac- teristic of regulation additional to ** restricted fluctuations." Thus, over daily intervals only fluctuations are evident, but over hourly intervals periodic factors are apparent. The latter reflect the fact that the dog does not sip water every quarter-hour or even 76 PHYSIOLOGICAL KEGULATIONS every hour, but waits longer intervals to raise the body's water content. A special study was accordingly made (fig. 46) to find how much and how frequently dogs drink, under four sets of conditions at one particular time of day. (l) With water continuously available, but without food, dogs drink within 2 hours only in one of ten tests (E). The single amount drunk is somewhat less than the inter- vening deficit of body weight. (2) With food, water is taken by dogs once to four times after each meal (A, B, C,) even though a meal of dry food be eaten every hour (b). (5) If water is not available for 1 or 2 hours, it is not drunk if offered at the close of either period. At the close of 3 hours (not shown), it is drunk in half the tests, and then in amounts equal to about half the deficit. (4) If the dog be warmed for 1 hour, thus increasing the deficit of water, water is drunk as soon as offered in every test but one, whether offered during the heating (D), immediately afterward (G), or 0.5 hour afterward. From this I conclude that a dog does not sip water at short intervals. Most water is taken, as is well known, shortly after food is eaten. But in the absence of feeding or heating, water is likely to be drunk about every 4 hours, in amounts less than suffi- cient to restore the body weight. However, the lapse of time is probably less closely correlated with drinking than the lapse of body weight is, for when deficit is hastened by heating, drinking occurs as soon as - 0.5% of Bq has been reached. That is the great- est change of water content that a dog usually allows without doing something to remove it. Whether the fluctuation of water content represents a load or a shift of content at balance, cannot be entirely decided. When a dog stands hour after hour in a stall, losing water by evaporative and urinary paths, leaving water untouched, I may conclude that either (a) the dog is running into negative loads, and the sensitivity of the processes leading to water ingestion is too low to act, or (b) the dog is staying in water balance, water not being required to replace that lost until food has also been taken or until the clock gets around to some other hour. On the latter criterion the usual losses of water are eliminations of excesses that arise as metabo- lism proceeds, and the sensitivity of the responses by intake is pos- sibly as great as of those by output. On the former criterion of water content the organism sacrifices water to the benefit of excre- VARIABILITIES OF WATER RELATIONS 77 tion, and intake of water lags behind other means of compensation. In this instance, as in very many others, constancy of all quantities is evidently impossible. I can decide to what quantity the organ- ism is apparently insensitive only in terms of some specified mea- surement. Actually the weight of evidence is that both a and b are partial factors, basal body weight changing in a trend while ± AW oscillates about it. In all this there is no implication as to whether the dog is better or worse served by having a small range of water contents, i.e., a stricter constancy of it. There is no evidence that great constancy is more fit, or that infrequent inversions of net exchange are cheaper, or whether some mean between them is optimal. The present concern is to find just what constancy prevails under arbi- trarily chosen conditions, when the state of the dog together with the conditions impinging preserve the content of body water within the fluctuations observed. Insofar as the fluctuations are limited both in amount and in time sequence, their evaluation describes the organism as a preserved unit. For often ^'constancy is in itself evidence that agencies are acting, or ready to act, to maintain this constancy" (Cannon, '32, p. 281). The physiological significance of variations in content is, I be- lieve, that the limitation of the variations measures the maintenance of that content. Whenever content tends to change, activities on the part of the organism intervene to oppose the change. If this could be said in mathematical language alone, many possible mis- understandings would be avoided. For, such a form of expression appears to hold fewest connotations and implications. On the other hand, numbers and symbols would convey little idea of an organism's constancy and maintenance, were physiological terms not placed in parallel. Other measures of water content might be used in place of (!) body weight, and each would fluctuate when observed at 24-hour intervals. (2) Chemical analysis of a group of dogs, killed one or more every day, would measure content. (3) Metabolic retention, measured as total gain of water minus total loss of water, would be another. (4) Sensible retention (intake of water as such minus output of water as such) could be measured in 24-hour periods. (5) The volume of distribution (see § 58) of a substance such as urea or sulfanilamide could be repeatedly ascertained on one individual. 78 PHYSIOLOGICAL EEGULATIONS In order that the variations found shall be clearly of physio- logical significance, the deviations among control measurements of each quantity recorded are appreciably less than the deviations to be evaluated. This stipulation probably rules out all analytical methods (2, above). The method of measurement is preferably such as not to modify the general physiological state of the animal without cognizance being taken of that fact ; this reduces the utility, for instance, of method 5. Successive variations of water content and their statistical parameters indicate in part what sort of ^'governor" controls the diverse processes by which certain departures in content do not persist. Each serves to characterize the maintenance of the dog's water content as a whole, treating it as a specific project in hy- draulic engineering. Some of the particular characterizations suggested are : the coefficient of difference between successive body weights at a series of time intervals ; and frequency of inversion in the direction of change of water content, or reciprocally the mean time that intervenes before each inversion or compensation ap- pears. Additional numerical factors may easily be devised, each adding to the completeness with which performance is evaluated. § 23. Vaeiations of turnovers Rates of exchange that prevail in those states and conditions selected as control ones likewise vary. The data to be mentioned -5 -4 -3 -2 0 +1 '2 +3 Water Con+en-fc Fig. 44. Simultaneous variations of rates of water drinking and of body weight (or water content). Water content is in % of Bq, zero of its scale being the mean of the 27 values of body weight. Measurements were made in 24-hour periods on 27 con- secutive days during which uniform food, containing additional 3.05% of Bq of actual plus potential water, was consumed daily. Dog B', 18.0 kg. New data. VAEIABILITIES OF WATER RELATIONS 79 are listed in table 12. Measured in 24-liour periods the ingestive water intake of dog B' (fig. 44) has a coefficient of difference (CA) of ± 12.2 of the total water gain. Meanwhile the urinary water output has CA ± 22.7. Here too, is evidence of activities that pre- vent unusually low and high rates of water exchange. Part of the smaller variation of intake may be connected with the fact that about half of the water (actual plus potential) came with the food, which was provided in constant daily amount. The free water alone varied by CA ± 22.2. Water Load Fig. 45. Eate of urinary water output in relation to total water load (body weight) upon two days in water balance. Dog C, Bo = 13,885 gm. Urine was collected in con- secutive 0.25-liour periods from a bladder fistula. Water was continuously available, but was refused throughout the measurements. M = mean rate. I = root mean square of differences (standard difference) in (a) 0.25-hour periods, (b) 0.5-hour periods, (e) 1.0-hour periods. New data of Kingsley and Adolph. In shorter periods (fig. 45) variations in rate of urinary output in individuals with fistulous bladders are nearly the same as in 24 hours. At 1.0-hour intervals CA is ± 16, at 0.25-hour intervals =i: 19. These values and an examination of figure 20 suggest that rates of urinary output are smoothed out over periods of time greater than 0.25 hour. Instead of jumping about at random, rates 80 PHYSIOLOGICAL KEGULATIONS tend to be similar to those just preceding and just following them. Inversions in direction of change (accelerations and decelerations) occur 11 times out of 29 possible ones (38 per cent). All changes of rate are gradual; they too look as though a mechanical "gov- ernor" exerted an inertia that prevented sudden fluctuations. Or, whatever is concerned in exchanges exerts influences that tend to be uniform and continuous. Trends endure for hours rather than either for quarter-hours or for days. Comparing paths of water exchange, I find that ingestive ex- change is less variable than urinary over periods of 24 hours, but Hours Fig. 46. Total water load (% of Bq) in relation to time in dogs nearly in water balance. The number of tests is indicated in each series. Water was available to the two dogs at all times except on dotted lines. A, one-fourth of daily food was eaten at zero time. B, one-twenty-fourth of daily food was eaten every hour. C, one-eighth of daily food was eaten at 0 and at 3 hours. D and D, control regime = no food or heat. E, heated by radiator during third 1.0-hour. F, no food or water for four hours, then presented with water. G, heated during first 1.0-hour (no water), then presented with water. New data of Eobinson and Adolph. in periods of 1 hour and 0.25 hour is much more variable (CA ± 75, ± 180) as ascertained in tests of the sort shown in figure 46. Evap- orative exchange varies about as much as urinary, within each of the time intervals mentioned. Each of the quantities concerned in water intake (table 1), and in water output, is equally capable of evaluation by its variations. In the end I would learn precisely which are the more uniform processes and which less uniform in the handling of water by the dog, and what periodicities are characteristic of each. Adjustment of water content is hy means of the exchanges rep- resented in the fluctuations that are being studied. Content is a VARIABILITIES OF WATER RELATIONS 81 subtrahend of gains and losses (§4), which in turn are sums of exchanges by several paths. The interrelations among the rates of exchange are indices to regulations. Does each rate vary at random? Or does path pi compensate for the vagaries of ps'? Or is pi positively correlated with ps at any one time? Thus, Rown- tree ('22, p. 131) says: ''The total output of water is determined by the total intake." Is there evidence that pi is an independent variable, while p2 behaves in accord with it ? In the data of figure 44 there is found no correlation between ingestive gain and urinary loss. In short periods of time also 3 p-c -I 0 +1 Wa-fcer Load Fig. 47. Frequency of occurrence of diverse water loads (% of Bq) when measured at hourly intervals (A) compared with rates of water exchange (% of Bo) at diverse water loads (B, C, D). A represents data from table 12, row 2. B, net rate in first 1.0- hour of recovery, from figure 16. C, total rate in first 1.0-hour of recovery, from figure 13. D, total rate in stationary state, from figure 29. (fig. 45) there is no correlation between the two, since ingestion did not occur. Even if there were a correlation, there is no way of ascertaining, I believe, whether pi influences p^, or vice versa, or whether both are coordinated by some other factors. Very often dogs exhibit rates of urinary loss greater than the average at 0.5 to 1.0 hours after each spontaneous ingestion of water. It is possible to say that increased intake results, after a lag, in increased output. The events form a sequence. Another correlation found is that ingestion follows low rates of loss. In the 82 PHYSIOLOGICAL KEGULATIONS first correlation, output is modified after intake has changed ; in the second, intake is modified after output has changed. Separation in time is also no guarantee that one determines the other. The variation found in any one individual that is maintaining water balance is to be related to the form of the equilibration dia- gram (fig. 29). Any departure of water content from the usual involves a load, and a load inevitably means a rate of intake unequal to rate of output. The variability found is an expression of the occurrence of each load ; some minimal departure from balance ap- pears before measurable steps for its correction go into effect (fig. 47). Thereafter the rate of the correction (recovery) depends upon the difference between rate of intake and rate of output. If a large difference of rates (large net rate) occurs at a small load, it pro- vides a rapid means of recovery, a deterrent to variations of con- tent. Quantitatively, the standard difference at hourly intervals of time (fig. 45) is 0.098% of Bq. At loads of ± 0.098% of Bo the economy quotients (rate of intake/rate of output) are 1.4 and 0.7 (compared with 1.0 at balance) as inferred by interpolation upon the diagram (fig. 47, C) for the physiological maintenance of water relations of a dog under the conditions specified. Variation of con- tent and equilibration of content are two aspects of maintenance, showing what does not happen and what does happen, respectively, when the water content is near balance. The avoidance of extreme rates of exchange has been studied with respect to certain other components and species by Gasnier and Mayer ('39). They defined several useful measures of con- stancy in turnover, (l) Precision is the ratio of gain to loss in a chosen period of time. Here this ratio is termed economy quotient. (2) Fidelity is the difference (latitude) between largest and small- est precision, or presumably any other measure of distribution of precisions. (5) Sensibility is the percentage of successive observa- tions of rate between which the sign of the difference has changed (inversions have occurred). (4) Rapidity is the shortest interval of time at which the gain most usually equals the loss, or over which the precision is near to 1. These are ready methods other than first differences of working with data upon variability of exchanges in dogs that are maintaining water balance. Each of the four might be illustrated from the data of figures 44 and 45. Each term has here a specific definition that may not be confused with some other definition or connotation of the same word. VARIABILITIES OF WATER RELATIONS 83 Departing from conditions that I chose to regard as standard ones, I may measure variations in water content when the total exchange, (turnover) of water is enhanced, as in physical exercise (fig. 48). It appears that water is not only more liberally ingested, but more exactly restores the body weight, in a hot climate than in a cool one. In other words, a high turnover now maintains more uni- form contents than a lower turnover. In a like manner a variety 6 Water Load Fig. 48. Simultaneous fluctuations in rates of water drinking and in water load (or body weight, in % of Bo) in a dog (16.0 to 18.9 kg.) walking on leash in two tests in different climates and drinking ad libitum after each 1.5 hours. In the hot desert the dog made up 96 per cent of the total weight losses by concurrent drinking; in the cool climate 63 per cent. Data of Dill et al. ('33). of conditions might be studied; in each one the constancy with which water content is maintained would be ascertained quantita- tively. It is still unknown whether the variability of water ex- changes increases with the variability of the environment. Similar- ities and differences among individuals in any one set of conditions also remain to be found. In rating the accuracies (fidelities) of maintenance no assump- tion is implied as to whether the dog functions ''better" by having 84 PHYSIOLOGICAL REGULATIONS more constant rates of water exchange than less constant ones ; that is a question for investigation only after some criterion of "better" to which even the dog will agree has been discovered. In brief, the rates of exchange that occur under standard condi- tions for water balance vary by measured coefficients, rates through two paths fluctuating simultaneously by diverse amounts. Avail- able data do not indicate that adjustments are more regular or ac- curate by gain than by loss, or vice versa, nor that control is exerted upon intakes more or less vigilantly than upon outputs. Both are coordinate in the maintenance of water content. Only in the average of successive periods of time does total gain equal total loss. § 24. Variations in rates at diverse loads Recognition that contents and rates each vary while the dog is in a state of water balance allows the introduction of variabilities into the equilibration diagram. First of all, the point Wo on the abscis- sae (fig. 47, C) is actually not a point but a zone. Since periods of 1.0 hour were used in the measurements of rate that go into it, this zone has a width of o or ± 0.10 per cent of the body weight. It may be supposed as a crude guess that the variability of any body weight that enters into the estimates of water load is at least this large. Second, the ordinate rate of turnover (Rq) is a zone about ± 16 per cent of 0.25 %/hr. or ± 0.04%/hr. in height. Rate and load together give a rectangle of random variations 0.08 %/hr. X 0.20%, that usu- ally includes 2/3 X 2/3 = 4/9 of the observations. Third, each series of rates for which averages have been plotted, such as figure 2, has a variation peculiar to itself. Every point in every graph might therefore be rendered as a rectangle or ellipse. These rectangles or ellipses might also be employed, with modifiers for numbers of observations, as measures of error and of significance. One question that may be decided is whether the variation of rates at several water loads is proportional to the mean rate or whether it is independent of the mean. The former is approxi- mately the case, and for this reason coefficients of variation are much more useful than standard deviations in characterizing it. Next, its magnitude may be found. Among 7 tests on two individu- als (B, fig. 2), in the first 1.0 hour after water was put into the stomach, the rate of total loss varied by C.V. ±: 24, of urinary loss varied by C.V. ± 27, or identical values. Further, selecting 24 tests VARIABILITIES OF WATER RELATIONS 85 on three individuals (all those having water deficits exceeding 2% of Bo, fig. 11), the rate of gain varied by C.V. ± 32. It may be con- cluded that ingestive gain and urinary output are managed with equal uniformity. There is here no evidence that drinking is less accurately controlled by the dog than other exchanges. Another question is whether the variation is different if instead of averaging those measurements of rate taken at one time I ascer- tain the distribution of rates at one load. The variability turns out to be similar. Further it would be possible to find the distribution of times or of loads at which uniform rates prevail. Measurements of variability in rates of exchange make possible the detection of unusual responses of the individual. How small a rate of water excretion, following the ingestion of 3% of Bo of water, shall be considered pathological? There are data of Falck (E, fig. 31) which, so far as the methods and conditions of measure- ment are described, duplicate those of Kingsley (fig. 2). But the rates of output are much smaller and the lags much greater. It is difficult to believe that three individual dogs used by Falck, con- sistent among themselves, were genetically different from three used by Kingsley, fairly consistent among themselves. Diversities of method, race, diet, conditioning and acclimatization that will produce such wide divergencies of result remain to be precisely investigated in some one laboratory. Again, ''distilled" water by stomach in quantities of 1 and 2% of Bo did not invoke diuresis, while tap water did (Hashimoto, '14). All distilled water is not similar. But few investigators would have considered it worthwhile to test the differences among samples of water, yet to disregard a difference in water administered is physi- ologically confusing. Unfortunately Hashimoto did not demon- strate that the fact of being distilled water was what mattered in those tests, nor did he show that the difference was a significant one. In general at diverse rates of water exchange, the variations are relative to the mean rates. In recoveries from water deficits the initial water intakes vary by C.V. ±: 32. In recoveries from water excesses the initial outputs vary by C.V. ±: 24. Hence the responses to negative and to positive water loads are of equal accuracies. Unusual responses may be adjudged by reference to these manifest variabilities. 86 physiological regulations § 25. Summary Variations of water content and of rate of water exchange in the dog are indications of the latitude permitted in the regulations of these quantities. Fluctuations outside the usual are opposed by activities or processes that tend to restore water balance or flow. Often the fluctuations found are not random in their sequences. Differences among successive values are influenced by durations of period of measurement, numbers of individuals, conditions of ob- servation, and other factors. From any one set of measurements, however, accurate characterizations of regulatory activities, with respect to water content or to rates of water exchange, are obtain- able in terms of statistics of distribution. Variabilities of water content or exchange may be correlated with variations in the environing conditions, and in states of the dog both present and previous. Evidence exists especially in water ingestion, as part of turnover, of long-time oscillations. The fluctu- ations permitted characterize, equally with the mean position of each content and rate of exchange, the ''normal" or balanced state. There is no evidence that during turnover water intake sets the pace for water output or vice versa, or that one path of output compensates for irregularities in another path of output. During recoveries, rates of ingestive gain in water deficits are about equal in variability to rates of total loss in water excesses. The variabili- ties in usual individuals being known, unusual features of compen- sations of water content in other individuals may be detected. This marks the end of the formal account of the water-time re- lations of the dog. Even within the limited selections of variables and of types of water loads the account is far from exhaustive. The dog's recoveries from sudden excesses of water and re- coveries from deficits of water were studied in relation to amounts of water load (± AW). Total exchanges, net exchanges, and par- titioned exchanges were each covariates of load, their regressions forming equilibration diagrams. The responses were evaluated in terms of modification ratio, velocity quotient, tolerance, promptness of modification, maximal rates, and completion of recovery. Sta- tionary states of water load were similarly investigated. It was found that excesses of water in the body were eliminated, almost exclusively through the kidneys, at rates proportional to the load. Deficits of water were dispelled by ingestion of amounts that were VAKIABILITIES OF WATER RELATIONS 87 slightly in excess of the believed load and proportional to it (chap- ter II). Recoveries from numerous types of water excess and of water deficit were compared, each type representing diverse states of the dog and diverse conditions for recovery. Several criteria were found whereby a water load might be distinguished from a new water content in water balance. It was suggested that all modifi- cations of water exchange and of water content may be classified qualitatively, while within each class quantitative diversities pre- vail. These types represent the physiological states of the dog, the diverse accompaniments of the water administration, and the paths by which water is administered and returned. Selection of those environments in which water losses are minimized and access to water is easy, may be of significance in maintaining the character- istic water content (chapter III). Physiological variation of water content in dogs serves as a measure of its regulation. With respect to short periods of time, successive variation often differs from random variation, denoting a temporal characteristic of possible governors. Several methods of evaluating and comparing the variabilities and their sequences are suggested, both for contents and for rates of exchange. Rate of water gain is not more variable than rate of water loss in periods of 24 hours, but in periods of 1 hour gain is much the more variable of the two. No further evidence was found that content is regu- lated more accurately either at gain or at output, and compensa- tions for departures from balance occur equally at both. In ex- changes during recoveries the variabilities are relative to the mean rates rather than absolute (chapter IV). Chapter V WATER RELATIONS OF MAN § 26. As the physiological pattern of the dog with respect to water exchanges has been described, so it is possible to character- ize quantitatively the water exchanges of man. In early crude studies it was implicitly supposed that all mammals are alike and that approximate facts about one species would hold for all. Now, it is easily demonstrated that the similarities, extensive though they be, do not extend to all characteristics of water relations. In fact, it becomes apparent that precision of description differen- tiates species and individuals physiologically with respect to these very characteristics. One description, even when thorough, can- not suffice for two descriptions. Descriptions of two or more species allow the identification of uniformities as distinct from diversities. Does a man compensate for disturbances of water content just like a large-sized dog? Does identity of organs of exchange mean that functions are quantitatively alike? If not, are the rates of water exchange correlated with other differences between the species? Proceeding with the investigation in somewhat the same man- ner for man as for dog, I analyze analogous data on water ex- changes. At first the data are restricted to the same four kinds of variables (water content AW, water exchange SW/At, time t, and velocity quotient 1/At). The responses are described follow- ing single ingestions by mouth, repeated ingestions by mouth, and desiccations by privation. From these, an equilibration diagram is obtained. § 27. Single ingestions by mouth Tolerance curves (fig. 49) for excesses cover longer times than in the dog. Or, in the larger species the half -life of the water load is greater. The curves are little modified when alimentation and absorption (see fig. 62) are taken into account. The coefficient of difference in rates of total water loss between successive control periods of 1.0 hour each is ± 13 per cent of the mean rate or ±: 0.020% of Bo/hour, and in rates of urinary excre- 88 WATER RELATIONS OF MAN 89 Hours Fig. 49. Course of sensible water load (% of Bo) after diverse amounts of tap water were ingested. Man. All tests (as indicated) in which the same quantity of water (relative to body weight) was ingested are averaged. l = half of administered load was returned. New data of Kingsley and Adolph. tion is ± 21 per cent of the mean rate of ± 0.012% of Bo/hour. With these as criteria of significant increases in rate, the latent period of diuretic response may be said to end when the urinary 0 12 3 4 Hours Fig. 50. Eate of urinary water output in relation to time. At zero time diverse amounts of tap water were ingested; all tests in which the same quantity of water (relative to body weight) was ingested are averaged. Same tests as in figure 49. 90 PHYSIOLOGICAL KEGULATIONS rate exceeds the control by double this deviation. The latency varies from 0.3 to 0.7 hour (fig. 50). The maximal rate of water loss is regularly attained later with large administered loads than with small (fig. 50). Completion of diuresis may be said to occur either (a) when the rate of urinary water loss once more comes within ± 42 per cent of the rates characteristic of control periods at the same hour, or (b) when the rate decreases by less than it fluctuates in successive periods of measurement. The times in hours required to complete the diuretic response are related to the volumes administered (AWc) by the equation tc = 1.3 + 1.0 AWe. Similar rules hold for diverse fractions of unloading, such as, for the half-life of the sensible water load. The volumes of water finally eliminated (returned) differ from the volumes ingested by a constant positive amount, and not by a uniform fraction of the initial load. Part of this volume of reten- tion vanishes when a preparatory water administration just pre- cedes the test. Three manners of computing the volumes elimi- nated yield nearly similar returns : (a) total water loss, (b) urinary water loss, (c) total or urinary water loss in excess of that elimi- nated in equal control periods, either at some arbitrary time or at completion of diuresis. If one must know how large a drink of water can be disposed of in 3 hours, the first comparison is used. If, however, the body is allowed unlimited time in which to adjust its water content, the second or third might be preferred. Losses by the several paths are found not to increase significantly with water load. It has been previously concluded that all the water ingested by a man in water balance is returned as excess urine by the time the diuresis ends (Adolph, '21). The evidence now available is more extensive and precise; from it (fig. 49) the conclusion appears to be that in man as well as dog the urine that issues at rates above control ones is significantly less than the water administered. This may mean that some of the water drunk takes the place of that which is being ''catabolized" or made available for excretion from other sources. It may be supposed that the man or dog is never '' saturated" but at best maintains himself in slight deficit, and so retains small absolute amounts of the water ingested. Other data (A, fig. 51) indicate returns in urine alone that equal WATER RELATIONS OF MAE" 91 the loads administered. How to obtain one or the other result at will is not known; a priming ingestion 3 hours previous to the test somewhat augments the return, however. Initial rates at which excess water is eliminated increase with water load up to a maximal rate (fig. 52). The later rates are Hours Fig. 51. Course of sensible water load (% of Bo) after a single ingestion. Indi- viduals with two types of hepatic disease (jaundices) were compared with control indi- viduals. Only in one, obstructive jaundice, was the excretion of excess water in urine significantly retarded. A, 12 control individuals; B, 7 patients with diagnosis of ca- tarrhal jaundice; C, 4 patients with diagnosis of obstructive jaundice with cancer of the bile duct. Data of Abe ('31a). more rapid, and still later rates decline. Once the rates of elimi- nation have begun to decrease (fig. 53), the rates are significantly independent of time and are related to load. Net velocity quotients (1/At) are highest at moderate loads (about +1% of Bo). The maximal values attained indicate what 92 PHYSIOLOGICAL EEGULATIONS +• a % 3o Mon 1 1 1 1 1 1 second one hour •x" 0 A / • _ O y • • 1 firs-t one hour i> 1 • A 1 * 1 1 1 o +1 +2 *3 +4 Mean Sensible Wa'ter Load Fig. 52. Eate of urinary water output (% of Bo/hour) in relation to mean sensible water load (% of Bq). Each point represents the total urine put out in one hour in a single test of figure 49. Two subjects are distinguished by triangles (K) and circles (N). For each test a single ingestion of water occurred at zero time. the organism can do in recovering from a single ingestion of water if given time to get under way and if given an optimal load. Values of 1/At may, of course, be of largest magnitude when values of SW/At are not maximal. The sort of data discussed above is confirmed in a less sys- tematic way by early studies of Falck (1852), Ferber (1860), and many others. Those studies differed mainly in furnishing mea- surements only at 1-hourly intervals, in not relating exchanges to body weight, in employing only one subject, and in measuring urinary output but not total output of water. Falck demonstrated that the time relations of the compensatory urinary output differed but little when water was administered by rectum instead of by mouth. O QC Sensible Wa+er Load Fig. 53. Eate of urinary water output in relation to sensible water load after single ingestion of water. The same tests are averaged as in figures 49 and 50. WATER RELATIONS OF MAN 93 While augmented excretion is proceeding in the presence of water excess, the opportunity to drink water is consistently re- fused. The only known accession of water is the slow formation by oxidations; its rate is uninfluenced by water excess (Carpenter and Fox, '30). In general, the rates of elimination of excess water by man are smaller than by dog (relative to body weight), and are slower in accelerating. Eates are less nearly proportional to loads. The amounts of water returned in diuresis are in both species less by an absolute volume than the amounts administered. It may be stated that after water excess the total water losses do not differ significantly from the urinary water losses. § 28. Repeated ingestions by mouth What happens when the accession of water to the body con- tinues? Water is ingested in 8 to 16 equal portions at intervals of 0.25 hour (Kingsley and Adolph, new data). Latent periods for appearance of diuresis vary widely, being very long at small rates of ingestion. Times required to reach maximal rates of diuretic response also are diverse in much the same proportion. Ulti- mately stationary rates of elimination of water are attained ; and at some rates of ingestion a steady state of content also is main- tained for an hour or more. In these tests the three factors of intake, output, and load are physiologically fixed with relation to one another. Even in moderate loads the rate at which the body excretes water appears to reach a limit (figs. 54 and 55). But temporarily at least this rate of excretion may exceed the rate of intake (Wendt, 1876). Velocity quotients (1/At) in steady states of ingestion and load resemble those that prevail during decreasing rates of water ex- cretion after single ingestion, if compared at equal loads. The quotients decrease with load, whether the loads be total or sensible, and whether the rates of exchange be total or urinary. After ingestion of water ceases, rates of elimination gradually decrease, with diverse periods of persistence and diverse decelera- tions. After those states have prevailed in which elimination rates equal ingestion rates, cessation of diuresis requires less time than attainment of its maximal rate does. But where ingestion rates are very high, some hours may be consumed in completing the 94 PHYSIOLOGICAL EEGULATIONS ■P n :s ~o ^ 1 t 3 1- Z3 4- o o Q) +• O cc 0 Man +1 +2 Total Water Load +3 Fig. 54. Eate of total water output (% of Bo/hour) in relation to total water load (% of Bo). Stationary state. One individual in 5 tests. Each point represents the rate that prevailed during one period of 0.25-hour, beginning 1.75-hour after the first of ten equal ingestions at 0.25-hour intervals, and continuing thereafter. Open circles, during water ingestion; solid points, after ingestion ceased. Data of Kingsley and Adolph. elimination of excesses, indicating definitely that water loss lias previously failed to maintain proportionality to water load. The termination of diuresis appears to be a smooth and orderly event. The diuresis is at these times a correlative of the water load; as the water load disappears, excretion diminishes without measurable lag. Suppose, however, that the excretion were de- Sensible Wa-ter Load Fig. 55. Eate of urinary water output (% of Ba/hour) in relation to sensible water load (% of Bo). Three individuals in stationary state, 21 tests. Water was ingested at 0.25-hour intervals, usually in 16 portions but varying from 8 to 16. Solid points, maximal urinary rate in any 0.25-hour period before ingestions ceased; open points, maximal urinary rate (and coincident load) in any subsequent 0.25-hour period. Data of Kingsley and Adolph. WATER RELATIONS OF MAN 95 pendent upon the load a half-hour previously (that the response occurred with a lag such as occurs at the initiation of diuresis). Then the return in diuresis might easily overshoot the volume ingested; an initial water excess would give rise to an ultimate water deficit. Hence the rate of diminution (deceleration) of water output might conceivably be an optimal process. Were the deceleration faster, overshooting might frequently occur. Were the deceleration slower, the restitution of water balance would appear to be less sensitive to changes, and the tail-end of any diure- sis would be much prolonged. Unfortunately no quantitative expression can be given to this virtual optimum, the absence of other factors (as always in problems of optima) being only a sup- position. But I wish to emphasize that this deceleration is of consequence to the organism. Without presenting further data for man, I conclude that the rates of water loss are very similar during stationary rates of water intake and during recovery (beyond the first 1.0 hour) from single water administrations. The only significant diversities are in loads less than + 1% of Bq; in them unidentified factors exert their influences. But in both repeated and single administrations, the constancy of rate of output with all loads above + 1% of Bo in man is in marked contrast to the dog ; it implies a maximal capac- ity for elimination and an inconstant velocity quotient. It may be noted, however, that some human individuals have exhibited higher rates of elimination than any shown in figures 52 or 55 {e.g., Haldane and Priestley, '16) ; these individuals appear to augment their losses as the loads increase above -j- 1 or even + 2% of Bq. In any case the rates of water excretion are proportional to water load over a narrower range in man than in dog. <^ 29. Water deficits Water deficits are studied by the following procedures. The subjects deprive themselves of all liquids and of foods containing large proportions of water, but maintain adequate food intakes. The diet is not strictly constant in some cases ; in others it consists of uniform amounts of butter, sucrose, and casein. Usual activi- ties of the subject are continued. After measured losses of weight, water is offered either in ad libitum amounts or in uniform amounts at ad libitum times. The quantities drunk are ascer- tained, and they represent the gains of water (fig. 56). 96 PHYSIOLOGICAL KEGULATIONS A^-4Wln.t.al = -2.06/o 0.6 Hours Fig. 56. Amounts of water voluntarily ingested (% of Bo) in relation to time after drinking was allowed. In Pj the deficit (-AW) was 4.48% of Bo at zero time; in A4 it was 2.06%. In Pj, 1.45 hours were required for half recovery, in A^ 0.75-hour. New data. The curves very roughly correspond to equations: AW = -0.63 log t and AW=-L22 1ogt. The ingestions are always found to be leisurely, after the first minute or two, and their rates decrease with time (fig. 57). At 1.0 hour the total amount drunk is usually only half or less of the orig- inal deficit of water as measured by body weight (fig. 58) . Nothing being known about the efficient causes of this characteristically slow drinking, other sorts of "reasons" for it may be imagined. 10 Hours Fig. 57. Eate of water ingestion (% of Bo/hour) in relation to time after drinking was allowed. Each rate is plotted at the middle of the period observed. Same data as in figure 56. WATER RELATIONS OF MAN 97 One such is that rapid drinking would allow the loss of the new water by diuresis, while slow drinking avoids such an effect of sudden accession. But when tested, the small diuresis that follows slow voluntary drinking is not significantly augmented when water equivalent to two-thirds of the deficit is forced down at once. At diverse water loads (fig. 59) the rates of gain differed in absolute magnitude. A high correlation between them is evident, and the mean rates of initial drinking are proportional to the deficits. +20 -ao ^ -80 Hours Fig. 58. Eelative water load in relation to time. The amounts drunk are computed as percentages of the deficits (loads) that initially prevailed for all tests in which the deficits are 2% or more of Bj. The numbers in parentheses indicate how many tests are averaged. A, dog in laboratory; B, man in desert; C, man in laboratory. At 0.5- hour, the difference between desert and laboratory is shown by the t test of Fisher to be significant with a P < 0.01. Data of Adolph and Dill ('38), and Adolph ('39a). To say that the rate of water intake is a psychological measure- ment, neither adds to nor subtracts anything from the result. No matter how many conditionings and experiences may influence the drinking, evidently the choice of rates is not unlimited. Moreover, most features of ingestion (table 15) are reproducible not only in one individual but among individuals. The behavior concerned in water drinking belongs to the cate- gory of "operant" behavior (Skinner, '38), the responses of inges- 98 PHYSIOLOGICAL KEGULATIONS tion being measured without reference to any recognized stimulus. Instead they are correlated with water load ; alternatively, load can be inferred to accompany some unknown stimulus, as is frequently imagined. Gain of water by oxidative production is not modified by deficits of water in the body as indicated by the rates of basal oxygen con- sumption (on 7 days of deficits ranging from - 1.5 to -4.9 AW, in 3 tests of Pinson and Wills). In water privation the rates of water loss diminished signifi- cantly below those in states of water balance, when ascertained on Water Load Fig. 59, Eate of water drinking (% of Bq/O.S hour) in relation to water load or deficit (% of Bo). Intakes were each measured in the first 0.5 hour of drinking ad libitum at the end of a period of weight loss. A, regression line representing the group means (X) of 44 tests (10 individuals). B, line representing the 11 tests (Q) at -AW>2, in the hot desert (7 individuals). C, line representing the 10 tests (A) at -AW>2, in winter laboratory conditions (4 individuals). D, group of 12 tests (•) following physical exercise (1 individual). E, hypothesis that 8W/At=:-AW. New data. constant diet in 24-hour periods (Dennig, 1898, 1899; Newburgh and Johnston, '34). The diminution prevails also for urinary loss alone. It holds further for evaporative loss alone, in periods of measurement lasting 0.2 to 3.0 hours (fig. 60). During the first few hours of recovery by drinking, only variable and insignificant modifications of those urinary rates and evaporative rates are found. WATER RELATIONS OF MAN 99 It may be said that water ingestion in man is most rapid when drinking begins. Over a smooth course of intermittent drinking the rate diminishes for an hour or more. In contrast with the dog, drinking is very slow and prolonged (Adolph and Dill, '38), as 3 a O O I 1 1 J 1 1 > 02 " Man / 0.15 - / / "5 -3 ■ ■ o y - 0.1 o • • 0.05 a) u c N, • ( ) ® W, ® © 0 V Water Load Fig. 60. Eates of water output (% of Bo/hour) in relation to water load. Loads (% of Bo) are equivalent to decreases of body weight over periods of 1 to 4 days of usual food intake but inadequate water intake. All rates of output were measured over periods of 0.2 to 3.0 hours; total by loss of body weight and urine; insensible by basal loss of weight lying on a Sauter balance; urinary by collected output. Each letter designates one individual. New data on 4 subjects; and N 1 reported by Hall and McClure ('36). Urinary plus insensible do not here equal total because separate tests and conditions were concerned in each. though overshooting of content at balance were fearingly avoided. Water loss by the paths of urinary and of evaporative outputs diminishes significantly. Both the augmented drinking and the diminished loss contribute to the prolonged recoveries from states of water deficit. 100 PHYSIOLOGICAL REGULATIONS <^ 30. Equilibration and recovery Combining the data from single ingestions by mouth and from water privations, I obtain an equilibration diagram (fig. 61). The initial rates of recovery are here represented in the first 1.0 hour ; they might be represented over diverse periods of time. Recovery by drinking is most rapid within the first 0.1 hour, while recovery by forming urine is most rapid after 1.0 hour. The diagram indicates that water balance prevails at only one content. Small departures from the content at balance mean large modifications of exchanges, and ordinarily the body is either in small positive or negative increments at nearly every instant. c 0) o O Q: °4 1 N. Gam 1 1 1 1 1 Inge5"tive ~ \ Loss - ^Oxidative T.OSS Ny Evaporative __ /Gam ~» 1 1 1 =r- -3 -2 +2 ^3 Fig. 61. B„) -I 0 +1 Total Water Load Rate of water exchange (% of Bo/hour) in relation to water load (% of (equilibration diagram in first 1.0-hour of recovery). Data concerning ingestive gain are obtained from figure 59 in negative load, multiplying by 1.2 (fig. 58) to convert from 0.5 hr. to 1.0 hr. ; concerning urinary loss from figures 60 and 52. Total loss equals urinary loss, plus 0.07% of BoAour as the mean evaporative + fecal losses at all positive loads, they being independent of water load (HaU and McClure, '36). Gain of water by oxidation (0.02% of Bo/hour) appears to be nearly independent of water load whether positive (Carpenter and Fox, '30) or negative (Pinson and Wills, new data). When no factors shift the apparent balance itself, water content is maintained constant within ± 0.22% of Bq in 24-hour periods and ± 0.08% of Bo in 1-hour periods (CA, new data). The partition of the rates of exchange according to paths (fig. 61) again shows that restoration is predominantly by one path of gain (ingestive) and by one path of loss (urinary). Net velocity quotients are equal in positive and in negative loads (in the arbitrary initial period of 1.0 hour) only up to ± 0.2% WATER RELATIONS OF MAN 101 of Bq. At all larger loads recovery by gain is much faster than by loss. Although ingestive gains seem slow, they are at loads above ±2% of Bo faster than even the maximal rates of urinary losses (fig. 55). Corrections for alimentary absorption (fig. 62) probably in- fluence the absorbed water loads within 0.6 hour of any ingestion, but not later. Hence they are of no consequence in the equilibra- tion diagram at 1.0 hour, or in its corollaries or derivatives. I J +1.5 ^ S^ +1.0 o 3 -5 I +0.5 -P o 0 Fig. 62. Water passed through the alimentary tract in relation to time after water was drunk by men in water balance. A, mean of estimations, by 5 methods, of water absorbed from alimentary canal (Baldes and Smirk, '34). B and C, estimations of water disappeared from stomach, by methylene blue volume remaining (Christ, '26). D, estimations of water disappeared from stomach, by glucose volume remaining (Moritz, '01), points being the means of 12, 18, and 3 determinations on two subjects. E, fluid aspirated from duodenum, means of 5 measurements on 5 subjects (Baird et al., '24). Horizontal lines indicate the loads ingested. Mean body weights are assumed to have been 65 kg. The equilibration diagram indicates conditions for maintenance of water content in man and for recovery after disturbance of it. It provides standards by which the response of water exchanges by an individual, as a whole and by particular paths, may be evalu- ated (§33). Behaviors by which men choose environments that favor main- tenance of water content are well known, but no systematic studies of them are available. In a general way people do things that are 102 PHYSIOLOGICAL REGULATIONS appropriate ; sometimes by impulse and sometimes by experience. In desert journeys, they travel at night and in shade, minimizing the creation of water deficits through evaporative loss. They are said to become restless when in water shortage, just as infants become restless. Concerning water excesses where choices of envi- ronment were permitted, on the other hand, not even such crude facts are known. § 31. Comparison with dog The data interrelating the compensatory processes in man (fig. 61) are obtained under conditions that at present appear to be the same as those prevailing in dog (fig. 13). What similarities and differences occur among the results? These are first and pro- visional comparisons; ones made below (chapter IX) will concern several species more exhaustively. Heretofore results might have been generalized from one species; now distinctions may also be found. Initial states are studied under which increments of water con- tent are being dispelled after excess of water has been introduced by mouth or deficit created by privation. The differences are : (a) at any given excess of water (computed as % of Bo) the dog ex- cretes water (per unit of Bq) about twice as fast as the man. Maxi- mal rates of elimination by the dog far surpass those by man. If some other factor than body weight were used, the same conclusion would hold, for the factor cancels out by considering AW along with its correlative SW/At; in other words, the net velocity quo- tients (1/At) are greater in dog than in man. (b) In any deficit of water the dog ingested water much faster than man. The dog more than made up the whole deficit in 0.1 hour ; the man required some hours to do so. (c) In deficits the sparing of urinary loss is greater in dog than in man. This is related to the fact that dog's urine may carry much more solute, be more highly concentrated. Hence all modifications of water exchange are smaller in man than in dog. (d) All responses are slower, in initiation, in accelera- tion, in half -life, and in completion, in man than in dog. These differences are far greater than the differences among individuals (standard deviations) for one species or the other. Among uniformities between man and dog are found: (n) In the equilibration diagram the rates of loss are represented by a curve having three limbs. The middle steep limb shows rates of WATER RELATIONS OF MAN" 103 loss that increase with water excess ; at its right a limb shows maxi- mal rates changing little with load, (o) the rates of ingestion are approximately proportional to water deficits, and zero in water excesses, (p) The rates of loss and of gain are equal at one point of balance, (q) In excesses, loss exceeds gain; in deficits, gain exceeds loss; so that in both balance is regained, (r) Losses are slightly decreased in water shortage, both by urinary and by evap- orative paths, (s) Oxidative gain is approximately uninfluenced by water load. Both uniformities and diversities between two species of mam- mals having been found, it is worth while to inquire which uni- formities hold for others as well. The pattern of the equilibration diagram for dog and man may be found in many other species, but its quantitative features show wide diversities of magnitudes. <§. 32. Variations A more intimate notion of the exactitude of water balance of man is obtained by the study of fluctuations in content and in ex- changes. Near water balance intake is much more variable than output over periods of 0.25 hour (table 12), although one path of it (oxidative gain) is very constant under standard conditions. Often water content (body weight) slowly decreases from quarter- hour to quarter-hour until water is taken to restore it in part. No great differences from the dog in this respect (§22) have turned up. Stability of a physiological function is sometimes said to be a partial measure of, as well as a usual requisite for, the organism's success (in surviving). The criterion of success, however, may equally be otherwise and diversely chosen (leading to confusion of terms unless specially defined), since the small fluctuations of water content now discussed are not known to affect survival. Some kinds of ' ' success ' ' are dependent upon greater rather than less variability. In the social organism, too, great stability is not always beneficial to survival (Pareto, '35, § 2195). In addition to random variability, it is necessary to distinguish oscillations or rhythms of function, either with respect to time or to some other variable. In the case of water exchanges these periodic fluctuations have been measured in man under continued water drinking. In the average individual, urinary outputs (dur- ing uniform hourly intakes) are said to vary in four cycles per day 104 PHYSIOLOGICAL, EEGULATIONS (B, fig. 63). Some single individuals do and some do not show these cycles. One way of testing the significance of the data is to compare the variabilities in rates of output in consecutive hours with those in random hours, in second hours, or in third hours. By any test, the extremes of the cycles appear still to overlap to a probability (P) of 0.05. Since the apparent cycles coincided with habitual meal-times, although the subjects of the B tests did not eat, food relations were O.20- LJ o Houi^s of +)ie Day Fig. 63. Rates of urinary water output (% of BoAour) during steady rates of water intake in relation to clock hours. A, mean outputs in 35 tests on 35 men ingesting every hour 0.11% of B„ of water and 0.13% of Bq of food (actual plus potential water = 0.225%/liour). B, mean outputs in 51 tests on 41 men ingesting every hour 0.11% of Bo of water only. Mean body weights are assumed to have been 68 kg. Means (M) for the 24 or 27 hours and their standard deviations (a) are indicated. At six points on curve A the standard errors are indicated by rectangles. All are data of Gerritzen ('36). studied. The day's food was divided into 24 portions and a por- tion was taken with each hour's drinking water (A tests). The cycles were again possibly apparent, in nearly the same positions on the clock. Hence without further change of regime, which in turn would modify additional balances of the body, uniform rates of water exchanges at various hours cannot be expected. This conclusion means that control tests must be made at the same clock times as other tests in any study of water exchanges in man. WATER RELATIONS OF MAN" 105 Another situation in which water exchanges were measured is found when a man walks in hot and in cool environments (Dill et al., '33). In hot weather the exchanges of water (by evaporation and by drinking) are much larger. The progressive decrease of water content during them, though drinking is allowed, is also large as compared with the dog's in fig. 48. In man drinking is never sufficient to make up for the loss of weight that occurs in the same period of time. Hence no stationary state of water load is reached while the man walked in the desert, yet it is quickly reached in the dog. Variations, then, in the water loads and water exchanges of man are of two sorts, in rhythms and without apparent rhythms. Use of long periods of time in measuring exchanges, and of aver- age data concerning them, tend to hide both sorts. As with many other activities, what is usual at one time may be incongruous at another. The kinetic equilibration of water content is itself a mean in a set of relations, about which distinguishable modifica- tions oscillate. <^ 33. Characterizations and tests Given the usual situation that a whole equilibration diagram for all times elapsed cannot be worked out on each man to be studied, what particular tests will yield crucial information? And even when a diagram has been worked out, what points in it are likely to characterize the individual or the species ? In many physiologi- cal phenomena, empirical tests of this sort have been selected by convenience, usage, and experience, to differentiate experimental and pathological states. (1) Often, for instance, an individual is either too young or too paralyzed to drink water from a cup, but can swallow water when it is put into the mouth or can receive it by stomach tube. How shall I ascertain whether he is deficient in water content? One test is to administer 1 or 2% of Bo of water and see whether the rate of urinary output more than doubles within 1.5 hours ; or, how much of the water appears as urine within 4 hours. For most individuals the position of their water contents at 0 hour, relative to the content at presumptive balance 4 hours later, may thus be ascertained. A test of a similar sort has been in clinical use for many years, Volhard's "Wasserversuch." In its usual form a subject is given 106 PHYSIOLOGICAL REGULATIONS one liter of water to drink and the urine excreted is measured in hourly periods for 4 or 5 hours (fig. 51). Quite often no precau- tions are observed as to whether the subject had opportunity to be in water balance at the start. Hence if little of the water drunk is returned, the individual may be either (a) in initial water deficit, or (b) deficient in elimination. A deficiency of diuretic response to the water drunk is usually presumed to indicate (bl) deficiency of renal function. But it may indicate {h2) deficiencies in any other parts of the metabolizing and mobilizing tissues. Hence further tests are desirable to differentiate these states, such as the duration of the same diuresis ; but here I am not concerned with allocating the shortcomings between (hi) and (b5). State (b) can be distinguished from (a) by repeating the test 3 hours after the diuresis that results from the first administration of water. (2) The maximal rate of urine production after large inges- tions of water appears to be a useful characterization of individ- uals. This establishes either a single point or an equilibration diagram, or a tolerance curve that can be compared with the ranges occupied by other tolerance curves established after the same ini- tial water loads. For instance, in individuals having an admin- istered load of 4.4% of Bq, the fastest urinary output is found to be 0.53% of Bo/hour, while other individuals excrete 1.10% of Bo/hour (A, fig. 50). To eliminate that load, diuresis lasts about 12 hours instead of the usual 4 or 5 hours. (3) Further features of water recoveries may be compared by coefficients that have already been defined, such as economy quotient, ratio of modification, velocity quotient, and increment returned. Future investigation will undoubtedly lend meaning to several other quantitative characterizations within the four variables studied. Tests are lacking especially in water deficits. This state occurs unintentionally where there is some interference with volun- tary water intake ; in which case the most sensitive test, ingestion itself, cannot be used. Diminished urinary and evaporative losses are hardly significant without controls in the same individual. Additional indicators from outside the four variables then need to be employed. For use, each quantity to be compared requires measurement of variabilities. If the variability of one rate of exchange is as WATER RELATIONS OF MAN 107 large as C.Y. ± 80, that is still a standard. The number of states in which water contents and exchanges are altered also is large. In practice it may be easier to "force fluids" at random than to attempt the differentiation of those states. In theory the intimate connection of water with such a large number of states is of out- standing consequence. § 34. Diverse types of water load A few studies that furnish portions of equilibration diagrams under conditions other than those used above may be indicated. (1) Water in excess is ingested with particular quantities of four different foods (data of Petrilli) . The subsequent elimination of the water is less than without food, and lasts longer. Anyone can imagine where some of the water goes and where some of the delay may occur. (2) "Water deficits are created by physical exercise instead of by dietary privation. The deficit is produced in 1 or 2 hours in- stead of in 1 or 2 days, therefore. Subsequent initial ingestions (points D, fig. 59) happen to be in the range of those established for privation alone. (3) Water deficits follow moderate exertion in very hot atmos- pheres during 1 to 3 hours. Subsequent drinking is at mean rates somewhat higher (B, fig. 58) than in another state (C), and the rates differ significantly. The difference is chiefly at small deficits (fig. 59). It is possible that the excess of heat content added itself to the deficit of water content in arousing ingestion. (4) In warm environments the total water losses in balance are modified by the rapid evaporation of sweat. Even in negative water loads the rate of evaporative loss is high; also it is said that sweating is faster at high water contents of the body (Gregory and Lee, '36). But suitable quantitative data are lacking for the con- struction of an equilibration diagram. Concordantly, the paths of water exchange are very differently partitioned, so that urinary losses are relatively small except where diuresis has been aroused. Evaporative rates sometimes exceed even the most profuse urinary rates; individuals that have been observed under two conditions that regularly result in maximal outputs are able to put out water in sweat faster than in urine. The water deficit that usually prevails during daily life in the desert (Adolph and Dill, '38) is too small to measure. But it can 108 PHYSIOLOGICAL KEGULATIONS be estimated by indirect means, as follows. The mean rate of total water ingestion is about 0.26% of Bo/bonr, as compared with the rate in other climates of 0.13. On the equilibration diagram (fig. 61) this rate is found at a load of -0.2% of Bq, hence it is the approximate deficit. It is not outside the usual variation of con- tents found in 24-hour periods in other climates (table 15) nor out- side the standard of variation found in that climate. In the various states considered above, the content at balance (Wo) is believed to be approximately constant. Shifts in the posi- tion of Wo have been partially investigated in man, particularly in pathological conditions. Examples of such shifts (water reten- tion) are found in lobar pneumonia (Sunderman and Austin, '30), and in treatment with the drug phenylethyl hydantoin (Rockwell, '35). After apparently remaining at the new Wo for several days, the body reverts (at crisis) to the original Wo- <^ 35. Summary Investigations of water balance in man are slightly more lim- ited than in the dog, for in man responses in stationary state of water deficit are unknown. Quantitative correlations in the other states of water content are drawn. The same velocity quotients prevail, in the hours after the first, following a single ingestion of water, and after 8 to 16 rapidly repeated ingestions. Several means of testing and characterizing water relations are empha- sized : time relations, completeness of return, and maximal rates of exchange. Man uses all four of the possible means of recovering the usual water content, as does the dog. In deficits, gains are increased while losses are decreased slightly, thus sparing water ; in excesses, losses are increased and gains are decreased to the minimum of oxidative formation of water. These four modifications of activi- ties appear to constitute the armamentarium for any adjustment of water content. Rates of loss are slightly smaller per unit of body weight in man than in dog. In man the rates are less nearly proportional to water load, tending to prolong the life of larger loads. The most remarkable feature of recovery in man is the slow rate of ingestion after any deficit of water. Often it is asked, why is the drinking leisurely? No other peculiarity of metabolism or structure appears to be known that is especially related to this one. WATER RELATIONS OF MAN 109 Variations of exchanges in believed water balance may occur rhythmically as well as at apparent random. Diverse parameters are suggested for the characterization of unusual individuals or states. Differences between man and dog, in equilibration diagrams and their derivatives, may be said to be quantitative only. Quali- tatively all features are common to both species, giving rise to the expectation that other animals will exhibit similar patterns of water relations. Chapter VI WATER RELATIONS OF FROG § 36. If the study of water in organisms is to be general, the same variables that are measured in mammals require to be corre- lated in numerous other kinds of living beings. A frog has many functional differences from a dog, and I now inquire whether and in how far the pattern of water exchanges is different. Is the frog, an animal usually immersed in water, less concerned with constancy of water content? The species Rana pipiens is the one studied except where otherwise stated. A frog does not drink water by mouth, but takes water into the body through the skin. In older investigations {e.g., of Durig, '01) this fact was ascertained by blocking the gullet in various ways. Equally decisive is the fact that frogs in water gain weight con- tinuously and at nearly steady rates in successive short periods of time. When watched, no frog is seen to open the mouth or to swal- low water during the gain of weight. The absence of muscular movements in water intake seems to some persons to confer greater automaticity upon the process in frog than in dog. A frog puts water out from body through the cloaca, all of it presumably having passed through the kidneys. Only in types of water load other than those studied here does water leave at signifi- cant rates through the skin or any other path than the kidneys, so far as is known. § 37. Water exchanges Recoveries of water content are measured in frogs immersed in water at 20° C. with nostrils in air. Excesses of water result from sudden injection of distilled water into the peritoneal cavity. Defi- cits are previously produced by 1 to 24 hours of evaporative desic- cation, the frog being temporarily out of water in air (Adolph, '39b) . Increments of body weight represent the loads of water. Figures 64 and 65 indicate the conduct of tests. When the frog 's cloaca is closed by ligation the weight increases, corresponding to the rate of entrance of water through the skin. Whenever the cloaca is later opened, urine escapes and the amount of it that has formed since ligation is thus ascertained. When the cloaca is not ligated, changes of body weight indicate net exchanges of water. 