LIBRARY 
 
 UNiVtf.'S!TY OF 
 CALIFORNIA 
 SANTA CRUZ 
 
YALE UNIVERSITY 
 
 MRS. HEPSA ELY SILLIMAN MEMORIAL 
 LECTURES 
 
 RESPIRATION 
 
SILLIMAN MEMORIAL LECTURES 
 
 PUBLISHED BY YALE UNIVERSITY PRESS 
 
 ELECTRICITY AND MATTER. By JOSEPH JOHN THOMSON, D.SC., LL.D., PH.D., 
 F.R.S. Fellow of Trinity College and Cavendish Professor of Experimental 
 Physics, Cambridge University. {Fourth printing.} 
 
 THE INTEGRATIVE ACTION OF THE NERVOUS SYSTEM. By CHARLES 
 S. SHERRINGTON, D.SC., M.D., HON. LL.D. TOR., F.R.S., Holt Professor of Physi- 
 ology, University of Liverpool. {Sixth printing.} 
 
 RADIOACTIVE TRANSFORMATIONS. By ERNEST RUTHERFORD, D.SC., LL.D., 
 F.R.S., Macdonald Professor of Physics, McGill University. {Second printing.} 
 
 EXPERIMENTAL AND THEORETICAL APPLICATIONS OF THERMO- 
 DYNAMICS TO CHEMISTRY. By DR. WALTER NERNST, Professor and 
 Director of the Institute of Physical Chemistry in the University of Berlin. 
 
 PROBLEMS OF GENETICS. By WILLIAM BATESON, M.A., F.R.S., Director of 
 the John Innes Horticultural Institution, Merton Park, Surrey, England. {Sec- 
 ond printing.} 
 
 STELLAR MOTIONS. With Special Reference to Motions Determined by Means 
 of the Spectrograph. By WILLIAM WALLACE CAMPBELL, SC.D., LL.D., Director of 
 the Lick Observatory, University of California. {Second printing.} 
 
 THEORIES OF SOLUTIONS. By SVANTE ARRHENIUS, PH.D., SC.D., M.D., Di- 
 rector of the Physico-Chemical Department of the Nobel Institute, Stockholm, 
 Sweden. {Third printing.} 
 
 IRRITABILITY. A Physiological Analysis of the General Effect of Stimuli in 
 Living Substances. By MAX VERWORN, M.D., PH.D., Professor at Bonn Physio- 
 logical Institute. {Second printing.} 
 
 PROBLEMS OF AMERICAN GEOLOGY. By WILLIAM NORTH RICE, FRANK 
 D. ADAMS, ARTHUR P. COLEMAN, CHARLES D. WALCOTT, WALDEMAR LIND- 
 GREN, FREDERICK LESLIE RANSOME, AND WILLIAM D. MATTHEW. {Second 
 printing. } 
 
 THE PROBLEM OF VOLCANISM. By JOSEPH PAXSON IDDINGS, PH.B., SC.D. 
 {Second printing.} 
 
 ORGANISM AND ENVIRONMENT AS ILLUSTRATED BY THE PHYSI- 
 OLOGY OF BREATHING. By JOHN SCOTT HALDANE, M.D., LL.D., F.R.S., 
 
 Fellow of New College, Oxford University. {Second printing.} 
 
 A CENTURY OF SCIENCE IN AMERICA. With Special Reference to the 
 American Journal of Science 1818-1918. By EDWARD SALISBURY DANA, 
 CHARLES SCHUCHERT, HERBERT E. GREGORY, JOSEPH BARRELL, GEORGE OTIS 
 SMITH, RICHARD SWANN LULL, Louis V. PIRSSON, WILLIAM E. FORD, R. B. 
 SOSMAN, HORACE L. WELLS, HARRY W. FOOTE, LEIGH PAGE, WESLEY R. COE, 
 AND GEORGE L. GOODALE. 
 
 THE EVOLUTION OF MODERN MEDICINE. By the late SIR WILLIAM 
 
 OSLER, BART., M.D., F.R.S. 
 