110 WATER RELATIONS OP PROG 111 Hours Fig. 64. Course of water load (% of Bq) in each of two frogs that were injected intraperitoneally (at the arrows) with distilled water. The cloacas were ligated at the time of injections and later opened (at the crosses) ; the first being ligated twice. The differences of body weight before and after the crosses measure the urine excreted since ligating. Redrawn from Adolph ('39b). ■*^o I -10 Hours Fig. 65. Course of water load ( % of Bq) after recovery commenced. Eana pipiens. Points and solid lines, total water load ; light dash Unes, gross load, which includes urine kept in the body by ligature on the cloaca. The difference between solid and dash line represents total output since zero time. Each line is the mean of 10 individuals grouped in order of the initial loads imposed. Further data of Adolph ('39b). 112 PHYSIOLOGICAL KEGULATIONS -30 -20 +10 +20 -10 0 Total Water Load Fig. 66. Bates of water exchange (% of BoAour) in relation to initial water load (% of Bo) (Equilibration diagram). Frog. Eates are all computed from the first 0.5-hour of recovery, each point representing the mean of 10 measurements on as many individuals. The square represents 10 individuals that did not survive the recovery. Eedrawn from Adolph (39b). In order to yield average data, tests are grouped in tens in the order of the increments of water initially contained in the bodies. The subsequent exchange is then (fig. 65) such that in every case the net weight of the frog tends toward the weight that prevailed before water was subtracted or added. But the gross weight (urine ■30 +10 +ao -20 -10 0 Total Water Load Fig. 67. Net exchange of water (% of Bo) in relation to mean water load. The exchange is that which occurred withiu the time indicated in hours of recovery. Each point represents 10 frogs, the same as in figure 65. Further data of Adolph ( '39b). WATER RELATIONS OP FROG 113 being retained in the cloaca) always increases, at diverse rates that climb with the deficit. When these initial rates are studied indi- vidually (Adolph, '39b), the standard deviations of rates among members of each group of ten are quite uniformly about ± 27 per cent of their arithmetical means. Urine formation is measurably present only in positive incre- ments of water content, and in negative increments down to - 10% 2 3 4 5 6 Waier Output Fig, 68. Water intake (% of BoAour) in relation to water output (% of Bo/hour) under varying water loads, within the initial 0.5 hour and the initial 1.0 hour of recovery. Each point represents the mean for 10 individuals, as in the same data of figure 66. The dash line represents the state in which intake and output are equal; this occurs only at water balance. of Bo (fig. 66). The initial rates are approximately proportional to the water contents of the body above - 10%. Measurements of urine production made over half -hour periods are less accurate than those made over longer periods ; nevertheless at each load the aver- age rate of urine production during the first 0.5 hour of recovery is found to equal the rate during 1.0 and 2.0 hours. Hence the intact frog shows no measurable lag in onset of diuresis following intraperitoneal injection of water. 114 PHYSIOLOGICAL REGULATIONS Initial rates of water intake and of water output together (fig. 66) constitute an equilibration diagram. The intake never falls to nothing ; even in the largest water excesses studied, its minimal rate persists. Urine production falls to zero in moderate deficits of water. Its rate is augmented less with excesses of water than intake is augmented with deficits. Since gain and load are little affected by the length of time elapsed as such, the diagram is not greatly modified at longer times than 0.5 hour (fig. 67). The net exchanges of water (fig. 67) represent the combined accomplishment of the total exchanges as modified at diverse loads. Since body weights are actually being measured it is directly as- sured that observed gains minus observed losses equal net gains. In 2 hours any deficit is paid off, but only half of any excess is cov- ered. >. OA Water Load Fig. 69. Net velocity quotient in relation to mean water load. Each point is derived from one group of 10 individuals in the first 1.0 hour of recovery. Load is % of Bo; velocity quotient is obtained from figure 67 by dividing the net rate of water exchange by the mean water load, = 1 Aour. For every rate of water output there is one rate of water intake under the conditions prevailing (fig. 68). The roughly hyperbolic form of the curve relating the two, emphasizes the reciprocal rela- tion of the two processes, as though a high rate in one excluded rapid activity in the other. Velocity quotients are larger in water deficits than in water excesses (fig. 69). Recovery requires only one-third to one-half the time after moderate desiccation that it requires after water injection. It is most prompt at small loads. Among diverse ex- cesses the net quotients are not much affected by magnitude of water WATER EELATIONS OF FROG 115 load, whereas the quotient for total intake (not shown) is propor- tional to reciprocal of load. Initial velocity quotients at small net rates of water exchange exceed those at large net rates (fig. 69). As, later in the same tests, the rates become smaller, the quotients remain significantly differ- ent (fig. 70). Hence the individual courses of recovery do not exactly recapitulate the array of initial recoveries at identical diverse loads. A few of the data shown for frogs are confirmed by independent measurements ; after desiccation by data of Durig ( '01) and Adolph ('32), and after water injection by sparse observations of Adolph ('27). 0.6 ^ 0.6 •i ^ O.d +> :> 02 0.5 1.0 1.5 a.Q 2.5 Hours Fig. 70. Velocity quotient (1/hour) in relation to time. Each point is referred to the middle of the period during which it prevailed. Each curve represents a group of 10 individuals starting with the water load indicated. From figure 65. A biologist unfamiliar with amphibia might expect a frog to gulp water by mouth after being highly desiccated, especially if water is made available to the tongue but not to the skin. Water is not ingested or swallowed. Dogs and men use the quick oral means of obtaining water, but frogs are constituted to sit patiently in water long enough for water to penetrate the skin. No cry of inaptitude in the frog's behavior would help investigators to ascer- tain whether or not the frog has maximal fitness in its water ex- changes ; swallowing of water on the part of an animal that has no way of metering it might be more damaging than appropriate. The accuracy of recovery of water content is indicated by the 116 PHYSIOLOGICAL KEGULATIONS net body weights reattained. Thirty-three frogs that had been desiccated to diverse extents have CA ± 2.87% of Bo after 6 hours of recovery, and ± 2.11% of Bo after 24 hours of recovery. Almost the same difference (CA ± 2.22) is found among frogs kept in water balance for 24 hours, and again among frogs recovering from ex- cesses of water ; indicating that the over-all fluctuations of weight are not accentuated by intervening desiccation or hydration. More- over, the differences are in each of the three groups (deficits, ex- cesses, controls) positive and negative in equal numbers. The exchanges labelled "intake through the skin" may conceiv- ably be net intakes and not total intakes, since it is not known whether water goes out through the skin concurrently as more comes in through it. This possibility makes no difference in the present study, any more than output through the kidneys depends on whether there is ''reabsorption" as well as "filtration," and whether there are any other processes intermediary to excretion. No significant partition of water exchanges among several paths appears in frogs. Loss by evaporation is exceedingly small, for the frog is nearly completely immersed and breathes air saturated with moisture. Gain by metabolic formation of water amounts to less than 0.01% of Bo/hour at 20° C. These quantities cannot be represented at all on the coarse ordinates of figure 66. In brief, loads of water imposed on frogs by intraperitoneal in- jection or by evaporative desiccation, lead to net exchanges appro- priate to restoration of control contents. Many recoveries are complete in 6 hours, all are complete in 24 hours. At numerically equal loads, recovery by net intake is faster. In excesses, intake continues at about the usual rate characteristic of balance while output is appropriately modified; in deficits, output helps intake to compensate, output significantly diminishing while intake is in- creasing. Thus the frog uses three of the four possible modifica- tions by which its exchanges may lead to rapid recoveries of water content. § 38. Variations Above I mentioned the variations that occur in the control state. Single individuals are weighed at hourly (or other) intervals to measure how much the contents and exchanges of water vary, and body weights are found to fluctuate by CA ± 0.77% of its mean/hour {103 tests, 19 individuals). Errors of measurement are insignifi- WATER RELATIONS OF FROG 117 cant, for weighings repeated at 2-minute intervals differ by only CA ± 0.07. Water intake in successive periods of 1 hour varies by CA ± 32.3 per cent of the mean intake (45 observations). Water output in 1-hour intervals is believed to vary by about the same coefficient ; but exact data have not yet been obtained in successive periods. While successive hourly periods show differences of intake aver- aging 32.3 (CA), these same rates deviate from the mean rates of each individual by 31.1 (C.V.). In a random series CA equals C.V., as here ; hence a high rate does not tend to follow a low one, nor vice versa. If there were trends lasting more than one hour, C.V. would tend to exceed CA. In the initial stages of recovery, the water intakes being aug- mented as much as eight fold, the variation among the rates (C.V. ± 27) is almost equal to that in turnover (± 31.1). This suggests that relative variability overrides whatever absolute variability may be present. Or, the reproducibility of the rate is proportional to the amount of water being handled as intake. Variability of content in water balance may be considered in relation to mean rate of turnover. In one hour 1.66% of Bq is on the average taken in and put out again. The standard differ- ence of successive hourly weights (±0.77%) is 46 per cent of this ; and the mean difference is 37 per cent, mean difference being 0.798 X standard difference. In 24 hours the turnover is 40% of Bo, while the coefficient of difference (± 2.22%) in 24-hour inter- vals is only 6 per cent of the turnover. "Precision" as computed by the method of Gasnier and Mayer ( '39) is therefore 95 per cent in 24 hours but only 63 per cent in 1 hour. The variation of intake (32.3 CA) might correspond to an inac- curacy of ± 0.54% of Bo each hour. This alone almost covers the variation in body weight (± 0.77% of Bo). Similarly the variation among individuals of output alone (± 23.5 C.V.) is a factor in the content. The variation of content may be viewed as an interplay of both these variations of exchange ; added together, the most usual fluctuation of intake would be 1.66 X V 32.3^ + 23.5 VIOO or ± 0.65% of Bq. The value 0.65 being less than the value 0.77 there is no significant evidence of mutual compensation, in one-hour periods, between intake and output. Most compensations occur over longer periods of time, but within less than 24 hours. 118 PHYSIOLOGICAL REGULATIONS Hence, maintenance of water content may be regarded as the resultant of fluctuations in two paths of exchange, imbibitory gain and urinary loss. Within periods of one hour no adjustment of rate of gain to rate of loss is evident, but in successive hours rates of gain compensate one another. In 24 hours, variation of body weight is only thrice as great as in one hour, indicating that adjust- ments of rates of one exchange to those of the other have intervened to limit the variation found. § 39. Some other types of water load Frogs have been subjected to water loads in several diverse ways. What is the course of recovery in those other physiological states and environmental conditions? (1) Frogs that are pronounced dead, if the criterion of sur- vival be a subsequent muscular activity, gain water at the same rates as living ones. This is the case both near water balance (Brugsch et al., '28) and at extreme water deficits (fig. 66). Yet the rates at those two water contents differ by a factor of 8, as though the machinery of intake is modified in spite of the concur- rent failure of some other activities. Sudden stoppage of the cir- culation without other immediate injury (Adolph, '31b) likewise indicates that movement of the blood is not necessary in ordinary water exchanges. (2) Frogs that are kept out of water while provided with vari- ous water contents may also be compared (Adolph, '39b). From deficit no recovery occurs even in air saturated with moisture; instead a very slow loss by evaporation continues (which is inde- pendent of water load), the frog being slightly warmer than the air (Adolph, '32). Urinary loss is almost zero after the frog has been out of water for two hours. Since gain by oxidation is smaller than the slight continued loss of water, even in deficits water bal- ance is never precisely maintained. From water excess recovery is by diuresis only (fig. 71). The excess, administered by sudden intraperitoneal injection of distilled water, is to the inside of the frog seemingly the same as in the stand- ard (immersed) conditions ; yet the diuresis is much smaller. Urine forms at less than half the rates that prevail in frogs immersed in water that have similar loads (fig. QQ) ; but due to the fact that water also enters the immersed frogs, the net recovery by the frogs in air is slightly more than half as fast. Hence with the stoppage WATER RELATIONS OF FROG 119 of intake, excretion of urine is prevented by more than the intake missed; or paradoxically, water entrance appears to promote net water elimination. Intermediate conditions between immersion in water and free- dom from surrounding water were tested. Very shallow water or wet towels allow nearly as much water intake as deep water, but wet filter paper allows little or no water intake. Gradation of in- take by this means is difficult to secure ; the phenomenon is almost "all-or-none." It might seem justified to state that the "wetness of the skin" makes the difference between the recovery of the frog in water and the frog out of water. But this would credit to some one fac- tor a situation that has many possible factors. The physical dif- 0 +10 ^20 ^30 To-tal Water Load Fig. 71. Eate of water exchange (% of Bo/hour) in relation to initial water load (% of Bo), in frogs kept out of water in moist air. Intake is zero; output as urine is indicated by points each of which is the mean of 10 tests. Kedrawn from Adolph ('39b). ferences of hydrostatic pressure, temperature, texture of surface, and many others, are all present. The physiological differences may be myriad ; the differences are more than skin-deep, since urine is not formed in the skin. So the two types of load are described with fewest implications as belonging (b) to a frog sitting on a dry wire screen in air saturated with moisture instead of (a) to a frog sitting in a glass beaker with water 2 cm. deep. That a frog depends for constancy of water content upon being in a wet environment is almost obvious. How much time does a frog ordinarily spend out of water? Does it locate water more quickly when already desiccated? Is the movement toward water random or oriented? I have tested many frogs in water deficit to see how in the labo- ratory they find their way to water. In the conditions tested, they 120 PHYSIOLOGICAL EEGULATIONS found water only by random movements ; once they touched it, the depleted frogs stayed in it. Frogs that did not stumble into it died of water-lack within a few centimeters of a life-saving pool of water. I do not doubt that in outdoor life, frogs have cues that guide them to water; those cues differ from the many that were tested indoors. Though answers to none of the questions are re- corded for the frog, certain observations have been made upon another amphibian, Triton (Czeloth, '30). From a distance of 70 cm. the salamander moves directly toward a body of water. It stays in contact with wet soil when possible, and frequents air of high humidity without any liquid water being present. Neither ocular nor nasal senses are required to detect either liquid or humid- ity. All these actions are ones that result in preservation of the animal 's water content. (3) Water might be expected to move outward through the skin when an excess of water has been injected into the body. It does not ; there is not even a decrease in the rate of movement inward in the standard state (a). Of all the methods of influencing water exchanges of frog that have been tested, outward passage of water occurs only under the influence of solutes: (c) when a frog is put into ''hypertonic" solutions (Adolph, '31b), and (d) when a frog recovers (by being put into water) after a sojourn in a ''hypo- tonic" solution of sodium chloride (new data). In both these cases some exchanges of solutes are occurring. In a few words it may be said that dying frogs augment their water intakes as much as living ones of the same water content; frogs out of water lose excesses less rapidly than those immersed in water; and water does not move outward through the skin, ex- cept as it either evaporates or else enters a modified medium that may allow exchanges of solutes. Undoubtedly many other physio- logical states would likewise show water exchanges that differ from those found in the arbitrarily chosen standard type of water load. "§> 40. Summary The equilibration diagram for water in the frog (fig. 66) is characterized by a curve of losses that is proportional to water content between - 10 and +30% of Bo, and a curve of gains that has three limbs, being nearly horizontal at great deficits and in all excesses, and steep at loads between 0 and - 20% of Bq. Net gains WATER EELATIONS OF FROG 121 are more rapid than net losses. No data are available in steady states of water excess or deficit. Eandom variations preserve water content with a CA of ±: 2.22% of Bo in 24-hour periods. This indicates the accuracy with which intake equals output, for 40% of Bq of water is turned over in each of these periods. Desiccated dead frogs exchange water initially at rates equal to living ones. Frogs kept out of water recover from excesses more slowly than those in water. Water is not lost (excreted) through the skin, when evaporation is prevented, in response to water excesses. With an entirely different path of water intake (the skin) from that present in mammals, a qualitatively similar diagram is present except that intake through the skin persists when water contents are already excessive. Quantitatively, recoveries occur at rates roughly similar to those in man, and much slower than in dog. Recoveries from deficits exceed in rate the recoveries from ex- cesses, even as they did in man and dog. Quantitative comparisons with other animals will be shown later (chapter IX). Living in an aqueous medium does not mean that fewer or slower compen- sations are provided for the recovery of usual water contents, nor that the organism is careless of what the content is. Chapter VII WATER RELATIONS OF OTHER SPECIES § 41. The present account of regulations of water content might be limited to man, or to mammals, or to vertebrates. Such limita- tion would leave out myriads of other classes of animals ; I judge that more insight is to be gained by extending considerations to many kinds. Are regulatory processes present wherever they are sought? Do animals forestall the need for compensations of dis- turbed water content, by frequenting appropriate environments? Do some species depend on adjustments of output to do the same thing that other species accomplish by adjustments of intake ? Are special structures concerned in compensatory exchanges wherever such adjustments are found? Two kinds of interest (at least) might attach to data concerning water regulations in varied parts of the animal kingdom. General physiology might have its dream come true of knowing exactly how general each of the processes and correlations concerned in water exchanges actually is. Comparative physiology might fairly rate various phyla and species according to their means of disposing of water loads and the kinetics of their exchanges. The perfect dream might require data upon a thousand species. I believe that a fair outline of the features of all water regulations is obtainable from the fifteen or twenty species for which appropriate though partial data exist. My plan is to present briefly the pertinent information concern- ing water exchanges in each of those several species. Special methods are required for the study of each, and particular features are to be noted. Thereafter (chapter IX) , quantitative similarities and diversities will be ascertained. Comparisons are not limited to animals possessing any one common structure except ''proto- plasm"; wherever exchanges of water occur they either do or do not serve regulations, and in diverse degrees and patterns. From knowledge of water relations in dog, man, and frog, the following physiological arrangements might, as a first extrapola- tion, be expected in other animals : (a) a water content varying less than ± 3% of Bo at intervals of 24 hours, (b) a turnover, (c) a compensatory recovery by augmented rates of loss when excesses 122 WATER RELATIONS OF OTHER SPECIES 123 of water are present, (d) a similar recovery by faster gain when deficits of water are present. The following pages show that these possibilities are realized in a variety of animals, with the exception that in some no turnover has been demonstrated. Also, a greater variety of rates of exchange is uncovered than would thus far be predicted. § 42. Rabbit Water excesses introduced by stomach (fig. 72) give rise to diuresis in the rabbit (Lepus). If the data only of series C existed, it might be concluded that recovery from excess water is both slow and incomplete. In that series the diverse individuals and tests 3 3 o Sensible Water Load Fig. 72. Rates of urinary output (% of Bo/hour) in relation to sensible water load (ingesta minus urine) (% of Bq) in rabbit. A, mean in intervals of 0.5 hour, data of Heller and Smirk ( '32a, p. 18) ; B, mean in hourly intervals, 7 tests and individuals of Abe ( '31b, p. 414) ; C, mean in hourly intervals, 32 tests on 11 individuals of God- lowski ('30, pp. 87 to 92). Evidently conditions for prompt and unimpeded diuresis do not prevail in series C. Standard errors are shoAvn by brackets at four points ; J. = half of load returned as urine. are more consistent among themselves than in any other series, yet in 4 hours less than half of the excess water is returned as urine. Comparison with the rabbits of series A and B suggests that those of series C were in water deficit before the new water was admin- istered. In rabbit the acceleration of diuresis is about equal to that in dog and man, for maximal rates of urinary output prevail in the second hour after water is placed in the stomach. When the excesses are very large (AW + 30), rates of excre- tion up to 5.4% of Bo/hour are found (data of Misawa, '27), but 124 PHYSIOLOGICAL EEGULATIONS they are little greater than at AW + 5 (data of Frey, '07). Ex- treme water excesses are established niore readily in this species than in dog because vomiting does not occur. In deficits, attained after privation of water but not of food, the rabbit recovers its body weight in a few minutes of drinking (fig. 73), taking, on the average, amounts slightly greater than the initial loads (fig. 74). That would not have been surmised from earlier studies of Pack ('23); for in his tests the rabbits, deprived for several days of food as well as of water, ingested in 0.5 hour even less water, relative to body weight and to water load, than did men (§29). 0 0.4 0.8 1.0 Hour Fig. 73. Course of water load measured by water consumed (% of Bo) during recovery from water deficit in rabbit. In each of 10 tests one of two individuals has been deprived of water but allowed food for 24 hours before zero time, when water is freely offered. Bars indicate standard errors of total amounts of water ingested. Dash line indicates mean water consumed in 12 control tests at same time of day in which water but not food is allowed ad libitum without previous privation. New data. Turnover of water has been ascertained as 13% of Bq in 24-hour periods (Gompel et at., '36; Gasnier and Mayer, '37). Most of the water is contained in the (green vegetable) food ordinarily eaten. On dry food (oats) plus water ad libitum, turnover is consider- ably less (7.0% of Bo/24 hours; new data). Variability of water intake in 24-hour periods is in a selected instance (HI of Gompel et at., '36) smaller than has been reported WATER EELATIONS OF OTHER SPECIES 125 for any other species {± 5.2 CA). The consistency among rates of intake is partially related to the regime of green food, but dif- fers greatly among ten individuals on the same diet (Gompel et ah). The concurrent urinary output is somewhat more variable (± 11.8 CA daily in HI), as though drinking is of greater concern. In water content the daily variability of HI (15 days) appears to be CA ± 1.53; of individual Z4 (80 days), CA ± 1.49. The varia- tions are greater than in dog and man by about the same ratio as the turnover rates. 1 ■■ O I- 1 ■ ■ - T- - ■ -I ■ ■ I ■ 1 6 \ ° o Rabbit \ o \ -t- c 50 . \ o o ^ o t. • o \ u • \ o o \ ^ s- 40 " \ o \ ° \ s 30 - - \ 20 ■■ \ \ \ \ 10 \ X \ \ \ 1 1 1 1 1 1 1 N -80 -70 -60 -30 -20 -10 Fig. 82. (% of B„). -50 -40 Eate of initial gain of water (% of Bo/24 hours) in relation to water load Terrestrial snail, Limax, four species. For each point an individual weigh- ing 1 to 15 gm. is desiccated for 3 to 16 days; then placed for 24 hours where water is available. The line represents the theory that intake in 24 hours equals deficit. Data of Kiinkel ('16, p. 88 ff.). "§> 46. Helix Many species of the shelled pulmonate snail Helix were sub- jected to desiccation. These desiccations with privation of water often lasted for months, involving therefore deficits of other sub- stances as well as of water. ''If water and food are available to the snails after desiccation, they begin to eat only after they have drunk" (Kiinkel, '16, p. 86). In three species (fig. 83) the returns approximately equal the deficits. "The quantities of water that can be drunk by the snails depend on the water content that the animals have before drinking" (p. 51). Additional observations show that water is taken through the body (mantle) wall as well as through the mouth. Kiinkel believes that both paths are ordinarily used. By both paths water appar- ently ceases to enter when the control volume of the body has been approximately reattained. 136 PHYSIOLOGICAL EEGULATIONS The only known difference between Limax and Helix with re- spect to water exchange is related to the fact that the latter can prevent loss of water by evaporation, deficits arising but slowly during hibernation or non-locomotion. During hibernation the body fluid gradually increases in concentration of dry residue, re- covering its initial concentration suddenly upon emergence and shedding of the epiphragm (Holtz and Brand, '40). Though Helix loses water very slowly during inactivity, in ordinary existence its weight fluctuates by (CA) ± 8% of Bo in daily periods (Howes and Wells, '34). Water loss is therefore not 60 1 4 1 ■ ■ — 1 1 50 \ \ \ \ • u 40 \ • JC • \ o a -A \ ^ o \ JL \ V 30 \ ^ -t- e \ :5 •4- \ O \ , -60 -50 -40 -30 -20 -|0 Water Load Fig. 83. Eate of initial gain of water (% of Bo/24 hours) in relation to water load (% of Bo). Terrestrial shelled snail, Eelix, three species, 9 individuals. From each body weight (0.4 to 19 grams) the weight of the shell has been subtracted. Privation of water lasted 6 to 12 months; then water was available for 24 hours. Data of Kiinkel ('16, p. 120 ff.). slow during usual activity. Probably the fluctuations depend upon water intake, for when food but not water is supplied, Httle ir- regularity of weight is found. Helix, then, usually does not drink water so frequently as once a day, as though it is insensitive to deficits of AW - 8. That picture of inconsistency in Helix is not borne out by the extensive study of analyzable water contents made by Brand ('31), for it would have been expected from the above facts that individuals analyzed at one time would vary by (C.V.) ± 8%. If Helix be an animal that regularly has a wide range of water contents ; it is not representative of invertebrates generally. WATER RELATIOlSrS OF OTHER SPECIES 137 § 47. Insects The quantitative exchanges of water can scarcely be represented for a single species of insects, for appropriate data are lacking. The qualitative nature of the gains and losses of water, and the tolerance of water deficits, have been outlined for various insect orders by Buxton ( '32b). In turnover most gains are said to occur by specific drinking of water, though some kinds do not habitually touch free water. Most losses are believed to be evaporative, ex- cept those in water excesses. The impression might be obtained that the water contents are much less uniform in insect species having non-aqueous environ- ments than in frogs or earthworms. So far it has no quantitative ■•■100- +50 12 3 4 Hours Fig. 84. Course of water load ( % of Bo) after ingestion of rabbit blood. Bhodnius prolixus, 1 individual of 78 mg. at 18° C. The insect took 225% of Bo of blood, or 169% of Bo of water. The load plotted is water ingested minus urine expelled (sen- sible load) ; an additional 10% was lost insensibly in 4 hours. Data of Wigglesworth ('31a, p. 414). support. Variations among analysed individuals are not greater than in some vertebrate species and many invertebrate ones (see table 19). Much information indicates that insects do something about a water increment, whether it be a deficit or an excess. For the for- mer, there is no controlled evidence, only surmise, that drinking by mouth is increased by water deficits (Buxton, '32b, p. 277; Mel- lanby, '38, p. 399). For the latter, data concerning excretory re- covery of water content are shown in Rhodnius, a blood-sucking bug (fig. 84) after one sudden excess of water (along with other substances) has been taken in the form of rabbit blood. The water is presumably eliminated through Malpighian tubules. 138 PHYSIOLOGICAL REGULATIONS While many kinds of insects are highly protected at their sur- faces from rapid evaporative losses, others are known to be more vulnerable to such losses. This has given rise to the question, are insects able to choose environments that tend to minimize evapora- tion? There is no doubt that some species frequent atmospheres of a particular humidity ; indeed, structures that detect differences of humidity have been described in spiders (Blumenthal, '35) and in mealworm beetles (Pielou, '40). The clearest case of preference for high humidity is in the isopod crustacean Porcellio (Gunn, '37) ; it is indifferent to atmospheres of 65 or more per cent relative humidity, but in the region of half saturation distinguishes differ- ences as little as 6 per cent. In general the animal keeps moving rapidly in dry air, becoming quiescent only when it arrives in a moist region ; in addition it orients away from dry air if it suddenly comes into a boundary region. Some other species shun wet air (locust, Kennedy, '37; mealworm beetle, Pielou and Gunn, '40). In the cockroach (Blatta) the preference for moist air is mani- fest only after previous desiccation (Gunn and Cosway, '38). The correlation between behavior and water deficit then constitutes regulation in the sense of all the studies here considered. For, in the humid air the rate of water loss is believed to be diminished; hence the modification of behavior tends to preserve the body's water content. To some extent the ability of insects to move toward bodies of water has been investigated, and positive attractions have been demonstrated in several species (Turner, '24; Krijgsman, '30; Hertz, '35). This behavior which in some is dependent upon sight, is an integral part of success in gaining water to compensate for deficits. Amid the diverse studies of insects in relation to water, data are insufficient to furnish an equilibration diagram, but enough to sug- gest that each of the constituent compensations exists. Neither turnover, variability of water content in an individual, nor modi- fication ratio can be stated. Certainly behavior toward humidity of environment contributes to maintenance of content in some species ; the complete story requires the correlation within one species, of this mode of adjustment with the others. § 48. Phascolosoma Does an animal that lives in the ocean have an easier time main- taining its water content than others? The marine gephyrean WATER EELATIONS OF OTHER SPECIES 139 worm Phascolosoma becomes desiccated by sojourn in sea water more concentrated than that in which it lives, or hydrated by so- journ in diluted sea water. Upon return to normal sea water the body weight in each case reapproaches the control weight (fig. 85). The rates of water exchange gradually diminish (fig. 86), but it may be noted that in positive loads the rates after the first hour are more nearly uniform with time. At equal loads the exchanges of water, all of which occur through the body wall, are much more rapid as intake than as out- Gephyrean Worm Phascolosoma Hours Fig. 85. Course of net water load (% of B„) after return of worms to normal sea water from sea waters of other concentrations. Phascolosoma at 23° C. Each curve represents the average of 2 to 4 tests (5 individuals per test) the points being means of interpolated weights. Additional data of Adolph ('36b). put. In relation to water load, each gain appears to be continu- ously proportional to the gradient of osmotic pressure between outside (Pe) and inside (Pi) the body (Adolph, '36b). The gain follows the equation : SW/At = h V"1B (Pj - Pe) , in which the cylin- drical worm has weight B and length 1. The exchanges are at rates roughly proportional to B^''^ or ■\/lB (body surface area), in the range of computed areas from 4 to 20 cm.^ The coefficient h has the dimensions of permeability. 140 PHYSIOLOGICAL KEGULATIONS Evidence of various sorts indicates that dissolved constituents (ions) of body fluids and of sea water are in these tests not exchang- ing between body and environment. Hence desiccation in concen- trated sea water is equivalent to desiccation by evaporation; the recovery in sea water following actual evaporation was tested and occurred at the same rates as above. Excesses that were obtained by previously injecting distilled water also showed the same re- covery rates as those obtained by previous swelling in diluted sea water. Gephyrean Worm Phascolosoma 2 3 Hours Fig. 86. Rate of net water exchange (% of Bo/hour) in relation to time after return of worms to normal sea water. Phascolosoma at 23° C. Same data as in figure 85. While recovery by intake is faster than recovery by output, yet it is commonly believed that the same forces of water transfer through the same integument are concerned. Actually the kinetics of the two exchanges also differ, as well as the values of h in the above equation ; for during water gain h is constant, during loss Ji is not constant. Since I am describing the rates of water exchange, I preserve ± AW as abscissae (fig. 87). For other purposes, in view of the supposed relationship of BW/At to osmotic pressure in Phascolo- soma, l/(Pi -Pe) might be used as abscissae of figure 108 instead of ± AW; the internal osmotic pressure (P.), being in turn, com- puted from l/CB-b) in which B is the body volume and b an WATEK RELATIONS OF OTHER SPECIES 141 empirically ascertained parameter termed "non-solvent volume." By this device the lines A, A', B, and B' actually become straight ; strong suppositions may thereupon be formulated for identifying the rates of water exchanges with forces ordinarily termed osmotic pressures. Often, the "permeability coefficient" h is regarded as having some inseparable and indispensable significance in such study of water exchanges, but I find additional comparisons among diverse conditions and species to be also useful. LU J cn Water Load Fig. 87. Initial rate of net water exchange (% of BoAour) in relation to initial water load (Phascoloso^na) . For each point, 5 individuals (0.8 to 2.8 grams each) are transferred at zero time from varying concentrations of sea water to normal sea water. AA', exchange in first 0.25 hour x 4; BB', exchange in first 1.0 hour. The curves are drawn so that the ordinates (within each of the two segments) are proportional to the initial difference of total concentrations between worm and new medium. Additional data of Adolph ('36b). Phascolosoma adjusts its water content without, so far as as- certained, the use of processes that are known to require internal energy. The equilibration diagram is nevertheless similar to others that have been studied; it has only two lines instead of the four usually recorded, since it represents only net exchanges. No turn- over has been recognized, unless it be by kinetic interchange of water molecules. Corresponding to the absence of turnover is the small variability of water content (see table 12). The conclusion may not be drawn that all marine animals adjust their water con- 142 PHYSIOLOGICAL EEGULATIONS tents so as to equalize osmotic pressures between body and environ- ment, for sooner or later animals like the teleosts and decapods are studied that pump water, instead of (or as well as) pumping solute, out of or into their bodies. § 49. LUMBEICUS Earthworms live in an environment of earth mixed with fresh water ; in the laboratory they are kept in water without earth. The turnover of water is ascertained by the following steps : the worms are handled until constant (basal) weight is attained, then left un- touched for 1 to 5 hours ; reweighed, rehandled, and finally weighed. Hours Fig. 88. Course of water load (% of Bq) in earthworms immersed in fresh water at 18° C. Each series consisted of 5 individuals just injected with water, or rendered deficient by evaporation. Before each weighing, nephridial anl alimentary reservoirs were emptied by rolling the worms repeatedly. New data of Wolf and Adolph, being similar (in water deficits) to previous data of Wolf ('40a, his fig. 7). The loss of weight when rehandled represents a minimal value for the water excreted in the period of time elapsed, the evidence being that all of it is held in nephridial reservoirs and alimentary tract while the worm is undisturbed. Nine-tenths of the fluid lost is believed to issue through nephridia (Wolf, '40a) ; the total turnover is 2.7% of Bo/hour. Variability of water content is about ± 2% of Bo, as ascertained by weighing individuals at 24-hour intervals with reservoirs empty. WATER EELATIOjSTS OF OTHER SPECIES 143 This variability is no greater than the frog's (see table 12), though the turnover is almost twice as great. Negative increments of water are set up by evaporation from worms placed in air ; positive increments by intraperitoneal injec- tion of water. After each load has been established, the worms are allowed to recover in fresh water (fig. 88). The control worms of this series happened gradually to lose net weight ; in positive loads the net loss is linear with time but slightly faster than in zero load. In negative loads the net gain is at first very rapid, and far from linear with time. -20 -10 0 ^10 Total Water Load Fig. 89. Eates of total water exchange (% of BoAour) in relation to mean total water load (% of Bo). Equilibration diagram of earthworm in fresh water at 18° C. Each point represents 5 to 22 tests. Data were either in initial or later periods of one hour. First it was ascertained that the total turnover is 2.7% of Bo/liour (Wolf, '40a). Since in the turnovers later determined the output is 2.0%/hour greater than the intake (fig. 88), 1.0%Aour is subtracted from all outputs and added to all intakes on the right of zero load; and the abscissae are moved 1.0% to the left. New data of Wolf and Adolph. Since total gain and total loss at diverse loads were measured in the same individuals, a complete equilibration diagram is available (fig. 89). In the initial hour, compensation of each deficit is eight times as rapid as correction of equal excess. Most of the modifica- tion is in the rate of intake through the skin. Like the frog's, this gain does not diminish in excesses ; in excesses the small modifica- tion of nephridial output alone allows slow recovery. The earthworm, as was said, recovers from water deficits chiefly by faster intake through the skin. The rate of intake increases 144 PHYSIOLOGICAL EEGULATIONS many fold at deficits that do not even double the believed osmotic pressure of the body fluids (Adolph, '27b, p. 56). Diminution of water output through the usual channels contributes but a small portion of the recovery from deficits ; in excesses it alone is avail- able. In this there is no feature by which the annelid worm differs from the frog or snake in its ability to compensate for unusual water contents. § 50. BiPALIUM Bipalium, a triclad turbellarian worm, is common in green- houses where sources of water abound. Individuals weighing 0.3 to 2.4 gm. were desiccated by Kawaguti ( '32) and then placed in tap 20 -50 -^o -30 -20 Water Load -\0 Fig. 90. Initial rate of water intake (% of Bo/hour) in relation to water load (% of Bo). Gain was in first 1.0 hour. Worm Bipalium (triclad turbellarian), 6 in- dividuals weighing about 1 gm. each, previously desiccated. On wet cotton at 30° C. Data of Kawaguti ('32), water (fig. 90). No effect of body size is evident among the few rates of exchange that he reported. An unusual feature is that the velocity quotient (1/At) is greater at very low water contents (large deficits) than in moderate ones. The course of the intake of water is gradual, resembling the earthworm's, and apparently occurs through the entire body surface. § 51. Akbacia egg. Echinodeem eggs Unfertilized eggs of echinoderms furnish considerable informa- tion about water exchanges. In the measurements of Lucke and McCutcheon ('27) on sea-urchin Arbacia punctulata, eggs first WATER RELATIONS OF OTHER SPECIES 145 swell in diluted sea water or shrink in concentrated sea water. On subsequent return to normal sea water (fig. 91), mean rates of intake and of output are computed from successive measurements of diameter (d). The rates of exchange in this series turn out to be proportional to the differences of osmotic pressure between egg and medium. However that may be, for equal differences of it, intake is slightly sloiver than output, at all temperatures from 12° to 24° C. In this spherical organism the exchanges are probably propor- tional to surface area (nd^) and to the gradient of osmotic pressure, according to the equation: SW/At = /iV"d^(Pi -Pe). The coeffi- a 1500 cpIOOO 1 1 1 o Arbacia Egg / LosSy/h ■^^jGa in f 1 1 1 ^ I t lS 500 0) +» o Q. ^ -50 0 +50 +100 +150 Water Load Fig. 91. Eate of net water exchange (% of Bo/hour) in relation to total water load (% of Bo). Unfertilized egg of the sea urchin Arhacia punctulata in normal sea water at 19° C. Eate is computed as though it prevailed during the first 1/60 hour; in 6 eggs transferred at zero time from each of 5 dilutions of sea water (80 to 40 per cent). Eates of gain in recovery from negative loads were not reported but are believed to be about 20 per cent less than rates of loss at the same positive water loads. Data of Lucke and McCutcheon ('27). cient h represents permeability (Lucke et al., '31). When li is con- stant and turnover is nil, the net equilibration diagram looks much like that for Phascolosoma. The eggs of half a dozen other species of echinoderms have been studied in an analogous manner (Leitch, '31, '36; Fukuda, '35) in limited respects. For the most part, water exchanges have been measured when eggs are transferred from sea water into various dilutions and concentrations. By analogy with Arbacia it might be supposed that exchanges in recovery depend only on the gradients of osmotic pressure that prevail. The data at hand do not en- courage this assumption, since significant differences in h appear within the same species at transfer into 40, 50 and 60 per cent sea 146 PHYSIOLOGICAL KEGULATIONS waters. In the early efforts to analyze the factors of water trans- fer and to rationalize the quantitative relationships, it was impor- tant to minimize the irregularities found. In a later analysis it is useful to realize that one permeability coefficient h does not charac- terize all exchanges of water in any one species. § 52. Feeshwatek Zoothamnium Zoothamnium, a genus of ciliate protozoa, allows independent measurements of output through contractile vacuoles and of coinci- dent body volume (Kitching, '38). Moreover, both initial states and steady states may be investigated. Two species are examined, one of which lives in fresh water, the other in sea water. When individuals of the first species are put into one of two chosen media, either a negative or a positive water load is estab- 800 1 e 600 u 2 6 400 ZOO Intake Freshwater Ciliate Zoothamnium -30 -20 Output- MO Water Load Fig. 92. Initial rate of water exchange (% of Bo/hour) in relation to water load (% of Bo). Zoothamnium, freshwater peritrich protozoan, at 15° C. A, an individual returned from 0.05 M sucrose solution to tap water. B, an individual returned from 0.005 M sodium cyanide, which had inhibited contractile vacuolar activity, to tap water. Each point represents recovery during a period of 0.07 to 0.17 hour. Data of Kitching ('38). lished. On transfer from those media to the control medium (fresh water), water is exchanged at the rates shown in fig. 92. The total water loss is the volume measured as leaving the body through vacuoles, while the gain is the change of volume plus the water loss. Measured in change of volume alone, the rate of net gain in deficits is much greater than the rate of net loss in excesses. It is uncer- tain whether this contrast in positive and in negative loads is gen- WATER EELATIONS OF OTHER SPECIES 147 eral, since the means (agents) used to attain these loads could be considered as special and peculiar ones. This and the earthworm are the only invertebrate species for which both turnovers and complete equilibration diagrams, includ- ing total as well as net exchanges and loads of both signs, are avail- able at present, so far as I know. Relative to body volume, turn- over in Zoothamnium is a hundred times as rapid as in earthworm. <^ 53. Marine Zoothamnium In a marine species of the same genus, water output is measured in diverse stationary states of water content (fig. 93). Loads are 400 Zoothamnium 1 /■ 3 ■H ■3 o morinum J /_ |200 - / - o / ""^ o \o- 1 .. 1.. 0 +50 +100 Water Load Fig. 93. Steady rate of water output through contractile vacuoles (% of Bo/hour) in relation to water load ( % of Bo) . Zoothamnium, peritrich ciliate, at 15° C. Each point represents one test, of rate in diluted sea water, between two measurements of rate in normal sea water. Data of Kitching ('34). secured by keeping individuals in any dilution of sea water for an hour or more ; then body volumes and steady vacuolar outputs are ascertained. Since the body volume remains constant, rate of water intake equals rate of water output. Very often it is assumed that water continually leaks into such an organism, whereupon the vacuoles pump it out, relieving the body of fluid. The augmentation of rate, amounting to 20 times the usual, and the shape of the curve in figure 93, suggest that the rate of output is proportional to the reciprocal of the concentration of the medium, a fact which allows the assumption of leakage. But the effective concentration of the body substance in each state of 148 PHYSIOLOGICAL KEGULATIONS load being unknown, nothing can be done to test quantitatively the notion that the rate of exchange is proportional to the gradient of believed "osmotic" pressure; it is indeed improbable that the con- centration inside the body is the same in all loads. § 54. Ameba The belief is wide-spread that if Ameba regulates its water con- tent, then any animal can. In this freshwater rhizopod, deficits of water (-25% of Bq) are produced by immersion of individuals in 0.2 M to 0.4 M lactose-saline solution (Mast and Fowler, '35, p. 160 ; '38, p. 303). Recovery of volume, after return to the weak saline solution in which the Ameba proteus has been reared, proceeds dur- ing the initial 0.5 hour with velocity quotients of about 0.8/hour. In solutions of similar weak concentrations but of other ions, recov- ery is slower. Whether any of the recovery occurs by reduction in rate of output through the path of contractile vacuoles was not ascertained. The maximal rate of turnover measured in contractile vacuoles of A. proteus is about 24% of Bo/hour, with a mean at 23° C. of 10% of Bo/hour (Adolph, '26, p. 377). The rate of output through vacuoles varies with the salt concentrations of the medium (Miiller, '36, p. 360) ; it is not certain that the body volume is unchanged in these concentrations. Excesses of water may be introduced into Ameba dubia by micro-injection. According to Chambers and Reznikoff ('26) the water, even in an amount of one-third the body volume, mixes with the body substance. Measurements of subsequent vacuolar output (Howland and Pollack, '27) indicate faster excretion, lasting 0.1 to 0.3 hour, at augmentations 1.16 to 2.9 times the control rate. The response was observed three times in one individual after as many successive injections. Following an injection of one-half the body volume, an augmentation ratio of 3.9 was recorded. Apparently the rates of compensation of water content in Ameba are, like Zoothamnium, proportioned to its high rate of turnover. That freshwater protozoa are fully equipped for deal- ing with both deficits and excesses of water is worthy of explicit statement. § 55. Note on plants The water relations studied are equally prevalent in plants. Measurements of transpiration and of potometry now call for cor- WATER RELATIONS OF OTHER SPECIES 149 relation with water load, ascertained either by weight or analyzed water content. It is known, for example, that maize (Maximow, '29, p. 210), wheat (Vassiliev, '36), and many other species (Knight, '17; Pisek and Berger, '38) readily suffer reductions of water content, and that evaporative losses are then diminished. Gain by intake through roots as soon as water again becomes avail- able is presumably faster than in controls. Rates of turnover are often enormous, some plants exchanging their own weight of water every hour (Knight, '17). In aquatic plants with large fronds, half of recovery from desiccation occurs in a few seconds (Kalt- wasser, '38). It is, however, not proposed to analyze further any of the data concerning organisms that are ordinarily considered to belong to the plant kingdom. § 56. Summary Of the many kinds of animals whose maintenances of water con- tent might be studied, a small sample is available for future com- parisons. Only two species of invertebrates (Lumhricus and Zoo- thamnium) having turnovers, were examined for rates of exchange during recovery from both excesses and deficits. Both are as well equipped to correct water contents as any mammal. No turnover of water and no special structures for water ex- changes are found in some animals that live in sea water. In water balance any possible turnover is then by the imperceptible replace- ment of molecule for molecule. In water increments, adjustments still result in an equilibration. No species has been found without augmentations in rates of the appropriate net exchanges, whenever water loads occur. This fact strengthens the impression that the maintenance of water content in any species may be universally studied by correlating rates of net water exchanges with water loads. Very often ''water diuresis" is considered a wide-spread phe- nomenon. Extending the term to include all increases in rate of water output by any visible path when positive water increments prevail, actually I find data demonstrating it in 8 species of mam- mals, one species of birds (Burgess et al., '33; Korr, '39) four of reptiles (fig. 80; Burgess et al.; Boyd and Dingwall, '39; Friedlich et al., '40), one genus of amphibia, one species of fishes (see fig. 102) ; two species of insects (fig. 84; Lester and Lloyd, '28), one 150 PHYSIOLOGICAL EEGULATIONS genus of crab (Nagel, '34), one kind of annelid worm, and 4 genera of protozoa (fig. 93; Miiller, '36). Perhaps a few more species could with further search of published reports be added to that list. I, too, have a sort of faith that all species, and especially those with turnovers, put out water faster when the body has a water excess. That inference is useful, but is not demonstration. Sometimes ''water drinking" is considered to be limited to mammals, birds and reptiles. Though paths may be anatomically different, greatly augmented rates of water gain are found in all species of other classes and phyla in which intakes have been ob- served during recoveries from water deficits. There is nothing to compel physiologists to restrict the study of compensations by water intake to those animals that have a particular anatomical equipment. Eather, the evidence is that animals of all sorts ac- celerate their intakes in every deficit, and none trust solely to sup- pression of output. Chapter VIII EQUILIBRATIONS IN PARTS OF ORGANISMS § 57. The regulation of water content has now been examined in whole organisms. The individual is a convenient unit for study in that it usually maintains itself in a semi-isolated condition. The question arises: whenever any portion of an individual can be tested, are the same sorts of relations between its water exchanges and water contents found as in the whole body? Such portions (living units) are organs, limbs, tissue masses, cells. Do blood volumes, and fibroblast sizes, also tend to be constant, and are they corrected after disturbances of them? Equally well, aggregates of individuals might serve as units. They may be the populations of nests, households, farms, herds, towns, forests, continents. Nor may it be assumed that the group is the mere sum of its individual units ; for when associated, new conditions impinge upon the individuals. The water exchanges of a whole city could be measured after droughts and after rains, after interruptions of supply and after forced utilizations ; such data will not be presented here. In the study of parts, two sorts of physiological situation may be kept distinct. In one, the portion of an individual remains in place {in situ) and, although loaded directly, recovers while sharing its usual relations with other parts (§ 58 to <§ 62). In the other, to be considered thereafter, (§63 to ^66) the portion is isolated. § 58. Volumes in situ In the organism as a whole, the content of water was very often not identified chemically, but was measured as weight, or (in proto- zoa and in echinoderm eggs) as volume. In parts of individuals in situ the measurement of volume becomes paramount. This may be made more explicit by designating the increments by ± AV in- stead of by ± AW. Of the volumes that are measurable in organisms, many are volumes of distribution. A particular substance in known amount is added to or subtracted from the body by a particular route, and its increment in some tissue is subsequently ascertained. In a re- stricted sense a volume of distribution is defined (Dominguez, '34) 151 152 PHYSIOLOGICAL KEGULATIONS as ''the volume of body fluid dissolving the substance at the same concentration as the plasma. ' ' More generally, a component added may supplement some already present, or if subtracted may decre- ment that already present; this (positive or negative) increment spreads through a virtual volume of distribution. Going beyond the limits of body fluids as distributees, of substances (heat or pres- TABLE 5 Mean volumes of distribution, as measured in plasma or whole blood of three species of mammals, after introduction of excess substance into the plasma or into the stomach. The concentrations of the substances marked * were measured in whole blood; the others were measured in plasma Distribuend Dog Man Eabbit Source of data Vital red 4.8 4.3 3.1 Smith etal. ('21) Sunderman et al. ( '36) Utheim ('20, p. 387) T 1824 4.9 4.3 Gibson et al. ('38) Gibson and Evans ('37b) Trypan red 3.7 3.4 Oka ('38), Takahashi ('35a) Hemoglobin 4.5 Lee et al. ('22, p. 156) Carbon monoxide* 8.7 7.3 5.6 Smith et al. ( '21, p. 351) Douglas ('10) Boycott et al. ( '09) Sulfate 21.0 Lavietes et al. ('36) Sucrose 22.0 Lavietes et al. ( '36) Xylose 26.0 Dominguez et al. ('37) Harrison et al. ( '36) Chloride 27.0 28.0 Thiocyanate 32.0 24.0 35.0 Crandall et al. ( '34) Lavietes et al. ('36) Bromide 32.0 27.0 Brodie et al. ( '39) Iodide 37.0 Wallace and Brodie ('37) Glucose* 50.0 30.0 Wierzuchowski ('36) Sveinsson ('40) Creatinine* 63.0 Dominguez et al. ('37) Urea* . .. . 63.0 65.0 Painter ('40) Sulfanilamide* Painter ('40) sure might be distributed) as distribuends, and of plasmas as dis- tributors, I define a volume of distribution ( Vd) as a virtual volume computed to contain an added or subtracted quantity (AGo) in that increment of concentration (AGs/Vs) found by analysis of a chosen sample. Or, Vd equals AGo introduced into or removed from the living unit (of volume Vo or weight Bq), divided by AGs measured in the specified tissue per unit of its volume Vs. Usually Vd is EQUILIBRATIONS IN PARTS OF ORGANISMS 153 expressed as a fraction of the living unit; or Vd/Bq = (AGo X Vs)/ (AG3XB0). Volumes of distribution have a wide range when diverse com- ponents are administered (table 5). In general each distribuend has a smaller volume of distribution in man than in dog, and smaller in rabbit than in either. Most volumes of distribution have been identified with anatomi- cal ' * compartments ' ' by hypothesis only. A few of the substances commonly administered as indicators of volume are supposed to attain after certain intervals of time virtually equal concentrations in all the ivater of the body (urea, sulfanilamide). Others are thought to become distributed in equal concentrations in the water of extracellular localities (SON', Cr, Br', sucrose, SO*"), with the exception that erythrocytes at least also contain them. Thus by inference it is said that a certain procedure measures something that approximates ''plasma" volume; another, circulating whole ''blood" volume; another "extracellular" volume. In general it is a necessary precaution to specify : Vd of brilliant vital red volume at 0.1 hour after injection by vein ; Vd of sodium thiocyanate vol- ume at 3 hours after injection by peritoneum ; Vd of urea at 2 hours after ingestion by stomach. Perhaps no two methods of measuring volume are likely to agree : either the actual volumes of distribution differ, or some factors in the two procedures are systematic ones. A notion that may be derived is that one procedure is not "su- perior" to another (though the volume it measures be identical with volume found by other methods), but that each procedure measures reproducibly a unique volume of distribution. Evidences of regulation now to be mentioned concern either parts of organisms that are volumes of distribution or parts the volumes of which are measured by mechanical or optical methods. § 59. Blood and plasma The volumes of "blood" and of "plasma" are peculiarly suited to the establishment of quantitative excesses and deficits ; perhaps this merely means that addition or subtraction is little trouble. Many hypotheses have been erected concerning the lability of those volumes. Choosing increments of volume that directly concern the blood, I here study hemorrhage and transfusion. In rabbits, "blood" volumes are ascertained as volumes of dis- tribution of trypan red, divided by hematocrit ratio (fig. 94). The 154 PHYSIOLOGICAL REGULATIONS accompanying "plasma" volumes (fig. 95) are the same volumes of distribution of trypan red, injected anew at each measurement. Transfusion of citrated rabbit blood suddenly enhances both vol- umes ; hemorrhage suddenly diminishes them. Thereafter, changes are most rapid immediately following the experimental procedure ; indeed comparisons of the volumes actually injected or withdrawn with the loads measured by the trypan red procedure, show that the fastest adjustments of ''plasma" volume occur in the minute dur- ^^Q - 5 \ Tests V I 1 Rabbit _g QJ ^-20 - c ) 2 3 Hours Fig. 94. Course of volume load (% of Vo) during recoveries from the transfusion of citrated rabbit blood (open circles) and from hemorrhage (solid circles). Eabbit, mean body weight 2000 gm., 5 to 8 individuals and tests. The volumes concerned are the volumes of distribution of trypan red (given by vein) divided by the hematocrit ratios. The enveloping broken lines indicate standard deviations ; these are 2 to 3 times the stand- ard errors. Data of Takahashi ('35a, '35b). ing which the transfusion, and especially the hemorrhage, is ad- ministered. "Plasma" fully recovers its initial volume within 0.5 or 1.0 hour, while whole "blood" does not. During an arbitrary interval of time between the first and sec- ond measurements of trypan red volume, the rates at which each volume changes are compared (fig. 96). Following an excess of volume produced by transfusion, the "blood" volume decreases; following a deficit of volume produced by bleeding, it increases. From the total "blood" volumes and the total "plasma" volumes, EQUILIBRATIONS IN PAETS OF ORGANISMS 155 ^EO- E > +10 O E E 0 I -10 h q: — 1 5 Test -■■1 1 1 / / Rabbit 1 1 1 ) • 0 12 3 Hours Fig 95. Course of increment in volume of distribution of trypan red given by vein ("plasma volume") (% of Vo), during recoveries from tranfusion of citrated whole blood (open circles) and from hemorrhage (solid circles). Eabbit, same tests of Taka- hashi ('35a, '35b) as are shown in figure 94. The enveloping broken lines indicate standard deviations, which are here 2 to 3 times the standard errors. (U -H a -30 -10 -10 0 +10 +20 Volume Load Fig. 96. Initial rate of net volume exchange (% of VoAour) in relation to initial "blood" volume load (% of Vo). For each point 2 to 8 individuals were subjected to hemorrhage (negative loads) or transfusion (positive loads) ; and the changes of volume were measured over the period from 0.03 to 0.50 hour thereafter, and the period from 0.03 to 1.00 hour thereafter, respectively, as indicated in figures 94 and 95. The volumes con- cerned are: the volume of distribution of trypan red (termed "plasma"); this volvune divided by the hematocrit ratio (termed "blood"); and the difference of blood minus plasma (termed "cells"). Data of Takahashi ('35a, '35b). 156 PHYSIOLOGICAL REGULATIONS are obtained by difference the supposed total circulating "cells" volumes. Exchanges of "cells" are very much slower than of "plasma." Rates of recovery of volumes are derived from the differences between pairs of successive measurements ; but each of them varies in absence of load by a coefficient of difference of about ±: 3 per cent (table 6), and hence some of the data are statistically TABLE 6 Coefficients of difference (CA) in volumes of distrihution in three species of mammals. CA represents 1/^/2 times the root viean square of percentage differences between successive tests on 3 to 6 individuals Distribuend Number of paired tests ... Interval between tests, hours CA of ' ' plasma ' ' volume, % of mean CA of "blood" volume, % of mean Source of data Dog Vital red 4 0.32 1.11 1.77 Smith ('20) Man Vital red 5 170 3.47 (C.V.) Sunderman and Austin ('36) Eabbit Trypan red 12 0.33 3.08 2.84 Trypan red 6 1.0 2.91 5.26 Takahashi ('3.5a) insignificant. Nevertheless all average exchanges of "blood" and of ' ' plasma ' ' are of such a sign that recovery is occurring. Further, reciprocals of increments in concentrations of various other con- stituents whose circulating amounts are assumed to be fixed, such as hemoglobin, solute refractive index, and colloid osmotic pressure (Onozaki, '35; Nagaoka, '36), indicate proportional changes of water content of the same signs. The responses to loads are both qualitatively and quantitatively comparable to the net equilibration of water content or body volume in the dog (fig. 16) or other organism as a whole. Net equilibra- tion thus presents a like pattern, regardless of the fact that entirely diverse processes and tissues are concerned in the exchanges of fluid. No means has been found of ascertaining what over-all volume is both gained and lost during equilibration, or during turnover. Hence total rates are not recorded, as they are for the whole dog or rabbit, but only net rates of exchange. The rate of turnover might be enormous, for it is believed that plasma in every capillary both gains and losses liquid continuously (Schade, '27). Corre- spondingly the fluctuating variability of ' ' plasma ' ' volume, which EQUILIBEATIONS IN PARTS OF ORGANISMS 157 might however represent mere error of estimate, is considerably larger than for the whole animal. Attempting to identify some of the tissues concerned in restor- ing blood volmne, Oka ('38a, '38b) induced various surgical and toxicological injuries in rabbits, then repeated the transfusions (table 7). These injuries are known to affect the liver or the kid- TABLE 7 Rates of loss of "Hood" volume as measured by trypan red injection and hematocrit ratio, in rahhits transfused with citrated rabbit blood. In each series, 5 individ- uals received by vein 2 per cent of the body weight (Bo) of blood; mean body weight 1750 gm. The volume of distribution (Vo) of trypan red was meas2ired at 2 minutes {-initial load) and 28 min- utes thereafter. Standard error is here equal to 0.45 of the standard deviation (a). Data of OTca {'38a, 'S8b) State of rabbits Mean Vd before operation, % of Bo Mean Vd 46 hours later = after oper- ation, % of new Bo Initial "blood" load, % of new Vd Rate of loss from "blood," % of new Vc/hour a of rate Normal P-poisoned Chloroform-poisoned Bile-duct-ligated Nephrectomized Ureters-ligated 6.90 6.69 6.46 6.54 6.67 6.47 6.76 6.95 7.52 6.97 6.67 7.71 -h29.5 -f 28.8 + 26.6 + 28.8 + 30.0 + 25.9 13.3 15.6 34.0 -5.6 6.6 12.4 + 11.0 + 17.1 + 11.0 + 14.7 + 16.3 + 16.1 neys, whatever else they do. The augmentations in mean rates of volume exchange appear to be significantly greater in chloroform poisoning and significantly smaller in bile-duct ligation; hence effects are in opposite directions. It is hardly to be inferred, how- ever, that the liver alone is specifically inhibited by those agents, nor is the liver demonstrated to have a particular role in restoring ''blood" volumes that have been increased by transfusion. The events that follow when whole blood is added to the volume already circulating differ little, I gather, from those after some other form of water is added to blood, though the initial rates with which recovery occurs may differ for diverse fluids. Several added fluids were compared by Oka, with the result that the differences of exchange rates among them in the initial half -hour of recovery are probably insignificant. For the study of the equilibration of "plasma" volumes, which were measured (fig. 94) in response to whole blood excesses and deficits, the increments administered might equally be transfusions 158 PHYSIOLOGICAL EEGULATIONS of plasma rather than of whole blood, and plasmapheresis (without replacement of anything but corpuscles) rather than hemorrhage. Suitable data of those sorts scarcely exist, unless in dogs subse- quent changes of volumes be inferred from changes of concentra- tion (Weech et al., '33; Calvin et al., '33; Freeman et al., '38). In rabbit many independent data confirm the recoveries of ''plasma" and of ''blood" volume above described {e.g., Boycott and Douglas, '09 ; Nagaoka, '36 ; Oka, '38). In dog, man, and other species, partial data are not lacking. Those who are familiar with measurements of "blood" volume each have favorite researches that might be considered in this connection. What I would prefer is a thoroughly systematic study, with estimations of volume at initial intervals of only 0.02 hour, that does not yet exist. In a few words, the maintenances of "blood" and "plasma" volumes depend upon net exchanges of those fluids in appropriate directions. After an excess has been established, net loss is rapid and in proportion to the amount of excess. After a deficit has been established, net gain appears. "Plasma" is exchanged faster than blood "cells." Investigations designed to discover machinery by which these compensations operate have succeeded only in confirm- ing these bare facts for a number of diverse types of volume load, and in several species. §60. Arm The human arm is a suitable unit for measuring volume changes, if by a pressure plethysmograph the volume of tissues minus blood is observed at desired intervals of time. The rates of escape of tissue fluid that had previously accumulated under partial venous stasis are thus ascertained (fig. 97). The rate of decrease of vol- ume in the hydrated arm turns out to be a function of the amount of excess fluid present in it. No methods or data exist to tell whether the gain of volume to an arm previously depleted of some of the usual fluid would likewise be proportional to the deficit. So far as the evidence goes, the "fixed" tissue mass of the arm adjusts its volume after a disturbance of it, in the same manner as the blood of rabbit, or the whole man. <§. 61. Other organs Some organs of the animal body are geographically available for enclosure in oncometers and plethysmographs. Of the large equdijIbrations in parts of organisms 159 numbers of studies that have been reported upon the volumes of such organs, none indicates the rates of recovery of volume after diverse amounts of fluid have accumulated in them or been lost from them. Most of the changes measured are probably in amounts of intravascular blood contained in these organs ; suitable pressure plethysmographs remain to be applied to livers, kidneys, and muscles. Changes of volume as inferred from measured changes of con- centration in analyzed samples of these organs, would be significant 6 +2 0 Water Load Fig. 97. Eate of volume output (% of control arm volume /hour) in relation to initial volume load (% of Vo) in the human arm. The load has been imposed by previ- ous venous congestion, and the subsequent changes in "reduced" arm volume are as- certained in initial periods of 0.17 hour in a plethysmograph at 34° C. Data of Landis and Gibbon ('33). only if the changes are larger than the standard errors among samples. Those are usually very large. But occasionally the con- centration may be measured optically upon one portion of tissue in situ. By that device control data are ascertained on the same sample, avoiding errors of absolute factors and of variabilities among samples. Further advantages are found in studying recov- eries from those increments of volume that are predominantly in the tissue in question rather than in the remainder of the body; such was actually the case for blood, plasma, and arm. 160 physiological regulations § 62. Cells and nuclei It would be a pleasure to include in the present study changes in the volumes of cells within animals. Possibly recoveries of volume are ascertainable in circulating erythrocytes {in situ), but they have not been ascertained. Some sort of opacitometer plus enumerator is required to play upon the blood within the blood vessels, calibrated specifically for volume changes. Increments of volume found in drawn blood have probably occurred much more quickly than subsequent measurements can be made. In some tissues, areas of single cells (Hirsch, '37) and nuclei (Wermel and Portugalow, '35) may be directly measured. The water exchanges might then be ascertained immediately after the microinjection of fluid into or the emptying of material from those units. Eventually other subdivisions of single cells and of unicellu- lar organisms may be similarly observed. The rates of shrinking or swelling in nucleoli, granules, and other units recovering from various disturbances of volume (preferably those that affect the unit more than the remainder), may then be measured, and the equilibration of volume described. At present this measurement has proven feasible upon the germinal vesicles in ova of invertebrates. In a species each of echinoderm and of annelid eggs, volume changes during swelling of the germinal vesicles (nuclei) are measured in parallel to those of the whole individual {Asterias, Beck and Shapiro, '36; Cerato- cephale, Kamada, '36; Arhacia, Churney, '41). The interval of time between swelling of the Qgg and swelling of the nucleus is much shorter for the echinoderm eggs in sea water than for the worm eggs immersed in a modified Ringer's solution, although the time required to complete the swelling is about the same in all species. Possibly similar contrasts would occur during recoveries from water loads ; at least the measurements have been demonstrated to be feasible. § 63. Summary of parts in situ The studies of recoveries of volume in situ mostly concern volumes of distribution as measured by adding or subtracting components in the body. In one study, "plasma" and ''blood" volumes of rabbits are reduced by hemorrhage and augmented by transfusion. Thereupon exchanges appear that effect restorations of volume; "plasma" volumes recover much more quickly than EQUILIBKATIONS IN PAETS OF ORGANISMS 161 "blood" volumes. By transformation of coordinates, net equili- bration diagrams are obtained for these volumes. Certain experi- mental "injuries" to the rabbits modify significantly the rates of plasma exchange, possibly indicating what sites or intermediaries take part in the exchanges. In studies by direct mechanical or optical measurements, volume changes are ascertained in arm, in certain distinct organs, in cells, and in their subdivisions. Only for blood and for arm are concrete portions of the equilibration dia- gram definitely established. § 64. Isolated parts Another large field in the investigation of equilibrations of vol- umes lies in isolated parts of organisms. These parts may be living or dead, or both, according to diverse criteria that are arbi- trarily chosen; for, survival after isolation is inevitably a matter of definition. In general the increments of volume whose recovery is to be studied may be established (d) before isolation, or (e) after isola- tion. Thus (d), a dog or a frog may be loaded with water, and then its tissues may be isolated and allowed to equilibrate by ex- changing fluid with a selected medium. Or (e) the tissues or cells may be removed from the body while in water balance, thereafter loaded and then transferred into the standard medium for equili- bration. Blood or plasma, being liquid, might be studied under isolation by putting samples of either behind uniform or identical partitions of some natural or artificial membrane. While the samples are ex- changing water with large volumes of a selected liquid, the rates of transfer would be measured. Other body fluids may be studied in like fashion. In any isolated units, the number of arbitrarily chosen factors will be much larger than within the intact body where these factors are already fixed by the organism. One or another of those factors will seem to someone to be "unphysio- logical. ' ' § 65. Isolated muscles Muscles isolated from frogs are studied by both method (d) and method (e). Positive loads are produced by injecting water intra- peritoneally and allowing either 0.5 or 1.7 hours for the water to distribute itself; then the muscles are isolated (method d) and 162 PHYSIOLOGICAL EEGULATIONS placed for recovery in a standard Ringer's solution. Duplicate muscles are analyzed with respect to dry residue, to find what their mean water content might be. Negative loads are imposed by first isolating the muscles (method e), and allowing them to desiccate by evaporation. At zero time they also are placed in samples of the chosen Ringer's solution. The tolerance curves (fig. 98) show adjustments toward a com- mon final weight. But the final weight is higher than the initial one, 0.4 0.6 Hours Fig. 98. Course of total water load (% of Bo). Isolated thigh of Bana pipiens, at about 21° C. The initial weights (Bo) are those found after dipping once into Einger's solution; later weights are after sojourn in Einger's solution since zero time. The lower three series (8 or 9 thighs each) have been previously desiccated in air, sub- sequent to isolation; the upper two series (15 to 19 muscles each) previously loaded by injecting distilled water into the peritoneal cavity of the living frog. In the latter case the initial load is computed from analyses of 30 duplicate muscles by the relation AE - 100 (Do/do -100. Data of Wolf ('40b). being in the neighborhood of +9% of Vq. Only muscles having excesses of water greater than + 9% of Vo lose water to Ringer's solution after isolation. Whatever the load, changes occur in the initial 0.2 hour that are independent of the later trends. In the form of an equilibration diagram (fig. 99), the same data indicate that the volume toward which recovery aims, as well as the rate of exchange, is modified with time. A shift in final volume of the muscle as compared with its original Vo (upon which the numerical scale is based) follows the isolation as such (Parry, '36). EQUILIBRATIONS IN PARTS OF ORGANISMS 163 No way is known of rendering the separated muscle like the intact muscle in respect to volume adjustments. Ingrained in investigators is the desire to say : the changes of volume in blood and in muscle may all be predicted, for the force of osmotic pressure, which is concerned here, produces movements of water proportional to the gradient of concentration of solutes. Such a hypothesis appears valuable, for in part the data agree with it. But the data do not allow the conclusion that only osmotic pres- sure is concerned. Even after the initial 0.2 hour, many other pressures or forces may be changing along with the volume of the 20 c y o 10 ?30 5econd half hour +Z0 ■ZO -10 0 ^\0 Total Water Load Fig. 99. Eate of water exchange (% of BoAour) in relation to total water load (% of Bo). Isolated thighs of Rana pipiens, immersed in Einger's solution at 21° C. Zero load represents the thigh's weight when in the frog that is in water balance. Data (of Wolf, '40b) taken from figure 98. tissue, and any or all of these might give a like result. In general, one scientist may get a thrill at finding the data in agreement with some prediction, another at finding the unpredicted. Therefore the actual obtaining of any data, however futile and unnecessary to the one, will be a source of satisfaction to the other. Possibly a third might be distinguished, who is interested only in prediction, and who does not find it necessary to retard the production of hy- potheses by ascertaining what the facts are. I believe it is rather distinctive of isolated tissues that following isolation, progressive shifts occur in the water content at balance (Vo). Recoveries from increments are then aiming toward a new content, even as balances in control samples do. This may not be universal, however, and the ideal tissue for study of equilibration 164 PHYSIOLOGICAL KEGULATIONS of water content is one that gets over the stage of "feeling" its isolation, so that for a time equilibration is complete. At present no such tissue is recognizable. § 66. Isolated cells Isolated cells are susceptible of study by the same methods. With due attention to ' ' survival, ' ' freshly isolated blood cells and previously cultured other cells have been subjected to volume changes. But cells rendered visible by direct fragmentation of massive tissues are equally available (Shear and Fogg, '34). Cells already isolated are not considered here ; it is more convenient to classify unicellular s of all sorts, as organisms (§ 51 to § 54) rather than as cells. a. Erythrocytes. For the most part, the rates at which these cells change their volumes have been studied by opacitometers. No instances seem to have been reported in which the changes mea- sured represent recoveries toward the volumes characteristic of cells in vivo. One deterrent to obtaining such recovery data is the belief that they can all be predicted from measurements of rates of swell- ing and shrinking in initiatory states ; however, recovery requires specific measurements. Mammalian erythrocytes are believed to exchange other substances than water, whether they be kept in plasma or in other prepared media (Rous and Turner, '16; Davson and Danielli, '38). Diversities among species may be expected. b. Leucocytes under certain conditions maintain regular (spherical) shapes, and so their changes of volume may be followed with an opacitometer. The data are limited to swelling in hypo- tonic solutions, followed by shrinking as recovery supervenes (fig. 100). While it is possible that the absolute rates reported are limited by other factors than the cells themselves, for the present the exchanges may have been at the rates recorded. Loss of water (recovery) occurs somewhat more quickly than the initiatory change of swelling. c. Fibroblasts in plasma cultures have been measured by mi- crometry of single cells (Brues and Masters, '36), but not during recovery toward original volume. It appears that fibroblasts from four sources (chick embryo, rat embryo, mouse sarcoma 180, and rat tumor 256) exhibit nearly equal apparent permeabilities to water upon transfer from 0.15 M to 0.05 M sodium chloride solu- EQUILIBRATIONS IN PAETS OF ORGANISMS 165 tions. What would happen if put into solutions containing more of the constituents of plasma or of interstitial fluid, is not known. Some fibroblasts are observed to engulf fluid from their sur- roundings (Lewis, '31). This process of "pinocytosis" may be an integral part of the exchanges concerned in turnover of water. d. Cells in fragmented tissues of mice were studied by Shear ( '35) with respect to a particular kind of swelling by bulges. These bulges occur commonly in either usual sodium chloride or Locke's solutions. When later transferred to solutions containing gelatin 0 0.05 0 O05 QIC 0.15 Hours Fig. 100. Course of water load (% of Vo). A suspension of rabbit leucocytes in equilibrium with Einger's solution is put at zero time into a similar solution but 0.4 as concentrated. After 0.05 hour, salt is added to bring the solution back to 1.0 concen- tration. Water load is estimated from the photoelectric potential created by a beam of light passing through the suspension, after calibration. From Shapiro and Parpart ('37). Redrawn by permission of the Journal Cell. Comp. Physiology. or other proteins the bulges diminish and disappear, at rates that, unhappily, were not ascertained by micrometry. The vacuole-like bulges might be counted as part of the cellular materials or not, as one chooses. It is apparent that recoveries of water content may be studied in single cells. Permeability itself is a regulated property; even when fixed with respect to water, it allows the usual pattern of equilibration by modifications of water exchange. Among cellu- 166 PHYSIOLOGICAL REGULATIONS lar units in general, either the whole cell may be concerned in giving up excessive water, or special structures like vacuoles may segre- gate that water, and either the whole cell or its surface may in deficits take up water. Both exchanges may represent special arrangements for regulatory water contents. § 67. Summary How water is handled by living units, is ascertained in parts of organisms in which the volumes can be measured. Volume changes are treated as net changes of water content. Both volume and water content suffer increments of many measurable types, each representing one method of loading the living unit. In all the tissue masses or cells that have been studied under conditions appropriate for recovery toward balance of volume (as defined previously, '§19), initial exchanges are faster as the incre- ments are larger. The directions of the exchange of fluid are such as to restore the masses to the volumes characteristic of control masses. I conclude that portions of individuals show much the same pattern of net compensations that whole organisms do. Comparisons of exchanges by the diverse tissues and cells studied (table 8) are limited by the less accuracy of the data as com- TABLE 8 Comparisons among net recoveries of water content or volume in tissues and cells Living unit Type of load Initial period of re- covery, hours Maximal rate of exchange, %of Vo/hour Mean peak load (AV), % of Vo Velocity quotient, 1/hour Half- life of load, hours Source of data Plasma of rabbit Hemorrhage Transfusion 0.5 1.0 30.6 15.0 -11.0 + 9.6 2.8 1.6 0.2 0.6 figs. 95, 96 figs. 94, 96 Blood of rabbit Hemorrhage Transfusion 0.5 1.0 21.8 9.7 -22.2 + 13.7 0.98 0.71 1.3 1.0 fig. 97 figs. 99, 98 Arm of Stasis of 0.17 7.4 + 1.24 6.0 0.2 fig. 100 man vems Muscle of Isolated and 0.5 19.2 -22.8 0.84 1.4 frog desiccated Leucocyte Isolated in 0.01 6000.0 + 68.0 88.0 0.008 of rabbit diluted Einger pared with those in whole organisms. It is clear that the most rapid recovery is in the smallest living unit, the leucocyte. Veloc- ity quotients indicate that exchanges between plasma and the re- EQUILIBRATIONS IN PARTS OF ORGANISMS 167 mainder of the body, and between arm and remainder through transport in plasma, are more rapid than exchanges in isolated muscle that has no vascular circulation. All recoveries are without measurable lag. Whatever divisions of labor there may be between individuals and their component parts, both kinds of units (whole and part) undergo, each in its own way, adjustments of water content. The vicissitudes of existence may often be diverse for each kind, but also sometimes not. Hence it is possible to select like tasks of regu- lation for rabbit, for rabbit blood, and for rabbit leucocyte; for example, each might recover from a deficit of water of like magni- tude. Roughly, to return half its deficit, the whole body requires 0.02 hour (fig. 74), the blood requires perhaps 1.3 hours (fig. 94), and the leucocyte requires 0.008 hour (fig. 100). Data are not available for comparing recoveries from deficits of one identical type and size that in fact proceeded simultaneously. A protection to the part is afforded by the fact that it keeps its place in the whole body, just as the whole organism's water content is regulated to some degree by keeping itself in certain environ- ments. For, the body limits the loads to which the part is exposed. And, the body facilitates the exchanges of water by which the part recovers. The relations between the organism and its constituent cells and tissues, therefore, may be equivalent, in part, to those between the environment and the whole individual. Chapter IX GENERAL FEATURES OF WATER EXCHANGES § 68. The investigation is now provided with an array of ma- terials. These materials offer information, and disclose relations, that seem to explain in part what each kind of organism does with water. Each handles water in a special way, as though there is nothing like it, and as though machinery to handle it is indis- pensable to life. How particular species, and indeed, other living units, adjust and recover customary water contents, seems to have been ascertained. The next task is to seek among those materials the constant and the inconstant relations. Instead of letting each species furnish its own story, I now try to find what all species together can tell about how to handle water. Two inclusive questions can be raised. What are the general features of all regulations of water content? And what quantitative factors differentiate the maintenances and com- pensations in one living unit from those in another 1 Each question will be answered by asking a series of subsidiary questions. For the present, the relations to be explored are among the four sorts of variables constituting the water-time system. Three of these (load, rate, time) are measured as such, though often rate of exchange equals change of load in a unit of time. The fourth (velocity quotient) is a particular ratio between two others (rate and load). Additional variables that creep in, fall in the quali- tative categories of paths, species, and types of loading. Those uniformities to be found among species and parts are similarities of correlations among the four variables; they are to be listed here. Diversities among them are of two sorts ; one is quantitative differences in correlations among the four variables, which will also be sought. Such diversities are supplemented by a second sort, namely, other correlatives of the four variables, some of which will be ascertained later (chapters X to XII), additional data being considered to that end. '^ 69. Some limitations For comparing the water relations of two organisms of great anatomical diversity, such as a mammal and a worm, several dimen- 168 GENERAL FEATURES OF WATER EXCHANGES 169 sions of reference are available. Occasionally the absolute volumes exchanged are described. Or, the exchanges are expressed as frac- tions of the volume and the water content of the body, whereby other objectives are gained. Again, time is measured relative to some physiological process (physiological times) instead of by the same clock in all species. Diversity of conditions and states in each species constitutes a severe limitation to comparisons ; I therefore compare while insist- ing upon the tentative nature of the results. Particularly, com- parisons of water exchanges by isolated tissues with those of tis- sues in situ may seem fantastic. Probably this opinion is the safer one ; perhaps the same view is preferable throughout comparative biology and in all analogies to non-living systems. It is always a propos to say : in this seemingly neat comparison the diversities of surface area, or the presence or absence of a circulation, or the amenities of vitality, are disregarded. So they are, and I think something is disregarded of the hundreds of possible factors, in any conceivable comparison. Hence the investigator chooses be- tween making provisional comparisons and making none; if the former, it is invidious to say that one mode of comparison is better than another; it may rather be said that it is different. Quanti- tative comparisons have the probable advantages of being im- plicitly provisional, and of being guided by dimensions and other criteria away from some varieties of confusions. I may confi- dently expect that some of the features coming to light are inde- pendent of what the observer imposes. The utmost judgment will not insure like conditions for two individuals or species, or for two days. Diversity of states in one individual may or may not rep- resent 'inherent" variability, so long as there is no possibility of robbing the organism of all environment. Why not subject all species to the same conditions, it may be asked? Only in the abstract, might it be ideal to use identical means and criteria of water load in all the organisms that are to be studied comparatively. (1) Not all organisms can survive like treatment or like environment. Desiccations of Ameha and of Thamnophis by evaporation, though both technically possible, do not attain the same water deficits in a uniform time. Recoveries of dog immersed in water, Ameha in air, rabbit blood in distilled water, are equally unsuitable. (2) Not all organisms are fur- 170 PHYSIOLOGICAL REGULATIONS nished with homologous structures and paths. What can be counted as alimentary water in Arbacia egg, as urinary output in leucocytes'? (3) Even when organs are supposedly or actually homologous, their performances may be diverse. Thus, frog's urine is not known to be hypertonic to blood plasma under any circumstances ; dog's urine is ; both urines are produced by kidneys. In earthworms the output from nephridia may or may not be counted as urine, and as homologous with the renal output in dogs. Three complete equilibration diagrams were constructed for the dog (fig. 38), all of which are equally valid in representing this species in comparisons of water relations. In fact, at least six types of modification of water content were recognized; some in excess and some in deficit; and there was no unique connection between the particular excess and the particular deficit associated in any one equilibration diagram. Among such diagrams, three or more time relations were utilized: initial (of diverse durations), steady, and maximal rates of water exchange. In comparisons among living units, whether species or parts of individuals, these several distinctions at least may be maintained. Each of them avoids one kind of confusion. § 70. Some paeameters comparing exchanges Measurable gain was contrasted with measurable loss of water at each load. What methods of evaluating the contrasts between the two, and of exhibiting the modifications of each exchange sepa- rately, are useful? (1) Economy quotients, the ratios of gain to loss, characterize the responses to water increments in the several species (table 9). The fastest recovery from water excess means suppression of gain while loss is maximal ; and the fastest recovery from water deficit presupposes the reverse. The economy quotient evaluates how closely this possibly optimal relation is approached. The dog's exchanges show higher quotients at every negative load than the rat's, which in deficits surpass those of man, frog, and earthworm. In positive loads, values departing most from 1 might be sought. The quotients of dog and man are not most extreme in short periods of time, for in water excesses the frog's diuresis has less lag than the mammal's. By choosing diverse modes of comparison, a case can be made for ''superiority" of the quotient in almost any species. Economy quotients are available GENERAL FEATURES OF WATER EXCHANGES 171 H w hJ pq =0 < H \ K W o> fel 1 03 1 ^ «H o 5 O ^ .3 05 C3 CO til05 rH-^Ttf^t^t^tOCSW o >-i . ^ . . . .0) bii>j be bDbD be '-I bi) bb tjD bb bJD cpiMta uscq «c [^ 5C tc « «B crt 03 00 : ; ■* ; 05 4- 0 (M : to ■ 0 ■ d • d (M ■* 1— 1 I— 1 ; Oi CM + 0 : CO l>; t^ ; 0 ■ d d d ; 10 to ■ 1 «o -<*<;•<*< ; 10 T-H Tj< : t:^ 05 + O ; o 1000 ; in 00 t- d-d ■ 0 0 d • d d d oo CO t^C0 00l0i-(00C0«0(>TO 00 + ooo oo oo oo t^ oi 00 o < odd <^ d d d d d d d d T)< I— 1 cgtr~i-H t^W -rHOO OlO lO 1 + rH O 1-H O O r-<_^ i— 1 O 00 : ^3 o d d d d d d d d d ;h I— ( N O CO o o> cq 0^ 4-9 1 t^Sr-ids t-:^ ds i-i o3 iH < CO OOO o w oi in (£> 1 Oi 3 Tfi to 3 d 3 t>^ 3 ^' ^* ^' i-H (M rH r-l to O O ; 0 0 0 0 »n 1 t-^ 3 00 ^ t~ ^ (M 2 in CO CM -*< »0 ■ • CO rH " c^ ; 0 0 TjH I d 06 d 1 ■ ■ to ■* : 0 0 CI 1 I d d ; to (M ■■■•■••• iH ->oO 00 00-3 O''^ O ^^ 5- ^^ ^^ i^ ^ 2 ^ TS '-^"Sg'-l-S'-^-S'-^-SOrHd -t^v.S +^_2 ^_2 ■►^i^ -^^ -^^ -^ ^1^ 2| 21 £l £ £ 2 •r-l C^ ,jj .rH C5 -r-i C3 • iH C3 . rH -rH -rH Perio ^Fmcc PmPh FhPh PhPh ;=h Ph Pm to 'I-* ■ s « CO •"!* 172 PHYSIOLOGICAL KEGULATIONS only in those living units whose total gains and total losses are known. (2) Modification ratios. The responses to change in load, in terms either of total exchanges or of various paths, may be mea- sured by comparing the maximal and minimal rates of exchange found (table 10). Proficient adjustment calls for very rapid move- ment of water at one time, little at another. At a glance it is apparent that oxidative gain is no factor in recovery. Nor is either fecal or evaporative loss greatly concerned in restoring water balance in any species studied, or under any conditions TABLE 10 Bates of water exchange {per cent of Bq, in first 1 Jiour of recovery) when water load and at Species Total gain Ingestive and imbibitory Oxidative T.O. Min. Max. T.O. Min. Max. T.O. Min. Max. r>og Dog (stationary states) Man 0.25 0.25 0.15 0.06 0.06 0.02 6.2 8.5 2.5 0.19 0.19 0.13 0.0 0.0 0.0 6.1 8.4 2.5 0.06 0.06 0.02 0.06 0.06 0.02 0.06 0.06 0.02 Eabbit Eat Frog Lumiricus 0.54 0.68 1.7 2.7 0.07 0.10 1.5 2.6 6.4 10.5 13.9 19.0 0.47 0.58 1.7 2.7 0.0 0.0 1.5 2.6 6.3 10.4 13.9 19.0 0.07 0.10 0.01 0.01 0.07 0.10 0.01 0.07 0.10 0.01 Zoothamnium 230.0 230.0 255.0 230.0 230.0 255.0 TABLE 11 Modification ratios, augmentation ratios, and partitions of Species Modification ratio (Max./min.) Augmentation ratio Total gain Inges. + imbib. Oxid. Total loss Urin. Evap. Total gain Inges. + imbib. Oxid. Dog Dog (sta- tionary states) Man Eabbit Eat Frog Lumbricus ... Zoothamnium 100.0 140.0 125.0 90.0 105.0 9.0 7.3 1.11 OC OC OC OC 9.0 7.3 1.11 1 19.0 14.0 23.0 14.0 8.0 60.0 5.7 1.23 100.0 70.0 50.0 42.0 60.0 5.5 1.23 1.4 1.4 1.5 1.7 25.0 17.0 12.0 15.0 8.0 7.0 1.11 0.24 0.13 0.13 0.15 0.9 1.0 1.0 32.0 mo 13.0 18.0 8.0 7.0 2.4 0.0 0.0 0.0 0.0 0.9 1.0 1.0 1 GENERAL. FEATURES OF WATER EXCHANGES 173 listed; for each proceeds almost unclianged at diverse loads. In urinary losses (table 11) the several mammalian species have higher modification ratios and more extreme augmentation ratios than the frog and others. In rates of gain, mammals are outstand- ing in fitting intake to the state of the body. Those are two of many comparisons among the quantitative compensations in water loads. It will be recalled that total rates and turnovers are un- known for any parts (organs or cells) of organisms. (3) Variabilities of rates of water exchange in turnover (table 12) may be compared by coefficients either of difference or of TABLE 10 varies, in several species. Total and partitioned gains and losses are listed at no load (T.O.) extreme loads lotal loss Urinary Evaporative Fecal Source of data T.O. Min. Max. T.O. Min. Max. T.O. Min. Max. 0.25 0.17 3.3 0.08 0.03 3.1 0.15 0.11 0.15 0.02 figs. 2, 4, 12, 44 0.25 0.17 2.3 0.08 0.03 2.1 0.15 0.11 0.15 0.02 figs. 12, 23, 28, 24 0.15 0.09 2.1 0.08 0.04 2.0 0.06 0.04 0.06 0.01 figs. 50, 54, 59, 60, 61 0.54 0.16 2.3 0.31 2.2 0.17 0.06 fig. 74 0.68 0.40 3.0 0.24 0.06 2.5 0.40 0.30 0.40 0.04 figs. 77, 108 1.7 0.1 6.0 1.7 0.1 6.0 0.0 0.0 fig. 66 2.7 0.9 5.1 2.5 0.9 4.9 0.0 0.25 fig. 89, Wolf ( '40a) ; Davis and Slater ('28) 230.0 220.0 270.0 230.0 220.0 270.0 0.0 0.0 fig. 92 TABLE 11 water exchanges, in several species, computed from table W (Max./T.O.;Min./T.O.) Ingestive or imbibitory as per cent of total gain Urinary or vacuolar as per cent of total loss Total loss Urinary Evap. T.O. Min. Max. T.O. Min. Max. 13.0 9.0 15.0 4.0 4.0 3.5 1.9 1.17 0.7 0.7 0.6 0.3 0.6 0.6 0.3 0.96 40.0 26.0 25.0 7.0 10.0 3.5 2.0 1.17 0.4 0.4 0.5 0.25 0.6 0.4 0.96 1 1 1 1 0.7 0.7 0.7 0.7 76 87 87 85 99 99 99 0 0 0 0 0 99 99 99 98 98 99 99 99 99.9 32 32 53 58 35 100 93 100 18 18 45 15 100 100 100 94 92 97 98 83 100 100 100 174 PHYSIOLOGICAL EEGULATIONS variation, with respect to (a) total intake or total output, and (b) diverse paths of intake or of output. Rate of total intake is more variable, in successive 1-hour periods, than rate of output in dog and man, and probably in mammals generally. Variation of intake is about equal to variation of output in frog and perhaps in aquatic organisms generally. In the environments and conditions chosen for study there is also an inverse relation, it seems, between the variability of rate and the steepness of the correlation line in net equilibration diagrams (fig. 47). This steepness is also expressed in net velocity quotients of diverse species. Among species and environments there is a direct correlation between rate of turnover and variability of content (table 12). Or, the more water passes through the body, the more chance there is for net retention or expulsion of some of it (§ 38). Among mammals it is evident that large body size goes with less variation of content (in % of Bo), and less turnover (in % of Bo). (4) Maximal rates of exchanges, found at extreme water loads in each of the organisms studies, may not be regarded as absolute values for the species. Provisionally maximal rates of urinary output, evaporative loss, and ingestive gain are recorded in table 10. Table 13 compares the maximal rates observed in whatever species the states of water excess or water deficit were tested. Many possible dimensions of the body might be correlated with these rates ; the hope of finding a limiting factor in the structure or use of the alimentary tract or the urinary apparatus or any other one provision cannot be seriously entertained, for single ''bottle necks" do not often characterize physiological processes. (5) Ratio of rates occurring simultaneously in any two paths may represent the partition of flow (table 10) during turnover in balance as well as in recoveries. For example, wishing to charac- terize adrenalectomized dogs, Uyldert ( '38) showed that the ratio : urinary loss/ingestive gain, which was in normal dogs 0.46, was unchanged by unilateral adrenalectomy, but was increased to 0.64 by either bilateral excision of adrenals or by adding sodium chlo- ride to the diet. Thus, each specific purpose suggests a feasible method of com- paring rates of exchange. If one of them does not emphasize the unusual features in an experimental test or in a species, another will. Altogether it is unlikely that all methods will yield signifi- cantly identical results for two species or metabolic states. GENERAL FEATURES OF WATER EXCHANGES 175 > ^> ^ li PQ < H 00 » ■o =0 4> ?Sj o V. g 1^ re r^ -2 ^ o o O S ^ g rt - ^ . ^ Is! - O' ^ s si- rs 2£S ^ oo m , ^ r,3 fO -^_g Q 'O '*. & ^ o g - ^ cd cd ^ ct O) (B C3 O) c R a c:t crt in a rO is l-l ^ o <3J m WW^ o3 c3 go |d r§#* -^ r-i -^ (M ] rt ; fH a> 00 -M J1 > tu D ibC be '■ bD bDbUbtibDbtibDbDbDbllbDbtibD «3 i«3 « jqU efl qn Cfa C3 to tpl 53 t3 53 «3 -IJ >t3 Tt< 00 o eg _o CO o oi fo" o o o in o o o kJ 3 =« 2 ^ I— 1 iH ■ • • ■ i-l ■ 00 >0 CD rt< o dJ ft oo" CO o CO t--^ rate ake lypo- ial -M fi^ CO 3 00 ; Ci O O O s^" S '-' b c3 "^ <» r3 r-5 3 (M ■ CO idcqcvJ^------ = = = ^- a-ss- iH tD -fj ^< H 3 O CO in tc in • tS ,j3 o ; w' «c i>^ Oi t-i ira ; CO 00 ; Oi lo vn o CO M IIZ^ T-H OaiHCO :(Mr-l ;r-l-*lr-IO'-l S.S'^. (M i-H ^ - mal of e in liort o CO Maxi rate intak any si a> ft ^ ; oi CO ^ Oi ; 00 o ; o o o o ci C5 (M in o .^ -< ti ~ O) O) ."^ tuo bjo > 2 - : -M 4^ • I-H ■rQ .■e^ "^ ^ bD .2 i 1 bO -4- t> ■^ "S o £ -S .5? a c3 o ai h 0 5 ,£ •+--«e1so;sSeS^ioS3S C p: ^ a h^ s O c3 c3 bo °2 9 fl c« 3 3 ^ =H -4-' 3 3 a> • H bDbo OJ •rH 3 II 3 3 -i-j fifi general featuees of water exchanges 177 § 71. Time relations What can be learned by analyzing the time relations in re- sponses to sudden water increments? Initiation of response to load, signalized by acceleration of exchange, is (table 14) without measurable lag in water deficits of all species. In water excesses the responses are more diverse, they appear to begin most promptly where osmosis gives direct passage of water to and from the aqueous environment. Where fluids are excreted by recog- nized special organs, a few species as frog and Rhodnius appear to accelerate elimination without lag; actually there is a lag, of probable significance (0.04 hour), even in the frog's diuresis (Adolph, '36a). It is possible to guess the extended lag in mam- mals is a period in which fluid becomes preferentially distributed to tissues before much escapes from the body. The maximal response (rate) is achieved almost instantly in several species, especially in acceleration of intake. Where this is the case, time is probably not a distinct factor in the exchanges, except as the loads or other quantities change with time. Completion of recovery, just like the initiation and the maximal stage of recovery, is related to the load of water administered. One criterion of completion (table 14) is the state existing when the rate of exchanges has returned to within ± 30 per cent (see table 12) of the initial or the control rate; this defines the end of deceleration. The total time elapsed indicated how long the organ- ism is occupied in reattaining balance. The times required for measurable restitution of basal rates of exchange vary from 0.04 hour (Zoothamnium) to 70 hours (snake). Nearly the whole range is present in one species (snake) when gain is compared with loss. If the brevity of the time occupied in recov- ery is any indication of a crucial state for survival, then water deficit is the more ''dangerous" situation in all species tested. But I have no evidence that death is actually avoided by the great speed of the exchanges observed. Partial relations of the time for completion or half -completion of recovery (fig. 101) to the body size and to peculiar body struc- tures may be discerned. Thus, Arhacia egg finishes its adjust- ments rapidly, Phascolosoma, a larger animal, slowly. The circu- lation of body fluids is a correlative in transport of water, sup- posedly enabling the 15,000-gram dog to eliminate through local 178 PHYSIOLOGICAL KEGULATIONS « 8 e ?J t;j -^ ■^ o ^ iH « S H 's S i-i i« PQ < ^— V =0 H e !»■■£ s c r«i a, S ?> =« .2 1- ^_, a rt a •f-t ^ M (h ^ rt aj 03 a » "3 OS .rH t.^ |2 _ CD -w ^.tw >-*' ^w ^.yu wju ".jju ^.ou ^*ju qt uu ^**; uju uji w «fl «« te «c tp; w # o in CO t>. o o' o o o' o" o' 0000 in m cq co i— 1 Oi— lOiH o- o. o o in in in i-5 0000 0000. CJ r-5 to -^ to CO 01 CO + 1 + 1 + + + ++I + I + + ooooo 000 1 + + I + I + 02 ,a ,0 « • be ^^ 1? 0 ^ s be 0 -tJ bj ) ^ m ■ 0 00 be 0 „ •Z S a. be ^.^ 1^ cc rQ e e '«S « 0 be ^ ^ -t^.S 0 0 - rt p-^ 0 C ^ !« a, ;s 0 0 - S 0 ^ cS ci = 0 o.«J -S rS S ^-•S' a 0 0) ft ' >5 Ph w 0 OQ h Eh B^ f^ ttn KJ ^ ft; C M 1^ t-l GENERAL FEATUEES OF WATER EXCHANGES 179 organs, the kidneys, an excess even more rapidly than does the 30-grani frog. Accelerations and decelerations of exchanges of water may be compared quantitatively (table 14), in % of Bo/hour^. As men- tioned, acceleration is apparently instantaneous in some species, hence almost infinite. Deceleration appears to be slower than acceleration in all living units studied ; it is greatest in unicellular units. Hours Fig. 101. Courses of total water loads (% of Bo) during recoveries. Comparison of water tolerance curves in 6 species, 3 to 21 tests being averaged for each curve. At each time, the water load is taken as the mean difference of weight increments be- tween the loaded individuals and the control individuals that were similarly denied food but allowed water ad libitum. Dog, from figures 1 (A) and 10; rat, from figure 76; garter snake, mean from figure 81; frog, from figure 65; earthworm, from figure 88 and further data; Phascolosoma, from figure 85. For comparisons of tolerance curves, initial loads may be equated. Then the relative loads in relation to time (figs. 102 and 103) epitomize those comparisons of eliminations already men- tioned. From excesses of water, rats recover most rapidly of any vertebrates. Invertebrates without special circulatory and excre- 180 PHYSIOLOGICAL KEGULATIONS tory organs are in some cases slower than rats, but unicellular ones are faster. In recoveries, by intake, from water deficits (figs. 104 and 105) the contrasts are enormous. If in mammals absorptive rates be substituted for ingestive rates, the diversities are scarcely reduced. In aquatic invertebrates the body size is not the only correlative of rate, since Bipalium is slower than some larger species. Parenthetically, adjustments by water intake are further ana- lyzed (table 15), the data happening to concern mammals. All species are capable of extreme feats of drinking; how much they drink appears to be graded to the deficits of water imposed. Just 02 04 06 oa 10 12 Hours Fig. 102. Course of relative water load (% of initial load) after recovery began in diverse species of vertebrates that received 5 to 11% of Bo of excess water. Toad- fish, young Opsantts tau, of 6.3 gm. kept in sea water at 23° C, at zero time injected in- traperitoneally with 9.5% of Bq of distilled water; new data. Babbit, 7 individuals of 1800 gm. each given 5.7% of Bq of water by stomach; data of Abe ('31a) ; see also figure 72. Eat, 18 individuals of perhaps 250 gm. each given 5.0% of B„ of water by stomach; data of Heller and Smirk ('32a, p. 15) ; see also figure 75. Snake, eight tests on Thamnophis of 8 to 101 gm., given 10.9% of B„ of water by intraperitoneal injection; data of figure 80. Dog, three tests on one individual of 13,870 gm., given 6.1% of Bo of water by stomach; data of figure 1. as the initial rates of ingestion are directly correlated with - AW, so several of the other measurements listed (duration of draft, rate of swallowing) are found to be proportional to deficit. Among species, many relations to body size are apparent. The frequency of the gulps of which the drafts are composed is quite uniform, and the size of each gulp is closely proportioned to body weight. At one deficit the duration of initial drafts is roughly the same in all species, and hence the rate of swallowing is proportional to body weight. It may be said that the features of intake are geared to GENERAL FEATURES OF WATER EXCHANGES 181 Hour Fig. 103. Course of relative water load (% of initial load) after recovery from excess began, in diverse species of invertebrate animals. Zoothamnium put from 0.005 M NaCN into fresh water ; initial AW + 38 ; basal body volume 1.8 x 10"® ml. ; data of Kitching ('38, p. 145). Caudina put from 60 per cent sea water into normal sea water; initial AW + 15; body weight 27 gm.; 22° C; data of Koizumi ('32, p. 277). Arbacia egg put from 60 per cent sea water into normal sea water, initial AW + 52 ; average for 6 individuals, body volume 20.3x10-8 ml.; 21° C; data of Lucke et al. ('31, p. 416). Phascolosoma put from 60 per cent sea water into normal sea water, initial AW + 44 ; average for 5 individuals, AF, of body weight 1.9 gm. each; 24° C; additional data of Adolph ('36b). Fig. 104. Course of relative water load (in % of initial load) after drinking was allowed to begin following the periods of privation. Initial deficits of weight were 4 to 8% of Bo. One test in each of 5 species; Burro, Adolph and Dill ( '38) ; Dog, Adolph ('39a); Eat, new data; Man, new data; Frog, Adolph ('39b). 182 PHYSIOLOGICAL REGULATIONS * o O rh o iH 86,500 2.4 7 l-H 00 l-H in o d o QO_^ « ■ rH Ol 3 to CO * IB .a pq I— 1 o CO lO 1-1 O CO tH O O , '-s ^ -C d « CO < o u I— 1 o >o o o o JiPl5'- o (M r-l CO ; CO in TjH r-J CO COCON O OO ooj : O «0 t~ • iH i-l O CO O^ (MOJO ICOO O rH (M O OO g (M rH O «^ qq «5 O O 03 a^r^ O O CO ^ CO ^^ . CO Ph &0 o ft <0 ^ CO 00 O Ol CQ O .^o ,-^ otjh ot- oo-ioi 05ira odco o ^'o o-* i>^o t, ^-^.2 . cs •«*i O50i— ICO 1— lOi— 1 CI CO -t^- CO • p-l •^■•^ b^,'^ -^^ a -S -s -e cs c3"^gor,30 gi^o .3;■*_ ca !>; oa 05 N + d d d d d T-H d im' rS f^ in in CO Oi a> iH OJ 3 O ^ - CQ 3 O to 1 iH (M' d • 1-i r-l O ■ • = < M 1-1 -rh in oi w 3 o in 3 oa (M_ :: o <» oq 1 1-! i-!o 1— iT-H I— loo'' CO 00 CO 1-1 05 3 O 1© 00 t- 1 • ■ ■ ■ ■ 1-1 d r^ d o o ■ (M Tt< in CO 1—1 O to 00 00 w 1 '■ i-i o o o o ■* : I : • ; o ; TjH O CO oa (M O ■* 00 O Od 1 ■ d ■ 1-i O O r-i O OJ s - C^ S-i^tHt-tH^-^H '+J So>^oooooooooo =4-1 o o P4 • rH C3 -4-3 tH Cj C^ -r-l C5 .rH .-H -M t-* -i-I pi^EH0Q^p4P^pHpHf^pHP^PHF^ OQ i <: 5 _<» g '3 00 O ft ?^ S S « « § CQ ■U ' ' ' 't o .^ 1 "iH ^ \^ Earthworrn____L;_ ■ — - J2x- -V '~l2]2!III;:=— _ G - N. \ Goin *- -v^ N. Phascolosomq,.'''' ,.---'' Loss "^v, \^ ^.■''' Garter SnaJ o S &1 a - a o t- ^ o o o t^ 00 O^fOt-XrHxX O 00 lO ■<^ ko oq CO ■>* 05 1— I l>. co' CO 10 CO ^' 06 O O05t^ t-ot^ -TtHr^ «>O«0l0l0l>-C0-#0i(MC0l0?DTj.■ Tji Tt? CO 06 o OiTtHMcOrtHOOlOO l"-; l-^ cc O 00 CO 00 O CVI i-H rH (m' r-I T^ CO i-i iHOWOirHaOlOWTjHt^OOSOqcOO N o 00 CO ci to' o o TJ^■ t^ 00' iH ; oi ooooi-oDcoooi-i-^oo5oSo§ot:S a o r;* 3 TO SI o IOO]COCOt^OOa5a5«Or-l,-HTtHO "J ■* ■* rH r-l 1— I Sh ^h ;h +i -(J +s ■*^■^^ua]^^!nu^^!^ 02 « <3 SI, ft ^ s 2^-S 5; T- be fto s S s ^ s ^ -P 00 ^ -• ?5 '*' -a § -s -2 .H s i .§ 2 g « e s I 198 PHYSIOLOGICAL REGULATIONS o «Q e s e-2 ft^ s =y •'^ -2 S O CS C «*H -rH 'O c3 ^ O f> a> ra oo o j3 ci rl 03 \^ .A >-> S^r!^ t^ §5^, OJ g PI c3 p< S-. CD -*^ ^^ ^ CO f2 rii -IJ Pnr-] £0 "^ S •" — ' tg ra ti ^ "£^ ^ . . 'C OJ fO O TT : K c —I o 2 «! 03 'O oi o Oft =5 . S &E O j f^S CO N '-^ .■S CD Sh rt .S MpqWpqfi OJiHOCIOi— l-^OqcOOlNlM COCOt^ 50iHTtH0505i-lTj(COTlr Iff ■*" o" t}h'~ IM t^ TJH OO t- TjH lO t^ lO ^O !>; O lO O 50 t-- cd ■*' r-I t^ CO CO eg o 00 Oa CO o O Tl? 05 O o o o o o o o o o o o o o o o o o o o o o o oooooooo lO ICi >0 Tfi CO lO eg w i-j lo ^_ CO o o o o o o" OOOOOOOr-l i3 rQ -1-S +i o ;• Jii c3 03 c3 be W pq Ph fl 2 « '2 r^ ^ -^ O '4i 00 gl, 6C 5 O S GENERAL FEATURES OF WATER EXCHANGES 199 .-H *: Tj( £0 , ^ fl _ S o S -^ _^S ^^^ 0000000000000000000 rHTjJo'*rw<^ioooooooooo<>ioo'or-ici'<*^coco «D00CO(Mi-Hi— l -oi? ^ 2 ^ S B ?^.2; •S 5 f^ "oo 200 PHYSIOLOGICAL REGULATIONS gain of water equals its loss, such a state can be any one of a vari- ety. For instance, the cow may be lactating or not (Atkeson and Warren, '34) ; the horse may be working or not (Zuntz and Hage- mann, 1898) ; the man may be on a high-protein diet or not (Vozarik, '06). But none of these states makes the mammal's turnover equal to the frog's; only forced drinking does that. The rates of turnover per unit of body weight are greater in aquatic species than in terrestrial ones; in smaller than in larger kinds ; in freshwater than in marine organisms. Among inverte- brates, rates of water turnover are strictly known only for some arthropods, protozoa, and the earthworm. Proportional to turn- over, but regularly lower, are the rates of loss in weight in terres- trial organisms recently denied access to water. These rates of loss may be substituted for rates of turnover, being considered subminimal values of turnover. In that way values for numerous further species, especially of insects (Buxton, '37a; Gunn, '37), become available. Extremely low rates of turnover are found, as 0.03% of Bo/hour in mealworm larva, and 0.01%. of Bo/hour in the tortoise (Testudo) of Benedict ('32), rivalling MacDougal's ('12) cactus that lost 0.0004% of Bo/hour. Granted that the losses are evaporative, it is possible to compare the protections against evap- oration enjoyed by these species, and per unit of believed surface area. Then the mealworm larva becomes equal to the tortoise, losing about 0.0003 grams/square centimeter and hour. These particular species have the lowest known requirement for water. A deduction from the tenets of physical science would be that any turnover of water occurs only with the transformation of some energy. Otherwise water would move through a circuitous path (through the organism) without fall of potential. Though gain might be passive, loss would require work, or vice versa. Only species without turnover may be suspected of degrading no energy for water exchanges. Yet many have been the attempts to believe that all the processes concerned in water metabolism occur with- out energy exchange. It is sometimes forgotten that an oso- mometer also transforms energy in approaching equilibrium by transport of water. Rates of water turnover (in water balance) may be partitioned among diverse paths (tables 20, 10, and 11). GENERAL FEATURES OF WATER EXCHANGES 201 § 75. Tolerated loads What ranges of water contents are compatible with recovery? The equilibration diagrams are constructed for limited increments of water, but most of them presumably could be extended to larger deficits and excesses. For some species the limits of water content that each survives have been measured (table 22). Perhaps the TABLE 22 Tolerated water loads in diverse species, in % of Bo Species Dog ( c ( ( 1 1 Man Rabbit i t Rat Cat Guinea pig Mouse Vole Deer mouse Chicken Pigeon ThamnopMs Anolis Sceloporus Phrynosoma Chrysemys Bana pipiens < < Bufo Amblystoma Plethodon Triton Salamandra Limax Blatta Tenebrio larva Popillia larva Melanoplus egg Placobdella Lumbricus Allolobophora Phascolosoma Arbacia egg Bipalium Zoothamnium Ameba Muscle of Frog Skin fragment of Frog Heart " " Heart " of Chick Deficit Excess Source of data -20 Falck and Scheffer (1854a) -24 Pernice and Scagliosi (1895) + 21 Falck (1872), by vein + 19 Harding and Harris ( '30) >-10 Dennig (1899) + 9 Helwig et al. ('35, '38) -32 + 39 Pack ('23), Misawa ('27) -36 New data -31 Czeruy (1894) + 24 Rowntree ('26) + 20 < ( -24 Hall ('22) -32 < t -31 ( ( -41 Pernice and Scagliosi -44 Schuchardt (1847) -46 Falck and Scheffer >-31 > + 18 New data, fig. 80 -46 Hall -48 ( i -34 e I -33 < < -41 1 1 >+35 Adolph ( '39b) -59 Rey ('37) -47 Hall -43 Caldwell ('25) -51 Rey < i -76 Kunkel ( '16) -42 Gunn ( '33) -52 Hall -40 Ludwig ( '36) -33 Thompson and Bodine ( '36) -70 Hall >-49 > + 24 Wolf ( '40a) and new data -60 Schmidt ('18) -69 Hall -31 + 98 Adolph, fig. 87 -35 + 155 Lucke, fig. 91 -46 Kawaguti ('32) >-23 >+39 Kitching ('38) -48 Mast and Fowler ('38) + 50 Howland and Pollack ( '27) -35 Wolf ('40b) -43 Morosow ('31) -70 (< -75 < ( 202 PHYSIOLOGICAL REGULATIONS most questionable data are those in water privation ; several days, even months, may be occupied to induce the deficit, and food is often refused by desiccated mammals and others, leading to greater deficit of weight but relatively less deficit of water than would occur from lack of water alone. Diverse increments of water are tolerated by the various species. More factors than the proportion of water to other sub- stances in the body are at stake. Time is one of these ; with slow desiccation (-40% of Bo in 120 hours) frogs withstood much more loss than with rapid desiccation (- 12% of Bq in 0.7 hour) (Almeida, '26). A thorough study of the numerous factors modifying the survival after deficits and excesses of water would be an extensive investigation. It is noteworthy that dogs are reported to survive the physiologically ''crude" procedure of injecting water intra- venously almost as well as (repeatedly) ingesting water (Falck, 1872 ; Rowntree, '23; Chiray et al, '38). Tolerated loads might alternatively mean those increments limited by some event other than survival and recovery. Toler- ances according to the criterion of vomiting, or convulsions, or anorexia, or anatomical lesions, or continuance of diverse func- tions, are of equal interest though automatically restricted to cer- tain phyla. The difference between tolerance against loading and tolerance against survival is illustrated by the fact that dogs with extreme contents of water may become unable to excrete it rapidly (Harding and Harris, '30) ; the excretory processes are tempo- rarily depressed, yet after a time the processes may recover and the dogs survive. The prevention of loads by various means is a part of the regu- lation of water content to be considered here. A dog often resists the establishment of large excesses by vomiting ; that resistance is of great moment in avoiding loads that might not otherwise be tol- erated. And the dog resists the occurrence of large deficits by having a body surface that minimizes evaporation. In both cases, prevention accomplishes the same sort of stabilization of water content that recovery (compensation) accomplishes; in both the rates of exchange are the tangible evidences of processes concerned. § 76. Summary Beyond this chapter the formal study of water in animals will no longer be limited to the four sorts of variables and their quan- GENERAL FEATURES OF WATER EXCHANGES 203 titative relations. Therefore the four may at this point be charac- terized briefly. They are : (1) Time (t). Events are followed, beginning either with the time at which water balance is first disturbed or the time at which recovery is permitted. Completion of recovery is marked by restoration of water balance, in which state the rates of gain of water equal the rates of loss of it. Time intervals (At) are diverse periods of time selected for study, differing not merely in clock units, but also in physiological significance. Initial periods of recovery are particularly distinguished from indifferent (steady) periods and from periods of maximal rates. (2) Velocity quotients (1/At) are the reciprocals of time inter- vals. They characterize processes of water exchange, being ob- tained either by dividing rates of exchange by load, or from the analysis of exponential curves of load during recovery. (3) Water increments (AW). Excesses and deficits of water contained in living unit are measured in diverse ways, often as increments of body weight. Definitions of control content are set up, both in relation to water exchanges and in relation to non- aqueous materials of the body. Diverse means of establishing increments are distinguished. Extreme increments are barely tolerated loads of water, as judged by survival or other criterion. Only moderate increments prevail in ordinary fluctuations of con- tent measured at successive equal intervals of time. (4) Water exchanges (SW/At). Gains and losses of water are total, partitional, or net. The economy quotient is the ratio of gains to losses, departing extremely from 1 whenever water bal- ance is lost. The ratios of maximal to minimal rates indicate the modifiability of total exchanges and of their separable paths. Equilibration diagrams, by comparing rates of exchanges with water contents, indicate the relative responses to diverse displace- ments from water balance. At balance the turnover rates are characteristic for the species and its physiological state. During recoveries of water content, rates of exchange are traced in respect to time. For several species and several types of disturbance of water content, numerical comparisons, in each of the diverse dimensions, indicate the activities concerned in recovery. Time intervals (At) and the corresponding velocity quotients 204 PHYSIOLOGICAL KEGULATIONS may be classed as follows. Each is measured by clock in relation to physiological events. (A) Initial interval after increment is imposed, or after recovery begins. It may be the first one hour (Atj), or the interval for half the load to be dissipated (Atn). (B) Uniform interval, often in a stationary state (Atg). (C) Interval containing maximal rate of exchange (Ati,). (D) Interval including complete recovery (AtR). (E) Instantaneous intervals along the time axis (AtM). Types of increment of water content (AW) may be grouped. Each is measured on one species by one method, such as body weight, liquid volume, ratio water/dry weight. (A) Net deficits, (a) By water privation, (b) By hypertonic solute, (c) By previous exosmosis. Etc. (B) Net excesses. Each may follow single or repeated admin- istrations of water (a) By mouth or stomach, (b) By injection : by vein, under skin, into peritoneal cavity, (c) By previous en- dosmosis. Etc. (C) Net random changes, (a) Control conditions, (b) Un- usual temperatures, (c) Unusual physical exercise. Etc. (D) Gross exchanges, ± SW per unit At. Each may or may not be corrected for control rates. According to paths as follows : (a) Gain: alimentary, metabolic, cutaneous, (b) Loss: urinary, vacuolar, fecal, evaporative, (c) Both: osmotic. The bases for these classifications are, as in all classifications, those of dimensions and of utility in thought. While AW and At may be varied in infinite types and intervals, only two algebraic signs (deficit, excess) prevail in AW and only one sign in At. The varieties (types) of AW and of At are not infinite in number but correspond to procedures and conditions for observations and measurements. In other words, types of load and classes of time intervals are treated as discontinuous variables, leaving the four continuous variables to be graded according to their numerical values. Each category could, of course, be subdivided further or in a different fashion. Thus, with respect to urinary loss, one nephron could be studied ; with respect to metabolic gain, the water formed from one precursor or in one organ could be measured. In general, while the four sorts of variables are distinguished by their physical dimensions, the subdivisions depend on physical and GENEEAL FEATURES OF WATER EXCHANGES 205 chemical conditions, physiological states, anatomical units, and physiological outcomes. The number of varieties of water load explicitly mentioned in the dog alone is about 15 (chapter III). The number of varieties of At considered is about 12, even though an arbitrary man-made clock is used in obtaining them all. Among 15 kinds of AW with 12 kinds of At, a product of 180 indicates the combinations that might be studied if I were concerned merely in rearranging coordinates. Uniformities among the water relations studied are features common to all the organisms investigated, while quantities limited to alimentary tracts or to osmosis do not apply in organisms devoid of them. In this way the account comes to consist of the materials of general physiology, and only further inquiry will show just how general the present conclusions are. Some of the recognizable features are : (1) The species studied lose water faster when water content is high and gain water faster when content is low. {2) Rates of gain and of loss are equal at one water content (balance), to which the organism recurs after each disturbance and at which alone exchanges rest. (5) Rates of water exchange are correlatives of water content. They are not always functions of the time elapsed since displacement or since recovery began. (4) Incre- ments of water content exceeding 10 per cent of the water already present are tolerated by all species tested. Certain other generali- zations are mentioned above in § 72. Chapter X SOME OTHER CORRELATIVES OF WATER CONTENT (IN DOG) § 77. The maintenance of water content has now been examined in relation to water exchanges. Ordinarily a gradual turnover of the body's water is occurring; extraordinarily one exchange (in deficit the gain) augments while the other exchange (in deficit the loss) often diminishes. Where separate exchanges are not ob- served, net exchanges are still of the same character as where simultaneous gain and loss were separately examined ; from which I infer that the pattern is the same in both cases but is incompletely visible in one. Perhaps the limitation of the study to water exchanges nar- rowed the view as much as it intensified it. Are exchanges the only processes that vary with water content? They seemed to explain how content is maintained and restored; but many other variables may or may not be equally concerned in water regulation. The possibilities can be tested only by broadening the horizon. Beyond the intensive investigation of the relations among the four sorts of variables within the water-time system, the inquiry might be pushed out in any of a large number of directions. A few topics that will be followed are : (1) Diverse correlatives of water content or load (chapters X-XII) ; (2) Equilibration and variability of diverse components (chap- ters XIV-XV) ; (5) Interrelations among components in balances and equili- brations (chapter XVII). These directions depend upon arbitrary choices, and there is at present no evidence that thorough study of any other combinations of variables would be less fruitful. The purposes in mind are to see how precisely water balance and its regulation can be defined ; and then to compare many regulations with that of water in order to find how general the features of equilibration, regulatory be- havior, and variability may be. Water content, which has been scrutinized particularly in rela- tion to water exchanges and to time, is equally related to many 206 COERELATIVES OF WATER CONTENT 207 other types of variables. How these many correlatives vary to- gether will be shown by a study in which ± AW (water load) is kept as one variable throughout, letting one after another of the many physiological quantities that are measured in relation to it, come under review. Each of them is a part of the complex that goes with water load. The study will be designed to show whether anything that changes with water load inside the dog furnishes a key to understanding what prompts recovery in the body as a whole. Ideally, the water load might be of one type in all divisions of the study. Experiments would be set up with dogs deprived of water as defined in § 13 and with dogs administered water by stomach as defined in § 12. Then, within the limits of the station- ary state of water load, diverse analyses and measurements would be made. No such ideal program has been carried through; and not from the difficulty of planning it with singleness of purpose, but from the realization that a strictly steady state with respect to all measurable components {e.g., chloride content of body, heat content of body) does not exist. Practical difiiculties of other sorts creep in, and for the present it seems sufficient to indicate, by means of the partial data from diverse sources that are available, what relations are known to exist. First, within one species (dog) several types of changes in water content will be considered. Those correlatives that are peculiar to one type will be detected, and hasty generalizations as to what characterizes all water loads may be avoided. § 78. Volumes of parts It is not hard to suppose that endless changes occur in diverse parts of a dog with each load of water. Point is given to the search for those changes as soon as it is asked: Where is the excess water deposited, or whence is the deficit withdrawn? Are there depots for water; what does a reserve of water look like? And, are some tissues especially sensitive to water loads'? Evidently volumes of tissues have to be measured, and not in relation to the tissue's own exchanges (as in §58) but in relation to the whole body's water content. At least three general procedures are suitable for measure- ment of increments of volume in any one portion of the dog's body. {a) The difference of net weights or volumes of parts secured at 208 PHYSIOLOGICAL REGULATIONS autopsy (table 23, column 2) from individuals subjected to water loads and from those not loaded. If F is the fresh weight (wet weight) of the part, then 100(Fi -Fo)/Fo = AV is the change of volume. But the weights of any one organ relative to the control body weights among different autopsied individuals vary (Engels, '04; Stewart, '21 ; Sato, '30) to a degree that makes many compari- sons of organ sizes under diverse water loads of little statistical significance, (b) In one and the same individual, some organ sizes may be mechanically compared before and after water is adminis- TABLE 23 Volume (A.V) and dilution (A-E) in various tissues of growing dogs when water-loaded by consuming diets deficient in water. Values in parentheses are clearly insig- nificant (P > 0.05) ; some others may he also Tissue Falck and Scheffer (1854b) Schiff et al. ('25) Garo- feano et al. ('25) Hamilton and Schwartz ( '35) AV A/D A/D A/D A/D A/Cl A/Na A/K A/base Number of analyses Whole body.. Blood 1 -21 + 8 -20 -19 + 7 -15 + 21 + 4 -27 -20 — 7 1 -19 -19 _11 -24 -11 - 2 + 0 - 2 -12 - 2 - 9 2 -26(?) 