 RESPIRATION. By J. S. HALDANE, M.D., LL.D., F.R.S., Fellow of New College, 
 Oxford, H on. Professor, Birmingham University. 
 
RESPIRATION 
 
 BY 
 
 J. S. HALDANE 
 
 M.D., LL.D., F.R.S. 
 
 FELLOW OF NEW COLLEGE, OXFORD 
 HON. PROFESSOR, BIRMINGHAM UNIVERSITY 
 
 NEW HAVEN 
 YALE UNIVERSITY PRESS 
 
 LONDON : HUMPHREY MILFORD : OXFORD UNIVERSITY PRESS 
 MDCCCCXXVII 
 
COPYRIGHT 1922 BY YALE UNIVERSITY PRESS 
 PRINTED IN THE UNITED STATES OF AMERICA 
 
 First published, May, 1922. 
 Second printing, January, 1927. 
 
THE SILLIMAN FOUNDATION 
 
 IN the year 1883 a legacy of eighty thousand dollars was left to the 
 President and Fellows of Yale College in the city of New Haven, to be 
 held in trust, as a gift from her children, in memory of their beloved and 
 honored mother, Mrs. Hepsa Ely Silliman. 
 
 On this foundation Yale College was requested and directed to estab- 
 lish an annual course of lectures designed to illustrate the presence and 
 providence, the wisdom and goodness of God, as manifested in the 
 natural and moral world. These were to be designated as the Mrs. Hepsa 
 Ely Silliman Memorial Lectures. Tt was the belief of the testator that 
 any orderly presentation of the facts of nature or history contributed to 
 the end of this foundation more effectively than any attempt to empha- 
 size the elements of doctrine or of creed ; and he therefore provided that 
 lectures on dogmatic or polemical theology should be excluded from the 
 scope of this foundation, and that the subjects should be selected rather 
 from the domains of natural science and history, giving special promi- 
 nence to astronomy, chemistry, geology, and anatomy. 
 
 It was further directed that each annual course should be made the 
 basis of a volume to form part of a series constituting a memorial to 
 Mrs. Silliman. The memorial fund came into the possession of the Cor- 
 poration of Yale University in the year 1901 ; and the present volume 
 constitutes the fourteenth of the series of memorial lectures. 
 
PREFACE 
 
 WHEN Yale University invited me to deliver the Silliman Lectures 
 for 1915 I was asked to deal with the physiology of breathing 
 and include a general account of the long series of investigations 
 with which I had been associated on this subject and its practical 
 applications in clinical medicine and hygiene. Owing to the war I 
 was unable to give the lectures in 1915, but in 1916 delivered four 
 lectures which dealt only with some of the more general conclu- 
 sions to which I had been led, and were published early in 1917 
 by the Yale University Press under the title ''Organism and 
 Environment as Illustrated by the Physiology of Breathing." 
 
 The war has greatly delayed the appearance of the present 
 book, which treats the physiology of breathing fully in accordance 
 with the original plan. I have, however, abandoned the lecture 
 form, and what I had written four years ago has had to be largely 
 recast owing to the rapid advance of knowledge. The book is not 
 a mere compilation, but contains much that has never previously 
 been published, and is an attempt to give a coherent statement 
 and interpretation of what is known of the subject at present. 1 
 fear that I may sometimes have unwittingly overlooked observa- 
 tions by others which would have added completeness to my 
 account. Yet I hope that what may have been lost in this way 
 will be made up for by the fact that the book embodies the results 
 of a continuous series of investigations leading to very definite 
 and consistent conclusions. 
 
 About the middle of last century the younger physiologists 
 broke away from the vitalistic traditions which had been handed 
 down to them, and set about to investigate living organisms 
 piece by piece, precisely as they would investigate the working 
 of a complex mechanism. This method seemed to them to promise 
 success, and was popularized by such masters of clear and force- 
 ful expression as Huxley. It is still the orthodox method of physi- 
 ology, but the old confidence in it has steadily diminished in 
 proportion as exact experimental investigation has shown that the 
 various activities of a living organism cannot be interpreted in 
 isolation from one another, since organic regulation dominates 
 them. The keynote of this book is the organic regulation of 
 breathing and its associated phenomena. 
 
viii RESPIRATION 
 
 The mechanistic theory of life is now outworn and must soon 
 take its place in history as a passing phase in the development^ of 
 biology. But physiology will not go back to the vitalism which 
 was threatening to strangle it, and from which it escaped last 
 century. The real lesson of the movement of that time will never 
 be lost. 
 