1 -18 -15 -12 -29 -13 6 -20(?) 6 6 -33 -22 -37 6 6 -19 (+12) -24 - 9 Muscle Rkin -31 -41 -54 -32 (-27) -19 -32 Skeleton Liver Brain Kidney -11 -14 - 8 -16 -15 -11 -34 -24 -26 (-14) -11 ± 0 (-19) -15 (- 7) Thyroid Hypophysis ... Ovary trated to the whole body, such as spleen (Barcroft et al., '25) and liver (Reid, '29). Ordinarily anesthesia and other conditions are introduced that need to be specified. This procedure has not yet contributed data in those types of water load in which other cor- relatives were studied in the present investigation, (c) Volumes of distribution may be ascertained repeatedly in one individual. Of particular interest is the congruence of increment in volume of distribution (AVd) with the load of water in the whole body (AW). Among individuals without water load, two measured volumes (fig. 113) appear to be correlated not only with body size, but rela- tive to body size with one another. As measured, the ''plasma" COKRELATIVES OF WATER CONTENT 209 Thiocyanate Volume Fig. 113. Volume of distribution of the dye T 1824 ("plasma" volume) (% of body weight) in relation to volume of distribution of thiocyanate ("extracellular" volume) (% of body weight). Dogs under control conditions; 15 individuals, averages of 73 measurements. Open circles, males; solid points, females, of which 4 are repre- sented by 2 points each. Data of Gregersen and Stewart ( '39). (T 1824) volume increases with the "extracellular" (thiocyanate) volume. Male individuals have consistently higher ratios than females. -15 -10 -5 0 +5 +10 +15 Total Water Load Fig. 114. Increment in volume of distribution (% of Vo) in relation to mean total water load (% of Bo). Dog. Line QQ' = load would be equal in partial volume and in whole body. SCN -yoZ. = thiocyanate ("extracellular fluid") volume in 14 individuals deprived of water and food; data of Gregersen and Painter ('39). T 1824 vol. = dje ("plasma") volume in same 14. V B i;oZ = brilliant vital red ("plasma") volume in 10 individuals repeatedly given water by stomach for 8 hours; data of Greene and Eown- tree('27). Hb vol. = dissolved hemoglobin ("plasma") volume in 5 individuals given water by stomach during 1.5 hours; data of Lee, Carrier and Whipple ('22). CO vol. = carbon monoxide ("erythrocyte") volume in same 5 tests. Hb+CO roL = hemoglobin plus CO ("circulating blood") volume in same 5 tests. 210 PHYSIOLOGICAL EEGULATIONS In dogs subjected to either positive or negative water loads the reputed plasma volumes (Hb, VR, T1824; fig. 114) change by larger fractions than the body as a whole. Unfortunately the water load is not too accurately known in any of the animals represented ; but there are additional data (in water deficits) to confirm the con- clusion that presumed plasma decreases in volume by more than its proportional share (Korth and Marx, '28; Keith, '22). The volume of "erythrocytes" in circulation (CO), as measured inde- pendently by carbon monoxide distribution, does not change ap- preciably with water excess ; leaving the ''combined blood" volume (Hb and CO) with less load than the whole body. The "extra- cellular" volume (SCN) in deficit decreases somewhat more than the average water load (AW). A favorite prediction would be that all volumes of distribution increase and decrease in proportion to the body load of water. So far no one volume is found exactly to fit this notion ; the data are not extensive enough to decide for others than those here shown and for Vd of bromide (Brodie et al., '39). There seems to be no doubt that most of the tissues whose incre- ments of volume have been thus crudely measured, contain extra water when the whole body has an excess, and are depleted of water when the whole is in deficit. The modifications of water content in some tissues are numerically equal to load of the whole, but in other tissues are distinctly greater or less. Plainly the body's load of water is unequally distributed through its parts. § 79. Water contents (dilutions) of parts To avoid possible confusions, I separate the results based upon volumes measured in a manner capable of determining absolute quantities, from those in which changes of concentration are ascer- tained. For these particular tests, excesses of water are produced within 2 or 3 hours, a time too short to allow many chemical con- tents of the body as a whole to change. The data are all obtained between meals, and diet need not then enter the picture. On the other hand, deficits of water are produced by partial privations last- ing many days, as illustrated in the data of Falck and Scheffer (1854b). Two young dogs of the same litter were analyzed (table 23, column 3), one at 76 days of age, the other at 104 days of which the intervening 28 days were spent on a diet of dry "Zwiebach." CORRELATIVES OF WATER CONTENT 211 Hence factors of growth in size, diet, time, age and constitution entered into the differences of the two individuals. Each dog was killed, its organs were excised and weighed, and the proportion of water in each organ was ascertained by a particular analytical procedure of drying. The changes of body weight and therefore of total water load (AW = -21) between "control" and "desic- cated" are known; what is the change in each portion of the body? First it is necessary to decide how the analyzed water contents of the same tissue in each of the two individuals shall be compared. The data known are: wet weight (F), dry weight (D) and water weight (E = F-D). The ratio E/F is tabulated by Falck and Scheffer ; but F is itself changing with water privation. The known fraction that is believed to change little or none is D ; or, compara- ble samples of tissues are those in which Di = Do. Hence the ratio E/D is the basis of comparisons ; in order to make its increments comparable to increments of body water (AW) as first defined, it may be corrected to fresh weight and % of Fo by multiplying by 100 Eo/Fo. 100Eo/E,_Eo\/Eo_ /B\Do \_ Fo U Wo)iw,-^^^[F,i)rV-^^ This is the amount of water added to or lost from 100 units of original weight. The increment is all water, however much non- aqueous weight may be included in Fo. The ratio AE has the same dimensions as AW and AV, for in both cases the water lost or gained is in per cent of the unit analyzed. Since AE may be of many types, the type of measurement from which AE is derived is more specifically designated by the abbrevi- ation A/D, or dilution of dry substance. Both measures, AV (== AW in this case) and AE, can be compared for the whole animals of Falck and Scheffer (1854b) : AV = AF/Fo = -^^^^^^ = -20.78%ofFo(orofBointhiscase) 3178.72 gm. / u v u / AE = A/D = -2p^^ = - 17.39% of Fo (or of Bo in this case) The comparison indicates that the dog deprived of water lost non- volatile materials too, and the two measures of water loss differ (by 18 per cent of their mean). A similar comparison may be made for each tissue mass as weighed and as analyzed (table 23, columns 2 and 3). Measurements of dilution are not limited to analyses by volatili- 212 PHYSIOLOGICAL EEGULATIONS zation of water. Any ratio between a changing fraction and a more fixed fraction, is suitable, and is a type of the general class AE. Thus, instead of total solids (D), the content of chloride (CI) or of nitrogen (N) or of protein (Pr) may be substituted. The non- protein fraction need not be ascertained as such, since it is always equal to (F - Pr). Hence, if Pr is the per cent of protein by weight, AE = A/Pr = 100 - Pr. /100^_ 100^) /lOO^ ^ ^^ /Pr, _ ,\ V Pri Pro // Pro \Pri / A non-chemical measure may equally serve, such as excess specific gravity over that of water, excess refractive index, or increment in depth of color. Then refractometer reading of tissue minus re- f ractometer reading of water = RI, a measure of the concentration of non-aqueous constituents, and From the data of table 23, it is concluded that in water deficit diverse tissues lose water to different extents. While the whole body loses 18 to 20 per cent, the skin, hypophysis, and sometimes muscles, lose more. Most other tissues lose less. The four investi- gations of tissue dilutions of dry substance (A/D) that are com- pared, do not show close quantitative agreement. In one investi- gation (columns 6 to 10) five measurements of dilution were made upon each tissue sanaple. These also show little agreement, indi- cating that electrolytes as well as water are redistributed. In each of the latter tests, 4 to 13 days elapsed in water privation ; the body weights were not reported in detail, leaving no adequate criterion of body load of water. Also in water excess, produced in morphinized dogs by infusion of 0.10 M solution of sodium chloride by vein, both relative and absolute increments of water were studied in various tissues (Engels, '04) (see table 29, column 2, § 91). The tissues modified more than the average for the whole body are the same as those modified in water deficits. Often it is supposed that some tissues act as depots, from which water is mobilized in deficits, to which water is brought in excesses (Magnus, '00). How shall a depot be defined? Evidently it has to do with storage of water. Shall I look for a tank (as the camel is said to have), that fills up whenever water content of the body exceeds a certain value and that empties at some lower values (lines MM' or NN' in fig. 115) ? If I do, I find none in the dog. Or, COREELATIVES OF "WATER CONTENT 213 perhaps the gates of the depot do not open quite so suddenly and completely (line 00'). Again, there is none of that sort for water. Maybe any tissue that takes up more water than the average for the body (line PP') will be counted as a store. Of that there is evidence, as in the volumes of distribution of SON, T 1824, and VR in figure 114, and in the tissue dilutions of tables 23 and 29, but no one part of the body seems to stand forth as a distinct store. All tissues share the water load in diverse degrees. If AV > AW de- fines a store, is the tissue whose AV < AW an anti-store"? It seems to me that for water, at least, the notion of depots is not substanti- ated sufficiently to use the term ; the increments found in each tissue correspond rather to equilibria of partition. 1- o ^»5 ' ?7 A / y H/Q' — , IM' -a a -J J J-IO 0/ Mi // / k 1 / -10 -5 0 +5 +10 Water Load (Body), AW Fig. 115. Diagram of possible relations of water load in a tissue to water load in body. Various lines are described in the text. In brief, methods are available for ascertaining by how much a given tissue exceeds in water increment the water increment of the whole body. In this way the distribution of water is found to be uneven. Every analyzable unit and every volume of distribu- tion is a compartment, but no very large ratios of partition appear between the units. § 80. Dilutions of blood and plasma The circulating blood may play a peculiar role in recoveries of water content in at least two ways. It is undoubtedly the chief vehicle of redistribution and equilibration among tissues. Is it also a messenger whereby kidneys and other tissues become aware that a load of body water exists % Blood and plasma are analyzed in dogs loaded with excess water 214 PHYSIOLOGICAL EEGULATIONS Hours Fig. 116. Increment in dilution (% of initial) of plasma and blood, and total water load of body (% of Bo), in relation to time after a single ingestion (6.2% of Bo) of water by stomach. Dog. Means of three tests, 2 individuals. Blood was heparinized as drawn: 1/T, increment in ratio of non-erythrocyte volume to erythrocyte volume (hematocrit) ; 1/Hb, increment in reciprocal of hemoglobin concentration (colorimetry of CO-Hb) ; 1/Db, increment in ratio of water to dry residue in whole blood. Plasma separated from this blood: 1/Dp, increment in ratio of water to dry residue; 1/EIp, increment in reciprocal of refractive index of plasma minus refractive index of water at 17.5° C. ; 1/Clp, increment in reciprocal of chloride concentration of plasma. New data of Adolph and Kingsley. by stomacli (figs. 116 and 117). Within the first hour after sudden administration, dilutions are not parallel to body load ; thereafter they are. After 1 hour, plasma dilution (fig. 118) is related to body load of water independently of time, and also independently (within the range of variability prevailing) of whether single or repeated ad- ministrations of water be used. The latter is remarkable in view +15 -P c 0) D09 ^>,:.-'-'-'iycip ^\ Loqc I/Dp-- l/Hb Hours Fig. 117. Increment in dilution of plasma and blood, and total water load of body (% of initial state), in relation to time after giving the first of 10 portions of water at 0.25-hour intervals by stomach (10% of Bo altogether). Dog. Six tests, 3 individuals. Measures of dilution are the same in figure 116. Data of Adolph and Kingsley. CORRELATIVES OF WATER CONTENT 215 +CU ' ' ■ ■■ 1 • +• D09 ii+16 - • 1- • / OJ / a_ / J.^12 - • 'A 1— t / ' • ^ » / / 0 ••*■ — -* •. Q. / 0 c^8 - * * • / o * • / 0 3 '' / 0 °M •. •'/ • • a - o 0, /• - o ^ • ^ 0^ \ ~ 0 Tote Water +6 Load Fig. 118. Plasma dilution, measured as % increment in 1/EI as previously defined, in relation to total water load (% of Bo). Dog. Individual C solid points, individual G' open points; 50 observations in 10 tests. All samples were taken at more than 1.0 hour after first administration of water. (The correlation coefficient is +0.66.) The regression line as drawn indicates that the plasma was diluted 2.31 times as much as the body as a whole. Data of Adolph and Kingsley. of the somewhat greater rates of urine output that follow single ingestions (figs. 7 and 27). In a similar manner, each of the dilutions measured in samples of blood is correlated with total water load (table 24). The in- tensity of correlation is indicated by the coefficient r, while the mean relationship is indicated by a proportionality, or slope of the regression line, fitted by least squares or other approximate method. Fig. 119. Increment in dilution of plasma or of blood, and total water load of body (% of Bo), in relation to time. Dog in control tests without water load. Means of 4 tests on 2 individuals. Measures of dilution as in figure 116. Data of Adolph and Kingsley. 216 PHYSIOLOGICAL EEGULATIONS Both the correlation coefficients and the control samples taken (fig. 119) agree upon the fact that some of the measures of dilution vary in partial independence of water load. These are the mea- sures upon whole blood (A/T, A/Hb), and plasma chloride (A/Clp). Other measures in plasma (A/Dp, A/RIp) vary no more than (C.V.) ± 3 per cent when body load does not change (fig. 119), and show correlation coefficients greater than + 0.5 with modified body loads (table 24). At 1.0-hour intervals the coefficient of dif- ference (CA) is ± 1.3 for A/RIp and ± 2.5 for A/Dp. These two are evidently the types of dilution upon which to rely for accurate reflections of body water load. Dog in water excess. TABLE 24 f individuals, 11 tests, 51 observations, and Kingsley {'40) New data of Adolph Covariate, AE C.V. of 21 con- trols at 1-hr. intervals Mean of 51 incre- ments per cent of mean control r with water load AW Regres- sion, AE/AW r with plasma dilution A/RI Regres- sion, AE A/RI Total water load, AW - 0.39 + 2.3 + 2.8 + 3.8 + 2.2 + 4.8 + 5.6 + 31.0 -^26.0 + 3.14 -1- 7.8 + 7.1 + 5.0 + 6.2 -f 11.9 + 9.6 + 507.0 -MOIO.O 2.31 2.39 2.54 2.05 127^0 295.0 + 0.66 0.43 Plasma dilution, A/RIp + 0.66 + 0.58 + 0.48 + 0.45 (-0.10) (-0.05) + 0.72 + 0.66 Plasma dilution, A/Dp Plasma dilution, A/Clp Blood dilution, A/Db • Blood dilution, A/Hb Blood dilution, A/T Rate of total loss Rate of urinary loss + 0.94 + 0.61 + 0.83 + 0.22 + 0.45 + 0.42 + 0.48 1.03 0.83 0.81 1.11 0.89 51.0 117.0 As in water excess, so also in water deficit the relation of total water load to various blood dilutions is ascertained (figs. 120 and 121). During 3 days of partial water privation the plasma dilu- tion decreases steadily and in parallel to water load (weight) of the whole body. Upon sudden ending of the privation, water is drunk ad libitum (fig. 120), whereupon the amount ingested exceeds the amount lost. When drinking is limited to the amount of weight deficit (fig. 121) the changes of plasma dilution are again in marked contrast to those of fig. 116. {a) At 0.2 hour a slight concentration of plasma regularly occurs, (h) Dilution proceeds much more slowly than without previous deficit, (c) Even at 1.0 hour dilution is scarcely maximal, (d) The plasma dilution is greater after pri- vation than without it, though the same amounts of water are ingested. Correlations, within these tests of privation and recovery, be- COREELATIVES OF WATER CONTENT 217 -3-2-10 I a Days Hour^ Fig. 120. Increment in dilution of plasma and blood, and total water load of body, in relation to time of water privation (3 days), followed by sudden voluntary ingestion of water amounting to about 6% of Bo. Four tests, 2 individuals. Measures of dilution as in figure 116. Data of Adolph and Kingsley. tween blood dilutions and water load, are less intense (fig. 122). Temporal sequences become important elements, for several days are represented in any one test instead of several hours. 3 2 10 I Days Hours Fig. 121. Increment in dilution of plasma and blood (% of initial), and total water load of body (% of Bo), in relation to time during water privation (3 days), followed by voluntary water ingestion (0.2 hour) of about 3% of Bo of water. Two tests, one individual. Measures of dilution as in figure 116. Data of Adolph and Kingsley. 218 PHYSIOLOGICAL KEGULATIONS The same data might be used to correlate rates of change of plasma dilution during recovery with the dilutions prevailing. Each dilution has the dimensions of relative volume, so that the correlation obtained is a net equilibration diagram. This would resemble, in another type of water load, the relations described in §59. That some measures of blood concentration may be converted by computation into others by means of equations or equivalents is generally recognized. A familiar one is the transformation of re- fractive indices and specific gravities of serum or plasma into pro- or 0 -5- 3 Q _ E _o CL 10- ■8 1 ■ ... J , - Dog \ ^ 6d '^ 1 1 ' • — • • • 1 • • • 1 . - -6 -4 -2 0 Total Water Load +2 ■^4 Fig. 122. Increment in plasma dilution, measured as increment in 1/RI (% of initial), in relation to total water load (% of Bo). Two dogs. The points of test 6d are connected; 6 tests in which most water was withdrawn from the diet for 2 or 3 days; then water was suddenly drunk and the blood was sampled repeatedly. Samples taken within one hour of drinking are not included in the correlation ; the correlation coefficient of the remainder is + 0.68. Combining this figure with the data of figure 118, the com- bined correlation coefficient is + 0.87. Data of Adolph and Kingsley. tein concentrations of serum or plasma (Neuhausen and Rioch, '23 ; Weech et al., '35). In general the familiar instances of conversion are those in which chemical entities are largely known. It may be recognized just as readily that any one measure of concentration is equivalent to (accompanies) another, providing their coincidence is measured under and limited to prescribed physiological condi- tions. A case is worked out in figure 123 ; once such a correlation and regression are known, it is quite superfluous for most purposes to measure the same plasma in both ways. The regression line there drawn does not pass through the origin. From this it may CORRELATIVES OF WATER CONTENT 219 be inferred that water loads do not merely dilute the solids of plasma in a lump, the increment of refractive index per unit of dry weight differing slightly for various solutes. Rather, diverse con- stituents are diluted unequally, as is to be expected from the fact that their volumes of distribution are not all alike. Again, at diverse + AW, blood chloride dilutions (fig. 124) and serum electrical conductivities (1/ECs, fig. 125) are measured, and their relationships are ascertained graphically or by equation; usually AE = cAW. Such relationships are often termed ' ' empiri- 44 1 1 r 1 1 i \ r '■ ■ 1 1 r - ■ ' 1 D09 Plasma °/ -1.349 V H-4I- •. • / o • "~/ cr o A c ■XI o -1348 i ^yi - a> CL e • / D: y»* Q^38- ^ J A' ' 1 a e to -1347 1 oj^-5° -■ a -J •0 o*o J?8^ .^ Q. > °%Y°o, ^35- F - 4-^ -1346 2 ■+- ^0 )^"°° - xs (D V' a cr /a Q) / Q: o / 0 q:32- . A < r- / /' II -1345 y ~ 2 n / 1 1 1 1 L 1 ._! 1 1 "0 .064 .070 .076 .082 .088 Plasma Concentration = Dry Weight /Wet Weiqbi Fig. 123. Eefractive index of plasma (RI) in relation to fraction of solids in plasma (D/F). Heparinized plasma of dog. Two individuals (C, open points; G', solid points) in 21 tests; given water by stomach, or deprived of water with constant diet, or neither. The ordinates on the outer RI scale, readings for plasma of the dipping refractometer at 17.5° C. minus the similar reading for distilled water. The correlation coefficient is + 0.963 ; the least-squares regression line shown does not pass exactly through the origin. Data of Adolph and Kingsley. cal"; they are just as valid as "physico-chemical" ones, the chief differences being that they are more difficult to predict, are worked out in vivo instead of in vitro, and sometimes have larger devi- ations. This, however, is the sort of correlation that exists among all the quantities that are each separately related to the body water content. Under the specified conditions, variables AM, AN, AP, and 220 PHYSIOLOGICAL KEGULATIONS AQ are each correlated with AW. Then, still under these condi- tions, AM and AN, AP and AQ, etc., also vary in parallel. Con- versely, AM, AN, AP and AQ characterize the state of the organism with respect to water load, and AW may in turn be found from its correlation with them. All this is implicit in table 24. The relation of hemoglobin dilution of the blood to water load of the body as a whole (fig. 124) might be an especially useful one, for having once established the relation, A/Hb = 1.03AW, data re- corded in the past would become available for the study of water loads in which body weights were not measured at suitable inter- vals, but in which hemoglobin dilution was measured, as was the +8 M^ +16 Total Water Load Fig. 124. Increment in dilution of whole blood (% of initial) in relation to total water load ( % of Bo) . Dog. Eepeated administration of water by stomach. Triangles and dash line, dilution of whole blood chloride, in two individuals of Underhill and Sallick ('25). Circles and solid line, dilution of hemoglobin, in four individuals of Underhill and Sallick and in two individuals of Greene and Eowntree ('27). Occasion- ally 2 or 3 determinations were made during one test. case in figure 125. Before further conclusions are drawn, however^ it may be recalled that dilution of hemoglobin is in other studies (table 24) not accurately related to water load. Increments in hemoglobin dilution might also be expected to parallel increments in "red cell" volume in water excesses, whereupon it is found (fig» 114) that at + AAV = 8 the volume of distribution of carbon mon- oxide ( Vd) is not significantly different from the Vd at AW = 0. The changes undergone simultaneously by the various concen- trations measured are diverse. At a chosen time, some, as A/Clg (fig. 125), indicate more dilution than the increment in volume of the "plasma" (fig. 114). Others indicate about the same or less dilution than the increment in volume of "whole blood," but not all dilutions measured in whole blood do so. Many plausible ex- CORRELATIVES OF WATER CONTENT 221 cuses have been put forward as to why changes of dilution do not parallel changes of volume. Each dilution, like each volume of distribution, is a separate and distinct measure bearing an indi- vidual relation to body load of water. It is possible also to infer how much of each non-aqueous sub- stance {e.g., chloride) leaves or enters the circulating blood (a) by comparison with the dilution of some one {e.g., protein) that is in those circumstances believed not to leave or enter, or (b) by com- •^40 0 +4 +8 ^\^ -^16 l/Hb— percent Fig. 125. Increment in dilution of serum or plasma (% of initial) in relation to simultaneous increment in dilution of hemoglobin in whole blood (% of initial). Dog. Since A/Hb is approximately equal to AW of the whole body (fig. 124), the abscissae here might be considered as total water load. 1/VisCs is the increment in reciprocal of serum viscosity; 1/ECa is the increment in reciprocal of serum electrical conductivity; etc. Q represents hypothetical equality of dilutions between ordinates and abscissae. Each point is the mean of 2 to 10 (usually 8) analyses on as many individuals that were given water repeatedly by stomach. Data of Greene and Eowntree ('27). parison with the changes in one of the measured volumes of distri- bution. In the intact body, only such relative estimates are avail- able. The above studies are confined to water increments by water privation and by administering water by stomach or vein. Com- parisons may be obtained in blood and serum of dogs that have been water-loaded by any other means. A favorite type of load that has been investigated is in the state following intestinal obstruction (Haden and Orr, '23). Others are the loss of gastric juice (Gamble and Ross, '25), loss of pancreatic juice (Gamble and Mclver, '28), 222 PHYSIOLOGICAL EEGULATIONS and loss of intestinal juice (Herrin, '35). In them also negative dilutions of serum-protein nitrogen, of blood hemoglobin, and of other measures, are found. But the relative values obtained among several sorts of dilution are not like those here reported. Volumes of distribution of the water load (or should they be termed dilutions of distribution?) allowably may be computed from the increments of some dilution, such as refractive index of plasma (fig. 126). They show that just after a load is administered, time is a large factor, in both positive and negative water loads. After a sufficient period (one hour) the volumes of distribution of water are all much less than 100 per cent of the body (fig. 118) ; this fact +10 Dog 9 25/ / / > — 'A 1 i / /- 1 5 0 Tota 1 u laier +5 +1C Load -10 Fig. 126. Increment in dilution of plasma as reciprocal of excess refractive index (% of initial) in relation to total water load (% of Bq). A supplementary grid shows volumes of distribution of the added water in plasma, as % of the body weight. The course of simultaneous points is traced in averages of three types of test. One hour after ingestion began is marked (/). B, ten successive administrations of water by stomach, 6 tests from figure 117. S, single administration of water by stomach, 3 tests from figure 116. T, single voluntary ingestion of water by mouth, 4 tests from figure 120. is confirmed by fewer like data of Abe ( '31a) and of Hatafuku ('33a). It indicates again that the plasma undergoes somewhat more change than the body as a whole ; the positive increments of water being apparently distributed through a volume less than that occupied by water already present. In the tests represented in table 24 some nine quantities are measured simultaneously. By correlating these two at a time a fairly complete study is accomplished; though triple and higher multiple correlations still remain open. Of the bi-correlations, only rate of water output in relation to dilution is shown graphically CORRELATIVES OF WATER CONTENT 223 (fig. 127). One of the questions at stake is, whether plasma dilu- tion is so closely related to rate of excretion that the dilution could inform the kidneys how much water load exists. The correlation is particularly significant when a wide range of water loads or plasma dilutions is investigated; the coefiScient of correlation (+ 0.48) is higher than for any quantity measured in correlation with rate of excretion, except water load itself. It is not the only 2.8 2.4 3 o 2.0 y 1.6 1.2- o (D 0.8 0.4 0 Doc Q", ur 8o- -* 00 » — O" -12 16 +20 -8 -4 0+4 -8 +12 Plasma Dilution (l/R.l) — percent Fig. 127. Eate of urinary water output (% of Bo/hour) in relation to plasma dilution as 1/EI (% of initial state). Two individuals C'(0) and G'(A) ; the 9 tests of figures 116 and 117 plus 7 more tests in which water was given by stomach, together with the 6 tests of figures 120 and 121 in which water was first denied, are represented. Two of these tests (on dog G') have the successive points connected by lines. Excluding the 6 water privation tests, the correlation coefficient in positive loads is + 0.48. correlation, and to designate plasma dilution as the cause of diu- resis would be misleading. To conclude, on the contrary, that "changes of blood concentration and volume are scarcely detectable in diuresis" (Adolph, '30, p. 63), merely because circumstances may be found in which the correlation is poor, is erroneous. To conclude that only a single factor of any kind governs the rate of water excretion may be equally blinding. 224 PHYSIOLOGICAL. EEGTJLATIONS On the whole, correlations between blood dilution and body load of water become apparent when large loads are provided. They are clearest for dilution of total substance in plasma {e.g., refrac- tive index). Many dilutions are colligative, but with slightly diverse ratios. The ratios here recorded indicate volumes of dis- tribution of the water increment amounting to about half the body weight. Further, dilutions are correlated with rates of urinary output of water. Accordingly, blood serves as a distributor, putting all other tissues in an equilibrium of water partition with it, and also furnishes in its dilution a possible stimulus to kidneys, mouth, and other organs, informing them what water content pre- vails in the body. § 81. CONCENTKATIONS OF URINE Urinary concentrations are closely related to the body's water load (fig. 128). Since these relations are widely recognized, they 1.06 4- O X 1 1.04 h o QJ Q_ [00 -4 0 +4 +8 Total Water Load Fig. 128. Specific gravity of urine in relation to total water load (% of Bo). Dog, on constant diet. Each dot represents a separate day; in negative loads 3 individ- uals deprived of water, in positive loads 2 other individuals given repeatedly water by stomach. Each cross represents a mean sp. gr. for a one per cent interval of load. Further data of Adolph ( '39a) and of Kingsley. are likely to be thought of as only qualitative. Actually a single measurement of urinary specific gravity may do more to identify the existence of a water increment than any other one measure- ment, particularly if other measures are not already controlled by a series of like data on the same individual. Quantitatively the correlation is blurred by the fact that in all positive loads the COKRELATIVES Or WATER CONTENT 225 samples of urine vary around one concentration and in all negative loads around another, with a steep transition between. Many other concentrations (chloride, urea) and ratios (chloride/creatinine) of urinary constituents yield equal correlations with water load. Over all, one conclusion emerges, namely, that in water diuresis, rate of output of water augments more than of other substances. In the long run, little else slips out of the body with it; its excretion is highly specific. § 82. Concentrations of other body FLuros Digestive juices are collected from fistulas while dogs are sub- jected to diverse types of body load of water. It is believed that the total particulate (electrolyte) concentrations in most juices are equal to those of blood plasma simultaneously. Some at least of those constituents {e.g., bicarbonate) that increase in concentration in deficits of water, do so in both blood and juice (Herrin, '35), The drained juices, wherever available, therefore serve as auto- matically extruded samples of analyzable body substance. They are equivalent to any other sort of sample in indicating a part of the distribution of water increments. <^ 83. Other compositions In the whole body of the dog with water load, few changes in the total amounts of materials other than water have been measured. To ascertain them, the whole body need not be analyzed, for with greater accuracy the increments in content of those materials are ascertained from determinations of intake and of output of them. Nitrogen is said to be characteristically lost in water privation (Straub, 1899; Spiegler, '01). It is uncertain whether the amount lost from the body is related to the duration of the water deficit more than to the extent (- AW) of it; in all cases the negative load of nitrogen (-AN) amounts to less than 1 per cent of the nitrogen in the body. Retentions of nitrogen are sometimes found, espe- cially in short periods of water privation ; these are related to the retention of urea and probably of other substances at low rates of water excretion (§84 (4) ). Phosphorus is reported to be retained with nitrogen (Landauer, 1894; Straub, 1899). Water excess leads to depletion of several measured chemical constituents, as nitrogen, phosphorus, chloride, creatinine, and acid (Heilner, '06; Underbill and Sallick, '25; Greene and Rowntree, 226 PHYSIOLOGICAL KEGULATIONS '27; Brull, Poverman and Goffart, '36). The most extreme deple- tion is that of chloride, which amounts to over 10 per cent of the body's supposed content of it after water equal to half the body weight has passed through the body during one day (Underbill and Sallick). The depletion of some or all constituents may be more closely related to the modified rates of water output than to the water content of the body at some one time. When water excesses are maintained for months very large deficits of chloride possibly result. Wolff ('35) reported that a 6.3-kg. dog was depleted of 0.65 equivalent of chloride during 157 days of excessive water administration. But I judge, from data of Wier ( '40) that a dog of this size ordinarily contains only one-third that much chloride to begin with; hence Wolff's account of chloride exchanges may have had a systematic error. Is there evidence that in water excess constituents are accumu- lated in the body or its parts that are not retained in water balance? In water deficit is there any augmented intake of other substances than water? I know of nothing to show clearly that such occurs, though in the shifts of physiological activities that accompany water loads there may easily be some. Fixed tissues sampled during water loads have been analyzed very infrequently for other constituents than water. Significant depletions in electrolyte contents (relative to content of dry ma- terial) of muscle and of skin are demonstrated in water deficit (table 23). If the loss of sodium or of chloride were proportional to that of water, the dilution, which is the ratio of water to electro- lyte (E/Na) would remain constant in the tissue. It appears that no tissue lost an electrolyte in so great proportion as water. In water privation chloride is not lost so extensively from the whole body as from the muscle; hence what chloride may have left the muscle may be in part translocated elsewhere. The study of compositions might be a mere exercise in correlat- ing, were there not special questions in mind. The chief one is, has the water-loaded body all its non-aqueous constituents in their usual amounts and places, all ready to soak up the water that will be gained or lost in recovery? For the most part, neither lack nor excess of water greatly upsets the dog's content of any other con- stituent. So there persists a full framework that may serve to indicate when the original water content is restored. correlatives of water content 227 § 84. Correlated metabolisms and behaviors Thus far, diverse changes of composition and physical charac- teristics have been mentioned in relation to the body's water con- tent. Another large group of correlatives are rates of physiologi- cal activities; they are set apart from compositions in that they involve intervals of time. The ones considered in the dog concern : (!) outputs of alimentary glands; (2) food intake; (5) energy transformations; (4) renal outputs and clearances; (5) neuro- muscular activities. A great many other quantities might be of equal interest if adequate data be correlated concerning them. (1) The outputs of water in the dog's alimentary glands have been studied in saliva, gastric juice, and bile. Saliva is secreted only in minute amounts ordinarily ; in tests some standard stimulus such as pilocarpine or food is given to increase output of saliva. The stimulus is given to the dog once a day (or once an hour) while water privation is decreasing the body water content progressively through several days. Since the changes of body weight are not recorded, rough estimates of water deficit might be made, from the time elapsed and the water missing in the diet. In each series of experiments the saliva is collected from submaxillary glands alone. With each of five stimuli a decrease of salivary output with deficit of water content is found (Crisler, '28; Gregersen, '31; Barron, '32). At positive increments of water the rates of salivary secretion have been recorded less precisely. With very large administra- tions of water, "unstimulated" salivation is reported as excessive (Rowntree, '23 ; Underbill and Sallick, '25). Possibly this produc- tion of saliva is correlated with vomiting (Rowntree) ; it is appar- ently reinforced by additional stimuli such as spinal irritation or pituitrin (Theobald, '34). The loss of saliva from the mouth may then be almost as effective as the outputs of urine in relieving the dog of the water load. But these events occur only in certain circumstances and in extreme excesses of water content. Vomiting is a direct means of refusing water ; of decreasing the intake of it. It is probably more useful to regard it so than to classify vomiting as a means of output from the body. I believe it has not been shown that water introduced by a non-alimentary route is appreciably lost through the alimentary tract. Gastric juice output has been measured in relation to water excesses (Pavlov, '01; Lonnquist, '06; Foster and Lambert, '08; 228 PHYSIOLOGICAL KEGULATIONS etc.). It increases both when food is given with the water and when not. The increased rate of secretion might be more closely related to the water load either of the body or of the alimentary tract (of the stomach itself). Whether the rate of juice output from the walls of the stomach that are in contact with water given, is the same as from pouch separated from the water, is not certain (Sutherland, '21). While the stomach produces more fluid, it also produces more acid and chloride (Ivy, '18). Bile output through a fistula is apparently not consistently affected by water excesses (Bidder and Schmidt, 1852, p. 162; Snell and Rowntree, '28). The composition of the bile is not signifi- cantly changed. Intestinal juice, pancreatic juice, buccal mucus, and colonic ex- cretion have probably not been measured in terms of output rates, in water loads such as are being studied here. It has not been demonstrated that a quantitatively reproducible correlation exists between water load and rate of any one alimen- tary output. Many physiologists are confident that passage of water into the alimentary tract "stimulates" the output of several juices. Existing data show extreme irregularities of performance, even with uniform procedure. On top of this is the uncertainty as to what numerical water loads (of absorbed water) prevail during and after the administrations and privations of water. The near- est thing to a correlation at present is between salivary output and size of water deficit. One of the motives in obtaining those mea- surements was the hypothesis that shortage of saliva might be fur- nishing a tangible signal to other organs and tissues of the dog, leading to compensatory responses. But since shortages of water have been observed in several diverse fluids and tissues, there is less occasion to think of any single one of them as providing a message. (2) Food intake is greatly reduced in water privation, as was noted in early experiments {e.g., Pernice and Scagliosi, 1895). This is not true on all diets nor at all loads, and quantitative studies are needed. With the anorexia have been correlated qualitative decreases of gastric motility (Rose et al., '31) originally on the assumption that a ''direct connection" between them exists. No criterion appears to have been proposed in physiological science for distinguishing a direct connection from an indirect one. Perhaps a useful statement of the situation is that by habitual mental associ- COERELATIVES OF WATER CONTENT 229 ations some physiologists are ready to make ''derivations" {sensu Pareto, '35, p. 508) relating anorexia and gastric motility, which they do not make relating gastric motility and water content of the body. It may be pointed out that going without food is a possible means of conserving water to the body. In extreme water deficits, food that is eaten is actually regurgitated and thus rejected. With- out food, less water is excreted, both in urinary and in evaporative paths (Morgulis, '23, p. 133, p. 266) ; and the water already present more nearly suffices to maintain the gradually shifting balance of water. (3) Energy transformations. In the body as a whole, energy metabolism is most often studied in the particular stages of energy transformation represented by oxygen consumption and by carbon dioxide output. No change in over-all rate was found in 3 hours after placing warm water in the stomach in amounts up to 2% of Bq (+2 AW) (Bidder and Schmidt, 1852; Eubner, '02; Lusk, '12). But after administering loads of + 10 AW by stomach, Heilner ( '07) found increases of 10 and 18 per cent in rate of carbon dioxide elimination. Rates of oxygen consumption increase only at certain times during water excess, indicating therefore less relation to water load than to interval after introduction. Increases of 20 to 600 per cent may occur for short periods after fluids isotonic with blood are rapidly introduced by vein (Davis, '35), but disappear entirely before the load has been eliminated. Significant modifications in respiratory quotient with water loads of the type discussed here have, I believe, not been found. (4) Renal clearances, or ratios of the rate of excretion of any constituent to the concentration of that constituent in blood or in plasma, are greatly modified with water content of the body. For the most part the correlations already made are with rate of uri- nary water excretion (Austin et at., '21). However, having estab- lished the relation between AW and AW/At, perhaps choosing the steady state for the comparison, I indicate that the relation of load to the renal clearance (fig. 129) is equally significant with that of rate. With this information at hand, it next appears possible that the increased output of diverse substances at the same plasma concen- tration of them can account for the depletions of nitrogen, phos- 230 PHYSIOLOGICAL EEGULATIONS 0 +2 +4 +6 Water Load Fig. 129. Eenal (urinary) clearance from plasma of creatinine, inulin, and urea (% of Bo/hour) in relation to sensible water load (% of Bo). Dog. The urea clear- ance was transformed, by means of figure 26, from a relation to rate of urinary water excretion as averaged in data of Dominguez ( '35, p. 537) to an approximate relation to water load. The ratio of creatinine clearance to urea clearance was then taken from Shannon ( '36b, p. 210), and of inulin clearance from Shannon ( '35 and '36a). phorus, and some of the other materials with water excesses, as noted in § 83. An actual example of the effects of water excretion upon urea content may be taken from data of Greene and Rowntree ('27). Dog 183 weighed 7.0 kg. ; taking 66 per cent of this as the volume of distribution (Painter, '40) of urea, which was 6 mM in whole blood, some 28 millimols were present in the body initially. This dog may have formed 9.3 millimols of urea per hour. If the rate of urea formation and the volume of distribution be constant, but the clearance of urea be augmented, almost doubled at times, during TABLE 25 Computation of extra urinary elimination of urea that might occur during maintained water excess in which urea clearance is doubled. Dog Hours Urea produced, m. mol./hr. Urea excreted, m. mol./hr. Mean urea content of body, m. mol. Urea content at end of hour, ni. mol. 0 9.3 9.3 28.0 28.0 1 9.3 15.6 23.4 21.7 2 9.3 13.1 19.7 17.9 3 9.3 11.1 16.7 16.1 4 9.3 10.3 15.5 15.1 5 9.3 9.9 14.8 14.5 6 9.3 9.4 14.2 14.0 CORRELATIVES OF WATER CONTENT 231 6 hours of copious, uniform water administration, the following values result (table 25). At the end of 6 hours the urea concen- tration in the blood was actually 3 niM, corresponding to a urea content of the whole body of 14 millimols. This agrees with the amount expected (last column) upon the assumptions mentioned. The agreement makes it probable that the rate of urea formation itself is actually unaffected by water excess. Another notable out- come of the computation is that the urea content of the body ap- proaches (at only half the control content, or Au-50) a new balance, in which elimination of urea equals production of urea. Chloride in plasma behaved similarly in the same test. 6i Hours Fig. 130. Rate or concentration (in arbitrary units) in relation to time after water ingestion. One test on dog E of Shannon ('36b, p. 208). Water was administered by stomach at zero time (5.0% of Bo) and at 5.15 hours (1.5% of Bo). Urea Clearance = urinary clearance of urea from plasma, in units of 2.0% of BoAour. Ureap Con- centration = concentration of urea in plasma, in units of 0.8 millimols/liter. Urea Rate = rate of urinary urea output, in units of 0.10 millimols/hour. Water Rate = rate of urinary water output, in units of 0.4% of Bo/liour. To conclude that the kidneys are responsible for the depletions of solutes in + AW is like blaming the pumping station for the great flow of water from faucets near dinner time. In water deficits the lesser renal clearances by themselves favor retentions of solutes. The above relations for urea are shown in figure 130 as they change with time during single loadings and unloadings of water. Renal urea clearance there temporarily augments by 50 per cent. Since water excesses are measured ordinarily as sensible loads, the 232 PHYSIOLOGICAL EEGULATIONS rates of urea clearance may bear a different relation during rising water loads than during falling loads. Further, at one load the urinary rates are often higher following sudden accession of water to the body (§12) than during decrease of load. Rates of excretion of urea (Marshall, '20) and chloride (Kings- ley) suddenly augment, when the rate of water excretion augments following ingestion of water. Chloride is eliminated four times as fast as without water ingestion, then its output subsides before water excretion diminishes ; as though the sudden increase of water excretion caught the processes of solute excretion napping, and some time were required to exact the separation of each solute from the flood of water. The ratio of solute output to water output is accordingly found to be smaller in proportion as more of each has been just previously excreted. In brief, while the content and elimination of urea and other substances are being influenced by water content and exchange, the concentrations of many or all substances in blood and other tissues are simultaneously affected. Evidently (1) the changes of concen- tration noted in section 79 are not solely those of dilution; (2) time factors loom larger than load factors in some of them; (5) other methods than those measuring rate of urea excretion alone are required to ascertain whether the production of urea is affected by water content ; similarly for any other component. (5) Neuromuscular activities have been noted only in qualita- tive fashion. In extreme positive loads of water the dogs of Rown- tree ('23) exhibited muscular twitchings, hyperirritability, as- thenia, ataxia, vomiting, coma, convulsions, and death. Each is a type of behavior capable of quantitative study at known water loads ; several may hasten the recovery of the animal toward water balance. Individual reflexes and responses are equally susceptible of evaluation, and might be found modified at moderate water loads. In extreme negative loads, still fewer observations are recorded. Anorexia, vomiting of food, restlessness, ataxia, and death are known. Dyspnea is prominent (Pernice and Scagliosi, 1895) but panting in response to concomitant heat is slower to start (Greger- sen, '31 ; Dontas, '39 ) . All instances in which ' ' thirst drives ' ' have been induced in dogs may furnish material for the present study. "Water load, in turn, is a proper measure of the intensity of these urges or drives to activity. COERELATIVES OF WATER CONTENT 233 Certainly the presence of too much or too little water in the dog disturbs the bodily functions in many measurable ways. From the even tenor of maintenance, departures are visible in the alimentary secretions, food ingestions, energy transformations, renal clear- ances, and diverse neuromuscular activities. Several of these modifications are distinctly correlated with the suddenness of water loading. Which of them come as shocks and which are compen- satory can be decided only in part, when judged according to their promotion of recovery of water content of the whole body. § 85. QUANTITATIVIE CHARACTERIZATIONS OF WATER LOADS Increments in diverse volumes, concentrations, and rates of functioning in particular tissues, now appear correlated with water loads of the whole body. Each of these is evidently interrelated with all the others; and as long as only one variable (AW) is, or possibly two variables (AW and At) are, alone ''independent," the relations among all the correlatives can with ease be described quantitatively. To do this, tables 23 and 24, and numerous two- dimensional graphs such as figure 124, are combined in an align- ment chart (fig. 131). The latter emphasizes the concept that no one of the quantities plotted is of more consequence than any other ; all are coordinate states or happenings. To the increments measured might be added the modifications in a variety of further quantities such as pressures (in arteries, cranium, spinal fluid, Rowntree, '26), physical condition of skin, and rectal temperatures (Keith and AVhelan, '26). Each of the modifications mentioned and most of their combi- nations, may appear in many types of water load. Speaking of man. Underbill and Fisk ( '30, p. 348) state that: "Such apparently different chemical conditions as Asiatic cholera, infant diarrheas, intestinal obstruction, influenza, war gas poisoning, and extensive superficial burns, exhibit as a common symptom a marked concen- tration of the blood. ' ' Measures of blood concentration have been preferred above all others to characterize states of water load. Data concerning them are obtained more easily than concerning any other correlative except body weights that have yet been used. It might be inferred from the above statement that any and every kind of blood concen- tration is modified. Haden and Orr ('23) say more specifically that ''the characteristic features of dehydration are: increased 234 PHYSIOLOGICAL, REGULATION'S blood viscosity, increased plasma non-protein nitrogen, increased protein destruction, decreased oxygen capacity, decreased blood volume." They quote specific researches on each point. But those features are not common to all types of ' ' dehydration. ' ' The type that most interested Haden and Orr was the state of upper intestinal obstruction with vomiting and inanition. An example of -C "^ +> > ©-Tr :=* a? > > ^ "^ — o o CO O 3 -o c cr ^ DC ' 4J 3 o' o 1 b o o l§ ii >» >• E o c^ g "5 "5 "o > i. •t^ J •(O O^ Csf 0^ CsT N C4 (M ~~ . (ft . vrt <.. OTN a-\ -1 1 1 Dog Pup 1 ■ •> Ha ^ \ - •Al. . o oo o o ■ oE: 1 1 1 o 1 1 V m 1 0 10 zo 30 80 700 Days after Birth Fig. 136. Duration of the half -life of total water load in relation to age. Nine- teen young dogs of 5 litters, and adults (N). Solid line represents the mean for all litters, dotted lines represent 2 separate individuals of litter H. New data. and amounts determined by some agency outside itself ; regulations of water and other substances develop before weaning is possible. Similarly, water drinking is not demonstrable at birth. A very young pup that takes milk from a bottle does not take water from a bottle, indicating that the two ingesta are distinguished. The newborn is not induced to take water by depriving it of water for three hours, nor by administering a concentrated salt solution. The first drinking of water as such I found at 15 days of age. So, compensations by diuresis and by drinking appear at about the same age. At first the drinking is slow; and the correspondence between the water deficit and the amount taken voluntarily is not close at 15 days of age, while at 35 days of age, hence before as well as after weaning, the adult speed and accuracy have been acquired. 268 PHYSIOLOGICAL REGULATIONS Equilibration diagrams, then, change as the pup develops (fig. 137). At birth, turnover of water (per unit of body weight) is double the adult's, and in water loads the modifications in rates of exchange are small. With surprising suddenness, rates both of gain in deficit and of loss in excess attain high velocity quotients. Only upon a few days of the pup 's life can intermediate stages be found. Thereafter, the time required for recovery from water load is independent of body size. Turnover, however, slowly di- minishes (as the ratio of surface to mass decreases). The tran- sition to compensatory processes makes clear how inadequate is the provision for adjustments in the newborn pup; and, by contrast, Loss 0 +2 Total Water Load +4 +6 Fig. 137. Eate of total water exchange (% of Bo/hour) in relation to administered water load (% of Bo). Equilibration diagrams. Young dogs in the first 1.0 hour of recovery. In excesses, pups up to 15 days of age had their bladders emptied by punc- ture through the body wall; at later ages micturition was spontaneous. Solid points represent newborns at 1 to 5 days of age, crosses those 7 to 15 days of age, and open points those at 16 to 32 days of age. New data. Adult dogs may be compared from figure 13. emphasizes what the pattern of equilibration enjoyed by the adult accomplishes. In man within the first days of life the rate of water turnover increases daily (Drossel, '29). The rate of urinary water output progresses from 0.02% of Bo/hour in the first day, to 0.24% of Bo/hour in the sixth day (Gundobin, '21, p. 363). Of course this change is related to the interruption of nutrition at the metamor- phosis of birth, and the gradual increase in intake of food (and water) during the first week. Throughout the first year the rate of urinary output subsequently diminishes in relation to body weight, being nearly proportional to B°'^, when judged from Gun- dobin's data. CORRELATIVES OF WATER CONTENT AND EXCHANGES 269 When excessive water is ingested by the human infant the pro- portion of it returned by renal diuresis is small in the first months (Ohlmann, '20; Lasch, '22). It is not certain that the transforma- tion of compensatory adjustment is as sudden as in the dog. Only after 6 months of age in man is the return as prompt and complete as in adults. There is some evidence that at 5 to 10 years of age the urine excreted within 3 hours markedly exceeds the water given, which is not the case in adults (fig. 49). In these measurements, a number of unrecognized differences other than age may exist; also, differences such as size of ingestion, or promptness and ease of micturition, may be legitimately included in a single variable. In frogs {Rana esculenta), where metamorphosis and respir- atory modifications are proceeding apace (table 30), age is less TABLE 30 Bates of water turnover at diverse sizes and ages in Eana temporaria at 18° to 22° C. Anus and mouth were ligated for 3 to 5 hours. Data of Bey ( 'S7, p. 1132) Number of tests Mean body weight, gm. Mean body surface, cm.2 Eate of water intake Stage per unit weight, %/hr. per unit surface, cm./hr. 1, tadpole 1, * 5 5 3 3 3 25 0.70 0.64 1.97 1.94 1.92 65.0 5.5 5.3 11.0 13.5 12.2 130.0 1.5 3.1 1.4 2.9 2.8 0.92 0.0018 0.0037 2, hindlegs 4, forelegs 5, metamorphosed 6, adult 0.0022 0.0033 0 0039 0.0047 * Mouth not ligated. correlated with body size ; no single rule is apparent in the rates of water turnover recorded. Physiological metamorphoses inevita- bly accompany aging in many or all species. Based on body weight, the turnovers decrease with age in dog, man, and rat, but not in bovine and frog. In water increments the rates of intake or output, the time relations of exchanges, and the velocity quotients also show diverse modifications at different ages. This fact may adequately prevent anyone from viewing regulations as fixed properties of unchanging organisms. (2) Among species. Correlations with body weight furnish a wide opportunity for comparisons among species. Data upon rates of water exchange are available in mammals of very diverse sizes for (a) total water intake (or turnover) in water balance and (b) maximal water intake after hypophysial injuries (fig. 138) ; (c) 270 PHYSIOLOGICAL EEGULATIONS urinary water output in ordinary water balance, and (d) maximal water output in water diuresis (fig. 139). Large numbers of measurements have gone into the determinations of means or maxima in each category. In a previous chapter these values of rates of water exchange are considered as species differences with- out reference to body size. Here they are found to be related with body weight according to allometric equations. The conclusions indicated among mammalian species are: {!) Total intake is roughly proportional to B°-^. (2) Maximal intake +4 CP +3 D -P C V- O :5 M- o Q) -P O cc o +2- +1 - 1 1 — 1 ■ -I - 1 / o "a. o - o X LU " y 7^ - o y < / +> 3 / §. / -a o £ / cc ^y fx E ,2r o 1 2. 1— 1 • — 1 1 — 0 I 2 3 4 5 6 7 L09 Gm. Body Weight Fig. 138. Eate of total water intake in relation to body weight among diverse species of mammals. Lower line, mean turnover rates in control conditions (from table 20): Eto = 0.010 Bo-88. Upper line, maximal rates after injuries of hypophysis (from table 13): E„„ = 0.033 Bo.97. is nearly proportional to B^° (Eichter, '38). (5) Ordinary urinary output is proportional to B°^ (4) Maximal urinary output is nearly proportional to B°-®. (5) The ratio of ordinary total intake (or equal output) to urinary output, ranges from 2 for small spe- cies to 5 for large ones. {6) The maximal urinary rates equal 32 to 20 times the ordinary rates, the augmentation ratio being less in small animals which have turnovers already large in proportion to body weight. (7) Maximal intakes after hypophysial injuries, and maximal outputs after water ingestion are nearly equal. (8) CORRELATIVES OF WATER CONTENT AND EXCHANGES 271 Maximal intake rates equal 5 to 17 times the turnover rates (= aug- mentation ratios). These correlations might be used to predict by interpolation the probable rates of water exchange in other species. I do not recom- mend prediction as a substitute for measurements of them. It is apparent that time of recovery from water load is independent of body size ; rat and man remove loads equally rapidly. Data upon water exchange in other classes and phyla of ter- restrial organisms apparently are not available, with a few excep- .^^a 1 1 1 1 1 1 E 1 y $/ +" +3 _ y^ / - 3 ^ ^r / "^ CL '^y y /^ c -P 3 o ^0-^ y y^ c ^ o ON y 10 o -C UJ 0) y^ / +> -5 -Q a Q y y^ >- o o c O Oc y o, D y -X c o >^ =- 0 3 o y / o 2: n 'h 0) X o -P / ^ ^y -t-" - X _ o ' / or , o 1 \ 1 1 ON o 0 7 12 3 4 5 6 ~^ Log Gm. Body Weight Fig. 139. Eate of urinary water output in relation to body weight among diverse species of mammals. Ordinary rates are in water balance (from table 28): IIto = 0.0064 Bo- 82, Maximal rates are during continued forced administration of water by stomach (from table 13): E„ai = 0.26 Bo.ts. tions as pigeon (Falck and Scheffer, 1854a) and chicken (Korr; Hester ei al., '40). Among aquatic animals (fig. 140) rates of turnover are proportional to B°'^^ and maximal rates to B°-^^. This confirms the more approximate conclusion of Putter ( '26) that water exchanges are proportional to body surface area. At equal body weights the turnover rates of these aquatic animals are roughly identical with those of small mammals of figure 138. A rule enunciated by Eichet (see Morgulis, '23, p. 91) is that the intensity of vital functions depends directly upon the size of the 272 PHYSIOLOGICAL EEGULATIONS organism. I presume that an intensity of the sort meant is mea- sured by the rate of water exchange ; and there the rule holds. Of course, when exchanges are measured as fractions of the body weight per unit of time, all the rates diminish with greater body size ; actually, they vary as some power of weight that is less than 1. Since all measures of size are correlated with others, any one of them that can be accurately and conveniently measured is preferred for use. Usually that one is body weight. 1-2 Log Body Volume— ml. Fig. 140. Eate of water exchange in relation to body volume among many species of aquatic animals. Both scales are logarithms of either ml. or ml./hr. Maximal rates (from table 13) are either for output in positive loads (solid triangles) or for intake in negative loads (open triangles): EB,ax = 0.50 Bo.66. Turnover rates (from table 21) are either for marine species (solid circles) or for freshwater species (open circles) : Eto = 0.016 Bo-75. (3) Additional features. Special structures and shapes, such as the ears of the rabbit and the wings of the bat, may modify the correlations of rates of water exchange with size among species, as in every other relation of body size. No attempt is made here to sieve out the partial correlations of water exchanges with pul- monary ventilation rates, kidney weights, oxygen consumption rates, and a host of other quantities. Two or three more selected variables will, however, be followed. COREELATIVES OF WATER CONTENT AND EXCHANGES 273 (a) It is recognized that the exchanges of water as described depend upon absorbing and excreting surfaces, conveying and circulating bodies, coordinating pathways, and many other parts. The presence of probable oversize (factor of safety) of each, is itself related to body size among species. Rapid movement of internal fluids is often absent in small species, such as eggs, certain worms, and some sessile organisms ; without convection, the move- ment of water is often presumed to be limited by processes of molecular ''diffusion." With convection inside the body, the in- tegument may be so thick and immobile as to limit osmosis. In gen- eral, how large might an organism be before osmosis through its permeable tissues would limit its water intake or output? Compu- tations might be predicated ; the only experiments seem to be on the frog (of 32 grams weight). When blood circulation was suddenly stopped, water output ceased, but intake of water from a dilute salt solution was unchanged in rate (Adolph, '31b). When stopped ("death") by previous desiccation, even maximal water intake was undiminished in initial rate (fig. 66). (b) Among terrestrial animals it is widely recognized that turn- over of water is directly related to turnover of food. Sometimes it is specified that the nitrogenous fraction of the food is a correla- tive (Babcock, '12) ; at other times that rate of energy transfor- mation is important. In mammals most water is regularly drunk with or following food ingestion. For mammals the rate of water intake during daily or longer periods is correlated with the potential energy of the food taken during the same periods. The ratio of the two increases with body size (fig. 141) ; possibly the ratio should be between the logarithms of the two intakes, making this relation also allometric. The two species having least ratios are inhabitants of deserts. Since water formed by oxidation is nearly the same per potential Calorie of all foodstuffs, the diversity among species is wholly in the portion of the water ingested as such. (c) For many of the same species of mammals the water output may be partitioned by paths (fig. 142), the total output being taken as equal to the intake. The fractions put out in fecal, urinary, and evaporative channels are those indicated. The evaporative frac- tion is nearly constant per unit of potential energy at all body weights ; this is often spoken of as the proportionality of rate of vaporization to rate of energy transformation. Its range of vari- 274 PHYSIOLOGICAL. EEGULATIONS ation is apparently as large in one species as among species. In small species nearly all water turned over is therefore expended in evaporation. Most of the diversity with body size is in the urinary fraction. Fecal outputs become large only in large species ; these species are ruminants, and all ruminants are large. Conversely, few species not ruminants are eventually available for further comparison within that range of body sizes. Man happens (accord- ing to fig. 142) to put out the largest fraction of water loss as uri- nary, being at a size between the energy evaporators and the roughage eliminators. 1.2 ■08- ^04 E 0 -T- r 1 Co-t Dcg D^ To-bak /6Rabbi+ '"fiC^Mouse Inges-fcive Dipodomys °RI. - - Oxidative 1 1 1 1 i -H 0 Fig. 141. 2 4 6 Log Gm. Body Weii^h-t (B) Total quantity of water ingested and formed for each Calorie of food ingested (and a suggested partition of it) in relation to body weight among diverse species of mammals. The line corresponds to the empirical equation, gm./Cal. = 0.17 + 0.173 log B. The data are from table 20. The elephant consumes about 2.3 gm./Cal. according to Benedict ( '36) ; this point (not shown) would be far above the line. Figure 138 relates rate of water exchange to body weight alone. These are species that have turnovers ; it is only an inference that species not yet measured will fit, within the present range of vari- ations. Figure 141 includes rate of energy exchange, as well as rate of water exchange and body weight. This is a mutual relation among three quantities ; there is no evidence that one of the three sets the pace for other one or two. It would be possible to add further correlatives and in the end to build up an inclusive descrip- tion of many relations to water exchanges among mammals, and among some other restricted groups of animals. In this brief fashion I have tried to present certain materials CORRELATIVES OF WATER CONTENT AND EXCHANGES 275 of ontogenetic physiology and of comparative physiology of water exchanges. Size and age are convenient correlatives of rates of exchange, for each brings all its other correlatives into the picture. Clearly, at birth equilibration of water content is not, in dog or man, of adult dimensions; it is transformed later into the recog- nized pattern of the adult. Regressions of size with rates of water exchanges in turnover and in maximal compensations, both total 1.2 a; 0.8 u _o o ZJ D. O0.4 S- 01 -p o £ - - /A Fecal / y^ \v - / Urinary \^ . . y oRat o Man_2^^::;:il5_ - White flouaeo . ^Rabbit „ o Elephant Fa-t riouie 0 1 riarmot Evaporative 1 1 1 +■ o 0 2 4 6 Log Gm. Body Welgh-fc (B) Fig. 142. Partition of rates of water output by paths, for various species of mam- mals, relative to rates of energy transformation. Each species is represented at its body weight. Total output (turnover) is taken as equal to the total intake of figure 141. Partition is relative to the total water loss, as listed in table 16 and in additional data of Benedict ('38). and partitioned, turn out to have the same sorts of significance as regressions of size and age with energy exchanges. Many data fit allometric equations of relative magnitudes. Presence or absence of internal convection, and types of metabolism of nitrogen, energy, and roughage, are partial correlatives that may be distinguished in the relations of water exchanges among species. <§i 97. Temperature Are the water exchanges as measured also influenced by body temperatures that prevailed? In ''warm-blooded" animals, in- creases of body temperature above the usual are accompanied by enormous increases of rates of total water exchange. Considera- 276 PHYSIOLOGICAL, KEGULATIONS tion of these may be deferred to a later chapter (XIV). In '^ cold- blooded" animals, rates of water exchanges may be studied over ranges of as much as 35° C. With respect to turnover, data on frogs show "temperature coefficients" for ten degrees C. up to 2.6 (Krause, '28). Whether maximal rates of intake or of output in water loads increase proportionally would be interesting to know. Season makes a difference in turnover rates at equal tempera- tures, according to DeHaan and Bakker ('24), winter frogs show- ing half the rates of summer frogs. This may involve many differ- ences of metabolism that have not yet been correlated. It may be remarked that analyzable water contents of frogs also vary with temperature and with season (Donaldson and Schoemaker, '00; Gaule, '01), corresponding to upward shifts of equilibrated body weight (Bo) at temperatures below 9° C. (Adolph, '27a; Rey, '38, p. 1113). Rates of water exchange were compared at five temperatures in Arbacia eggs (Lucke et al., '31). For swelling in 60 per cent sea water, the rate of entrance of water and the permeability coefficient increased 2.7 fold for ten degrees C. rise in temperature. For shrinking (recovery) the same factor was 2.4. This means that the equilibration diagram (fig. 91) increases its ordinates in that proportion. If close comparisons of water relations among species are to be made, a single body temperature might be chosen for many species, but no one temperature would be suitable for all. Warm-blooded animals automatically fix the temperature of the body ; cold-blooded animals are subject to conditions imposed by the investigator within limits. Whatever temperature is imposed is inevitably open to objection on some other grounds. Hence it may be explicitly stated that most of the quantitative comparisons of water ex- changes would be different at other temperatures from the ones represented, and that a particular relation between exchanges and temperature characterizes each species. §98. Races Do racial differences also matter? Closely related species or subspecies of animals frequent diverse environments, foods, and climates. Attempts have been made to relate whatever racial diversities prevail to the rates of water intakes in deer-mice. Com- parisons may here be limited to simultaneous tests under identical CORRELATIVES OF WATER CONTENT AND EXCHANGES 277 conditions (table 31). It turns out that two species of deer-mice differ significantly in voluntary water intake, though reared in the laboratory upon constant diet. The difference was 32 per cent of the mean water intake, and thrice the probable error of each mean. Two subspecies of either species, however, drank like amounts of water. Such a difference between species calls attention to but one of the large number of ''constitutional" factors related to water exchanges in any class of animals. So far as the meager facts indi- TABLE 31 Rates of intake of free water compared in five subspecies of deer mice (Peromyscus) . Paired tests between and within species. All individuals were bred in captivity. Data of Boss ( '<30) Subspecies Number of indi- viduals Water intake in % of Bo/hour Mean Pairl Pair 2 Pair 3 Pair 4 P. m. s. 28, 24 29 24 27, 20 28, 21 0.73 0.86 0^83 0.05 5 0^5 0.49 0.04 0.05 4 0.79 P. TO. g P. m. r 0.67 0.67 0.83 P. e. e 0.52 0.48 P. e. f 0.49 0.18 0.000001 6 0.49 Difference 0.21 0.00002 5 0.28 Probability No. of tests 0.000000 5 to 10 cate, races and subspecies (as classified on the usual criteria) are not sufficiently different physiologically to exchange diverse amounts of water in turnover. <§. 99. Summary It turns out to be possible to deal with many variables either separately or together, when all relate to some one or two variables (SW, AW/At) that are kept continually in view. The procedure is to classify correlatives according to dimensions and in other arbi- trary ways. Further, within the classes certain factors are mutually exclusive {e.g., two diets, two atmospheres), and hence each of those could be considered independently of its alternatives. No basis is found for believing that variables not given special consideration are qualitatively different from those mentioned. For the most part the variables studied were thought of earlier in the development of physiological science, or were recorded handier. So far as I can judge, size of body, or rate of nitrogen turnover, is just as ''intimately" related to rate of water exchange as is refrac- tive index of blood plasma, or body water content itself. But the 278 PHYSIOLOGICAL EEGULATIONS roles of the former in recovery from water loads is not so easily hypothesized. Each variable also has interrelations with all the others. There is no evidence that an organism is physiologically compartmented to the extent that water exchanges are independent of functions that do not have the name ' ' water ' ' attached to them or implied in them. In chapter IX and previous ones I attempted to limit the num- ber of variables considered, to those classified as constituting the time-water system. There, other variables were recognized only in order to label the conditions of measurement, or else to lay them aside for future consideration. In chapters X to XII certain of the other variables are introduced in the form of correlatives of that first system. The study of correlatives may serve to inter- relate many or all the known measurements that are recognizably concerned with water in animals. Forces of several physical varieties are undoubtedly concerned in the water exchanges of animals. The difficulties of measuring them with assurance of completeness has so far prevented the con- struction from them of the analog of an equilibration diagram. Grouped factors in water exchange, such as permeability, are within limits useful in representing comparative water exchanges. Body size and age, and their correlatives, serve as backgrounds against which to view the developments of functions of exchange and compensation within the species, and the comparisons among diverse kinds of animals. Factors of both environment {e.g., tem- perature) and heredity distinguish the manners in which mainte- nance and adjustment of water content are achieved. Chapter XIII WATER BALANCES AND EXCHANGES. RECAPITULATIONS § 100. The materials thus far presented, though serving special interests, may also be considered preliminary to more general con- clusions now to be drawn. These conclusions are partially limited by the voluntary choices of variables and the more involuntary availability of data. But within these limitations a variety of comparisons and a considerable number of inductions may be made. This chapter, in fact, attempts to systematize the materials in such a way as to provide a summarized picture of the physiological regu- lation of water content in animals. Were the study limited to water relations, the presentation might close with it. § 101. Classification of variables The scheme followed up to this point may be viewed in terms of variables, classed as follows, largely according to their dimen- sions. Separate divisions are maintained for the quantities most stressed; numerous distinguishable sorts are lumped together in the other divisions {cf. § 5). (1) Rates of water exchange or volume changes (Rw or Ry). (2) Living units; species, parts of organisms (U). (3) Paths of exchange of water or volume (p). (4) Changes in properties of whole or of parts (s). (5a) Types of load of water or volume (f ). (5b) Amounts of water content or volume (W or V), or load (AWor AV). (6) Time and time intervals (t or At). (7) Velocity quotients for water volume {k). (8) Changes in metabolism and in behavior (M). (9) Forces, permeability, size, age (N). First I describe how each of the divisions named, when related to others in the list, forms part of the picture of regulation. When the picture is completed, the order in which factors are considered will have had very little influence upon the relations, comparisons, and uniformities found. The sorts of variables mentioned may be grouped into continu- ous and discontinuous factors. For, quantities such as water con- 279 280 PHYSIOLOGICAL, EEGULATIONS tent and time may be divided into infinitesimal gradations ; while paths of exchange and types of displacement are more limited in number and kind, and are usually thought of as qualitatively di- verse. So, (2) to (5a) are at least in part discontinuous variables, while (5b) to (9) are usually quantitative factors. Some belong in both groups. Thus, changes in properties or in age are continu- ous with respect to each component or individual but discontinuous with respect to the many individuals, constituents, and tissues in which they appear. In the numerous cases of this sort, other classifications might be profitably explored. Again, some of the variables that can be fixed by an experimenter are often thought of as independent ; those that brook no external selection may be termed dependent. Then (5a) to (6) are usually chosen at the will of the observer, and (7) to (9) follow from the physiological constitution of the organism. Though the distinction is not a biological one, this grouping of variables is useful in that it facilitates quantitative treatment of the coordinated properties. By the methods of nomography, it is simpler to represent on paper any number of dependent variables, provided the number of so- called independent variables is reduced to one or two, than to deal with a greater number of the latter. Accordingly, fixing (2) and (6) so that they do not vary, (5b) and many other quantities can be represented (fig. 131). And in a succession of such nomograms, several selected types of displacement or of conditions for recovery (5a) and ages (9) might be portrayed. In this way an epitome is obtained that describes quantitatively a physiological pattern. Having already discussed (section 11) the interrelations of the four variables of the '^ water-time" system, (1) (5b) (6) (7), and treating at least two of the four as dependent variables, I can now show some of the relations among the remaining six sorts of varia- bles. In the list of variables there are many possible subdivisions ; of these some are mutually exclusive, and may be contrasted, while others are duplicate ways of classifying. The object in the following section is to indicate general features and contrasts, and not just to represent the numerous transforma- tions of coordinates that could be made. Each sub-heading em- phasizes a single one of the variables named. § 102. Scopes op the classes of variables (1) Rates of ivater exchange (Rw) during recoveries may be WATEK BALANCES AND EXCHANGES 281 grouped into total, partial, and net. They are either gains or losses. Acceleration of intake or output (table 14) might be a measure of the readiness with which the living unit responds to loads. Wherever latent periods are absent, accelerations are, of course, enormous. In other instances (all of which are water outputs) accelerations are less, and occupy varying periods of time up to 2 hours, whether gross or net exchanges be considered. Decelerations fall into similar classes, for rapid decelerations usually follow rapid and prompt accelerations ; where accelerations are less, decelerations are also less (table 14). Maximal rates of exchange are, within limits, direct functions of water increments. Actually, few species show any limits to the increase of net intake or of net output. It might be said that the ''bottle necks" of water exchange, if any exist, are never com- pletely crowded. Equilibration diagrams for water (SW/At at various ± AW) are of one type, provided maximal rates of exchange at any one load are chosen for correlation. (2) Species and parts (U). For present purposes it is often sufficient to group species as is done by taxonomists, and parts as is done by anatomists. Wherever physiological similarities ap- pear, however, alternative bases for classifying are indicated. Thus, among insect species, larvae often belong to utterly different categories with respect to water balances and exchanges from either embryos or adults. Innumerable bases could be found for lumping individuals, or organs, or cells, measurements upon which shall be averaged. At present many body builds, many muscles, many red blood cells are studied as though they constituted a homogeneous population, in spite of the fact that in the future, distinctions between gastro- enemii and tibiales antici will probably become significant. In the same arbitrary way, I separate parts in situ from parts isolated. Fifty years ago it was usual to mix observations on anesthetized rabbits and on unanesthetized men. Tomorrow, correlations may be further limited to one age, litter mates, one individual, one diet, one cage. (3) Paths of water exchange (p) may first be identified anatom- ically (table 21). In many species and even phyla none is known. None appears to be separately distinguishable in the exchanges of 282 PHYSIOLOGICAL EEGULATIONS most parts of organisms, other than the general surfaces of those parts, be they lymph vessels, connective-tissue sheaths, or nuclear membranes. But for cerebrospinal fluid of dog and man (Weed, '38), intraocular fluid (Robertson, '39), and other units, it is gen- erally believed that special surfaces of one-way exchange exist. Where paths have been identified it is possible to make con- tributory studies upon specific organs. Thus, the kidneys are often believed by vertebrate physiologists to offer an understanding of many aspects of water regulation. Possibly some information concerning the elimination of water excesses as responses to water loads will eventually come from their separate study; little that relates to loads is recognizable now in analyses of renal function. Ordinary turnover rates are not often regarded as properties of particular organs ; is the distinction between diuresis and normal output made by organisms or by physiologists ? (4) Changes of properties (s) with water loads are as numer- ous as the physical and chemical procedures for measuring them. Very many contributions of biophysics and biochemistry furnish data for this particular sort of physiological investigation. Con- tents, pressures, volumes, and dilutions may be distinguished. Actually most measurements of composition now available concern the distributions of water loads within organisms. No doubt a great many properties that have not been measured, also change with water increment; often they were not measured because no theory was conceived to indicate that there might be a particular connection. (5a) Types of water increment imposed by the experimenter have diverse consequences. Classification of types might recog- nize: environment for recovery, diet, procedure for loading, path of loading, stationary loads. Diverse tolerances, time relations, recovery quotients, and paths are found in the responses to various types. A tentative conclu- sion is that time relations are the most variable features of the responses, among factors so far considered. For ultimately re- covery occurs, and largely by the use of some single path of ex- change, in response to most types of water load. (5b) Water contents and various volumes, modified experi- mentally, are evaluated relative to initial or control weight or volume of distribution (AW or AV). In most cases the changes in weight or volume, or in water content relative to analyzed dry WATER BALANCES AND EXCHANGES 283 content, are rapid enough so that other changes of composition are much smaller than those of water. (6) Time (t) is an extremely large factor in determining the state of the organism with respect to water (fig. 106). The curves of change in water content after displacement have in common the trend toward Wo; but the time scales of these changes are very diverse. Initial, steady, maximal, and fractional hours are distinguished. "Latent" periods are absent in all water intakes (table 14). They are also absent in water outputs by blood of rabbit, by Plias- colosoma, and by Arhacia eggs, living units in which exchanges occur across the entire surface. Where present, these periods are hypothesized to represent delays (a) of translocation, (b) of arous- ing responses to water load, (c) of paths or processes of exchange, or (d) of development of some unmeasured mediate quantity with which the measured load is correlated. (7) Velocity quotients (k) are computed from rates and incre- ments (SW/At-r- AW), or from the exponential curves of water load in time. Often the quotients for net exchanges are nearly uniform over a wide range of water increments, at some selected time after the increment had been imposed. The values of the quotient are various in diverse tissues and species (table 16). All values found at maximal rates are above 0.3/hour, except in water excess of the snake. This means that recovery of water content at balance occurs in all those species in 0.1 to 4 hours ; no load lasts appreciably longer, as would be the case in some organism indiffer- ent to water or unprovided with means of detecting and correcting aberrant water contents. (8) Changes of metabolisms (M), or of rates with which proc- esses go on in the organism, furnish numerous correlations with water content. Of these, clearances through vertebrate kidneys have received most study ; next in quantity of information are total rates of oxygen consumption. Studies of neuro-muscular behavior name thirst as an urge that appears in negative water loads, but furnish few measurements of the activities that go with it. Con- vulsions, vomiting, and anorexia have not been quantitatively cor- related. Among some one or few activities of organisms, such as clearance, excretion, chemical transformation, and movement, it is useful to find in what range of water increments, in what types of water load, and in what conditions they attain particular rates. 284 PHYSIOLOGICAL REGULATIONS Then, so far as the one kinetic variable (water exchange) is con- cerned, all those increments and conditions look quantitatively alike. It may seem fantastic to homologize all the conditions under which a diuresis of a certain magnitude occurs, or all those under which a 3 per cent increase in dilution of plasma exists ; yet just this sort of common response is a basis for analyzing physiological phenomena, I think. (9) Other correlatives (N) such as size, forces, age, sex, per- meability, of the living unit each show quantitative relationships capable of extensive study. Most large species are terrestrial, but it is possible to select equally sized organisms in fresh water, in sea water, and on land ; both their turnover rates for water and their maximal rates of water exchange are then found to be but weakly related to type of environment. For the most part, rates of water exchanges are in proportion to powers of body weight varying from W' to B^°. <^ 103. Interrelations op the variables If variables are chosen so that not more than one of them need be treated as independent, the interrelations among many may be represented in a single diagram (fig. 131). In similar diagrams the quantitative differences among species and tissues may turn out to be almost infinite; these differences serve to characterize each living unit. Uniformities are equally evident ; it is easier to state these uniformities as contours in diagrams than to put them in words. A nomenclature might then be invented by which to designate the shapes of the contours, similar to that used for classi- fying finger-prints. When the pattern of interrelations shall be known for a number of species, comparison of quantities and classi- fication of qualitative combinations after the manner of § 72 can be carried out. But whereas there only three variables (load, gain, loss) entered into 25 qualitatively different combinations, here seventeen variables may yield thousands of relative types. So long as investigation consisted in measuring the simultane- ous changes of any two factors, an enormous number of studies (2""^) would be made, since the number {n) oi quantities and condi- tions seems semi-infinite. Even in 2'""^' papers the interrelations in water exchanges would scarcely have been touched, for the num- ber of combinations of three variables at a time turns out to be about 3'""^', etc. Temporarily, investigators gain satisfaction from finding qualitative answers, from making two measurements at a WATER BALANCES AND EXCHANGES 285 time, and from inferring' that what is true in rabbit holds equally in man. In time, however, the wishful reasonings by which the particular pairs of variables were chosen will probably disappear, and the remaining {n-2) quantities become conspicuous by the absence of data concerning them. While figure 131 furnishes an exact description of the water relations in one kind of organism under one set of conditions, the scheme of description is general, for it applies to any organism, group, or parts. From many such descriptions the range of values over which the quantities vary may be found. No other test of the generality of correlations with water load seems to exist. <§ 104. Proceduees or steps used The first step in any investigation is to select kinds of observa- tions or measurements that are to be made ; very often, of course, that is done subconsciously. The second step, grouping those kinds into classes of variables, seems to be more successful if done with all the consciousness that can be brought to bear. Thereafter, the correlations among classes of variables are made in the light of whatever hypotheses prompted the investigation, largely by stand- ardized methods. A set of measurements, tests, and conditions are chosen ; they are classified ; and the inevitable coincidences among them are ascertained. Any desired degree of completeness in characterizations may be attained, such that finer and finer con- trasts can be found among diverse species or individuals or states. The specific steps in arriving at this stage of physiological description of water relations may be set down as follows : (1) Select by some reproducible criterion a group of individuals or parts for experimental tests (U). (2) Select one or more modes of imposing water excesses or deficits in graded amounts (f ) . Fix the conditions for maintenance of the individuals and for recovery from the increments. (3) Follow in time (t) the changes of water content or volume (W or V), and hence the rates at which these changes occur (Rw)- Ascertain their deviations. (4) If desired, identify and measure separately the exchanges of water or volume in particular paths of gain or of disposal (p). (5) If desired, identify and measure separately other changes in the organism (s, M, N) at diverse water increments and times; changes of composition, structure, partition, rates, and frequencies 286 PHYSIOLOGICAL REGULATIONS of diverse processes, behaviors, and others. Ascertain their deviations. (6) Set down the quantitative changes that are measured and their interrelations, remembering that water content is the quantity whose maintenance is to be examined. (7) Note the unique and the general characteristics of each system (group of variables) described. (8) If desired, find whether some of the changes vary progres- sively with repeated imposition of the same water increment (ac- climatization, facilitation). There is nothing very peculiar about this procedure ; it is a par- ticularized and explicit form of any plan of experiment. It em- phasizes the fact that materials and conditions are selected, the precision of time and of quantities, and the notion that what the experimenter does is much the same, regardless of the hypotheses that he entertains. In a primitive stage of physiology, an investi- gator of water relations might stop with stage (3) or (4) or (5), and might be content with a qualitative answer to a single question. At present, the most complete quantitative description is required to answer any comprehensive question concerning water relations, forming a durable part of a maturer science of physiology. In the future, still further steps may be added. <§s 105. Outline of water relations Using the specific materials in earlier chapters, I can now sketch the investigation of water maintenance and its correlatives without the details. This recapitulation in general terms allows emphasis on uniformities found among many instances, and disregard of elements that vary with species or parts, and with conditions. At the outset, ivater increment or load was defined as the differ- ence of water content in the organism or part in two physiological states, namely, test and control. In general, control states were arbitrarily chosen, but a further definition of them was found in the existence of water balance (equality of intake and output rates) in what are believed to be standard, usual, resting, or occasionally basal, situations. Sometimes the biologist has difficulty in choos- ing ''natural" conditions for an organism or its isolated parts, but he can usually define the conditions existing. With water increment were correlated diverse variables that had been measured simultaneously or in other relation with it. Of "WATER BALANCES AND EXCHANGES 287 these the rate of water exchange was most frequently ascertained ; it seemed to be specially related (by virtue of dimensions and common component) to water increment and content. Sometimes the rate of exchange was known only as net flow; at other times both gain and loss were simultaneously measured, and either in total or in paths. The symmetry of gain as contrasted with loss was stressed. Water content was readily disturbed under controlled and mea- surable conditions by two general procedures : (a) stopping the con- tinual gain or the continual loss, (b) imposing extra gain or loss. Each load was initiated either gradually or suddenly, and then either released from further interference or partially continued, and often in such a manner that an approximately stationary state of water increment prevailed. Accordingly, the several rates of exchange might remain stationary within chosen limits of time ; otherwise the temporarily modified rates of water exchange were measured. It turned out that net rates of water exchange were markedly different at diverse water increments. Within the limits com- patible with life, net gains appeared in negative increments (water deficits) and net losses in positive increments. Hence the exchange was always of a type that dispelled the increment, thereby accom- plishing recovery, and compensating for the increment of water in the animal. Net exchanges were zero solely when intake equalled output, hence at water balance. At other contents high intakes accompanied low outputs, and vice versa; gains and losses increased together only in what were believed to be "forced" situations. When organisms or their parts were allowed to recover, the time courses of water exchanges were followed. Sometimes the initial rates of exchanges were the most rapid ones ; at other times delays occurred. As the increment diminished with time, the rates of ex- change were modified according to the relation of increment to rate in steady states. Eventually the net rates decreased to zero, at which time it could be said that recovery ceased. The return of water up to this time was then comparable with the increment originally present. Diverse types of water increment were produced by various means, recognizing specified conditions under which recovery occurred. In general, increments were absolute (such as are usually measured by changes of weight or volume), or relative to designated constituents (such as are usually measured by changes of a concentration). The same increment might be absolutely negative and relatively positive. 288 PHYSIOLOGICAL REGULATIONS Behavior toward water in environments was of two sorts : cer- tain animals were able to locate water, usually without seeing it; and certain animals frequented moist air instead of dry. In both sorts, the preference for water or moist air was exaggerated when the body was in water deficit. Statistically significant preferences for environments that tend to diminish water losses and favor water gain thus play a role in the maintenance of water content. Not all species are known to show evidence of preference for water in environment ; some may accept whatever comes, just as tissues and some parasites in situ do. Variabilities of water content and of water exchanges measured the net results of regulation. One individual undergoes continual fluctuation of each, but the narrow or wide limits within which it varies indicated how sensitive the system of organism plus environ- ment is toward increments of water. Frequencies of reversals, precisions of rates, and durations of movements, served to char- acterize the processes of maintenance. So long as the study of water was confined to the factors of increment and time, variables of four classes of dimensions were dealt with: content (AW), exchange (SW/At), time (t), and velocity quotient (1/At). For each species and set of conditions, certain numerical characterizations among these four were selected, which by their uniform relations allowed many comparisons. These were : Rates; turnover, initial, stationary, minimal, and maximal (tableslO, 13, 15, 21). Economy quotients (table 9). Modification ratios (table 11). Augmentation ratios (table 11). Tolerance curves (fig. 106). Tolerated loads (table 22). Variabilities; successive, individual, within species (table 12). Acceleration and decelerations of exchanges (table 14). Periods ; latent, recovery, to maximal rates (table 14). Half -life (any fractional-life) of load (§ 7, table 32). Precisions of turnover (§23). Completenesses of recovery or return (§7). Partitions of exchange (tables 10, 11, 20, 17). Equilibration diagrams (fig. 110). Usefulness of comparisons made according to these criteria is illustrated in the various tables. In general the comparisons drawn WATER BALANCES AND EXCHANGES 289 were among (a) species, individuals, tissues, and cells; (b) types and agents of increment and conditions of recovery; and (c) paths of water exchange. The preferred characterizations were such as were demonstrated to apply to very diverse sorts of living units, almost regardless of their special structures and functions. Less general criteria were available when additional correlatives were taken into consideration. Thus, all organisms exhibit water ex- changes after a water increment is imposed ; but only some exhibit exchange through integumentary paths, or in proportion to exposed area. TABLE 32 Adjustments of water load. G = gain, L = loss Species Dog Garter snake Eat Phascolosoma Earthworm Frog Man Eabbit Arhacia egg (data of Lucke) Zoothamnium (data of Kitehing) Net gain in deficit Hours for half return 0.02 0.04 0.15 0.4 1.0 1.4 1.6 0.04 0.1 0.02 Modifications G, (L) G G, L G G, L G, L G,(L) G, L G G, L Net loss in excess Hours for half return 1.5 4.0 1.8 1.5 1.6 2.3 2.1 1.8 0.04 0.2 Modifications G, L (G),L G, L L L L G, L G, L Within one class of animals showing structural similarities, e.g., vertebrates, water increments are accompanied by character- istic changes in particular tissues. In vertebrates, some of the modifications found are in : (4a) Tissue volumes, volumes of distribution; (4b) Tissue concentrations ; (4c ) Amounts of components in tissue and body ; (4d) Pressures within tissues. (8a) Rates of exchange of components ; (8b) Rates of transformation of energy; (8c ) Rates of circulation of blood and other fluids ; (8d) Behaviors and other physiological activities ; (8e) Responses to excitations, agents. The diverse sorts of modifications in tissues coincident with incre- ment in a single variable (AW) constitute a whole pattern of strains or displacements ; the organism is then out of kilter until recovery 290 PHYSIOLOGICAL KEGULATIONS has occurred. Of these correlatives all appeared to be equally significant, for no means was found of distinguishing direct ties or nexuses from others. All quantitative relations therefore equally characterized the water relations of an organism or its parts, and whatever was neglected through convenience or ignorance left the description of the physiological relations incomplete by that much. Possibly all of the changes are resistances to further increment and promoters of recovery. A considerable number of modifications in composition and metabolism were actually and quantitatively correlated with water increment (± AW) ; extensive data were for the dog, represented in figure 131. In that correlation, ± AW has the arbitrary status of an independent variable ; but the relations are reciprocal among all the quantities present. Finally, the role of forces in water exchanges, of bodily dimen- sions, of developmental ages, and of other factors of heredity and environment were considered. Each category emphasized par- ticular aspects of the adjustments of water content. § 106. Ufifokmities The qualitative results arising from the investigation outlined are indicated by the specific conclusions now set forth. Each is an induction from data already presented ; whenever exceptions to the statements made shall turn up, they will be noteworthy and useful. (1) The living units studied respond to each change of water content by (among other things) modifying their rates of water exchange. (2) The modified exchange is of such a sign that excesses are eliminated and deficits are paid off, when conditions allow. (3) Upper and lower limits of water content exist, compatible with continued life and observed activity; outside these tolerated loads, recovery of water balance usually does not occur. (4) The variability of water content of one individual in stand- ard conditions is related to the amount of modification of exchange rates that occurs per unit increment of water content (fig. 47). (5) Among differentiated channels or paths of water exchange in any species, gain by only one path and loss by only one path is markedly modified with water load (table 10). (6) In no case is water found to be manufactured from other chemical substances at faster rates in response to water deficits. WATER BALANCES AND EXCHANGES 291 (7) Time relations and rates of recovery may be similar even though paths and structures employed are diverse (fig. 111). (8) Rates of water exchange are faster with larger increments than with moderate ones. But the total times occupied in recovery (and the velocity quotients) are nearly independent of increment. (9) The amounts of water returned before the control rates of exchange again prevail, do not exceed the water loads administered (figs. 1 and 49), when compared with equally timed control individuals. (10) Recovery is faster, and often more prompt, in negative increments of water than in positive (fig. 106). (11) Species or units that eliminate excesses rapidly, also pro- vide for restoring deficits rapidly. Though diverse organs be con- cerned in gain and in loss, the two are proportioned to each other. (12) Tolerance curves are usually such that net exchanges are most rapid in the early portions of recovery, diminishing as loads are discharged. Most loads are therefore exponential with time. (13) High rates of turnover accompany high variabilities of water content at successive times (table 12). (14) Intake by mammals is usually more variable in periods of hours than in periods of days. Output, however, is equally variable in periods of diverse durations (table 12). (15) No one kind of change other than augmentation of water exchanges, is known to occur in all organisms studied, in invariable correlation with increments of water content. (16) But in vertebrates, for instance, certain changes of com- position, rates of other exchanges, activities and behaviors are found to accompany water increments under specified conditions (table 29). (17) Maximal rates of urinary output are nearly proportional to body weight, among the species of vertebrates studies (table 13). (18) Evaporative loss is a larger fraction of total loss in small mammals than in large ones. Oxidative gain is a larger fraction of total gain in small mammals than in large ones. (19) Rates of exchange of water are among many animals pro- portional to the 0.6 to 1.0 power of body weight (figs. 133 to 140). (20) In mammals in which turnovers in water balance have been measured, the maximal rates of intake or of output observed (fig. 138) are 20 to 30 times the rates of turnover in the respective individuals. 292 PHYSIOLOGICAL REGULATIONS (21) Animals frequent environments that furnish water to them, or that minimize loss of water. Both behaviors are exag- gerated in water deficit. (22) In water increments of the whole body, plasma is diluted twice as much as the whole body (dog). Whole blood dilutions are less certainly correlatives of load. (23) Volumes of distribution of several injected distribuends also change markedly more than does the whole body (fig. 114). (24) Most tissues analyzed share in the distribution of excesses and deficits of water, at least in four species (table 29). In diverse species no one tissue preserves its water content more consistently than others. No outstanding depots of reserve water have been identified. (25) Hence very many living units, whole individuals, their parts and aggregates, of diverse species, may be studied with re- spect to water loads ; and similar patterns are found in their main- tenances and recoveries of content. These conclusions are arrived at from correlations which are explicit here. Further tests of their generality lies in additional information. At present none is known to me that contradicts any one conclusion. But generalizations are always provisional; it is not improbable that organisms exist which fail to fit some of the specifications of those studied thus far. It does seem to me im- probable that any species of organisms exists whose individuals are wholly without means of adjusting their water contents ; and regulation of this function may be one of the many requisites for survival. Quantitative conclusions do not lend themselves to statement in words ; they can be fully reviewed only by reexamining the diagrams and tables. § 107. Diversities Comparisons among species and among parts of organisms that were modified in water loads, revealed features of difference that are not known to be mere corollaries of structural difference : (1) In frog, earthworm and ciliate (in fresh water) the rates of water gain are no less in water excesses than in turnover. (2) Some whole aquatic animals (Phascolosoma, Arbacia egg) manifest no turnover of water ; in them recoveries of water content are accomplished through exchanges of novel sorts, instead of through quantitative modifications in paths already operating. WATER BALANCES AND EXCHANGES 293 (3) Many diversities are related with body size, age, surface of exchange, possible forces of exchange, structural differentiations, and other features of the units whose content is being equilibrated. (4) Though paths of water exchange differ, no one kind of path is regularly (inherently) slower or faster than other kinds, or its exchanges less or more accurately related to water load. (5) Species and living units may be arranged in various series, depending, for instance, on velocity quotients in recoveries. For gains the order of rates is : dog, snake, rat, plasma, Phascolosoma, earthworm, man, frog, muscle. For losses : plasma, dog, rat, earth- worm, Phascolosoma, man, frog, muscle, snake (figs. 106 and 101). Such series are often qualified by the fact that water excess was produced in each unit (dog, plasma, frog, and muscle) by slightly different procedures (administration by stomach, injection of citrated blood by vein, injection by peritoneum). Further restric- tion in type of load, however, would not permit comparison of whole rabbit with its plasma or its leucocytes ; even injection or previous osmosis does not allow, in diverse units, identity of loading. As in all scientific investigations, tentative comparisons are drawn, pro- visional generalizations are obtained, without assurance that an- other generation of physiologists will find as much satisfaction in these particular relations as in some others. The above diversities among species are the materials for study of the comparative physiology of water relations. All the quantita- tive materials of tables 8 to 16 are the outcome of that study, which can scarcely be summarized more briefly. Equilibration diagrams, having a uniform plan, fall into classes according as one or two kinds of exchanges are modified in positive and in negative load (fig. 112). Their quantitative study tells how much modificatioil in each kind occurs, perhaps as evaluated by augmentation ratios. Rough comparisons of water exchanges have been made in the past, but those with which I am familiar do not appear justified. ' ' The amphibian is unable to preserve his water content . . . inde- pendent of that of his outer world," says Cannon ('32, p. 283). That is no distinction of amphibia or of aquatic animals, for all spe- cies need access to water at some times. Even amphibia do not wait for water to come, but seek it (Czeloth, '30), specific behaviors being indispensable means of preserving water content. ''The urine out- put of most dogs is less readily increased on water ingestion that is the case with man," imply White and Findley ('37b, p. 747). 294 PHYSIOLOGICAL REGULATIONS That is contrary to data presented above. And so for further statements, including some that I have made. There is no evident hmit to the variety of ways in which living units adjust their water contents. Varied though the compensa- tions and behaviors contributing to water adjustment are, the out- standing fact is that every unit that has been investigated gives evidence of special activities, one of which is modification of water exchanges, that result in maintaining water content more constant than when those activities are experimentally prevented. The apparent relations among those activities and modifications, both qualitatively and quantitatively, is set forth in a manner that coordinates the materials of the comparative physiology of water. § 108. Agents and types op load A great many agents regularly modify water relations, whether they are widely recognized as ''hydrators" and "dehydrators," or not. The measurable water increments (AW and AV) observed in animals under the influences of some few of these agents were already studied. But other agents abound, and many of them would scarcely be proclaimed as producers of AW or AV. Ex- amples are : diuretics, diaphoretics, cathartics, blood substitutes, secretagogs, hot atmospheres, and low oxygen tensions. Whatever designations such agents may have, a common property is to influence water content or volume and water ex- changes. Hence the modifications in the body and its parts at any water load may be compared quantitatively in the same manner as above. The effects of several agents may thus be characterized with respect to any and all the modifications of volumes, concen- trations, and metabolisms that can be measured. Something of this sort has been done for pituitary extracts, for instance. With respect to water output after water administration, the times to maximal rate of urinary output, or the times for elimi- nating half the water increment (mid-excretion point, index of diuresis, half -life) are compared (Burn, '31; Heller and Urban, '35; Martin and Herrlich, '39). Equally useful in those assays would be any of the parameters listed in § 105. Particularized tests of the organism's status with respect to water may be useful in the comparison of all conceivable agents, whether those who make the tests are interested in the agents or in the organisms. Such would be: {!) Total water output, or total WATER BALANCES AND EXCHANGES 295 intake, in 1, 2, or 4 hours. (2) Volume of distribution of the dye T 1824 at 1 hour. (5) Clearance of inulin or urea through kidneys at 1 hour. (4) Specific gravity of urine at 1 hour. (5) Rate of flow of saliva at 1 hour. Most of these tests mean more if the same individuals are similarly measured in control periods. Or, the problem might be to test the "efficacy" of blood substi- tutes. For that purpose, measurements of SV/At after the infusion of standard volumes of each may be selected. In order to charac- terize further the maintenance of blood volume, I set up the hy- pothesis that the responses to subsequent hemorrhage have more to do with blood volume than the responses to physical exercise have. Hence, the investigator of physiological patterns compromises be- tween (a) being overwhelmed by the semi-infinite number of mea- surements that could characterize them, and (b) using hypotheses and other factors of convenience to select a few quantities that can be measured with accuracy enough to describe the peculiar pattern of the organism's states and responses. A substitute is ideally a com- plete replacement for the naturally occurring material. For blood, the substitute is such that among other things it becomes subject to at least some of the regulations and maintenances that preserved the original volume. Only in that way does it continue to fit into the living unit of the blood and the living unit of the whole body. While the influences of agents and conditions can be studied by means of tests on single individuals or their parts, comparisons of those influences among species and parts seem to me to require gen- eral modes of characterization. AVhere feasible, permeabilities and reluctivities may be computed under specified conditions ; these take into account the areas serving for water exchange and the pressures effective in moving the water. Where areas are un- known, other coefficients may serve. Where pressures are un- known, rates of water exchange may be compared at like water increments. ''Like" water increments may be those having equal ratios to body weight (as used in this study) or having any other common denominator that someone selects. The investigation of agents and types of water load therefore reduces itself to the repetition of one or many arbitrarily selected tests. But knowledge that the test is one aspect of a pattern of regulatory modifications of the organism, aids considerably in un- derstanding what the agent does to the organism. It is enhancing or inhibiting some identifiable portion of the living unit's adjust- ments of water content. 