 The book belongs to a transition period, but the transition is 
 forward and not backward. My treatment of the subject may 
 possibly be looked on askance in some quarters as reactionary: 
 for I have been largely influenced by the ideas and work of older 
 physiologists. If, however, I have gone backward, it is only to 
 pick up clues which had been temporarily lost; and all of these 
 clues lead forward forward to a new physiology which embodies 
 what was really implicit in the old. 
 
 The leaders of the mechanistic movement of last century got 
 rid of vitalism, but in doing so got rid of life itself. I have tried 
 to paint a picture of the body as alive. Though the picture is 
 imperfect, others will soon paint it more completely. The time 
 has come for a far more clear realization of what life implies. 
 The bondage of biology to the physical sciences has lasted more 
 than half a century. It is now time for biology to take her rightful 
 place as an exact independent science : to speak her own language, 
 and not that of other sciences. 
 
 The endeavor to represent the facts of physiology as if they 
 would fit into the general scheme of a mechanistic biology has 
 led, it seems to me, to the present estrangement between physiology 
 and medicine. Since the time of Hippocrates the growth of 
 scientific medicine has in reality been based on the study of the 
 manner in which what he called the "nature" ( <iW ) of the 
 living body expresses itself in response to changes in environ- 
 ment, and reasserts itself in face of disturbance and injury. The 
 underlying assumption is that organic regulation and maintenance 
 represent something very real, and that only through the study 
 of it can we recognize and interpret disturbance of health, and 
 effectively aid maintenance or restoration of health. I have en- 
 deavored to return to what seems to me the truly scientific Greek 
 tradition, and to give it a form which is not only consistent with 
 modern science and philosophy, but brings physiology and medi- 
 cine into that close and special relation indicated by the common 
 etymology of the words "physician" and "physiology." 
 
 Most of the investigations specially referred to in the book 
 have been carried out on man. It was only by human experiments 
 
RESPIRATION ix 
 
 that the almost incredible delicacy of the regulation of breathing 
 was discovered; and human experiments have revealed to us in 
 other ways how rough many of the experiments on animals, or 
 on "preparations" from the bodies of animals, have been. Organic 
 regulation, with its all-important relations to practical medicine 
 and surgery, was often entirely overlooked. I hope that the book 
 may contribute towards establishing human physiology in its 
 rightful place, which has been usurped too long by experiments 
 on fragments of frogs and other animals, or on the mere super- 
 ficial physical and chemical aspects of bodily activity. 
 
 I wish to offer my sincere thanks to Yale University for the 
 honor it has done me in inviting me to give the Silliman Lectures. 
 Between Oxford and Yale Universities there is a traditional 
 association, and to me in particular the association stands for 
 friendship, hospitality, and community of ideas. My only regret 
 is that in coming to Yale to lecture on the physiology of breathing 
 I seemed to be doing what an Englishman calls bringing coals 
 to Newcastle, since I had to refer so frequently to the results 
 reached at Yale by Professor Yandell Henderson and his pupils. 
 
 The book sums up the results of more than twenty years of my 
 own experimental work, thought, reading, and discussion. To the 
 old pupils and other friends who have worked and thought with 
 me, including friends in the mining and engineering professions 
 and in the Navy and Army, I wish to express my debt. Their 
 names are often quoted in the text, but I should like to say how 
 much I have been aided more particularly by Professor Lorrain 
 Smith, Professor Pembrey, Professor Boycott, Commander Da- 
 mant, Mr. Mavrogardato, Dr. Priestley, Dr. Douglas, Professor 
 Meakins, and my son. In connection with the Pike's Peak Scien- 
 tific Expedition, the results of which occupy such a prominent 
 place in the book, Dr. Douglas and I had the great advantage 
 of being associated with Professor Yandell Henderson and another 
 Yale graduate, Professor Schneider of Colorado Springs. The 
 book owes much to the talks we had on the Peak in the summer 
 evenings when our work was over and the lights were twinkling 
 over the prairie far below from Denver to Pueblo. 
 