296 physiological regulations § 109. General theories op water constancy A number of theories have been proposed concerning water in animals, as they have concerning every other aspect of organisms that has been observed, aiming to picture in one way or another the maintenance of constant water contents, (a) One theory states that forces related to protein constituents of tissues hold water there. ''Absorption of water by muscle is determined in the main by the state of the colloids contained in the muscle" (Fischer, '10, p. 74). (b) Another suggests that the water content of tissues is governed by the ratio of cholesterol to phospholipid in them (Mayer and Schaeffer, '14). (c) Again, hydrostatic pressures and partial osmotic pressures may be balanced when water is distributed equitably between a unit of tissue and its surroundings (Starling, 1896; Schade, '27). (d) Or, perhaps the total quantities of solute present in the whole body dictate the quantity of water held (Gam- ble, '29). Emphasizing water exchanges, (e) Rowntree ('22) said, "The total output of water is determined by the total intake," and further, ' ' The need of the body for water is determined largely by environment and metabolism." (f ) Reversing the terms, Richter ('38) stated, "Maximum intake may be determined by the maxi- mum capacity of the kidneys," while "it would seem likely that the maximum output is determined by the total fluid capacity of all the cellular spaces of the body." (g) Very often it is supposed that the kidneys watch over the water content of the body, while nothing is said concerning organisms that have no kidneys. These last two hypotheses seem to have been framed to apply to certain mammals. Explicitly limited is (h) the inference of Babcock ('12) that "the water requirement of mature animals that excrete urea, when at rest, depends chiefly upon the amount of digestible protein con- sumed." (i) Adolph ('33) thought the requirement of mammals might be predicted by adding together the factors of nitrogen, salt, and energy metabolisms. These last two ignore the general ex- perience that some other investigator will experimentally raise another factor to first magnitude, requiring another statement to express the relationship. In general, all these and many other views may partially cor- respond to facts, but not in the manner hypothesized, for all are aspects of a large picture that I believe is one of numerous interrela- tions. There is much evidence that no factor D is invariably cor- related with water content W, and much that many factors E, F, G, WATER BALANCES AND EXCHANGES 297 etc., are often correlatives. Any of the above theories may be use- ful as working hypotheses ; they have suggested experiments and measurements. But to suppose that a complete and permanent understanding of water regulation will be obtained by pushing the responsibility for regulation of A upon B which in turn needs regu- lation by C, and so ad infinitum, is encouraging an endless game of shuffling the factors. All theories actually take the form of mak- ing variable A dependent upon variable B. Sooner or later some other investigator suggests that B depends on A. Evidently no one theory satisfies either the data or the investigators. Eventually there arise one or more inferences concerning each of the many relations that are likely to be established. I know of no theory specifying the forces concerned in water maintenance that applies to all organisms and living units. I share the opinion that forces balance across the boundaries of all living units in stationary states. I too recognize partial similarities be- tween some of the sorts of forces present in living units and those in non-living systems. I see little probability of identifying all the forces present and assigning a proportionate role to each; or of separating forces concerned in water balance from those related to other components. And I see no way of feeling completely satified even though categories are assigned to each force present. The theory that all organisms are set to adjust their water con- tents by modifying their rates of water exchanges, appears to me to be as valuable an inference as any. For the cases investigated it has changed from a theory to a rule. The pattern in which the organisms are set, and the quantitative features of their adjust- ments, are revealed by measuring further the modifications of composition, of metabolism, of behavior, and of variability that coincide with diverse water increments of the organism. Very possibly there is no ultimate ''determinant" of water con- tent, short of the whole living organism and its environment ; for anything less is a very partial account of the adjustments con- cerned. If so, then whatever all living units have in common, all water regulations also have in common. In other respects very diverse relations may be shown, and no theory general enough to preclude other theories is to be expected. <^ 110. Summary The course of the investigation may be described as follows. At the start, particular procedures (especially the emphasis on 298 PHYSIOLOGICAL REGULATIONS description) were selected. Then data concerning water incre- ments and exchanges in the dog were studied intensively, and particular correlations between them were specified. After like data in other species were added, comparisons among them were drawn, together with a number of generalizations. Further, it be- came evident that parts of organisms, even single cells, could be treated in the same manner as whole individuals ; for each unit pre- served and recovered volume. The variable labelled water incre- ment was kept throughout, while diverse changes of compositions, metabolisms, and other physiological activities were correlated with it. Finally some general properties of organisms having increments of water or volume were enumerated and quantitatively compared. It might be asked: why was the maintenance of water in the whole body studied, instead of maintenance of water in the blood plasma, whose role as the ''internal medium" of the organism has been emphasized since Bernard (1859, p. 42) first pointed it out? I think it is clear that exchanges by gains and losses, behaviors toward environments, successive variabilities of content, and correlated events, are capable of much more complete and accurate study on the body as a whole. For, water could be measured as it went in and out, it and other metabolites could be located, sensory phenomena could be studied, and fluctuations from hour to hour could be evaluated for the whole; whereas for the plasma alone little quantitative accuracy has been achieved and few exchanges are known. These are the advantages that carried the investiga- tion of water regulation beyond the stage at which Bernard left it. Many more properties of water-loaded organisms will come to light when further observations are added, for no one investigation is likely to exhaust the possible interrelations. The pattern of interrelated quantities that has been ascertained serves provision- ally to describe the regulation of water content in the organisms studied. Once it is known what features are characteristic when water content is regulated, it becomes worthwhile to see whether other components of organisms are regulated in similar ways. Part B REGULATIONS OF SEVERAL COMPONENTS AND IN GENERAL Chapter XIV HEAT '^ 111. Introduction The investigation pursued thus far has presented, both in detail and in general, the physiological relations of water and of certain volumes. The relations are such as are known to be implicated in the maintenances of water content and exchanges in living units. The patterns found might be peculiar to water, or else some features might be common to many components. To find which, it is necessary to study other physiological components, and in con- siderable variety. To this end, the term component is not limited to chemical entities, but is extended to include all sorts of quantita- tive properties of living units. Perhaps anything that can be measured in the organism may be treated in a fashion analogous to that in which water and volume were treated. That too can be ascertained only by trying. The investigation now becomes an inquiry into maintenance and recovery of various functions of organisms, and ultimately of the interrelations among maintenances. The components chosen for intensive study (chapter XV) may or may not turn out to be a ran- dom sample of the very large number that can ultimately be stud- ied. Even though future physiology may discover types of regu- lation not yet observed, the relations described herewith prevail in the situations named. Certain choices could be exercised among the data available. All the components might be selected for one species (dog), or for one type of production of increments (exercise by running), or for one path of exchange (urinary). Instead I propose to utilize data that seem to me most adequate, distributed quite at random in these and other respects. In the end the study yields a partial picture of a block of physiological patterns that operate together in one individual. Heat is the first component to be considered, and the species chiefly studied are man and rabbit. How are heat exchanges and other properties modified when a man is warmer or cooler than usual? What is usual? 301 302 physiological regulations § 112. Maintenance in man The recognition of constancy of heat content of the human body is almost universal. Perhaps this content is measured, as body temperature, oftener than content of any other component. How are heat exchanges related to heat content? This might be answered by procedures familiar in the study of air-conditioning of a house. Observations of the thermostatic arrangements are alone not sufficient ; the behavior of the installed apparatus in the house must be examined. The whole house would itself be exposed to diverse external conditions when performance is being de- scribed. In the human body a similar study is a desirable pre- liminary to characterizing a state of fever, of hypothalamic lesion, or of heat stroke. That heat gain in the long run equals heat loss may be regarded as a prediction of the first and second "laws" of energy. None the less, it was accepted as a singular victory for thermodynamics whenEubner (1894) in the dog, and Atwater (see Lusk, '33, p. 120) in man, ascertained that the oxygen consumed and the carbon diox- ide produced in 24 hours of metabolic transformations actually yield energy equivalent to the amount of heat collected in a calorimeter. The known course of energy transfer predicts that heat is scarcely transformed into any other kind of energy without the presence of large differences of temperature. Of all forms of energy, therefore, the chances of completely measuring heat output are greater than of measuring other outputs. Turnover of heat in man amounts to about 1 Cal./kg. hr. in usual circumstances. Further, a huge number of factors have been identified as having some influence upon it. The balanced state of the body is any one in which rate of heat gain equals rate of heat loss. Though the actual heat content of the body might be unusual, rates of both total exchanges are then correspondingly greater than basal ones, and an unusual balance prevails. Hence it is possible to distinguish a usual balance from others. Usual balances are characterized by the presence of those rectal or oral temperatures found in a random population. The varia- bilities of these temperatures at a uniform time of day in a given set of conditions (table 33) are indicated by standard deviations, HEAT 303 1 i^ 1=1 1 Ol -5 s in CO "^ CO C«5 o !M eg ^N c ^ ci _• a 05 CO ^ §00-^ Oi «o tH r^ CO ., tn O CO r-- OJ + -5* ^CO t~ -^ rt'« P ro ° O c a ph i 2 oT ^2; Ah O « o s» W O 304 PHYSIOLOGICAL REGULATIONS in both dogs and men, of about ± 0.20° C. It is well known that a daily cycle shows changes of more than this. If the individual were exposed to random conditions instead of selected ones, greater or smaller variations might be found. In ascertaining a man's body temperature one virtually asks whether his body is keeping within two or three times the standard deviation for the popula- tion. If not, two further questions arise : is he out of balance, or, is he maintaining an unusual (non-basal) balance? Both ques- tions can be answered either by repeated temperature measure- ments that indicate whether net gains or losses of heat are occur- ring, or by complete calorimetry of both heat production and heat loss. It is amazing that few human individuals have been discov- ered who consistently differ in rectal temperature from the mean of all individuals. But the hourly variations in one individual are greater (standard difference ± 0.13 to ± 0.28° C.) in newborns than in adults {± 0.10° C.) (Eaudnitz, 1888). It may be noted that in one individual on different days the rate of heat production under ''basal" conditions fluctuates by coeffi- cients of variation of ± 3.5 per cent (table 33). Among individ- uals of one body weight and height the coefficient is ± 2.9 per cent (Berkson and Boothby, '38), values which show that unknown cor- relatives are insignificant in subjects selected as "normal." The maintenance of constant heat content may depend upon either (a) continual equality between gain and loss, or (b) succes- sive compensations of one by the other. If gain and loss differ by more than the errors of measurement, within periods of (say) one hour or less, it is natural to ask, which of the exchanges has the more uniform rate, and, does the uniformity of rate indicate a less pliable process? Quantitative answers do not appear to exist for man, but will be given below for rabbit. § 113. Heat loads (Max) Departures from the usual balanced state are produced by acci- dental or deliberate changes in body and surroundings. The in- itiation of heat deficits by cold media was systematically studied by Lefevre ('11). A man is placed in a water calorimeter (his p. 94) or in a wind calorimeter (p. 103), and the rates of sensible heat loss (non-evaporative loss) are ascertained (table 34). In water, as much as 5.7 Calories/kilogram of body weight are lost in 0.2 hour. During the exposure the rate rapidly diminishes until a sta- HEAT 305 tionary rate of loss, 21.7 Cal./kg. hr., prevails. This represents an increase (augmentation ratio) of about 22 times the control rate of heat loss. The resistance to heat loss might be computed in arbitrary ways ; in any case it changes during the loading. By a usual defi- nition the resistance is the reciprocal of rate of loss per unit tem- perature difference between inside and outside. Lefevre (p. 498) used as '^ coefficient of resistance" the ratio of heat production to heat loss, but this is the definition of the economy quotient (§13). TABLE 34 Bates of heat loss during loading, Cal./lcg. hr., in various species; indicating how quicMy deficits of heat may be imposed. Data of Lefevre {1897) Species Body weight, kg. Rate in water at + 5 °G. Extreme rate in other conditions Rate of turn- over Maximal augmen- tation First minute Sta- tionary Conditions First minute Sta- tionary ratio, station- ary Dog Man Monkey ... Pig, small ... Rabbit ... Guinea pig Chicken 8.0 64.5 2.7 4.2 4.5 0.7 2.5 2.3 72 97 169 255 51 174 77 72 25.0 21.7 50.0 63.0 17.4 67.0 36.0 31.0 -t-l° C. air + 5° C. air -8° C. water + 5° C. air -18° C. water 267 96 6.4 4.4 82.0 8.7 33.0 2.0 1.0 4.0 4.0 3.4 3.9 3.0 3.0 12 22 20 16 10 17 12 Duck 10 Negative heat loads of other types are produced by rectal irriga- tion, or by drinking ice-water, or by surrounding the body with ice. All are like baths in eliciting local deficits of heat that are later distributed. Positive heat loads are rapidly acquired in physical exercise, diathermy, intense radiation, stoppage of heat losses, and other circumstances. For accurate measurement of the heat load, not only the total heat lost but also the total heat produced during transition is ascer- tained, for in any transitional state the load is the difference be- tween the integrated gain and the integrated loss. Alternatively, heat load is ascertained as the product of increment in tissue tem- perature and of heat capacity. In the whole man, various tem- peratures appear in diverse portions of the body, and approxima- tions of their distribution (Burton, '35) are necessary to evaluate the heat content of the body. This measurement is less practical 306 PHYSIOLOGICAI. EEGULATIONS than the first, both during periods of rapid change, and especially where one volume of tissue (superficial) is being cooled while another volume (deep) is not being cooled. Acclimatizations have been recognized, whereby exposure to given conditions yields smaller loads after repetitions than before. Changes of body heat content are thereby resisted. A man im- mersed in water at 5° C. for a few seconds subsequently diminished the rectal temperature by 1.0° C. After 3 months of training, twice as long an exposure secured less decrease of rectal temperature (0.1 to 0.7° C). Less heat was also lost in the later exposures (Lefevre, 1894). In periods of physical exercise in hot atmos- pheres, progressively faster losses of latent heat occurred, with smaller accumulations of heat in the body, during the first ten days of life in the desert (Adolph, '38). Even basal heat production appears to become slightly slower every summer (Gessler, '25), when the organism partially transforms itself from a frequent regulator for deficits to a more frequent regulator for excesses. All these are acclimatizations. It might be possible to derive an '' absolute" value for the heat content of man; this would ideally be the total heat lost in cooling to absolute zero, a determination not likely to be made. Such a value would be useful in comparing heat increments with incre- ments of other components, for all changes of heat content might then be fractions of the content that prevails during heat balance. Some value for this content (in Cal./kg.) could be assumed, such as 309° X 0.83 X Bo, the coefficient 0.83 being the mean specific heat capacity of human tissues, and 309° K. the usual average tempera- ture of the body. Assumptions concerning heat capacities may if desired be avoided in studies solely of net exchanges, since the rela- tions here considered hold equally if AT (increment of mean tem- perature) is substituted for AH (increment of heat content per unit of body weight). Rates of change of either local or integrated tissue temperature (BT/At) then replace the rates of heat exchange (SH/At). In brief, heat loads may be estimated with almost the accuracy of water loads. Special circumstances of heat production and heat distribution are recognized as opposing the experimental establish- ment of either deficit or excess. The oppositions and the sharp discomforts incident to heat loads are evidences of the narrow range within which heat content is usually preserved. HEAT 307 ^ 114. Recoveries (Man) In tests now to be considered, excesses of heat develop during physical exercise in hot atmospheres. Thereafter the net rates of heat loss are studied during later recovery indoors, where no solar radiation prevails and the air is cooler (fig. 143). In the first 0.3 hour the net loss is slow, heat production probably decelerating gradually ; thereafter load diminishes with time up to 1.0 hour. In other tests, deficits develop during the drinking of ice-water, and other adjustments are occurring initially (fig. 143). 0.2 0.8> 1.0 0.4 0.6 Hours Fig, 143. Heat load (increment in Calories per kilogram of Bo) in relation to time during recovery. Heat loads (increment of rectal temperature x 0.83) were imposed either by walking in the hot desert or by drinking ice water; recovery consisted in resting indoors in an air temperature of 30 to 31° C. Several tests that started with about the same load are averaged, the number of tests included being indicated. Five subjects (A, D, H, K, P). Further data of Adolph ('38), and of Pinson and Adolph ('42). The rates of net exchange are clearly related to the diverse heat loads (fig. 144), gain of heat augmenting in deficits and loss aug- menting in excesses. In still other tests total exchanges are measured instead of net exchanges. After moderate cooling (fig. 145) the rate of heat pro- duction is no greater than before ; shivering is usually absent. In this and similar tests, therefore, recovery occurs by temporary diminution in rate of heat loss. I draw the same conclusion from 308 PHYSIOLOGICAL REGULATIONS -12 -10 -8 -6 -4 -2 0 Hea+ Load Fig. 151. Eate of heat production (Cal.Ag- hr-) in relation to heat load (Cal.Ag.) in rabbits of diverse ages. Each point represents several individuals that cooled during the 2 to 3 hours required for the measurement of oxygen consumption, except in group Gr where rewarming was proceeding. Points are roughly connected in successive age groups designated A to F. Ages are numbered in days after birth; from data of Ging- linger and Kayser ('29, p. 740). Additional adult curve (G) from figure 149 (data of Gasnier and Mayer, '35). Horizontal lines indicate range of heat loads prevailing during the measurements, as computed from rectal temperatures (T°) alone; (T°-39°) xO.83, duction and heat loss are augmented in heat excesses brought about by confinement in warm air. Thus, although the turnover of heat in rabbit is thrice as rapid as in man, no marked differences are known in the variability of heat content and of heat exchanges, nor in the modifications of heat exchanges in the presence of heat loads. But in rabbit variations 318 PHYSIOLOGICAL KEGULATIONS of exchanges are more fully known, and extreme deficits of heat have been examined. Heat regulations are acquired after birth in the rabbit, neither gain nor loss being subject to effective self- modification at birth. § 116. Comparisons among species The properties of the equilibration diagram for heat in man and rabbit having been established, a few quantitative diversities in other species and a few general features may be mentioned. In Hours Fig. 152. Course of heat load during recovery, in various species of vertebrates. Eecovery began at zero time, after heating or cooling by various means. Heat load = ° C. difference from control x 0.83. Turtle, Gopherus, 2.4 kg., new data. Snake, Boa, 6 kg., data of Benedict ('32, p. 68). Lizard, Heloderma, (Gila monster), 1.2 kg., new data. Man D, Homo, 77 kg. from figure 143. Man P, 77 kg. new data. Mouse 122, 0.040 kg., and Mouse 118, 0.023 kg., Mus; data of Chevillard ('35, p. 1046). Eabbit, Lepus, data of Gasnier and Mayer ( '35) ; cf. figure 149. the actual tests an individual during loading accumulates heat or cold faster or slower than usual ; thereafter initial or control con- ditions are restored, allowing heat exchanges to proceed as they will (fig. 152). The selection of control conditions by experi- menters is arbitrary ; tests which allow the animal to select its envi- ronment from among a number available (Fraenkel and Gunn, '40) are to be recommended. Certainly environmental conditions affect heat exchange, just as they do water exchange. Since no HEAT 319 method is possible of doing away with the environment, the organ- ism's provisions for dealing with heat loads are, I believe, mani- fested most clearly when the environment is uniform throughout one series of tests. In a more extensive inquiry, many quantita- tive relations would be established for one organism recovering in each of many environments. In daily life the organism is continu- ally shifting its rates of heat exchange to cope with complicating circumstances as they ari^e. Most organisms (poikilothermic) attain temperatures approxi- mating those of the environment or parallel to them, and often they show curves of cooling or warming partially similar to those of inanimate objects (fig. 152). In others (homeothermic) the net exchanges are sometimes opposite in direction to that toward cus- tomary inorganic equilibrium. In poikilothermic species rates of heat exchanges are still of consequence in maintaining a gradient of heat. Terroine and Trautmann ('27) ascertained the relations be- tween rates of heat production and air temperatures in each of 12 homeothermic species. Each showed a minimal rate at some inter- mediate (''neutral") temperature. The rates could now be re- lated to heat loads if data upon mean body temperatures had also been obtained. Each of those 12 species has its own relation be- tween augmentation of heat production and increment of tempera- ture, some species being more responsive than others to equal heat increments. That might be a basis for visualizing a phylogenetic gradation of regulations. Maximal rates of heat production in heat loads divided by minimal rates are ratios of modification. Under the name of "metabolisme du sommet," they were found by Giaja ('25a) to be 3.2 to 4.2 in the mouse and four other species, but by Herrington ( '40) only about 2.2. Even in monotremes, in which the heat con- tents are irregular, augmentations of that order are found in low air temperatures (Martin, '02). It may be said that the mono- tremes' inexact maintenances of body temperature are not the out- come of mere inability to modify their heat productions. Portions of further equilibration diagrams might be established for other species, particularly the dog (Wada et al., '35 ; Heming- way, '38; Shelley and Hemingway, '40) and the mouse. The latter species recovers from temperature deficits more rapidly than larger mammals (fig. 152). During recovery, heat is produced by 320 PHYSIOLOGICAL REGULATIONS oxidation much faster than in heat turnover, and heat is lost mark- edly slower (Chevillard, '35, p. 1046). It is said that heat may be lost faster in the dead mouse than in the living one (Pincus et al., '33). Control of the mouse over its temperature develops during the first ten postnatal days (Pembrey, 1895; Sumner, '13; Ging- linger and Kayser, '29). In the 2-day-old mouse compensatory heat productions and their accompanying muscular movements CPincus et al., '33; Stier, '33), are maximal in environments of 22° C; in still cooler ones the recovery becomes slower and less successful, just as in the rabbit (fig. 149). Guinea pigs and some other species, in contrast, are at birth fully armed with compensa- tions for heat increments. Coefficients may be computed and compared in order to charac- terize the convection systems, heating systems, and refrigerating systems of animals. They are convenient parameters of conduc- tion, permeability, emissivity, vaporization, radiation, and other quantities. Some of these coefficients are constant under limited conditions. But the so-called physical and chemical processes in- volved are by no means fully identified ; so far as I can ascertain, there is no relation between the constancy of a parameter of heat exchange and the fact that an imputed process in it has a name. All the methods of comparing water exchanges appear to apply also to the study of heat exchanges. Among species of diverse body sizes and shapes, rates of heat production in turnover may be set forth (Lambert and Teissier, '27; Benedict, '38, p. 131), and related to supposed body surface area and to many other quanti- ties. Velocity quotients, modification ratios, economy quotients, tolerated loads, tolerance curves (fig. 152) and equilibration dia- grams (fig. 150), are all applicable for comparisons. These quo- tients, ratios and diagrams would be required for thorough charac- terizations of modifications in thermal adjustments of animals such as are produced in diverse lesions of nervous tissues (Brooks, '35; Clark et al., '39) and many other states. Moreover these quanti- ties in heat exchanges are directly comparable with the correspond- ing quantities in water exchanges and other exchanges (see table 40), for most have the same dimensions. Differentiations may be made between increments of heat and changes of heat balance. For, balance having been defined as a point where gains equal losses, all the criteria of water balance apply to heat balance. Heat load (AH) without change of heat HEAT 321 balance (Ho) does not stay constant in any circumstances that al- low of recovery. In practice, endless arguments arise as to whether in a state of hyperthermia (fever) or hypothermia (1) the "thermostat" of the organism has changed its setting or {2) there is an aberrant imbalance between gain and loss of heat. Once criteria such as those suggested have been adopted, the question can be answered by calorimetry in every instance. By speculation, one chooses between (la) assuming the existence of a second regu- lator to ''reset the thermostat," which presumably means there are two or more superimposed thermostats, and (lb) regarding a shift of heat balance as a concatenated member of a large group of interacting quantities. Enough has been set forth to indicate that investigation of heat regulation yields the same sorts of quantitative relations as does water regulation. In particular, the tolerance curve and the equilibration diagram relate the rates at which restoration occurs to each increment of heat load. Evidently this sort of physiologi- cal pattern is not limited to biochemical entities such as water that have particular kinds of absorptive and emunctory organs at their disposal. Heat is an entity of a different category, with entirely different arrangements physically, anatomically, and physiologi- cally concerned in its exchange. When similar patterns emerge among supposedly dissimilar processes, it becomes profitable to take the tentative view that diverse kinetic equilibria can be studied independently of the specific forms of machinery present. § 117. Summary Maintenance and recovery of mean body temperatures (heat contents) may be described by means of relations of heat exchanges to heat contents. Partial data exist for man, rabbit, mouse, and some other species. Variability of heat content may be viewed as a resultant of the variabilities of heat gai^is and heat losses. There are slight indications that heat losses may be more con- cerned in compensating for the unsuitabilities of heat gains than vice versa. Further, losses by vaporization may through succes- sive modifications compensate for ordinary fluctuations in sensible losses, and vice versa. Equilibration of heat content may be characterized in the same ways as the air-conditioning of a building that has an installed thermostatic equipment. Variabilities in diverse conditions, and 322 PHYSIOLOGICAL REGULATIONS modifiabilities in diverse states of heat load, are described. Matur- ing of the animal's heating and cooling system during early post- natal life, and changes of the system after experimental lesions, may be similarly characterized. Diversities among species are related to body size, supposed phylogeny of temperature maintenance, the expenditure of water for cooling, and the use of oxidations for warming. Quantitative differences consist in the relative modifications elicited in heat gains and heat losses by a given load of heat. The relations of heat in certain animals may be described in the same terms as the relations of water. But whereas mature mammals and birds manifest fixed patterns of compensation and behavior toward heat, many other animals are indifferent to heat until extremes are reached, whereupon they shun environments that impose those extreme increments of heat content. Both kinds are successful in preserving heat content, but those with the nar- rowest deviations of content (homeotherms) also modify a larger number of interrelated variables in accordance with content. Chapter XV DIVERSE COMPONENTS § 118. Total substance A variety of properties and constituents of organisms are now to be considered with respect to maintenance. Total bulk of the body is one. Are deficits of body substances compensated by in- creased intakes, or by decreased eliminations, or by both? In studying the maintenance of total body weight, all substances, in- cluding water, are measured in the exchanges. 04- UJ 0.3- 0.2- 0.1- - 1 1 1 1 ^ \ nta kes 1 / - ^-^ B •^/ / - A /^^^ Output^ Rabbit - \ 1 1 1 1 -12 -10 -8 -6 -4 Weiaht Load -2 Fig. 153. Eate of weight exchange (% of Bo/hour) in relation to weight load (% of Bo). Eabbit. Two deficits of body weight were previously produced by partial (A) or complete (B) privation of food while water was available. Each point represents one day during progressive recovery of 6 or 8 rabbits that were allowed a uniform food ad libitum for 0.25 of the day. Rate of output was computed as rate of intake minus rate of net gain in weight. Data of Maclagan ('37). Data for the rabbit record total intakes, and relate the intakes to increments of weight. Elimination rates are computed from intakes minus weight increments (fig. 153). Deficits are estab- lished by complete or partial abstinence from food (either includ- ing water or not), excesses are established by overfeeding or forced intake. The relations shown are far from familiar to all who eat. Ee- covery from deficit in the rabbit may occur not by augmentation of 323 324 PHYSIOLOGICAl. EEGULATIONS intake but by suppression of output. However, in other tests, moderately increased rates of intake are recorded by Maclagan ('37). In another species, the dog, augmentations of intake occur after each deficit (fig. 154) ; in still another, the rat, food missed is never made up (fig. 155). Recoveries of net substance may be ascertained from body weight alone (fig. 156). In all four species reported, the curves are exponential and asymptotic. The velocity quotients are here in- Days Fig. 154. Increment of body weight (% of Bo), and relative rates of intake, in relation to time during and after (A) total privation for 48 hours, (B) forced intake by stomach. Dogs vrere allowed unlimited dry food and water except during the 48 hours. Amounts forced are indicated in block at lower right. Four tests on 3 indi- viduals (A) and five tests on 2 individuals (B). New data. verse to the initial (administered) loads, and there is nothing to indicate that they vary from species to species. Even when fed limited amounts of a constant diet, animals gain body weight faster in the initial part of recovery from food deficit than later, this gain corresponding to the greater retention (or the lesser elimination) of substance (dog, Morgulis, '28). Hence re- covery involves here, and in other known instances, decreased rates of elimination ; anabolism proceeds with less wastage. During a single meal the rate of ad libitum food intake of the rat DIVEKSE COMPONENTS 325 (fig. 157) is proportional to the amount still to be eaten. The diminishing rate also corresponds to a weight deficit that diminishes exponentially with time. Rather similar data exist for cats that lapped up their food (Bousfield, '33). The maximal loads possible in privation of food vary with initial states of the individual, rates of depletion, and other factors. Deficits of body weight that are barely tolerated probably do not differ significantly among the mammalian species that have been Days Fig. 155. Increment of body weight (% of Bo), and relative rates of intake, in relation to time after (A) total privation for 48 hours, (B) forced intake by stomach. Eats were allowed unlimited food (dried milk) and water except during the 48 hours. Amounts forced are indicated in block at lower right. Seven tests on 7 individuals (A) and five tests on 5 individuals (B). New data. In A the rats consumed no more food or water per day after deprivation than before deprivation; the dogs (fig. 154) did. studied. Extreme loads are : dog, - 63% of Bo (Howe et al, 12) ; cat, -54 (Sedlmair, 1899) ; rabbit, -49 (Rubner, 1881) ; and man, -45 (Luciani, 1890). Immature rats appear to withstand prolonged partial feeding that limits their weight. Subsequent recovery, if allowed at early ages, occurs with faster net gain in weight, yet with slower intake of food, than in control individuals (Jackson, '37). 326 -4Q Rabbit B PHYSIOLOGICAX. REGULATIONS 1 1 1 1 — 1 40 &> \^ 160 "200 Ete 280" Hours Fig. 156. Body weight load (% of Bo) in relation to time, during recovery in 4 species of mammals. All were previously deprived of food but allowed water; during recovery, food was allowed generously but was probably artifically limited in some re- spect in each experiment. 1 = half of load returned. Man, 2 individuals of Howe, Mattill, and Hawk ('11). Dog, 5 individuals of Morgulis ('28 and '29). B and A, rabbit, 8 and 6 individuals of Maclagan ( '37) as in figure 153. Eat, 11 individuals of Hitchcock ('26, p. 216). Forced overfeeding has been less frequently studied. The rates of recovery from it have now been ascertained. It is known that combustions as well as excretions proceed more rapidly (specific dynamic action, luxus consumption). Food is often refused (figs. 154 and 155) and absorption is said to be less complete. I find no -30 -20 Total Substance Lood Fig. 157. Eate of intake of total substance or food (% of BoAour) in relation to load of total substance or body weight (% of Bo). Eate of intake is given in terms of the number of pellets of food eaten when the rat pressed a lever to obtain each one. J, mean of 8 rats, K, mean of 5 rats, that were almost completely deprived of food until the initial 1.0 hour in which the test was made. Data of Skinner ('38, p. 397). DIVERSE COMPOlSrENTS 327 measurements of food intakes after prolonged food excesses, nor precise information upon the tolerated loads attained by overfeed- ing. Altogether, however, the form of the equilibration diagram is the same for total substance (fig. 158) as it is for water. At zero load, turnover rates of total substance show small vari- ability. Illustrative data are for the rat, in which species the amount of food consumed varies in daily periods only 10 per cent of the mean. It is sometimes said that food intake is habitual, and this expresses in one phrase the fact that many undefined factors are present in maintenance of every species, and recognizes a gov- 6 h X 4 - (2 o 2 O o a: ( 1 ^test5 1 1 1 1 ^^^^^ ^7 \ \ Rat / / - Gain "^ H \ \ \ / / / \ \ \ \ / / - \ \ \ \ \\ / / / - \\ / / 2 2- Loss \ l^^s---^ Loss - Gain 1 1 1 \^^ n. -20 -10 0 +10 Load of Total Substance Fig. 158. Initial rate of exchange of total substance (% of Bo in first 1.0 hour) in relation to imposed load of total substance (body weight) (% of Bo). Rats were deprived of food and water for 0, 1, 2, or 4 days, or given excess of food and some water on one day; then were allowed to recover. Most of the exchange is water. At the dash line, exchange during the hour would equal load. New data. ernor-like action in all of them. The epithet habitual does not recognize the actual fact that a rat is not fooled into continuance of eating when food has already been introduced by stomach or peritoneum. It is well known that in deficits approximate balances of intake and of output may be struck that are much smaller than turnover at zero load. Such rates may be visualized in figure 153. But it may not be assumed that load is stationary if a rate of intake is artificially adjusted to equal an expected rate of output. For in that state, rate of output may reach some other lower value. Rather, the situation in both deficits and excesses is delineated (for 328 PHYSIOLOGICAL REGULATIONS the dog) in figure 159. Total energy is measured in place of total substance; the two are equivalent, since no qualitative choice of foods is allowed. Intake, which is controlled by the experimenter, varies from 2 to 5 Cal./kg. hr., without much discrepancy between it and output. Outside these rough limits, in low intakes the output is at a constant minimal rate ; while in high intakes, the dog con- tinues to eat all the food given, and retains a considerable propor- tion instead of expending it. In contrast to water, storage of com- ponent is a prominent feature, modification of output playing a smaller but also significant part in adjustment. O O u F 4- 1 - 1 1 ■ 1 ■■ -■ 1 - ^ / ^^ o ,^^ • Dog - ,-^^^ / - / - / - / / , 1 1 1 1 012345678 Total Enerqy Intake— Caly1\q. hr Fig. 159. Rate of total energy output in relation to rate of total energy intake. Meat was given in varying amounts to a dog kept in a calorimeter at 20° to 30° C. Each point represents a period of one day; individual II closed circles, individual III open circles. Along the dash line output VFOuld equal intake. Data of Rubner ( '02, p. 109 solid points, p. 115 open points). Duration of each allowed rate of intake influences the output. For carbon (fig. 160) the intake and output are proportional to the total substances. During each day that an intake greater than the turnover of former maintenance is fed, the output increases until finally the output becomes equal to the intake. Given time, therefore, a new balance of exchanges is struck, and a new content of total body substance appears to be maintained. This fact ac- counts for much of the confusion prevailing as to how the content of total substance (or of total potential energy) is maintained. Par- ticularly in the dog, there often seems to be a discrepancy between the amount ingested and the amount "required." The answer is that minimal maintenance is a balance, within the accuracy of usual DIVERSE COMPONENTS 329 measurement ; while ad libitum maintenance is also a balance, more expensive to hold. Many days are required to demonstrate a bal- ance at either, after a transition of intakes is allowed. A consider- able range of carbon contents is thereby permitted, such as was not found for water or heat. For nitrogen, one part of the story is very similar (fig. 160), namely, the gradual adjustment of output to equal intake. But in contrast to carbon exchange, when nitrogen intake ceases, then its output becomes very low. Hence, over a wide range of intake rates for nitrogen, output rates approximately equal them (fig. 161). 0/fr02 <0.3 £ 1 ^ o UJ "5 0.1 o 0 0.15 0.1 -005- ^ 0 -o— C Intake — ^ N Intake -> ■N Output Doc) \. S 0 48 \9^ £40 96 144 Hours Fig. 160. Rate of exchange of nitrogen and of carbon in successive days. A dog of 5.86 kg. was deprived of food for 2 days, then fed an excessive quantity of meat (500 gm.) each day for 6 days, then starved for 2 days. Rates of outputs gradually adjusted toward the rates of intake. Data of Rubner ('02, p. 246). Yet, exact balances occur at only one point for each individual. The fraction of any excess nitrogen taken in that is actually stored is small, and the rate of unstoring it during deficient intake is small, as compared with carbon or with total energy. Variability of con- tent is least for that component in which modifiability of exchange is greatest. Where constancy of output is maintained (energy), storage is provided, with wide range of contents. Very many studies, especially of nitrogen metabolism, have ignored the pre-existing content of nitrogen in the body, and worked entirely with rates of exchange (fig. 161). For example, Voit and Korkunoff (1895) tried to find the minimal intake of meat for 330 PHYSIOLOGICAL REGULATIONS balance. Starting with 0.015 gm. X, kg. hr., they gave slightly more on successive days until output was no longer greater than intake. But meanwhile the nitrogen content of the body was being depleted ; the deficit so obtained may or may not have been serious; and acclimatization of output (as in fig. 160) may have been slow or fast. Balance finally occurred at 0.028 gm. X. kg. hr. (C, fig. 161). Hence in nitrogen exchange, the time factor is large, for weeks may be required to establish stationary rates at one bodily content. OilOr To+al Ni+roaen |n+ake— qmYkg hr Fig. 161. Eate of total nitrogen output in relation to rate of total nitrogen intake. Points of nitrogen balance, indicated by 1, fall on line D, equality. Solid circles (curve C) dog II of Eubner ('02), cf. figure 159, fed meat only. Squares and solid triangles. 3 dogs of Michaud ('09). Open circles and triangles, dog of Voit and Korkunoff (1895) ; curve B (triangles) indicate exchanges when fat and carbohydrate supplements are given in large amounts, the difference from curve A (circles) being "protein sparing." Periods vary from 1 to 14 days at diverse points. Several ordinary criteria of balance may be satisfied at any of many nitrogen contents, for approximate equality of intake with output occurs in a great range of contents. Hence the emphasis is placed on the minimal nitrogen tunwver, the "practical or physiological minimum." This is distinct from output in deficits of nitrogen ("experimental minimum''). Actually nitrogen intake is much more variable than water intake; the range of approximate bal- ances that can be forced to occur for water, do frequently occur for nitrogen. In dogs there appear to be no measurements of volun- DIVZESE COMPONENTS 331 tarv protein intake independent of total intake, and in most practi- cal situations there is little opportunity for an animal to select nitrogen intake independently of total energy intake. With all this in mind, it is scarcely worthwhile to believe that present data yield more than a sketch (fig. 162) of a critically derived equilibration diagram for nitrogen. The fact has been estabhshed that certain particular substances are selected for ingestion by rats when these substances are de- pleted. Such substances are water (Eichter, '36) and vitamin B (Eichter et al., '37b, *3Sb). The increased "demand" for calcium following parathyroidectomy (Eichter and Eckert. '37a) is mani- Total Nitrogen Load— qm^'Vij Fig. 162. Eate of total nitrogen exchange in relation to total nitrogen load. Xi- trogen equilibration diagram. Dog. Only the qnantitr of whole diet was varied in each individual, not the eonstitutents of the food, which differed in A, B from those in D, E. A, four weeks of recoverv following partial privation of an adequately constituted diet for nine weeks, individual of 13.94 kg. of Morgulis ( '23, p. 266) . B, previous balance of same individuaL C, output in dailv periods of no food intake, individual of 5.S6 kg. of Bubner ('02, p. 246). D, output in successive days of e:seessive food intake, same individual. E, output in successive days of excessive food intake, dog ill of 4.1 kg. of Bubner ('02, p. 113). fested by 3 to 6 fold increases in calcium lactate ingestion. Sodium chloride is taken 6 times as rapidly as usual after adrenalectomy (Eichter and Eckert, '3Sa) and during pregnancy (Barelare and Eichter, '3S). Specific "appetites" of this sort have hardly been measured in other species. Although in everyday life the animal can do Uttle to unmix its foods, upon the specific appetites appears to depend the maintenance of particular chemical constituents of every individual. Excesses of food lead to faster recoveries of some constituents 332 PHYSIOLOGIC-U. EEGTLATIOXS of it and slower of others. Hence recovery of total substance depends on what particular materials are ingested. In a mixture, water is eliminated most rapidly when in a certain excess relative to other materials ; nitrogen next, chloride next, and carbon last. Other substances could be intercalated in this list. Each compo- nent has a velocity quotient characteristic of it, and these quotients may be investigated either in ingestions of more or less pure com- ponents, or in mixtures. In general, also, diverse quotients are found for one component with each mixture, and this is a matter for further study. It is a plain fact that every organism tends to have a constant amount of material in it. In some protozoa accession of material leads to reproduction of more individuals (Adolph, '31a), restoring each to a size near the mode. In mammals, accession leads to faster disposal, at rates differing for each constituent (energy, carbon, nitrogen, water). Deprivation of material leads to extraordinary efforts to obtain it, and especially to obtain the very constituents that are lacking. TThereas the human individual sometimes relies upon the bath-room scales to guide his rate of intake, organisms generally temper their exchanges so as to maintain a uniform body weight without any visible measuring device. § 119. Glucose ix dog One of the first components for which a maintenance of con- stancy was realized to exist (Bernard, 1855; Chauveau, 1S56) was sugar. Deprivation of carbohydrate did not much decrease the blood's concentration of sugar; feeding nothing but sugar, even infusing a solution of it, did not much increase it. The identifica- tion of glycogen as one of the forms of storage for carbohydrate led to the dim realization that reversible equilibria buffer the changes in actual glucose content of the body. From this prototype of reactions through which excesses and deficits tend to be ab- sorbed, has grown the knowledge of diverse paths of disposal for sugars, as well as for innumerable other components both chemical and physical. To obtain quantitative data upon unanesthetized dog, two types of glucose increment are used; excesses are produced by continuous intravenous infusion of glucose, deficits by previous intravenous injection of insulin. The glucose content of the body as a whole is computed on the assumption that the volume of distribution of glu- DIVERSE COMPOXEXTS 333 cose is 50% of Bo (Wierzuchowski, '36) under all the conditions measured. Equally informing, and perhaps more direct, would be to correlate the concentration of reducing substances found in whole blood with the rate of its change ; then, however, the partition of disposals would require further assumptions before common dimensions would be shared. Glucose was infused at constant rates, five of which are repre- sented in figure 163. At the smaller rates, loss equals gain after Hours Fig. 163. Glucose load (increment in gm. of glucose/kilogram of Bo) in relation to time. Glucose was infused intravenously for 6 hours at constant rates that are in- dicated (in gm./kg. hr.) hx numerals on each curve. Loads are computed from blood <:oncentrations assuming a uniform volume of distribution of 50% of Bo. Data of Wierzuchowski ('36, p. 322). the first hour, so that loads are stationary. But at higher rates glucose accumulates. The rate of total loss (disposal) of glucose remains steady after the first hour (fig. 164) ; but at infusion rates of 8 and 9 gm./kg. hr. the rate of loss does not equal the rate of .gain. The loss is partly in urine (fig. 165), though almost none is so lost at the lower rates of administration. "When rates of loss are correlated with loads (fig. 166), it is evi- dent that output attains a limiting rate when 8 gm./kg. hr. are 334 PHYSIOLOGICAL KEGULATIONS Hours Fig. 164. Rate of glucose disposal in relation to time. Infusions continued during 6 hours at rates that are indicated by the dotted lines and by the numerals, in gm./kg. hr. Rates are inferred from changes of blood sugar concentrations during periods of ■J to 2 hours, each rate being indicated at the middle of the period. Same data of Wierzuchowski ('36, p. 322) as in figure 163. infused. This corresponds to a load of 7 gm./kg., and at lower loads the rate of loss is nearly proportional to load. In excesses, three paths of disposal are measured: amounts excreted, amounts oxidized, and amounts disappearing but not ex- Hours Fig. 165. Rate of excretion of glucose in urine, in relation to time. Glucose was infused continuously at rates indicated by numerals, in gm./kg. hr. Data of Wierzu- chowski ('36, p. 314). DIVEKSE COMPONENTS 335 creted or oxidized (fig. 166). The proportions represented by these paths differ enormously at various glucose increments, represent- ing quantitatively facts that are loosely known. At small incre- ments practically all glucose is transformed, presumably to other forms of carbohydrate. While the rates of synthesis and oxidation are limited to about 3.6 gm./kg. hr., excretion is not limited but aug- ments in proportion to the glucose load. One might envisage diffi- cult situations for the organism if (a) there were no excretion of glucose, (b) there were excretion of glucose at small loads (^=no threshold), or (c) combustion of the extra sugar were not carried i: £" 6 I (D Z - -■ ■ 1 ' ' o_J. 1 A.--^ ~ u Dog y^^O / rf Excre+eol in ur\ne / ^ — ~ • -•_ V "^ ■ — • / Stored not as qlucose I-'-a-T— =^ , Oxidlzeid 1 1 +2 + 4 +6 +8 +10 Mean Glucose Load -qm/ kg. Fig. 166. Rate of net glucose disposal, and its partition, in relation to mean glu- cose load in approximately steady states of glucose excess. Dog. Glucose was injected intravenously at constant rates during 6 hours. The last 3 hours were taken in each test, 29 or 48 tests being included in the 9 means. Glucose load was computed from blood sugar concentrations and a Vd of 50% of Bo. Hence an increment of 1 gm.Ag- or 0.1% of Bo = 2 gm./liter of whole blood; and since the initial content was 0.97 gm./liter of blood, 1 gm.Ag- or 5.55 millimols/kg. = 206 per cent of the control content (Go). Data of Wierzuchowski ('36, pp. 314, 322, 327; '37b, p. 153). out. As it is, there are arrangements for meeting exceedingly large influxes of glucose, and it has often been remarked that the organ- ism has reserve means of meeting situations so extreme that they may have never occurred in non-experimental situations. How extreme these are, may be realized from the statement (Trimble et al., '33) that glucose is ordinarily absorbed from the alimentary tract at a rate of only 1 gm,/kg. hr. In Wierzuchowski 's experi- ment, the dog speeds up its rates of disposal of glucose even to the point where excretion of it is faster than removal of it by synthesis. Total exchanges and turnovers are subject to small uncertain- 336 PHYSIOLOGICAL REGULATIONS ties as to whether all the oxygen apportioned to carbohydrate represents glucose burned. In addition, whatever the portion of glucose that passes to carbon dioxide, much more glucose may be continuously formed and transformed in many processes of inter- mediary metabolism, and the turnover of glucose recognized for the whole body may be but a small part of the sums of all the local for- mations and transformations of this chemical compound. The dog's turnover has been estimated at 0.25 gm./kg. hr. (Reid, '36), and is the same in depancreatized individuals. Glucose content, at least of the blood, is less variable than con- tents of many other components (table 38). According to data of TABLE 38 Sugar concentrations in Hood under standard conditions, made on a separate day Each observation was Species Dog Man Rabbit Number of individuals 1 23 23 141 100 27 Number of Mean, C.V., observations gm./l. % 48 0.97 6.2 44 0.94 4.6 5 each 0.94 4.7 23 0.94 6.4 141 0.97 7.2 157 1.16 10.5 100 1.05 16.1 85 1.18 18.2 1000 1.24 11.5 Source of data Wierzuchowski ('37b) Okey andRobb ('25) i c Pierce and Scott ( '28) Eadie ('23) Clough et al. ('23) Scott and Ford ('23) Scott ('27) Wierzuchowski ( '37b) almost no variations are detectable in a dog at hourly intervals, nor even at the close of a day of loading and unloading of glucose. Adrenalectomy does not increase the vari- ation of blood sugar concentration from day to day (data of Zucker and Berg, '37). Among species, the rabbit (table 38) in contrast to dog and man, shows higher content and greater variability. That contrast represents a difference in maintenance ; yet hereto- fore rabbit and dog have been studied indiscriminately. The data yield a glucose-time system of four variables (AC, SC/At, t, 1/At). At diverse loads and rates of loading, character- istic latent periods, times to maximal rates, and initial rates of disappearance are found (figs. 163 and 164), all of which are com- parable to the temporal characteristics of the dog's water-time sys- tem with continued administrations (§12). The velocity quotients (1/At) for the total glucose exchange (fig. 167), are more constant with time in slow infusions than in rapid. Evidently after the first DIVERSE COMPONENTS 337 Fig. 167. Glucose velocity quotient (l/hour), in relation to time after continuous infusion of glucose began, at rates indicated by numerals (in gm./kg. hr.). Data of Wierzuchowski ('36), from figures 163 and 164. hour the state is a steady one (fig. 168), the velocity quotient for total glucose disposal is independent of time; and this quotient diminishes progressively with load to a much greater extent than the velocity quotient for water disposal. While rates of excretion of glucose increase with loads (fig. 166), the rates of synthesis and of oxidation are independent of loads above 2 gm./kg. of body weight. The patterns of these paths stand in sharp contrast; for here are two paths for disposal (rate of synthesis, rate of oxidation) that do not vary with load, while another (renal excretion) is mainly dependent on it. The latter is such that clearance from blood, and velocity quotient, are approxi- Glucose Load — qm/kg. Fig. 168. Glucose velocity quotient (l/hour) for net disposal, in relation to glucose load. Dog. The triangles represent the period 1.5 to 3.5 hours, and the circles 3.5 to 5.5 hours after continuous glucose infusion began. Same data of Wierzuchowski ('36). 338 PHYSIOLOGICAL, REGULATIONS mately constant at all loads above 1 gm./kg. That load corresponds to what was once termed the "assimilation limit." Some parallel data have been obtained by the entirely independ- ent method of analyzing total carbohydrate in the body (fig. 169). Many points of technical difference may be noted between this set of tests and the previous ones ; this set nevertheless indicates the slowness of the exchanges of total carbohydrate, especially with excretion blocked, as compared with those of glucose alone. Re- moval of pancreas does not limit the "ability" to utilize carbo- hydrate, but limits those high rates of utilization to greater loads. 0.8 1 o o cc .Glucose Load~ qm/kg.of Body 8 \Z Blood Sugar Concentration qm/l. of Blood Fig. 169. Kate of carbohydrate loss (in gm./kg. hr.) in relation to mean glucose load. Dog under nembutal anesthesia after "evisceration," and privation of food for 3 previous days. Glucose was infused by vein continuously for 2 hours before the test began and during 4 hours in which the carbohydrate disposal was being measured. Each point represents one individual, samples of whose tissues were analyzed before and after the 4-hour period. Note that ordinate scale is ten times that of figure 166. Here the loss is probably all by oxidation. N, control; P, pancreas also removed. Data of Soskin and Levine ('37). Deficits of glucose are induced by injections of insulin (fig. 170). Thereafter glucose is gained (net) and more slowly as zero load is approached. Absence of medullary portions of the adrenal glands interferes but little with the course of recovery from deficits of glucose. The slower exchange after high doses of insulin may be accredited to persisting action of this agent in tending to remove DIVERSE COMPONENTS 339 glucose. Possibly the curves would lose their convexity if the per- sisting action were discounted. Taken together, figures 166, A, and 170, A, constitute a diagram of net equilibration. But the scales of the latter figure must be reduced to one-twentieth their present size in order to match those of the former figure. The slopes near zero load are about equal in the two curves, meaning that small excesses and small deficits are adjusted with equal speeds. Only, the range of deficits that is tolerated (and chemically possible) is very small compared with the range of excesses. 0.4 -'03 -OZ -0.1 0 Mean Glucose Load Fig. 170. Kate of net glucose gain (gm./kg. hr.) in relation to mean glucose load (gm./kg.), as found in the blood of dogs previously subjected to intravenous injections of insulin, A and A', 17 and 12 tests with 0.1 unit insulin/kg- B, 5 and 8 tests with 0.5 unit/kg. C, 11 and 15 tests with 1.0 unit/kg. Solid points, bilaterally medullecto- mized; open points, unoperated. Data of Zucker and Berg ('37, p. 541). If total exchanges of glucose were estimated instead of net exchanges, the ordinate at zero load would be increased by 0.25 gm./kg. hr. It is unknown how total glucose gain varies in positive loads, and total loss varies in negative loads. Whereas water, heat, nitrogen, and carbon are probably lost only by modifications of their rates of elimination from the dog's body, the specific com- ponent glucose is disposed of by chemical transformation, a process which also makes its total loss difficult to measure. Similarly, in deficits the rates of net glucose gain reported might barely result from diminutions of total loss without any increases of total gain. Had the hlood been chosen for study in place of the whole body, 340 PHYSIOLOGICAL REGULATIONS rates of recovery in steady states of excess of blood glucose might be correlated, complete data for which are available. From other sources, recoveries after single administrations would be repre- sented by a series of ' ' glucose tolerance ' ' tests. For them, a known amount of glucose is suddenly injected into the circulating blood, after which the blood's concentration of sugar is measured at suc- cessive times. Such tests in dogs show that (1) disappearance of glucose is absolutely faster with larger doses, (2) disappearance rates diminish with time and with concomitant decrease of incre- ment, (3) disappearance rates are modified by removal of liver, pancreas, or adrenals, by anesthesia, insulin, phlorizin, renal liga- tion, and numerous other factors (Goldstein et al., '32). Many devices have been used to compare glucose tolerance curves. Rates of recession, slopes of return, hyper-glycemic areas, peak effects, and terminal levels, are all of service to the graphi- cally minded. ' ' Insulin sensitivity curves ' ' may equally be termed tolerance curves in glucose deficits, to indicate that they differ in component and sign from tolerance curves for other substances. Finding now that "tolerance tests" are particular instances of recoveries, all the rate and time factors and ratios found useful for water (§ 105) may be applied to glucose. Glucose, therefore, exhibits net equilibration of content both in stationary states and temporary states. Excesses are dissipated by several paths, and the proportion lost by each path varies with the load. Deficits and excesses of equal magnitude are re- turned at about the same rates. Promptness of recovery as ex- hibited in tolerance curves in positive and negative loads is matched by the smallness of variabilities of content found in single indi- viduals and in populations. § 120. Caebon DioxroE in man Carbon dioxide (plus bicarbonate, etc.) may be thought of as a constituent of the body as a whole. Exchanges of carbon dioxide in mammals represent chiefly losses by elimination through lung alveoli, competing with gains by internal production. Increments in the carbon dioxide content of the body are roughly estimated from the extra carbon dioxide that has been eliminated (a) in forced breathing and (b) in recovery from equilibrium with high tensions of inspired CO2. It is found that 0.2 hour or more is re- quired in man to attain a new stationary content of this substance, the net retention or elimination approaching some asymptote. DIVERSE COMPONENTS 341 The equilibration diagram for carbon dioxide (fig. 171) shows enormous responses to increased contents. Neither in increased nor in decreased contents is there a known change in rate of inter- nal (chemical) production. Recovery from deficit is limited to sup- pression of loss alone. This fact was described by Haldane and Poulton ('08); after forced breathing, apnea persists until the Q) O^ E o -E O X s. LJ -C CM IT> o o i. V 47). It means that preference for moisture is greater in water deficits, even as drinking is faster in water deficits. Very often it is said that an animal in water deficit has "drive" or "thirst drive." That term denotes a hypothetical state of the animal, and connotes a virtual force that expends itself in seeking water. It is comparable to inferring that a force of "osmotic" pressure moves water into the body, and if I were to call it "drink- ing pressure" no fault could be found. But what is directly measured is not drive, but either water deficit or water drinking. The intermediate term can be dispensed with, and water load be used to describe the physiological state. In brief, behaviors that diminish losses or enhance gains in defi- cits, and those that enhance losses or diminish gains in excesses, protect against inconstancies. Rarely is any one behavior con- tinuously operative, so long as many components are being pre- served by a single individual. But, however modified by condition- ing or intelligence, appropriate behaviors are found widely in organisms that observably respond to environments. § 133. Sequences in time By comparisons and inductions from data such as figure 176, sequences of load are found to fall into a generalized tolerance dia- gram (A, fig. 180). The scale of times as well as of loads varies enormously for diverse components, and somewhat for diverse species. Five periods or phases may be distinguished, in each of which a different physiological state (load) prevails. (I) Control or initial state, in which load is zero. (II) Transitional or loading state, in which an increment or load is gained under the influence of component, agent, or conditions. (III) Stationary state, in which load is relatively uniform; often conditions are constant. (IV) Recovery or unloading state, in which load is lost. (V) New control state, in which recovery is completed or nearly so ; it may or may not differ from the initial con- trol state I. UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 367 State I is characterized by the mean content (Co) and by its variability (oc). Descriptions of conditions that prevail enable the measurements to be reproduced. +£ 0 12 3 Tinne O' ' ^ Fig. 180. Diagrammatic representation of physiological states (in load and in rate of exchange) in relation to time. An agent impinges at time O and ceases to act at time O'. Five periods, representing five physiological states are distinguished : I, control or balanced; II, initiating or loading; III, stationary or loaded; IV, recovery; V, second control or recovered or balanced or final. Eates given represent gains (G) and losses (L) in positive loads ; in negative loads G and L would be interchanged. Vari- abilities of load (ac) from the control mean (Co), and of rates (oR, oG, oL) from the control means (Ro, Go, Lo) are indicated. Areas M = Nr:M' = N'' are each equal to Cm, the majdmal or stationary load. Portions of this curve are observable in the actual data of figures 21, 23, 143, 163, 164, 160, 176, 106, and others. State II is related to the initial and the stationary states, very often being represented by an exponential curve connecting the two. Its duration is sometimes longer when the stationary load is greater. In some respects it represents the resultant of what the 368 PHYSIOLOGICAL REGULATIONS agents and conditions do to the organism and of what the organism does in response. Partial generalized treatments of state II were given by Widmark and Tandberg ('24) and Burton ('39). State III may not appear in a given instance, for the agents or conditions that induced a load may be removed before it appears. Constancy of load means that rate of gain remains equal to rate of loss, and often state III is identified from rates of exchange. Otherwise the stationary character may be judged solely from the constancy of content. State IV is the portion of the sequence that has been studied most extensively in this investigation. It seems to represent most often what the organism does to adjust contents. Sometimes its slopes are the reverse of those in state II. Very often the agents that induced the load persist with diminishing effects for long periods of time (as when drugs wear off or atmospheres gradually cool), in which case recovery can hardly be regarded as an unen- cumbered activity of the organism. Sometimes there is no way of knowing how long agents persist (as when epinephrine is injected), any more than of knowing exactly when agents not followed by immediate consequences have become effective (as when water is ingested). In all instances the investigator may choose a time that can be recognized reproducibly and count hours from it, leaving to anyone the "interpretation" of intervening events. State V is sometimes known only as an asymptote ; sometimes it is assumed to be identical with the zero load of state I; some- times it is acknowledged or demonstrated to be different from state I. Very many physiological sequences are incomplete approaches to hundreds of successive stationary states ; these may be regarded as states Illd, Vc, etc., perhaps being distinguished only by whether the load has just increased (III) or decreased (V). Thus, an increment of blood volume, or a load rate of oxygen consumption in a gastrocnemius muscle (Keller et al., '30) varies enormously during each muscular contraction. Transitions and recoveries probably succeed one another with each posture and each move- ment ; and an asymptotic state is rarely reached. Data illustrating the loads and rates involved during transitions from one stationary state to another have been worked out with respect to rate of oxy- gen consumption in human exercise (Hansen, '34; Szwejkowska, UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 369 '38). Anyone can within a few minutes construct a curve (like A, fig. 180) for heart frequencies in man. With respect to rates of exchange, the sequences are bound up with those of load. The same five states may be distinguished in the exchange diagram (B and C, fig. 180) whether net rates, total rates, or certain partial rates are measured. In instances in which gains and losses have been measured, one exceeds the other during any transitional state and the converse relation holds during the recovery state. Moreover, since net rate integrated with respect to time equals load, the areas of difference between total gain and total loss are equivalent to maximal load, and very often are equal to one another (M = N). When M does not equal N, it is perti- nent to inquire in what identifiable respects state V differs from state I. Further consideration may be given to recovery states. Kates of recovery are rates of decrease of load. Some of the loads stud- ied follow exponential equations of the type AC = ae"*"', as was early shown by Michaelis ('07). Differentiating with respect to time, SC/At = R = -a/i;e-" (§71). The load throughout the five states might be analogous to the height of a ball thrown upward. The height attained depends on several distinct factors; the fall (recovery) exhibits accelerations free of the force of the throw. In ballistics the usual trajectory is a parabola ; in physiological tolerance curves it is often two expo- nentials. Certainly the relation between load and time is not an invari- able one for diverse components, nor for one component in many species. No augmentation of exchange may be evident for 0.3 hour (water excess in dog), or the load may be proportional to time (ethyl alcohol in dog), or no recovery may be measurable (ampu- tated leg in dog). The last part of recovery may take as long as the individual lives ; the first part may be infinitely slow or fast. In practice, recoveries are compared quantitatively either in a uniform fraction of the recovery {^C/a) or within an arbitrary interval of time during the recovery (At). With each set of data it is therefore necessary to designate the value of 1/a or of At that is used. Very often the half -life (tq) of the load {a = 2) is a suit- able means of comparison, and A: = In 0.5/tq. However, when, as is often the case, the entire curve of loads during recovery is known. 370 PHYSIOLOGICAL REGULATIONS and it is exponential, then the value of 1/At = /c is independent of the fraction 1/a or the interval At used. Figure 180 describes, therefore, the relations to time in any dis- turbance and recovery. A story concerning a particular compo- nent may be planned by measuring those quantities that will fit it. Conversely, all components may be compared by means of the numbers found for each of the curves and parameters represented. They seem to describe adjustments of any physiological compo- nents. § 134. Loads A load is defined as the deficit or excess of any measurable component in a living unit, relative to its content in a control state. A component is, in turn, any property of an organism that is sus- ceptible of measurement. What are the several methods of measuring loads? (a) Taking the organism or its part as it happens to be, the investigator keeps a complete account of gains and losses while the individual passes from state I to state III ; and later from state III to state V. At the end he decides whether state V equals state I, whether net gain or loss throughout state II equals net loss or gain throughout state IV, and whether some of the component escaped measurement by having been transformed. At whatever times gains and losses are ascertained, the coincident loads and rates are known. (b) By an appropriate kind of measurement that presumably does not interfere with the physiological state, repeated determina- tions of load are obtained. Examples of the kinds of measurement that are available are: (!) Body weight or volume or length as an indication either of total substance or of water content, depending upon the conditions in which the measurement is applied. (2) Determinations upon small tissue samples by a variety of physical and chemical and serological procedures, with or without either estimates of volumes of distribution or assumptions that volumes of distribution remain constant. (5) Introduction in vivo of for- eign indicators, either physical, chemical, or serological, usually with subsequent sampling of tissue for analysis. The changes of load with time being known, the successive rates of net exchange may be computed. Or, the changes of load with time and the rate of total gain being known, the rate of total loss UNIFORMITIES AND COMPAEISONS AMONG COMPONENTS 371 may be computed, or vice versa; for, the difference between total gain and net exchange is the total loss. It may be induced from equilibration diagrams that at two diverse loads, the same rate of exchange would not be expected to occur. Yet it is sometimes assumed that a measurement of rate of net loss, for instance of nitrogen during nitrogen deficit, is a measurement of turnover in balance. Wherever positive loads have been investigated, negative loads may be expected to be of equal interest. The history of physiology shows, however, that intensive studies of changes in one have not often led to inquiries into the paired state. This omission indi- cates that other considerations than quantitative symmetry of function have guided investigations. The range of loads observed in the study of each component is limited. Occasionally little experimental effort is put into obtain- ing extreme states (tolerated loads), while in some instances the production of injury (irreversible change) in the organism is ap- proached. The criterion of irreversibility is open to many differ- ences of judgment or end-point; perhaps the criterion of ''death" (judged by a chosen independent test) within a certain number of hours from the time at which the load is imposed, is a generally useful measure of tolerated load. Each component may then be varied within the range stretching from mean viability in negative load to mean viability in positive load. The magnitude of load in the stationary state or at the maxi- mum is as much a function of the rate of recovery as of the agent impinging. For, the maximal load is a sum of gain and of loss, and since rate of loss usually emulates rate of gain, it is often the case that no acceleration of gaining will long outdistance losing. An analogy is throwing a ball vertically; acceleration of gravity becomes equal to acceleration of lift at a very finite height. Overshooting of the content characteristic of balance, did not regularly occur during recovery in any of the components here reported (cf. Burton, '39). In quantitative comparisons of diverse components a problem is to find commensurate coordinates, for^ both loads and rates are measured in a variety of units (table 40), For certain specific comparisons, common units are available; calories, grams, chemical equivalents, weights, volumes ; these rep- resent appropriate particular equivalences of energy or of sub- stance. For other purposes it is preferable to record every load 372 PHYSIOLOGICAL BEGULATIONS ft^ p^ ^J" rii >> o -M a CO o a' s\ > p '- 3 O O ii H s ft P 3 t> O) O e o 03 ^ 2 rt 62 o ca ' irt .1—1 ^ . ^ 3lfc| .^.fl c ,q ^ ^ g M 3- ^ :r IT'S 'S ^ 2 -=! '2 K^^ Jiftftoj 03 aoj-r' rt, a o o ^ ;:2 S"'~=m^'^ ■ rH 1^ ns o -rt pj ."t:; .rt " -r-i «s "^ pm a ^r' £3 ^^ 1—1 v.^ ^1 •- o Fishma i Yant z and :3^ w O rrt ^ QJ ^^ 'g - ?r^ fl 03 m a ■* rt ^ P « rt o =ei:5. ^-^^ J2 2! ,vi ;^ '^"-1 O O rH O CO t^ 00 I— I Oi O O O lO t~ Ol -^ iH I O iq O O N 05 ; o o o o o o 0O-+I— lOOOOOOOOOO _,_, + ' '+'+-*< 1-i OS (M O O O O O O 2'=^ + + + + + id rQ ^3 O o3 rt 0 0)01 O tH ■ o o O OT 02 O O o o ^ ^ ■^ lO o O W OO > I— I o o o o o rHTt* ra ^3 c t: . . . 'O o bn !3p ti ti bb o ^ .. .. ^ hobX)bX)2 ® s « a a a ^- \ 'a 0 m./k l./k l./k Ab ats/ l./hi l./hr tuo <-J bjD a -^ be S be be bjci\ j3 '^ So f>^ ft ^ o a -M ?3 "' en ft (B pq-R — 03 i-ii a; •a T1 g rt rt 1^3 bei Khoo i-" S ID .rt O " M --3 o S^ f^ tH oo a." ^ a^ M rt i^ M «.s -a aj .H fS a ;3 p ^ oj , - o *: 03 *- OQODhH WOO UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 373 as a fraction of the body's control content. But this is difficult to accomplish for heat, temperatures, pressures, and electrical poten- tials, which either have no physiological zero or are regularly absent from the body. If fraction of control content could be used universally, it would have the advantage that every conceivable component, regardless of its physical dimensions, could be ex- pressed in it. Loads might also be equated in units of variability (o) or in units of daily or hourly differences (CA). This, the ''beta'^ method, would have the advantage of being independent of all physical dimensions, and the great disadvantage of requiring a special study of variability before any comparisons could be made. Could lethal extremes be used to equate loads to a common scale? The difficulties here are: (l) the physical scales would dif- fer from the physiological scales within the range of positive loads as compared with the range of negative loads; (2) the scales would differ for two types of loading of the same component, one of which killed at a less point on the physical scale; (5) no lethal load exists for some components; (4) the scale of loads would be revised with each statistical investigation of tolerated maximum; (5) in man the lethal limit could often not be determined, and in elephants would be rather expensive. On the whole, I see no probability of finding a biological means of making commensurate loads of all types and components. Loads from which recoveries occur evidently represent debts and credits in the organism. For some, equivalents are known that act as security for their discharge (phosphocreatine for oxygen, hydrogen ions for carbon dioxide, osmotic pressure for water). It is unnecessary to limit the measurement of debts and securities to chemical or any other variety of components. The security may be divided among many forms at one time. The combined states and activities of the organism represent this security, as evinced in everything that changes during recovery. § 135. Rates of exchange Rates of physiological activity are comparable for various com- ponents wherever the components have similar dimensions or known equivalences. Sometimes dimensions themselves can be transformed with high assurance of accuracy, as in the loss of heat (cal./kg. hr.) by evaporation of water {% of Bo/hr.), when the latent heat of vaporization is believed to be known. 374 PHYSIOLOGICAL EEGULATIONS The recital of the rates of turnover of diverse components, as in table 40, may be of interest in certain connections. Each mean rate, and the variability of each rate, characterizes the individual and the species. It establishes norms against which any unusual individual or an individual in any new physiological state may be compared, in much the manner that physical anthropology and clinical medicine practice. It is not always realized that most components concerning whose gains and losses little is known, still have turnovers. Con- stant potential, pressure, or concentration probably represents continual decay and replacement, though no prediction may be safe for all components. Prothrombin in dog's plasma decays after the plasma is isolated from the body ; in fact no organic compound in wet killed tissue is known to last indefinitely. Often intermittency of content furnishes a means of estimating turnover rates. Thus, arterial blood pressure falls during each diastole ; this fall multi- plied by the heart frequency, is the turnover of pressure. And so minimal values of the turnover of excitability in heart, peristalsis in gut, tension in single muscle fiber, and lift during walking are ascertained. In general, measurements of turnover depend upon one of the following methods: (a) flow across a boundary, (b) difference in flows to and fro (arteriovenous differences), (c) accumulation of marked materials, and (d) intermittent contents. All methods treat the rate as though stationary over some period of time. In the study of regulations of any one component the several rates of exchange relative to turnover rates are indicated by aug- mentation ratios and modification ratios ; they indicate the latitudes over which rates of exchange respond. The maximal rates of ex- change express capacities with which each organism is endowed. In particular, it turns out that for most components (table 40) the maximal rate of gain is similar to the maximal rate of loss. For a few components, economy quotients (ratio of total gain to total loss) are known. At extreme loads it is the rule that econ- omy quotients are far from 1. A ratio similar to the economy quotient was proposed by Verworn (1898, p. 487), termed "bio- tonus," representing the rate of assimilation or anabolism relative to the rate of dissimilation or catabolism. Metabolic equilibrium prevailed only when the ratio was 1. This limited form of econ- UNIFORMITIES AND COMPAEISONS AMONG COMPONENTS 375 omy quotient was not illustrated by quantitative data, and states in which it varies were not ascertained. The most efficacious recovery might be defined as that which exists when the economy quotient becomes infinity or zero ; in which case either loss or gain is completely suppressed, and net exchange equals total exchange. That rarely occurs, and with some excuse. Perhaps it can be supposed that complete suppression appears only where cessation of exchange can be managed without interfering with any other component (as, when ingestive gain of water ceases in the dog's positive water load). But for one exchange to con- cern only one component is the exception; further physiological properties are modified by the continued production of heat in posi- tive heat loads, and by the continued gain of carbon dioxide in posi- tive loads of carbon dioxide. Other cases of unforeseen interrela- tions give rise to more elaborate and less widely accepted excuses, with little descriptive foundation, such as could be given for the continued intake of water by the frog in positive water loads. To illustrate how unforeseen combinations arouse unwitting excuses, I quote a careful observer. "The Corixidae are air breathing," says Krogh ( '39, p. 118), "and one would expect their integuments to be practically water impermeable. Is it possible that they take up water and salt in such quantity with the food that an osmotic regulation becomes necessary?" In this one expression of sur- prise it is implied that oxygen exchange is highly correlated with water exchange through any one surface, that impermeability is less trouble than exchange, that every property of an organism has fitness, and that a physiologist knows what to expect more often than not. Yet no biologist mixes his data and his "derivations" to a less degree; most scramble them more. Not only is it difficult to discover a rate of exchange that cor- rects one component without affecting others; it is also apparent that the relative rates of exchange among components are compati- ble one with another. Heat loss by evaporation would be impossible in a steady state (fig. 48) if water gain were slower than water loss by this path. Even decelerations fit together ; a man would readily undercool (fig. 143) if his water and heat loss by vaporization did not decelerate before the heat load was entirely dissipated. Components that are loaded slowly also recover (unload) slowly in most instances. Thus, rectal temperature and heart beat 376 PHYSIOLOGICAL EEGULATIONS in exercise (man), lactic acid and oxygen in exercise (man), and excitability and lactic acid in stimulation of nerve (frog) constitute pairs that move at very different rates, but in nearly the same ratio in both directions. Either reversed processes or matched "gover- nors" may be at work in the two states (II and IV) of the organ- ism. Actually the processes usually do not occur in the same paths, for example the gain and the loss of water. A difficulty in phrasing a more general induction is that the maximum attained by the load often is itself bound up with compensation and recov- ery, the asymptote of loading (state III) being already an expres- sion of rapidity of recovery. Intermittency of rate occurs in a great many components (limb movement, sleep, reproduction), and for the most part it is ignored in their study. The dog accomplishes by intermittent ingestion of water exactly what the frog accomplishes by nearly continuous imbibition of water. Intermittency is a convenience to the organ- ism in permitting it to go about other activities which preclude the exchanges in question. The dog devotes perhaps 0.5 per cent of its life to drinking, 2 per cent to eating, 1 per cent to micturition, 30 per cent to sleeping, and so forth. Segregation of each allows greater variety of undertaking in all the remainder of the time. Rates of recovery in organisms are modified by many agents and conditions. Loss of water, gain of heat, and loss of lactate, are all accelerated by moderate physical exercise in man. Glucose removal by dog is changed by previous glucose administration, by deprivation of adrenal glands, of pancreas, and of liver. Carbon monoxide is removed faster when oxygen, or oxygen + carbon dioxide, is breathed. Results of experiments, treatments, and therapies may be suitably measured, of course, as modifications of loads (tolerance curves) and of rates (exchange curves). It appears to be exceptional, as was said, to find agents that modify one component only. Rates of exchange show acclimatizations. These are progres- sive changes in rates incident at given loads, that appear after con- tinued or repeated loadings. Time is required to bring them forth. Thus, repeated administrations of water to a dog result in greater rates of urinary output at the same water excess (Kingsley). The rate of heat loss increases upon successive days of exposures to hot atmospheres (Adolph and Dill, '38). Less chloride is lost through the skin in sweat upon successive exposures to hot atmospheres UNIFORMITIES AND COMPAEISONS AMONG COMPONENTS 377 (Dill et al., '38). The exchanges and velocity quotients are not, it seems, permanently fixed when the load first appears; a sort of conditioning is apparent, using that term either as an ecological or a psychological one. Occasionally such progression with time is termed adaptation, or accommodation; each term has its implica- tions which no one wishes to extend to all phenomena in which repetition has an effect. Here is evidence that patterns of main- tenance are not all preformed ; it is possible that exchanges of many components are changed with experience superimposed upon initial endowment. In general, most components of organisms appear to be pro- vided with outlets that pump faster in repletion, and with intakes that pump faster in depletion. Each of many physiological func- tions has flood controls and drought controls. Turnover is absent for those components whose content is zero, and for relatively few others. Where continuous turnover occurs, machinery for recov- ery is already in motion ; recovery is achieved ' ' by modifying the speed of a continuous process" (Cannon, '32, p. 181). § 136. Velocity quotients Instead of regarding each component and its compensations as discrete, I can next regard them as coexisting in one individual. I can then compare the adjustments of diverse components, and not merely in terms of what tissues handle them, but also quantita- tively, and chiefly by means of velocity quotients. Velocity quo- tients have the same dimensions as constants of chemical reactions, and wherever they are constant with time, correspond to reactions of the ''first order," whether or not chemical transformations are involved. Their numerical values indicate how many times over, a load of the given amount would be disposed of in one hour. Diverse components in one species (table 40) of which the extreme loads studied are less than the tolerated loads, show an enormous range of values of this quotient. In man the extreme quotients differ by a factor of 10^ ; more extreme quotients may be found in still other components, as would be apparent if the recovery of excitability in nerve (16,000,000/hr., <^ 128) were included in the comparison. Grading components according to the velocity quotients shown in their adjustment, I note (table 41) that respiratory exchanges and heart frequency recover most rapidly. The elimination of 378 PHYSIOLOGICAL EEGULATIONS certain electrolytes and the replacement of excised tissues are the slowest processes. Net exchanges of heat and of water are about equally provided for. About the same order among components is represented in the quotients available for two species. Physiologists know that it takes longer to recover one's nitro- gen balance than to recover one's breath ; that an excess of bromide may last for weeks and an excess of lead for years ; while an excess of CO2 lasts for a few seconds and a deficit of excitability for a millisecond. In poorly defined terms it could be said that the fast- est adjustments are the most urgent; and for some of the corre- sponding components there is evidence that loads of them interfere TABLE 41 Rates of recoveries in dog and in man. Data from tables 42 and 40. Velocity quotient - rate /load = 1/hour Component Net exchange 1/At in dog, 1/At in man, Heart frequency Water Lactate Carbon monoxide Creatinine Total substance Gain Loss Gain Loss Loss Loss Loss Gain 90.0 330.0 33.0 0.45 1.6 0.55 0.45 0.004 60.0 29.0 0.44 0.33 1.0 0.24 0.21 0.01 readily with many other processes. Stated more accurately, the velocity quotient is a suitable measure of the promptness with which recoveries occur. The data also indicate a relation between velocity quotient and tolerated load. Attainable loads of carbon dioxide and of oxygen are very small compared to water in terms of molecules, grams, or other usual dimensions. For those two components a small incre- ment in molecules is a large increment in per cent of the content at balance, however. Again, the ratios of turnover to content (table 40) are very large for those components having high veloc- ity quotients (oxygen = 12) ; this ratio may prove to stand in close proportion to the quotient. It may mean that high sensitivity to increment accompanies large change in relative concentration, such a change being secured by having the content small. Many ex- haustive studies are required to furnish the data from which such correlations can be adequately and broadly drawn. When quotients are considered in relation to loads, several UNIFOEMITIES AND COMPAEISONS AMONG COMPONENTS 379 varieties may be distinguished. The velocity quotient may relate rates to total loads (incremental k), or limit itself to any speci- fied range of loads (limited k). Its rates may represent the total, the net, or the partial exchange ; partial exchanges being as numerous as the measurable paths for the exchanges. Diverse time intervals are concerned wherever stationary states are not studied. Of course the varieties and sizes of the living units com- pared by the quotient are as large as taxonomy. As in all the rela- tions so far discussed, therefore, there is nothing universal about -I 0 +1 +2 +3 Load'Arbitrory Units Fig. 181. Velocity quotient (l/hour) of net exchanges in relation to load. Man. The loads are in diverse units, the abscissae being scale divisions upon the graphs from which these data are derived. Water, first 1.0 hour of recovery, figure 61. Heat, first 0.5 hour of recovery, figure 144. Oxygen, first 0.005 hour of recovery, figure 172. Carbon dioxide, steady state, figure 171. Lactic acid in blood, first 0.5 hour of recovery, figure 174. Heart frequency, first 0.005 hour of recovery, figure 175. Skin healing, 96-hour periods of recovery, figure 179. velocity quotients ; they are means of comparison among particular distinguishable functions. Total velocity quotients, where known, vary with load, usually approaching infinity at small loads and approaching stable values at large loads. Even net velocity quotients are for many compo- nents not constant at diverse loads ; hence some of the comparisons made are arbitrary in so far as the loads present are unequal or 380 PHYSIOLOGICAL. REGULATIONS incommensurate. Wherever commensurate, it is feasible to com- pare quotients at the numerically same load. Frequently the quotient happens to change but little (less than twice) throughout the range of loads investigated (fig. 181). Often a sharp shift in the value of h occurs at Cq. When the range of loads is limited to either positive ones or negative ones, no trends greater than a factor of 10 are actually found in net quotients. The reciprocal of velocity quotient has the dimensions of resis- tance, for the load may be classed as a potential and the rate as a flow. When h is constant with load, then, it appears that there is no greater resistance to the exchanges under large loads than under small loads. As for other parameters, the term resistance as here used may have little except its dimensions in common with the term as used in other sciences. TABLE 42 Initial rates of recovery in dog. Velocity quotient, Jc^, = 1/hour - 0. 6 9S /half -life in hours Component Net exchange Distributee Velocity quotient Source of data Heart frequency Loss Gain Loss Loss ( ( ( i Gain Loss i I 1 1 1 1 ( t 1 1 i ( Gain Loss " Gain Gain Gain 330.0 33.0 0.45 3.0 1.6 1.5 1.0 0.55 0.50 0.45 0.20 0.10 0.06 0.05 0.02 0.04 0.01 0.014 0.004 Brouha et al. ( '36) ; Dill Water Body Serum Blood i I i < < ( ( i Body Blood i ( I ( Plasma Body i i Bile flow Body et al. ('32) Adolph; fig. 10 Kingsley; fig. 22 Freeman ( '36) Eiegel('27);Hahn('33) Wierzuchowski ; fig. 166 Zucker and Berg; fig. 170 Calcium Lactate Glucose Carbon monoxide Galactose Creatinine Urea Propylene glycol Ethyl alcohol Brilliant vital red Plasma protein Total nitrogen Chelate output Total substance Stadie et al. ( '25) Bollman et al. ( '35) Dominguez et al. ( '35) Marshall and Davis ( '14) Newman and Lehman ('37) < < Smith ('25) Stanbury et al. ('36) Eubner; fig. 162 < ( Berman et al. ( '41) Morgulis; fig. 156 If velocity quotients are considered in relation to time, k is constant whenever an exponential equation represents the data: C = ae'^K In such instances the rate of exchange is at all times proportional to load (E/AC = A;). When k is not constant with time, values are compared within some particular interval. This may or may not be a clock interval ; it is possible to compare k in the first half of each recovery, or first 1/eth, or first 1/lOth. UNIFORMITIES AND COMPAEISONS AMONG COMPONENTS 381 If the half -life (tq) be ascertained, then A: = ln 0.5/tq = 0.693/tq. That happens to be the manner in which the quotients shown in table 42 are obtained. Hence the quotient indicates how much of the load has been dissipated in the interval chosen; its reciprocal indicates how long a time is required for completion (return) of a chosen portion of the task (load) incumbent on the organism. § 137. Paths of exchange Among diverse components it is possible to compare all those exchanged through any one particular channel, whether emunctory, synthesis, or alimentation. Here an approach is made to organ physiology, a science which seems at best to recognize a small part of the relations involved. Paths can rarely be distinguished for exchanges other than those by the body as a whole ; separable paths of exchanges by parts of organisms are almost unknown except for certain types of chemical transformations. Anatomically distinguishable paths (table 43) are of two kinds: TABLE 43 Some paths prominently concerned in the exchanges of several components in dog and man. Gain (G) and loss (L). Those paths in which modification is known to occur with load of the component in question are in italics Component Eectal (fecal) Respiratory (pulmonary) Excretory (urinary) Synthethie (chemical) Alimentary (ingestive) Water Heat L (G) L L L L L,(G) G L ""l L (L) "l (L) G G,(L) G,(L) Total substance L (G) G Glucose Carbon dioxide G,(L) G L G Oxygen Chloride (L) G,(L) Lactate G, L One-way paths (renal, most alimentary) allow only losses or only gains. Two-way paths (pulmonary, chemical) regularly exchange certain components in both directions ; however, a particular com- ponent ordinarily moves in only one direction. Reciprocating action allows certain economies of movement (breathing), and is inherent in respiration and in some chemical transformations. So far as is known the kidneys of mammals do not reciprocate, except if a loss of water is defined as a gain of total concentration, or the like. In contrast, the alimentary apparatus serves for losses by excretions, regurgitations, and salivations. Conversely, each path 382 PHYSIOLOGICAL KEGULATIONS of exchange is characterized by its role in respect to each com- ponent. Exchanges might also be classified as reversible and irreversi- ble. In a strict sense an irreversible process is one from which there is no recovery. When a man loses a leg, replacement does not occur; yet there is some healing and there are certain func- tional compensations. In a second sense many water exchanges are termed irreversible, for what goes in through the mouth ordi- narily comes out through other organs. Yet these organs are so coordinated that many ' ' errors ' ' of intake are accurately adjusted by rates of output, and vice versa. Few indeed are the instances where chemical reversibility regularly operates in living organ- isms ; yet whatever happens is reversed so far as the whole indi- vidual is concerned, for loss of any component is the reverse of its gain. To distinguish a "biological" reversibility is hardly neces- sary, since the living unit simply combines processes or rates of activity that the dead unit does not use. But, perhaps most physiological phenomena are just peculiar combinations. Organisms' gains or losses of particular components at measur- able rates, occur by processes that are not understood sufficiently to allow them to be classified with much finality among varieties of forces or energies recognized in physics or in chemistry; as diffu- sion, conduction, convection, radiation, chemical transformation, synthesis and decomposition. This statement is not at variance, I believe, with the fact that the major efforts of biologists of a gen- eration have been exerted in the hope of securing such identifica- tions. Consideration of the whole body, on the other hand, instead of portions of it such as the blood plasma or the kidneys, seems to be the key to measuring the relations in which tissues, organs, and individuals may be compared. Some rates of recoveries are probably limited by patently mechanical events, such as recovery of posture in swaying, recov- ery of limb position in running or boxing, movement of food out of the alimentary tract, change of lung volume. Similarly, in any vibratory or pendular movement the recovery may be restricted by a "natural" frequency. Such limitations are generally identified from comparisons with dead units as models. They emphasize to me merely that anatomical or microphysical provisions are some- times as crucial in the organism's life as factors that are less familiar. UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 383 Some physiologists would say that recovery is ''of course" rapid for those components that are handled by organs of large proportions such as organs of respiration. This view regards anatomy not only as separate from physiology but antecedent to it. Equally, the ample organ is caring for those components for which rapid elimination or absorption often arises. Again, in one organ such as kidney there is at one time rapid elimination of one component and slow elimination of another. No one knows whether ample provision for eliminating Ji invokes hypertrophic function with respect to any J2. Or, leaving organs aside, there is provision for the rapid production of glucose within a dog and not of raffinose. Shall the organism be said to lack inventiveness toward making the latter substance, or shall the actual provision be regarded as sufficient and any greater provision as encumber- ing? Questions of this sort lead to no decisive answers; only the description of the rates found while the investigator asks the ques- tions, corresponds to the facts known. A problem of the organism is how to regulate the most compo- nents, each in the largest range of rates, with the least machinery. Were a separate machine (organ) present for each, the burden of rarely active tissues would be enormous. By multiplying the uses of one structural unit, economy of body is greater. It was once widely supposed that one "organ" served only one function, though I presume one function might include its dealing with many components. The kidneys are examples of organs that regularly excrete scores of components, and each in an independent or nearly inde- pendent clearance. How the independence of one component from another is secured, no one has ascertained. To some small extent histological differentiation is correlated with diversity of the com- ponents handled, e.g., glucose is absorbed in proximal tubules of frog, chloride in distal tubules. But there are not so many visible differentiations as components. In fact, unforeseen (foreign) components {e.g., diodrast or phenol red in dog) are handled just as specifically as usual ones. Recovery from excesses of a hundred components occurs through one and the same kidney. "In a living organ we are dealing with something of which the functions, if we speak of functions, are endless, since the activ- ities are endless, constantly seeming to grow in number as we in- vestigate further. Its true function, to the eye of a physiologist, 384 PHYSIOLOGICAL KEGULATIONS is to maintain these endless activities in balance with the endless activities of other organs, and not merely to perform one specified action" (Haldane, '17, p. 85). Where paths are still less visibly specialized, as in capillary sur- faces and cell surfaces, it is said that "permeability" controls exchanges. It is often inferred that (a) permeability is fixed, and that (b) one surface limits all components. Neither of these in- ferences seems at present justified by adequate data. Rather, there is evidence that exchange of the one component in deficit or excess is modified without modification of exchange of many other com- ponents. This fact points to equilibration of the same sort as in whole organisms. It means that (a) conditions for exchange (permeability) are not fixed, (b) permeability is not just a limiting factor, (c) permeability is not all one function, and (d) permeabil- ity probably is not vested in an ordinarily pictured monolayer. I find no evidence upon which to base a conclusion that dis- posals by storage or by chemical tranformation are faster than or slower than disposals by elimination or by translocation. Even in the same species supposedly identical paths do very dissimilar things. Each component still requires study in its own right ; no rules have been uncovered by which the rates of exchange may be predicted. For almost any one of the components that have been investi- gated it is possible to distinguish two or more paths of simultaneous exchange. In those same components, however, only one path varies its rate of exchange with any one load. The impression might be gained that the constant exchange (turnover) is by unavoid- able paths ; whereas the particular path that interferes least with other components is the sole department of adjustment, suiting its rate of specific activity to the special task before it. In the case of glucose (dog) three paths of adjustment were distinguished, each of which varied its rates of exchange over restricted ranges of load, and no two were alike. Or, three disposals came into operation in diverse proportions ; synthesis, oxidation, and excretion. There is probable utility to the organism in this precise arrangement of functions and in the exact overlapping by which they share with one another and reinforce one another. The competition among the paths of an organism is a quantitative ordering of activities so that excretion does not seize that which synthesis will preserve, and vice versa. uniformities and comparisons among components 385 § 138. Types of loading The means by which increments are produced in organisms might merit an amount of study equal to that here put upon the recoveries from increments. Indeed much effort in current physi- ology goes into the search for methods and agents by which par- ticular components may be experimentally disturbed. The means (types) are in part peculiar to each class of organisms, and therein define its properties. Increments of many, perhaps all, com- ponents are to some extent avoided or resisted by organisms ; and one study that might be made, though not attempted here, is to find how few components suffer increment from each agent. An organism in balance is defined as one that maintains con- stant, within measured limits, the contents of one or more specified components. Persons who speak of positive and negative bal- ances, would in my terminology speak of positive and negative increments, or retentions and depletions. Similarly, those who speak of ''levels" of intake would mean rates of intake or of turn- over. "Levels" of concentration or of composition are, on the other hand, usually contents of the component specified. Increments or loads arise incidentally in numbers of situations in which the organism finds itself. Often loading is the organ- ism's reaction to recognizable stimuli, often not. Descriptively speaking, positive loads follow (a) forced gains at rates exceeding the rates of loss, (b) inhibited losses at rates smaller than the rates of gain, or (c) both. Negative loads follow the inverse conditions ; but where no turnover is present only (a) is possible. For some components only (b) is feasible. The load is maintained in a stationary state when forced gain is equated with concurrent loss, or vice versa. Data of physiological significance are obtained chiefly when the organism is free to man- age at least one of the two exchanges ; ultimately it then exhibits the load and rate characteristic of State III (fig. 180). When the ex- perimenter decides the rates of both gain and loss, the organism exhibits nothing but a changing load. When the experimenter decides neither the intake nor the output, the organism enter State IV and recovers (at characteristic rates). For some components the rate of some exchange in State IV is not measurably different from the rate of its free exchange in State III, at any one load. No method is apparent of predicting for which components or under which conditions this holds true. There 386 PHYSIOLOGICAL EEGULATIONS is no evidence that exchanges in the stationary state are any more or less characteristic of the organism's physiological constitution than the exchanges in any other states. But they have the great advantage, in common with exchanges in State I, of being indepen- dent of chronology or nearly so. For many a component it has been observed that the more of it the body has, the more of it is steadily exchanged. Or, within limits R is positively correlated with C. Independently investiga- tors rediscovered this for total substance, total energy, nitrogen, water, and carbon dioxide. Surprise would have been saved in (n-l) instances had there been any inductions concerning equili- brations in steady states. The rich organism is sometimes (a) one that has much, some- times (b) one that has credit, and sometimes (c) one that spends much, either by high turnover or by temporarily reducing its con- tent. All these meanings are disentangled by relating SC/At with AC and C. The loadings and unloadings of diverse components that have been studied in one species have not all been measured in the same individual, nor simultaneously, nor under like conditions. Hence it is by no means certain that they are comparable. Comparability becomes most certain by simultaneous measurements in the same individual (chapter XVII) of diverse components and of their exchanges. Then the question arises whether component Ji be- haves the same when J2 is also modified, and when agent Y2 is substituted for Yi. So a study of loading and recovery consists in mensuration of a group of concurrent changes. The changes found characterize the agent inducing the load, and the organism reacting. The distinc- tions among agents and among modified properties replace the customary generalizations from isolated observations made, with- out benefit of clocks, upon diverse species and under random conditions. § 139. Comparison of kinetic parameters It is desirable at this point to compare several terms that are currently employed in designating the exchanges of components by organisms and their parts. Names used for these quantities are metabolic rate, clearance, accumulation rate, invasion coefficient, permeability coefficient, velocity quotient. All are kinetic quan- UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 387 titles of recognized dimensions ; each has its connotations. What usefulness has each in describing phenomena of the sorts that are dealt with in this investigation 1 Metabolic rate is ordinarily the number of molecules or quanta appearing or disappearing in a unit of living material per unit of time (L^L"^T"^, or sL^^T'^). Instead of a unit volume or mass L"^, a unit of surface L~^ is often used, and occasionally a unit of length L"\ an individual, a population, a cell. In practice, metabolic rate is sometimes limited to processes that involve recognized chemical transformations; occasionally it means only rate of oxygen consumption. Originally, of course, it had little quantita- tive connotation. Often it is regarded as applying chiefly in stationary states. Clearance started from the specific definition : ' ' the volume of blood which one minute's excretion of urine suffices to clear of urea" (Moller et al., '28). The volume flow (L^T^^) is a virtual one and not a visible one, as is more evident in the older equivalent, urea excretory ratio: urea excreted in one hour's urine/urea found in one volume of blood. Usage already extends the term clearance to other volumes than blood {e.g., plasma, volume of distribution^ body) to other intervals of time than minutes {e.g., hour), to other disposals than excretion {e.g., chemical conversion), to other paths than urinary {e.g., hepatic, unknown) and to other substances than urea {e.g., creatinine). Hence investigators now speak of plasma clearance, galactose clearance, renal clearance, complete clearance, filtration clearance, tubular clearance. The original definition is no longer a sure guide to the meanings. Sometimes clearance is referred to a unit of supposed body surface; accordingly the dimensions are often L^L"^T"\ which overlap those for metabolic rate. In supposition clearances occur in stationary states ; in prac- tice they are measured as often in states of unloading. Originally clearance was distinguished by the fact that no overall chemical transformation but only translocation was known to be involved, the same substance being measured in urine and in blood. But more recently it is permissible to measure merely the rate of disap- pearance from blood {e.g., Bollman et al., '35), really a rate of decrement in concentration, and to call this (L^T"^) a clearance without recognizing the path of exchange. It is evident that no unanimity of usage prevails, even to the extent of preserving di- mensions other than the factor T~\ Clearance is nearly always a 388 PHYSIOLOGICAL EEGULATIONS rate of emptying, and that limitation necessitates a separate term for rate of gain. Accumulation rate is such a term. Ordinarily it is related to body mass and has the dimensions L^L"^T"^, but frequently it too is referred to unit of believed surface of exchange. Occasionally it is restricted to net exchanges, and to movement against a gradient of concentration. Invasion coefficient (Bohr, 1899) was devised to express the changes of phase (partitions) involved in intake of oxygen and out- put (evasion) of carbon dioxide. It recognizes the differences of partial pressures existing at two ends of a gradient, and includes ostensibly more processes than diffusion. It is the number of mole- cules passing down the gradient of partial pressure per unit of surface and of time (L^L"^ P'^T'^). Certainly the terms invasion and evasion have emphasized the direction more than any other factor of exchange. Permeability coefficient is usually defined so as to apply only to such movements of substance or heat along a gradient of pressure or concentration as are induced by the gradient itself. But it also recognizes the surface area of exchanges (L^L"^ P"^T"^). Some- times the coefficient (7?) and time (T) are lumped together into a ''minute number." Wherever the thickness over which the gradi- ent extends is also measurable, a coefficient of diffusion (Fick, 1855) may be computed (L^L"^ L^ P'^T"^). A coefficient of osmosis may be kept separate to designate the molecular movement of water. Velocity quotient has been used here in connection with phe- nomena of excretion. In all but name it has long been current in connection with changes of excitability (time constant), rate of relative growth (increment of size per present size per unit of time) and certain other physiological events (velocity constant). Possi- bly its virtual use is coeval with the expression of biological phe- nomena by exponential equations. It is the change of quantity in a given time per unit of quantity present (L^L'^T"^), or, by exten- sion, any T"\ It may be computed from any of the above para- meters of exchange, and often vice versa. Recovery rate has had no previous distinct vogue, though often employed qualitatively. It might be defined as the net rate of exchange relative to the load (L^L"^T"^). While the load is dimin- ishing it is not equivalent to metabolic rate, but relates the ex- change to the load of the component being exchanged, and not to the UNIFOEMITIES AND COMPAEISONS AMONG COMPONENTS 389 organism's mass as such; and is not concerned in stationary states. It is identical with a net velocity quotient in any State IV. This list of kinetic parameters that have already been used in biological work might be extended indefinitely. There may be no iron-clad objection to the formulation of quotients for each new dimension and component that can be studied. But, two practical problems arise. First, a new name or letter is customarily used to express relations found in each phenomenon analyzed; later it is usual for the phenomenon to gain membership in a much wider class. Then the parameter corresponding to it loses the original and special meaning. Seemingly this endlessly arising situation may be avoided by noting the dimensions of each parameter used and trying to use only one name for each combination of dimen- sions. That is very difficult to do consistently in one 's own investi- gations and still more difficult to plan for in advance in a demo- cratic world. Second, every analysis of biological phenomena involves connotations as well as quantities. Is it more useful to unite clearance and metabolic rate under one term, or to maintain a biological separation that has no algebraic justification? Both views have their merits, and in diverse connections each solution is useful, even in one investigation. ^Tiereas an engineer would speak of Flowi (px), Flowo (po), FI0W3 (pa) • • • , physiologists are prone to speak of excretion rate (A), utilization rate (B), distribution rate (C) and disposal rate (A + B + C). The latter list bristles with implications, implica- tions that unfortunately differ for each reader. Not being a reformer, I am content to point out the situation. Every observation calls forth a desire to distinguish it from every other and a desire to identify it with others. The observer strikes some sort of a compromise. At any one stage of a science, phe- nomena are on the whole being lumped as their identities become emphatic, and split as current needs indicate. Common dimensions are supposedly prerequisite for lumping, but do not compel it. § 140. Changes in tissues, other metabolisms, etc. The data presented in this investigation illustrate the widely recognized fact that any physiological state involves some modifi- cations in a number of components. Excess of glucose makes itself felt in heat production, circulatory changes, lactate metabolism, lipid deposition, shifts of electrolytes, water deposition, and water 390 PHYSIOLOGICAL EEGULATIONS excretion. Conversely, the multifarious contents and exchanges of any one component are manifested in loads of many other com- ponents. Hence it is possible to compare, for instance, the rate of water excretion or ingestion in relation to diverse loads of each of several components (§80). Various "reactions" of compensation and defense are commonly recognized in pathological science (Bloom, '40), each of them representing a load or a rate that accom- panies some particular type of disturbance of usual content. Several parts of one organism are simultaneously recovering from the load of the whole with respect, for instance, to water or to glucose. Every cell, tissue, organ, and organism has its own recov- ery rate; and comparisons among the simultaneously observed units may be as extensive as the data. Whereas blood and liver may be alike in rates of return of water load, they may be enor- mously diverse with respect to glucose exchange. Their diversities express in quantitative fashion those divisions of labor that have been the objectives of much study. For, "special functions in spe- cial tissues" has been a mainstay of physiological concepts since Galen. § 141. Species Data such as have been presented in previous chapters furnish materials for physiological characterization of individuals, species, and classes of animals. They also extend to cells and organs. "What are some of the manners in which they may be arranged to this end? It was found practical (<§ 70) to distinguish numerous species, in respect to a single component, by their rates of turnover, their rates of maximal exchange, etc. With data now at hand for sev- eral components in one species, further differentiation is in order. A list of components and exchanges such as is shown in table 40 characterizes, I judge, only one species. Certainly the list for man differs significantly from that for dog (table 42). I believe the relative quantities can suffice to show the difference, inde- pendently of the absolute or physical units there employed. Equilibration diagrams serve as epitomes of most of what is known concerning regulation of each component, and a book of such diagrams might be composed for one species. Or, adopting a uniform arrangement and set of scales for all the materials, several components may be put in direct comparison in one dia- UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 391 gram. Thus data are available for water (JSg. 29) nitrogen (fig. 162) and glucose (figs. 166 and 170) in the dog. Taken together (fig. 182), these net equilibration diagrams furnish a beginning of a physiological description of the species, for they deal with quantitative activities that occur in response to each component's displacement. In an anatomical description, the initial study might be the observation of four legs and a head ; later the relative sizes and growths of these parts are accurately ascertained. No less significant, I believe, are the relative magnitudes of the physio- logical processes that occur in one body. Plenty of "reasons" 10 LO 10 01. 1 1 1 1 — V^ter Dog Glucose — - 1 1 ■ T 1 Woter_,_- — — ' /■^ ' Glucose 1 1 1 1 : — ' Nitrogen 1 1 .1 1 -10 -6 -4 +8 +10 2. O +£ +4 -^ Load—qm/ka. Fig. 182. Eate of net exchange plotted on a logarithmic scale, in relation to load, for three components. Dog. All are gains in deficits and losses in excesses. All rates are measured in steady states. Loads are arbitrarily compared by weight/kilogram of Data are from figures 29 (water), 166 and 170 (glucose), and 162 (nitrogen). B might be hypothesized why an excess of nitrogen ''should" be eliminated less readily than an equal excess of water or of glucose ; none are set forth now. Here are the facts ; these are the scales upon which the dog handles the three components. Modification ratios, augmentation ratios, maximal rates, economy quotients, and velocity quotients all characterize the component as well as the species. Quantitative differences among components are dis- cerned, independently of whatever machinery aids recovery of each in the body. Certain comparisons can be made by denominating initial loads as 100 per cent (fig. 183), and following their subsequent history. 392 PHYSIOLOGICAL KEGULATIONS That plan makes evident the relative time scales in which recov- eries operate. Sometimes it is possible to limit comparisons to the shapes of tolerance curves, leaving the loads in whatever arbi- trary units are convenient for covering the range of known data. Recovery depends upon net exchange, for no amount of increased intake would overcome equally augmented output. Increments of some components are disposed of with great haste (oxygen, *\00t ■^50 +25 -50 -100 0 01 OA 0.6 Hour3 Fig. 183. Relative load in relation to time for nine diverse components. Man. Bromide, data of Hale and Fishman ('08). Iodide, data of Anten ('02). Water, from figure 51 (A) and figure 56 (P). Heat, from figure 143. Lactic acid in blood, from figure 173. Heart frequency, same data as are represented in figure 176. Carbon dioxide, approximate data of Adolph et al. ('29); and of Liljestrand ('16), cf. figure 171. Oxygen, from figure 176. Skin healing, data of Carrel and Hartmann ('16), cf. figure 179. carbon dioxide), others almost not at all (fat, calcium, lead). Correspondingly the content of carbon dioxide is exceedingly uni- form, that of fat not. Hence adjustments of diverse components are to be classed according as the responses to loads, expressed perhaps as velocity quotients (tables 42 and 40), are fast or slow. Lists of such quotients give the order in which recoveries appear UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 393 to the organism (table 41), the order being nearly the same in two mammalian species. Among anatomically more diverse species it is probable that somewhat greater physiological differences also will appear. When the compensatory equilibrations in an organ and in a whole individual are compared, it will probably be found that the latter has (l) adjustments to more components, and (2) greater independence of adjustments among the several components. Ade- quately to test (l) is difficult within a human lifetime. Statement (2) has not been rigorously tested. If the man receives water with any known combination or quantity of solutes, contents of water and solute are eventually restored. But no one organ in situ is known to recover fully without the whole body recovering. And no isolated organ so recovers, to present knowledge. Here are outlined two of the many quantitative methods of comparing regulations in the organism and in its parts. It is a temptation to suppose that the compensations of the whole body are prerequisite to the recoveries of each compartment in it. For instance, the maintenance of body B might enable maintenance of ''extracellular" volume V, it in turn enables maintenance of cell L, and it in turn enables maintenance within it of Golgi body N. The dog regulates heat content in B, but such regulation is un- known for any V or L of the dog's body. On the other hand, the properties by which N, L, or V exchange water with their sur- roundings may be just as prerequisite to B's water maintenance as the reverse. All the units are coordinate, so far as I perceive, each of them regulating its components in its own sphere and fashion; there is no tenable distinction between prerequisites and postrequisites. Comparative physiology evaluates the differences among spe- cies, and the same data that manifested the uniformities serve also to indicate the contrasts among living units. Quantitative differ- ences need not depend on anatomical or chemical differentiations, and may be finer distinctions than they. Whereas biochemical tables list the compositions, potentials, and catalysts of several species, physiological tables comprise their rates of change in selected states. Comparisons have in the past usually been limited to those between two species and in respect to one component in one load. Evidently any scheme that coordinates the information of numerically diverse loads, and is capable of serving for all spe- 394 PHYSIOLOGICAL EEGULATIONS cies, extends the field of work. Anyone may derive new views and concepts in comparative physiology by listing the accuracies of regulation (variability of content), or the modifications of ex- change, of some component in several species. § 142. Equilibration diagrams Having found equilibration diagrams of great use in represent- ing what is known of regulations by compensation, I summarize their characteristics. They are constructed as follows : (a) The increments of content were ascertained for some com- ponent in a living unit. These increments were termed loads (AC). (b) The content of the component in the same or another indi- vidual in a control state was taken as zero load (Co). (c) The rates of gain and of loss (oC/At) of the same com- ponent were ascertained at diverse loads. (d) The units employed in measuring C were either sensibly identical or believed similar in loads (AC) and in rates (SC/At) of one component. (e) When rates were proportioned to loads, velocity quotients (1/At) were obtained that were independent of the dimensions of C. (f ) The diagram correlating rates with loads was constructed in a uniform fashion for several chemical constituents, energies, forces, and frequencies of diverse organisms. (g) Quantitative comparisons were facilitated among many components with respect to the relative rates and velocity quotients found. The generalized equilibration diagram (fig. 110) has the follow- ing features : (h) G (gain) > L (loss) whenever AC is negative, and L > G whenever AC is positive. Each is of a nature that tends to restore AC to zero. (i) At balance G = L. This is found to occur in those states that were chosen as controls or zero loads. (j) At other loads than zero, either G has another magnitude, or L has another magnitude, or both. (k) The rate of net exchange increases as load increases. (1) The slopes (1/At) of the lines correlating net rates with loads near the zero load, measure the velocity with which any small load that accumulates is again lost from the organism. (m) The equilibration diagram represents a symmetry of proc- UNIFORMITIES AND COMPARISONS AMONG COMPONENTS 395 esses that restore the organism after negative and after positive displacements from balance. It represents also a contrast of total gains and of total losses at any one load. I think the physiological importance of the features found to prevail among all equilibration diagrams is best exhibited by sup- posing that the diagram turned out differently : (n) It might have occurred that neither L nor G varied with load. Then one value of C would be as likely to occur as any other ; no value would persist nor be approached in recovery, except temporarily and by chance. (o) L might vary, but equal G at all loads, with the same conse- quence. (p) In negative loads L might exceed G, and in positive loads vice versa. Then recovery would be impossible, for the content would move farther from Co- These possibilities were named in §72. (q) L might vary, but equal G at more than one load in the same individual. Then if once sufficiently displaced in content, a new balance would be maintained. (r) L and G might vary at random and independently, never being coordinated. Then recovery would be wholly left to chance, like an animal in an insoluble maze. It is apparent that there are numerous possible arrangements in the organism that would not automatically lead to recovery. That these other arrangements are not found, indicates to me that the relations represented are of consequence to the organisms con- cerned, that probably the species would not be here without arrangements that are successful and that insure survival. It was mentioned (§72) that total exchanges G and L might each be classed into 11 types of modification in relation to load. Together, 121 combinations of the two seemed possible; 96 were judged incompatible with recovery. Of the remaining 25 combina- tions only a few were found for the component water, but several have turned up in the study of other components. Reviewing the equilibration diagrams that have been presented, I find 6 of the 25 in actual operation, not counting fragments. These 6 are repre- sented in figures 13, 61, 72, 146, 171 and 172. When any other type of equilibration diagram shall be found, it will offer an interesting contrast with these. It is possible to suggest a few incidental features of physiology 396 PHYSIOLOGICAL EEGULATIONS that may have discouraged the previous realization of quantitative equilibrations: (1) Until recently, long periods of time were used in measuring metabolisms and other rates of activity. Hence re- covery of balance was often completed before the measurement of exchange was made. (2) Conditions of balance were preferred; turnover and minima were ascertained first. (5) Loads were mea- sured chiefly in blood. But the rates of exchanges by blood were too rapid to measure in some instances ; it was difficult to assmne that volume of blood was constant ; and only net rates of change in con- centration were ascertainable. (4) It was supposed that all the separable processes of exchange needed to be labelled before de- scription of the exchanges could be undertaken. For the most part, however, the materials were at hand, and a mere shift of inquiry finds new relations present among them. The specification of conditions for study, of course, suggests innumerable further studies by these procedures. Equilibrations would be concerned with all the capacities of living units to combat excesses and to make up deficiencies. In a general way several bio- logical sciences are mainly concerned with just these ; particularly toxicology, pathology, ecology, and physiology. One of the imputed virtues of a scheme like that of the equilibration diagram might be that one investigator need no longer be a specialist on one com- ponent. Rather, having studied the relations of Ji in one living unit Ui, he is now qualified to tackle any J in any U with which he is interested in becoming familiar. He may revel in any quantita- tive relations and not be confined to a single component. What quantities (components) in organisms are susceptible of study by equilibration diagrams ? I should predict that any mea- surable characteristic of any living unit may be correlated with the rate at which the same characteristic changes. Seemingly any pro- cedure for quantitative measurement, any organism, and any clock, are all that are requisite to the investigation of an equilibration. The devices or procedures employed are not limited to those familiar in any one field of biology or of non-biological science. Wherever there is a definable system, changes occur in it whenever displacements or increments of component occur. § 143. Summary of the variables studied For any one component, some nine sorts of quantities were investigated. These were roughly classified (§ 101) into three dis- UNIFORMITIES AND COMPAEISONS AMONG COMPONENTS 397 continuous variables: living unit, type of loading, and path of exchange ; and six continuous variables : load, time, rate of ex- change, velocity quotient, change in tissue (volume, dilution, amount) and change in activity (energy transformation, behavior). A useful summary of these variables is their definition, which may now be succinct, together with references to the foregoing use of certain ones. Living unit, U, the portion of organisms under actual observa- tion or measurement. Component, J, any measurable property of the living unit. Content, C, amount of component present in the living unit. Load, AC, increment of content; Excess, + AC; Deficit, -AC (fig. 110). Balance, Co, when C = 0. Time, t, hours by the solar clock. Rate of exchange, R, rate of transfer or change of component, SC/At. Gain, G; Loss, L. Total, net, partitioned. Loading, numerical increase of ± AC. Recovery = Unloading, numerical decrease of ±: AC. Recovery rate, net R during unloading. Turnover = Turnover rate, Ro, or R at Co, when L = G. Tolerance, sequence of loads. Velocity quotient, h, 1/At, R/AC. Augmentation ratio, Rm/Ro where R^ is measured total rate, usually maximal or minimal. Modification ratio, Rmax./Rmm. Economy quotient, G/L. Physiological state = State, combination of contents in the liv- ing unit, with emphasis on the one or more contents measured. Condition, any environmental situation of U believed to be reproducible. Stationary state, any state in which C does not change with time, hence G = L. Kinetic equilibrium = Normal, stationary state in which Co and Ro are present. Equilibration, group of relations of exchange to content, or rate to load. Body mass, B, measured weight or volume of U. 398 PHYSIOLOGICAL EEGULATIONS Volume, V, any volume in U measured by a specified procedure. Usually a volume of distribution (Vd) of a distribuend added to or subtracted from U. Dilution, E, reciprocal of concentration of some substance or physical colligative found in some part of U that was analyzed. It may be piously hoped that few inferences will be drawn from the terms used in this nomenclature, that would add to the mean- ings conferred by the methods of measurement as actually used. In so far as each term is an abstraction, it is only so by virtue of the fact that similar measurements were employed for similar data. If these actual methods be kept in view, it may be more difficult than otherwise to wander far into metaphysics from the phenomena and relations so described. <^ 144. General summary All the quantitative correlations in this investigation fall into a very few types. Each type includes those correlatives having cer- tain dimensions ; hence to summarize them it is only necessary to refer to the nine classes of variables (§ 101). Taken two at a time the correlations would fall into 36 types. In actuality those utilized were more limited than this; in nearly every case either time or load was one correlative ; hence, at most, 15 types prevailed. A few of these correlations were common enough to merit special names ; these were : Load vs. time, the tolerance diagram. ACoct. Ex- amples, figures 180 (A), 1. Rate of exchange vs. time, the exchange diagram. Rcxt. Examples, figures 180 (B), 2. Rate of exchange vs. load, the equilibration diagram. R oc AC. Examples, figures 110, 13. Particular portions of tolerance curves, or curves for particular components, are currently designated by special names. Thus, there are insulin sensitivity curves, contraction curves, elimina- tion curves, accumulation curves, and penetration curves. Simi- larly, particular kinds of exchange curves are diuresis curves, excretion curves, and age curves. It seems to me that by the classification of those and other vari- ables, and of those and other correlations, into a few types, a huge mass of data assumes manageable proportions. Instead of a be- TJNIFOEMITIES AND COMPAEISONS AMONG COMPONENTS 399 wildering miscellany the material becomes ordered and useful. Considered in this way all that has been presented may be viewed as requisite to comprehensions of organisms. Thereupon it be- comes possible to derive general statements concerning what organ- isms do to maintain their constant properties. It is clear that regulations have to do with what I have defined as contents (properties) of the organism. Content of each com- ponent varies within the limits fixed by the processes of its mainte- nance. Departures from this norm are forestalled by behaviors whereby the animal frequents environments that favor the norm, and by compensations whereby exchanges are modified in a direc- tion to restore the norm. Some processes are ready at all times to prevent or to correct departures of every component. A separate structure is not available for performing each such duty. All the processes have patterns that are similar in many features that have been mentioned ; in addition each exhibits quantitative differences peculiar to it. The species is physiologically characterized by that particular combination of these quantitative features which it embodies. Chapter XVII INTERRELATIONS AMONG COMPONENTS § 145. Up to this point I have studied each component as though it were independent of others. Though an organism evidently has many components under regulation, each could be examined sepa- rately. Had that not been possible, but little progress in their analysis would have been made by the correlation of so few vari- ables as were actually recognized. The more closely a physiologist examines what is going on in an animal, the more he realizes that any single component is not receiving the organism's whole ''attention." To find mutual influ- ences at work, he may measure adjustments of two or more com- ponents that are simultaneously disturbed. No new materials might be needed to show coincident recoveries among several components, for some have already been presented in the form of correlatives of water load (§ 80, § 85). There, a tendency pos- sibly prevailed to regard water increment as an independent vari- able and other components as dependent. A more explicit study of several components upon an equal footing, and the exchanges of each, may therefore serve to indicate what a bag of interrelations the organism is. Data appropriate to the study of great numbers of simultaneous properties and events are not difiicult to record, for tables may fully represent them. The desire to visualize more and more relations among them, however, drives me to explore additional methods of representation. The same sort of urge that compels an investigator to seek more data, here plunges him into a search for multiple relations. § 146. Heat load and water load in man Everyone knows that a quick way to get into negative water load is to overheat. Similarly a convenient way to get into negative heat load is to drink ice water. These two sorts of experimental situation may be considered in detail. Transitions from heat balance to considerable positive heat loads occur when a man walks for an hour in the hot desert (Adolph, '38, p. 492). During walking (transitional state), as heat 400 INTERRELATIONS AMONG COMPONENTS 401 accumulates in the body, water leaves the body; this I represent by plotting the heat load against the simultaneous water load (fig. 184, J). At the end of the period of walking, the man sits at rest in a cooler atmosphere indoors ; recovery of heat content begins a few minutes after the exercise and exposure to solar radiation cease. Later, recovery of water content begins, for now the subject drinks cold water as rapidly or as slowly as he desires it. In this period both heat load and water load are diminished, so that the initial or balanced state of the man is approached. Rate of Net Heat Exchange To"tal Water Load Fig. 184. Heat load (Cal.Ag- of body weight) in relation to water load (% of Bo). Man. The data are arranged to show the simultaneous courses of both loads (and of the net rates of exchange, per hour, of heat and water) during loading and recovery. Ex- posure in the desert conferred negative water load and positive heat load (two tests on 1 subject are averaged) ; ingestion of ice water conferred positive water load and negative heat load (3 tests on 3 subjects are averaged). In aU tests, recovery occurred in air of about 31° C. Each point is the mean for a period lasting 0.37 hour. Further data of Adolph ('38) and of Pinson and Adolph ('42). At every stage of the tests, the rates of net exchange both of heat and of water are, of course, known. When those rates are each related with the loads of the same component (fig. 184, M and P), portions of two simultaneous equilibration diagrams are ob- tained. Anyone may trace corresponding points on the two sepa- rate curves by following various dash lines from one equilibration curve to the other. In this manner the relations among four 402 PHYSIOLOGICAL REGULATIONS simultaneous quantities, two loads and two rates of exchange, are established. I now consider the second sort of experiment. Negative heat loads together with positive water loads follow the drinking of ice water (fig. 145). Recoveries of both components simultaneously are thereafter accomplished while sitting. Relating heat load with water load (K, fig. 184), I observe that the load of heat as estimated increases after the load of water begins to diminish. But later the load of heat diminishes the more rapidly of the two (as judged by the fraction of the load disposed of in a unit of time), so that heat balance is at the end of 2 hours more nearly restored than water balance ; the same fact appears in J also. At every stage of K the rates of net exchange of heat and of water are known ; indeed the total rates also were measured (fig. 145) but are not plotted here. The two exchanges, when each related to loads of their own component, yield two equilibration diagrams (N and Q). Thus two of the four possible combinations of positive and negative loads of two components are worked out; the other two an experimenter could secure by appropriate arrangements, such as ingesting hot water to make both loads positive, and injecting pilocarpine to make both loads negative. Further, each admin- istered load might be varied in diverse tests, particularly in rela- tive magnitudes of the two components, each yielding data for a ''loop" diagram. While the development of loadings might be regarded as controlled to a considerable extent by the experi- menter, the recoveries are not, in so far as they freely proceed in a uniform environment. What comparisons of simultaneous quantities are facilitated by the type of diagram shown in fig. 184? The following may be men- tioned: (l) loads of two components, (2) positive and negative loads of each component, (5) time relations of loading and unload- ing, (4) rates of net exchange of two components, (5) rates of load- ing and of recovery at one load, and (6) four quantities ascertained at one time (2 loads and 2 rates). Quantitative relations, in this as in every other diagram, in- crease in clearness as time is spent in working with them. The number of data contained in this figure is no greater than in a table, but more relations are shown than would be explicit by non- graphical methods. The four quantities recorded do not just INTEERELATIONS AMONG COMPONENTS 403 ''happen" to be simultaneous; rather their simultaneity manifests the quantitative interrelations among them. Meanwhile the spe- cific communications by which they are related remain unidentified. What is learned from simultaneous increments of two com- ponents that was not found in each component separately (figs. 61 and 144) ? (1) Recovery of water content is slower after excess of ice water than after excess of warm water (data of Pinson and Adolph, '42). (2) Recovery from deficit of water content is faster after the subject is overheated than after he is only deprived of water (fig. 58). (3) Maximal rates of total water loss during exercise in a hot atmosphere (loading) are greater than those in positive water load without exercise (unloading). (4) Recoveries of water take longer than recoveries of heat, in both of the combinations tested. Or, heat velocity quotients are greater in positive heat loads than in negative heat loads, they are greater than water velocity quotients in negative water loads, and in positive water loads. (5) During exercise in the hot desert (loading) the velocity quotient for water loss approximately equals that for heat gain. A corollary of this fact is that line J is straight, water loss being proportional to heat gain during the exercise. It is strictly not characteristic of all exercise tests in the desert nor elsewhere. (6) Rapid water loss (sweating) accompanies a large heat load and a rapid heat loss. Even if no one knew that vaporization of water is always accompanied by a loss of heat, the establishment of this relation would be just as significant, upon the criterion of description. The "willingness" of the organism to contract a water debt when in positive heat load seems to me of the same order as the ''willingness" to absorb the cold as well as the water from the alimentary tract into which ice water is put. The constitution of the organism and of the universe makes them equally inevitable, I believe. How reciprocal is the correlation between heat exchanges and water exchanges in the course of daily life? Negative water load develops where positive heat load prevails, but usually not vice versa, for negative heat load often induces negative water load too 404 PHYSIOLOGICAL REGULATIONS (cold diuresis). It may be that physiologists are merely more familiar with the instances in which heat load precedes water debt ; for it is easy enough to arrange for the inverse to happen by de- pleting water, then administering a pyretic drug to develop a positive heat load or pilocarpine to develop a negative heat load. Here it may be observed how arbitrary is the distinction between what the organism does during loading of its ''own free will" and what is ordinarily regarded as forced upon it. Each interrelation of water and heat is quantitatively char- acteristic of the organism. Those oxidative chemical reactions that produce 0.12 gm. of water with each calorie of heat, depend upon the constitution of the organism just as much as do the activities whereby during loading in the desert 13 gm. of water are lost while each calorie of heat is gained (line J). The only difference I discern is that the former ratio is less readily modified by condi- tions than the latter; it is more often seen in vitro. Actually a rather constant relation was found (Adolph, '38, p. 494) between heat retained and water debited. If each of these ratios had a name, chosen as acceptably as ' ' respiratory quotient, ' ' the relation would seem real to more individuals. When a quotient is bandied about, discussed, and abbreviated, it takes on an individual meaning. Numerous other data relate water loads (various body weights) with simultaneous heat loads (various body temperatures). Tests of McConnell and Yaglogiou ( '25) may be used to trace the first half of curve J during exercise indoors in still air or in moving air of diverse humidities. Data of Winslow et al. ( '37) and of Hardy and Soderstrom ( '38) cover some of the same and many other atmos- pheric conditions. Each set describes what happens in the human body when heat accumulates and water depletes. Unfortunate it seems that those observers did not investigate recovery states. In brief, two components may be experimentally loaded at one time. Then both during loading and during recovery the relative changes in their increments and their exchanges may be compared. In examples analyzed the increments are (i) negative heat and positive water, and {2) positive heat and negative water. Equiva- lences in rates of their exchanges indicate certain patterns by which ane component is preferentially treated or functionally linked in others. INTERRELATIONS AMONG COMPONENTS 405 § 147. Several components during physical EXERCISE (man) In many researches a number of physiological functions have been measured during and after physical exercise. A small series of such measurements has been analyzed above (fig. 176) ; it may be used for further correlations (fig. 185). Change in Rate of Cardiac Output +> 10,000 0 Load +400 Rote of +600 +800 Op Consumption § 6l~0 +200 Load Heart Frequency Fig, 185. Load in the rate of cardiac output in relation to load (a) in the heart frequency and (ft) in the rate of oxygen consumption; during and after physical exer- cise. Man. The simultaneous movements of two loads at a time are shown in the upper left diagrams. Below are two equilibration diagrams for quantities (a) and (6) ; to the right, one equilibration diagram for quantity (c) ; dash lines during gains, solid lines during losses (recoveries). Corresponding points are connected at diverse intervals of time. All loads are % of the initial rate of frequency, all changes in rate or fre- quency are in % of the initial rate or frequency /hour. The data are those of Hjin- Kakujeff ( '36) in figure 176. Three selected components are here interrelated two at a time ; rate of oxygen consumption, heart frequency, and rate of cardiac output. No objection has been found (§72) to treating rates of activities as components ; since rates are now loads, accelerations are their rates of net exchange. It happens that in these three com- ponents all the loads are positive. 406 PHYSIOLOGICAL, REGULATIONS At simultaneous points rate of cardiac output is correlated with heart frequency ("W, fig. 185). The points during exercise (in- creasing loads) and the points after exercise (decreasing loads) fall upon a single line W. The changes in rate of cardiac output (V) and in heart frequency (T) therefore form similar halves of two net equilibration diagrams. Both in transitional and in re- covery states the initial changes are the fastest. When cardiac output is correlated with oxygen consumption (X), the points fall along a very narrow loop. The shape of the loop is an indication of the great deceleration of oxygen consump- tion at the beginning of recovery (curve U). Again the initial changes of state are the fastest. The similarities among the three components are much greater than anyone would anticipate in figure 176. Not only the loads in relation to time, but the loads in relation to one another and in relation to their rates of change are of one pattern. But the mag- nitudes are very different when computed, as all of them are, in per cent of the rate or frequency in the control state. Possibly relative scales that would make the magnitudes equal are justified; further utility in such an equalization would be found if it turned out that in a variety of kinds and intensities of exercise and of other circumstances, uniform proportions prevailed among the three components. Figure 185 shows that cardiac output changes 3.9 times as much as heart frequency. Or, for each 1.0 per cent of increment in heart beat 3.9 per cent of additional blood circulates. This is the slope of line W. Similarly the slope of line X is such that cardiac output changes 0.44 times as much oxygen consumption. And, oxygen consumption augments 8.9 times as much as heart frequency. The three components are, indeed, commonly recognized to be related to one another by the fact that their ratios have names: (a) oxygen consumption/cardiac output = arterio-venous oxygen difference; (b) oxygen consumption/heart frequency ^= oxygen pulse; (c) cardiac output/heart frequency = stroke volume. Each of these ratios of absolute units varies somewhat during the exer- cise, since none of the three regressions is exactly linear and through the origin. The acceleration in exchange of the three components are simi- lar, within about 15 per cent, for uniform fractions of the loadings ; that is, the three dash lines of figure 185 are superimposable if INTERRELATIONS AMONG COMPONENTS 407 the maximal loads present in each are brought to a common scale. But the decelerations are probably significantly diverse ; upon the superimposed scales the deceleration of oxygen consumption is greatest and of heart frequency is least. Altogether, 6 variables are represented in figure 185, and simul- taneous points of all 6 may be found connected by light dash lines. Among varying numbers of variables from 2 to 6 at one time, the total number of combinations is 57; or among 2 at a time is 15. Whatever features are evident now that were not noticed in figure 176, are brought out solely by the transformations of coordinates. Many other quantities have been studied during and after phys- ical exercise. Indeed, the same tests that have just been analyzed furnish data for systolic and diastolic arterial pressures (fig. 176). Each of these is very different from the other 3 components; for both show negative loads at some portion of the test ; and none of the 7 additional bicorrelations of each with other single loads forms a line having a narrow loop resembling X or W. When other data concerning physical exercise are examined, the conditions and states and individuals concerned are, of course, dif- ferent from the above. Selecting two familiar components whose rates of change contrast greatly, I correlate heat loads with simul- taneous values of heart frequencies, using data of Christensen ('31). Two rates of work are represented in figure 186. In S a stationary state with respect to heart frequency is maintained dur- ing the interval from 0.2 to 1.0 hours after work began, but not with respect to rectal temperature or to heat load. In R, what looked at first (0.1 to 0,5 hr.) like a stationary state of heart fre- quency, later proved distinctly unsteady. Both tests show approxi- mately rectangular loops correlating the two loads, two of the four corners being at the start and the stop of exercise. A partial diagram of this variety was used by Dill et al. ('31, p. 512). The rates of loading and unloading contrast sharply for the two com- ponents. But each rate follows a similar course in loading and in unloading. Heart frequency augments and diminishes suddenly at start and stop, so that most of its transition and its recovery is completed in 0.05 hour. Heat load changes almost uniformly with time throughout the test, requiring about as long for its recovery as for its loading, and never reaching a stationary state. All the components that have been studied in chapter XV, ex- cept those of tissue replacement, have been at some time or other 408 PHYSIOLOGICAL EEGULATIONS measured during and after exercise in man. Each could enter a series of interrelations of the sort indicated. To insure a basis for comparisons it is desirable, of course, to secure all the measure- ments during one bout of exercise, or during a series of repro- ducible bouts. In the end, economy of representation of the data could be secured by constructing alignment charts : one chart for loads, one for rates of change, one for velocity quotients, or one for all combined. The components considered would include water, heat, total substance, potential energy, external work, glucose, oxygen, carbon dioxide, lactate, chloride, nitrogen, and as many ^24 *2.0 o O I "§^1.2 o _J t-oe (U X ^04 X /-I2 /-I 3 hr S"tart exerci stop exercise 1.0 hr^ o.aJ Ob-\ I -0.4 04- 01 hr-.^0.2 I -0.2 vLJ-Olh 0 ■^20 +40 +60 ^80 +100 +120 Heai-t Frec[uency Load — percent Fig. 186. Heat load (Cal.Ag-) in relation to heart frequency load (% of initial frequency) during and after exercise. Man, subject M.N., 72 kg., working on stationary bicycle. E, 86,000 kg.m./hour external work for 1.50 hours; S, 58,000 kg.m./hour for 1.00 hour. Heart frequency and rectal temperature were originally read every minute. Heat load is computed as increment of rectal temperature x 0.83. Data of Christensen ('31, pp. 463 and 464). others as could be handled. The limitations to the investigation would not be in securing the data but in grasping the interrela- tions present, for the number of correlations even when taken only two at a time, increases as [(n-l)^+ (n-l)]/2, or (n^-n)/2. The reward of the investigation would be a picture of how the organism coordinates a number of simultaneous processes in a combination that would snarl up many a man-managed factory indefinitely. INTERRELATIONS AMONG COMPONENTS 409 Physical exercise is, in brief, an example of physiological ac- tivity that involves numerous components. Two groups of them, examined with respect to loads and net rates during loading and during recovery, are found to be characteristically intercorrelated both in cases where nexuses are recognized and in others. Eecov- ery in each is relative to recovery in the remainder. § 148. Some other components In the past the physiological interrelations among components have been studied chiefly: (a) where one chemical compound is formed from others, (b) where cations and anions balance, or di- verse cations replace one another, (c) where energy is transformed from one kind to another. The studies may equally well be ex- tended to components where ' ' there is no reason to expect ' ' mutual associations. As soon as "empirical" correlations have been found, plenty of "reason" will usually be forthcoming. Many physiologists object to the correlation of data except in the light of some theory. To write A = f(B) is, they say, useless unless f is either a constant or a believed causal connection. But it is prob- able that any one relation, usually a partial correlation, between any A and any B, is eventually worthwhile for someone to investi- gate. When he feels the urge to find how A is related to B, that urge can be legitimately satisfied. Indeed, the manners and forces of interactions in organisms seem to be so various that I believe physiology cannot afford to wait for non-empirical hypotheses or visible "connections" to precede measurement. For, meanwhile physiology would be limited by the contents of other sciences. As an instance, relying on the knowledge of chemistry, no one has been able to guess yet how much water is retained in mammalian livers of any species for each gram of carbohydrate (glycogen) loaded. All correlations, then, are equally valid, according to any gen- eral criterion that has been proposed, other than opinion. A = f(C) has no more statistical significance than A = f(B), though it may have a higher coefiicient of correlation, or carry smaller probable error, or be known to occur in more species. Only by personal judgments and attractive hypotheses is a greater intimacy assigned to one relation than to another. I doubt whether it is any contribution to physiology to say that ' ' all measurements of the state of the body with respect to water must be relatively 410 PHYSIOLOGICAX, REGULATIONS empirical until the mechanisms of regulation are understood" (Adolph, '33, p. 349). Evidently a non-statistical kind of signifi- cance is referred to in the statement of C. Y. Cannon et al. ('32) that "If these results could be explained or predicted, reasoning from a physiological basis, they would be highly significant. Since no good physiological explanation can be made, the level of sig- nificance secured is only great enough to warrant further test." Wishing to find in what manner the activities of the organism are knit together, I do not think it desirable to suppress knowledge of any relation that comes to light. Were no other connections visible, body size alone would require that water content, water turnover, heat exchange, total substance, heart frequency, and many other quantities are related in the way they have been ascertained to be. They need no common organ of exchange or common physical characteristic to fix them at those particular rates. In attempting to ''explain" those observed cor- relations, one theory suggests one connection where a hundred probably exist. Experience teaches, I think, that biological rela- tions usually have too many factors to allow identification of pre- dominant or causal ones even if such exist. Having decided that any two kinds of measurement made simul- taneously upon one living unit may be usefully correlated, I might now analyze any of a semi-infinite variety of investigations. It seems appropriate to mention several that concern the component water, with which the present work started. Each will illustrate a system of variables manifesting an interrelation among compo- nents that are being simultaneously handled by a living unit. (1) In the dog, the continuous infusion of 1.1 M solution of glu- cose induced marked losses of water (experiments of Wierzuchow- ski, '36). The urinary outputs of water and of glucose were mea- sured during 6 hours of infusion, when for each gram of glucose retained about 24 grams of water were depleted. At the same time exchanges of oxygen, carbon dioxide and some others were ascer- tained, allowing interrelations to be established among 18 com- ponents. (2) In the dog in water loads, measurements of concentrations of diverse variables in blood have been presented (figures 126 and 117). Each variable may be designated as a distinct component, and in reality the data already used in other connections are now available for comparisons among components. Indeed, table 24 INTERRELATIONS AMONG COMPONENTS 411 quantitatively correlates 7 components that were measured simul- taneously. These interrelate the state of and what is happening in the whole with the state of its parts. In the same manner equilibra- tions also might be represented, as they exist simultaneously in diverse parts of one organism. They would indicate the competi- tive rates of recovery among the parts. (3) In man, water was withdrawn from the diet to varying extents, and in daily periods the net exchanges of seven electrolytes were measured (Wiley and Wiley, '33). Conversely, to constant diet and water intake each of four salts in equivalent amounts was added in a separate period (Wiley, Wiley and Waller, '33). Dia- grams may be constructed by plotting the body's load of one against the load of another, upon which each period traces a differ- ent loop. It was found that deficit of water leads to deficit of electrolytes, and excess of electrolytes leads to excess of water, a relation that is, however, only qualitatively reciprocal. In other tests on man, water and varying amounts of sodium chloride were taken by mouth either together or separately (Adolph, '21). The load and the rate of exchange, of water and of chloride, depended on the relative times and amounts in which they arrived in the body. Water might be depleted by salt excess when intake of water was denied. Evidently interrelations need not be limited to simultaneous states but may relate what follows with what precedes. Very many speculations suggest how electrolytes may be related among themselves and to water in the human body ; planned experiments can test all of them quantitatively. In a third set of tests on man, sodium chloride was withdrawn from the diet, while the exchanges of water and several electro- lytes were measured (Goodalland Joslin, '08; Stohr, '34; McCance, '36, '37). Deficit of sodium led to deficit of water and of some electrolytes ; numerous correlations are possible that are limited by the fact that each individual was tested differently. Going further, if the researches of Tuteur ('10), Wiley, and others had been planned to allow correlation with these latter ones, the same amount of manipulative effort would have produced results capable of a high degree of synthesis. There is no great trick to finding interrelations among com- ponents ; most researches in physiology that are at all quantitative, ascertain data suitable for the sort of study here illustrated. A point to be emphasized is that a constant proportionality, or an 412 PHYSIOLOGICAL REGULATIONS equation that relates two quantities under a variety of circum- stances, is not needed before the interrelations may be analyzed and represented as aspects of the organism's functioning. Count- less are the circumstances in which heart frequencies have been measured in man. All of these circumstances are associated with change in some other measurable component, and often with known loads of many components. Each load is a '^factor" in heart fre- quency; whether each is a separate factor is a matter for definition and for further study of relations. Hidden in the data of physi- ology are, I believe, hundreds of clear correlations that have not even been surmised, which nevertheless indicate how components are related in the animal body. The physiologist may take advantage of the generalization that every bodily function (component) is related to nearly every other. Keys et al. ('38) wished to select persons who would acclimatize advantageously to high altitudes. They, therefore, made eight random measurements upon each of ten men ; later took those men to the mountains for some weeks, and found which were able to do most physical and mental work there. Correlations between the initial measurements and the later performances revealed what combinations of physiological characteristics were best suited to life in the mountains. The important feature is that any set of tests may serve as a basis for discovering a given physiological aptitude. No one need know what constituent processes are at stake in the particular acclimatization; ''blind" trials upon random individuals will furnish the information needed for extensive predictions. In the end, physiology has a set of relations similar to what anatomy has in "Cuvier's Law." Cuvier (1827, p. 83) realized that he could pick out bones belonging to one species of animal from a mixture, and predict the conformation of the jaw and teeth from that of the claws. ' ' Every organized being forms a peculiar sys- tem of its own, all the parts of which mutually correspond." Simi- lar statement is frequently possible in physiology and in functional pathology, as has long been shown in medical practice. Just as a hypertrophied heart often accompanies unusually high arterial blood pressure, so deficit of salt often accompanies modified func- tion of adrenal glands. To say that one can see no ''connection" between what these small glands are metabolizing and what the whole body is metabolizing may be correct; but that statement blurs an interrelation, which holds even though in some bodily INTERRELATIONS AMONG COMPONENTS 413 states the glands are dispensable. Again, the rapid turnover of water in a frog, in quantitative contrast to that in a man, accom- panies formation of nrine always hypotonic to the blood plasma, immersion in fresh water, and absorbent skin. The frog's normal state depends upon them all, and the impression is gained that any other combination of properties would be incompatible. Interrelations and interactions among components that are under increment are properly connoted by the term syndrome. This term has the present advantage of emphasizing the total nature of the physiological state instead of some supposed mechani- cal relation among parts or components. In general, interrelations among components may be estab- lished and appreciated without knowledge of the lines of mutual communication or influence. Interrelations are the substance of physiological science. Several illustrations serve to indicate that the unforeseen character of the relations is so common as to be no barrier to the acceptance of all correlations that are statistically significant. Compatibility among properties is visible in every maintenance and recovery of those properties. § 149. General features In the data and interrelations here presented, certain character- istics may be discerned. What particular and what general con- clusions may be drawn 1 How may the methods illustrated furnish a synthetic picture of the physiological patterns of the organism? (1) First some characteristics of the diagrams here employed may be mentioned. Diagrams were constructed in which load of one component was plotted against load of another component (fig- ures 184, 185, 186). (a) The broader the *4oop" (ratio of its small- est diameter to its largest diameter) the greater is the difference in rates of net exchange of the two loads. Scales might be propor- tioned for this purpose by taking maximal loads as equal coordinate units. (h) When no *'loop" is present, correlation of the two loads is very high ; the correlation is then almost independent of time. (c) The departures from an oblique line connecting the points of initial state and of stationary or extreme state, are often equal upon the two sides of it. This symmetry means that loading and unloading of the two components bear similar relations to one another. 414 PHYSIOLOGICAL REGULATIONS (d) In equilibration diagrams, the greater the difference in rates of loading and of unloading, at any one load, the more is time a correlative and the less is load a correlative, of the rates of exchange. (e) Recovery in one component does not prevent recoveries in others, though it may diminish their rates. Discharge of one load does not often create large loads of another component. Were any one component predominant it might easily be that all sorts of disturbances (loads) would arise in an uncompromising restoration of that component to zero load. Perhaps such easy compatibility manifests niceties of coordination among the handlings of diverse components. Conflicts may continually arise : obtaining water may preclude cooling off, fleeing may inhibit excreting water, mobilizing glucose may involve destroying water. The solution of those con- flicts is the organization of the organism. Conceivably, interrelations are of two sorts. In one, Co is fixed for all components ; that is the case in the data presented. In the other, Co changes for one or more of the components. The diffi- culties of distinguishing whether Co changes or not, lie in the effort to define Co, not in the record of what happens. If it be said that final content (Cj) after f hours is the criterion of recovery without change in Co, then d = Co is the sign of it. If temporarily gain equals loss at a content other than the original Co, then that is the criterion that Co is not fixed. Such distinctions are required only when words are to be used to summarize the relations among com- ponents. (2) One question that has not yet been explicitly asked is : how specific is the interrelation between the component that is loaded, and the modification of exchanges of the very same component? The only adequate answer is to be derived, I believe, by a correla- tion of the following sort. Set down as abscissae all components whose loads are to be tested, and set down as ordinates all com- ponents whose modifications of exchanges are to be observed. See what exchanges occur, and by how much each is modified, over a wide range of loads of each component. To do this adequately, a diagram showing net exchange of each y against load of each x is constructed in each box of the chart. It turns out, for the com- ponents listed in chapter XV, that exchange of no ' 'foreign" component is modified as much as is the component that is under increment. For a remarkably high proportion of components, con- INTERRELATIONS AMONG COMPONENTS 415 siderable modification appears only in increments of y itself. The relation between load and corrective exchange is therefore unique and specific ; the very component that is present in unusual amounts is the one whose exchanges are most modified. The existence of that specificity was guessed in the first place; now that it rests explicitly upon a considerable array of facts, its non-random char- acter is more evident. (3) Having found that certain features become easy to grasp in diagrams of interrelations (load of one component plotted against load of another component), how can I formulate multiple corre- lations? Or is it necessary to stop with two or three components at a time? One step might be to represent an additional component in a 3-dimensional diagram. In another direction it is possible to proceed by combining com- ponents. Wherever two or more components have the same dimen- sions, their combination may be a sum or a mean; the values obtained, however, preserve the definitions given to each of the components included. Where the dimensions differ, restricted progress is possible, (a) The components may be redefined to make them commensurate in terms of control contents of each ; in terms of standard deviations of each; in terms of lethal loads of each ; in terms of selected coefficients of physical or physiological equivalences (§ 134). (b) A combined load may be represented as a multidimensional sum or product, e.g., x gm./kg. plus y cal./kg. plus z atmospheres; or v gm. X cal./kg. X atm. (c) Tables or alignment charts may represent an unlimited number of compo- nents even though each of them be visibly a separable variable. In particular, I believe all the changes measured at any moment during one bout of exercise may be counted up so as to represent the total combined state of the organism. Instead of the two com- ponents represented in fig. 185, some 20 or 30 separable components might enter into one quantitative statement of a physiological state at a chosen moment. Combined components actually occur in every science. For cer- tain of them names exist. Any solution is such; cardiac output (heart frequency X stroke output) is another. Water plus certain solutes is plasma; plus heat is warm plasma; plus blood cells is whole blood. So ''biological" terms as well as non-biological ones are often known whereby to designate combinations of properties. For numerous other combinations terms will be forthcoming when 416 PHYSIOLOGICAL REGULATIONS the study of relations is undertaken. Indeed, it is not necessary to recognize whether or not a component is compounded of other ele- ments. A quotient or a subtrahend may be regarded as a combina- tion in this general sense, equally with a multiplicand or any other f(Ji, Jo). A correlative of two or more recognized components sometimes answers the definition of a resultant, and usually of an * ' emergent. ' ' Not only may several loads in an organism be considered as a combination, but their simultaneous evaluation constitutes a state- ment of net physiological state. The various modifications thus assess the organism as a whole, as well as they assess individual components. The extent to which this estimate can now be carried is not great; more needs to be known concerning physiological ratios of each component. For certain components, physical and physiological equivalents are established: 1 gm. of water vapor- ized = 0.58 Calories absorbed ; 1 gm. of fat burned = 9 Calories of heat produced ; 1 calorie transformed = < 0.25 calorie of external work. Many other equivalents are more limited in their applica- bility, as their physiological ratios vary of tener with circumstances. The assets of the organism in any two net states have been in fact but very partially compared. I believe physiology cannot avoid regarding many combined components as one net component. The emergent of one combi- nation becomes in turn one of several elements in the next combina- tion, and this process may go on indefinitely. Permeability, exci- tation, and respiration are names of such multiple combinations. Or, in a plane or solid figure the manner in which the state of the organism moves hither and yon might be represented. I could use more than three dimensions if I could grasp the simultaneous meaning of more coordinates. Familiarity and practice with such numbers of variables might conceivably cultivate such a grasp. In many correlations it is customary and convenient to distin- guish independent variables from dependent ones. The distinction is created by the experimenter or recorder of relations, and, so far as I see, has no counterpart in the organism. Arbitrarily the in- vestigator selects certain conditions for the organism, and for the time being these are described as independent variables. But in another kind of test the organism may select its own quantities of each same variable, in accord with its physiological states. Arbi- trary definitions that are required for the representation of results INTERRELATIONS AMONG COMPONENTS 417 are not to be attributed to the components represented. He who creates variables in a separate order of independence is one who also terms those variables ''causes." Only by virtue of having tested some variables as if independent, does it appear justified to imagine they are causes. Methods of representing many physiological variables are now very primitive, and I hope that more promising ones will be found than the crude diagrams herewith suggested. The organism is apparently not limited in performance by human inadequacy in recording the multitude of its relations. (4) Another approach to interactions of physiological proper- ties is as follows. A great variety of loads and conditions might be studied in their interrelations if a measure of "value'' is se- lected. Familiar examples are the commercial yields of agricul- tural crops (Fisher, '37). The monetary value of yields equal in weight varies from time to time with economic situations ; similarly any scale of physiological value varies with the appraiser. Diverse measures of value emphasize man-power, longevity, enjoyment of leisure, meat or milk production, piece work, mental stability, physical fitness, or any other better or worse defined criterion. Medicine and surgery have an almost singular measure of value in the survival and comfort of the patient. In Arbeitsphysiologie, to operate machines of certain specifica- tions, measurable muscular forces and energies are expended in particular manners. Physiologists choose criteria as to what are the immediate and ultimate effects of the work upon the human organism. Not content with rate of oxygen consumption, it is usual to measure heart frequency, cardiac output, arterial pres- sures, coordination and speed of movement, reaction times, effects of giving sugar, effects of rest periods, effects of room tempera- tures, and many other ''factors." Multiple indices are devised whereby each measurement and test takes its share in rating the individual. Some scientists have a tradition of "varying one factor at a time"; others vary many simultaneously. Those who have con- sidered the matter quantitatively point out that the latter "ar- rangement possesses two advantages over experiments involving only single factors: (i) Greater efficiency . . . and, (ii) Greater comprehensiveness" (Fisher, '37, p. 110). "If the investigator, in these circumstances, confines his attention to any single factor, we 418 PHYSIOLOGICAL REGULATIONS may infer either that he is the unfortunate victim of a doctrinaire theory as to how experimentation should proceed, or that the time, material or equipment at his disposal is too limited to allow him to give attention to more than one narrow aspect of his problem" (Fisher, '37, p. 101). Adopting some components earlier discussed, an investigator of athletic accomplishment might choose ten loads for simultaneous study, as follows : {!) Water load, 0. (2) Water load, +0.5% of Bo. (3) Heat load, + 0.2 Cal./kg. (4) Heat load, + 0.5 Cal./kg. (5) Glucose load, 0. (6) Glucose load, -0.1 gm./kg. (7) Glucose load, + 0.2 gm./kg. (8) Heart frequency, + 20 per cent of resting, induced by epinephrine. (9) Heart frequency, +40 per cent of resting. (10) Heart frequency, + 60 per cent of resting. Of the different combinations of the n loads taken 2 at a time, in this case 45, 8 are mutually exclusive by definition, so that 37 ex- periments at least are to be performed. In each state the human subject runs 1 kilometer in the shortest time he can; these times are correlated with each of the ten loads. For a more complete answer, enough repetitions will be made to establish the statistical significance of each combined physiological state in the perform- ance of running. This is the explicit form, in the language of in- crement, of myriads of investigations. In general, to evaluate n states characterized by diverse simulta- neous loads, a plan is set up in which n{n-l){n-2) . . . /I. 2. 3. . . . total combinations are studied. States that represent two or more loads belonging to one component are precluded; leaving in the above example 133 experiments. In each the investigator ascer- tains the time required to run, and thereafter applies multiple correlation, or some other technique of measuring association, to the results, finally deriving a value for each of the combinations of components. The usual scheme of combining loads is the Latin square, which arranges all the possible combinations in a certain sequence that INTERKELATIONS AMONG COMPONENTS 419 secures maximal randomness. Many physiologists consider any such formal plan a discouragement to their sentiments and pleasure in work. At the same time it may be realized that ' ' as the art of experimentation advances the principles should become clear by virtue of which this planning and designing achieve their purpose" (Fisher, '37, p. 9). The design of experiments now looms large in physiology as in other sciences. It may be said that the analysis of simultaneous interrelations among components exposes quantitative competitions during re- coveries. The load of each component modifies exchanges of itself more than exchanges of other components. Components are found combined in multiple ways, each combination constituting an entity that may still be represented as a single component. All components and combinations may be compared in terms of their effect upon some common index of "value." All enter into an account of the actual physiological state of the living unit. § 150. Meanings of interrelations So far it might be supposed that the study of related components is simply a matter of trying to represent a number of events that "by chance" are simultaneous. But it was noted early (<§ 80) that any change in one component inevitably leads to changes in other components. It may now be inferred that this extensive associa- tion is a basis of recovery and of physiological organization. Numerous are the instances in which one component does not change without modification of others. If water is removed from the organism, either solutes are removed too or osmotic pressure is increased. If the heart beats more frequently, either the cardiac output increases or the stroke output decreases. It might be pos- sible to dissociate any two given components in the organism, but I infer that during life one given component is rarely if ever dis- sociated from all others. I draw the induction that many com- ponents are interrelated or tied to one another. This mutiplicity of connections has the effect of anchoring each component ; when- ever the component is disturbed (loaded) the resistance to change of the other components is met. A conception reached is that of a large stress-strain system. Displacement of any content puts under strain not only the com- ponent itself but all the components that are associated with it. In addition, other components serve to locate the zero load of the 420 PHYSIOLOGICAL. EEGULATIONS first one, for, the fact is that the original zero load is again pre- cisely located after recovery. I infer it is only the quantities (contents) that are not changed that are able to 'inform" the processes of recovery how much to return. If any component were independent of all others, probably no means would be left the living unit of ''knowing" where balance lies. Undulations of diverse components similar in time are also evi- dence of interdependences. Such rhythms may be "spontaneous" or induced. The interrelations are made known by the demonstra- tion of what components take part in the one rhythm. When first observed the undulations are usually regarded as random, later they are said to be physiological. Often, I believe, rhythmical changes are stabilizing something; perhaps even concerned in maintaining constant a function that itself shows no periodicity, as in the uniform flow of blood with intermittent heart beats or uniform flow of heat with fluctuations in peripheral blood flow of man in every minute (Burton and Taylor, '40). The facts lead to the general proposition that probably no one physiological function can be different from what it is without many others being different. Each change involves a ''network" of many unforeseen changes. To find a new equilibration or bal- ance for one component means securing some arrangement for bringing all the others to new balances too. The shift of Co might be sudden and complete, as in some components at birth; or pro- gressive, as during ontogeny or acclimatization. A living unit may be viewed as such a system of unexceptionally related quantities. The components are each subject to certain ties and relationships in space and time. The existence of inter- dependencies among them is a matter of fact ; that all components have interrelations can never be demonstrated except with ex- trapolation; that, further, these interrelations are the visible ele- ments of recovery and of stability probably remains a hypothesis. Such a view may be compared with the implicit supposition of many physiologists that an organism is a conglomerate of chiefly uncorrelated happenings. For each function it is assumed there is a regulator or key or master lever. All or any components might then be at designated loads, at one time or separately or in any combination. This supposition could be supported by finding some load of a component that shows high constancy unaccom- panied by changes of other compounds (after exhaustive search). INTERRELATIONS AMONG COMPONENTS 421 Observation of uniformities indicates, on the contrary, that corre- lated happenings are found in most instances where they have been looked for. That realization is altering the course of physio- logical science. I conclude that where recovery and constancy of a component occur, there are interrelations with other components. The equili- brated organism is probably one whose n components are all asso- ciated in such a way that loadings of even a considerable number mil not destroy the ''memory" of that net state of the organism, to which all properties trend when recoveries intervene. I conclude that any two or more simultaneous changes in com- ponents, when correlated, show evidence of inherent interrelations. Examples such as water content and heat content, or various properties in physical exercise, furnish coefficients and ratios of preferential change. Diagrams that correlate several components, and resultants that combine components, are of practical use in dealing with numbers of physiological variables. All those vari- ables enter into one general account of the pattern of the organism and the means by which its properties are fixed. Chapter XVIII CHOOSING PHYSIOLOGICAL VARIABLES § 151. As the facts and relations have accumulated in this in- vestigation, certain rules for their selection gradually appeared. I was not consciously formulating such rules; rather the materials themselves seemed to move into place, and I then began to wonder by what tokens they gravitated. Given the objective of finding regulatory processes of kinds that could prevail in all animals, the steps by which that search was accomplished became partly auto- matic. Now that the objective is achieved, I believe it is worth while to turn aside long enough to examine the general import of these rules of procedure. This is not a treatise on methods. Yet, every scientist has methods. They are equally effective and decisive in the substance of his results, whether or not he recognizes and classifies them. Methods are usually gained by unconscious experience and by the examples of others, almost never by precept. Scientists find strange those procedures they do not habitually use. Sometimes suspicions are aroused by them. To understand all investigations, it is necessary to admit that ways of doing things other than those sanctioned by custom may be justified, even be successful. For, to limit the solution of inquiries to the sort of contribution to science for which tastes are already cultivated would be to exclude all efforts to see beyond the limitations of present outlook. <^ 152. Peoceduees How did this investigation actually proceed? Though most of the steps taken were unforeseen, there was a pattern among them (§104). What ones of the procedures might illustrate general usages in scientific work? Would they be useful in planning investigations, in so far as investigations are explicitly planned? A basis of this inquiry consisted in distinguishing among physi- ological variables. Each variable was some quantity that could be identified in a defined and reproducible manner. Since the identity of the quantity depended upon the operations by which it was mea- sured, each variable had dimensions that in part at once offered a 422 CHOOSING PHYSIOLOGICAL VARIABLES 423 practical classification of it among its fellows. Thus, there were loads (AW, AC), each distinct with respect to component, species, living unit, and other particulars. There were rates of exchange (SW/At, SC/At), each distinct (in addition) with respect to paths of exchange and unit of time involved. There were components, liv- ing units, analyzable tissues, modes of chemical analysis, rates of physiological activity, volumes, concentrations, various ratios. Any quantity that could be adequately defined and classified might appear on the docket of data available. Among these data, further choices were made, taking two vari- ables at a time from among those quantities that had been mea- sured simultaneously. The choices were guided by the desire to learn how exchanges are concerned in maintenances of content. At first, loads and rates were correlated; AW was related to SW/At, AC to SC/At. Later, AW was successively related to each of many varieties of AE, AG, AH, etc.; to SC/At, SG/At, SH/At, etc. The variables and the relations among them having been classified, living units were compared within one class of relations after another. Without indicating many of its subdivisions, the whole course of the investigation is represented by the following outline : Ji. Component Water (and Volume). Chapters II to XIII. Ml. The four variables. Chapters II to IX. Gi. Species Dog. Chapters II to IV.