 Readers will easily see how many gaps remain to be filled up. 
 To fill these gaps the observations and experiments required are 
 not yet available. The words of Hippocrates are as true now as 
 when he wrote them more than two thousand years ago :6/3ibs flpa- 
 
 Xvs, rj & Ttyy-n ^ K P^ 
 OXFORD, MAY 1920. 
 
x RESPIRATION 
 
 Owing to the aftermath of the war there has been considerable 
 delay in printing the book, and meanwhile a good deal of new 
 work has appeared on the subjects of certain chapters. Where 
 this could not be incorporated without serious recasting in the 
 proofs it is referred to in addenda to the chapters in question. 
 
 MAY 1921. 
 
CONTENTS 
 PREFACE vii 
 
 CHAPTER I. HISTORICAL INTRODUCTION i 
 
 Early theories of respiration, i. Boyle and Mayow, i. Black and Priestley, 
 2. Lavoisier's interpretation of respiration and the source of animal heat, 2. 
 Mayer and the source of vital energy, 3. Discoveries as to the composition of 
 animal food and excreta, 3. Discoveries as to the blood gases and the part they 
 play, 4. Theories and discoveries as to physiological regulation of vital oxida- 
 tion, 4. Work of Liebig, Voit, Rubner, Pfliiger, and others, 5. Discoveries 
 as to physiological regulation of body temperature, 6.-^-The "energy require- 
 ments" of the living body, 6. The problem of regulation of breathing, 8. 
 The respiratory center. Work of Legallois and others, 8. The vagus nerves and 
 breathing. Work of Hering and others, 9. Chemical excitation of breathing. 
 Work of Rosenthal and others, 10. Theories of "vagus apnoea" and "chemical 
 apnoea," n. Conclusions as to various chemical and other excitants of breath- 
 ing, 13. Criticism of these conclusions and starting point of the investigations 
 described in succeeding chapters, 14. 
 
 CHAPTER II. CARBON DIOXIDE AND REGULATION 
 OF BREATHING . . . . . . .15 
 
 Effects of varying proportions of CO 2 and oxygen on breathing in man, 15. 
 Importance of the alveolar air, 16. Method of sampling the alveolar air, 17. 
 relative constancy of the alveolar CO 2 percentage, 19. Effects of varying 
 oxygen percentage of the alveolar CO 2 percentage, 20. Effects of varying COa 
 percentage in the inspired air on the alveolar CO 2 percentage, 21. Extreme 
 sensitiveness of the respiratory center to variation in alveolar CO 2 percentage, 
 22. Apnoea after forced breathing is due simply to lowering of alveolar CO 2 
 percentage, 24. Constancy of mean alveolar CO 2 pressure in spite of great 
 variations in rate and depth of breathing, 27. Rise of alveolar CO 2 percentage 
 during muscular exertion, 2g. Effects of varying barometric pressure on alve- 
 olar CO 2 percentage, 30. Constancy of alveolar CO 2 pressure with varying 
 barometric pressure, 31. Individual differences in alveolar CO 2 pressure, 32. 
 The anatomy of bronchioles and alveoli, 33. "Alveolar air" is air of 
 Miller's "air-sac" system, 35. The "effective" or "virtual" dead space in 
 breathing, 35. Great variations in effective dead space with varying depth of 
 breathing, 37. "Alveolar" and true respiratory quotients, 38. Errors due to 
 ignorance of the variations in the effective dead space, 39. Gas pressures of 
 alveolar air and arterial blood, 41. Question as to varying composition of 
 air in different alveoli, 42. General conclusion from Chapter I, 42. 
 
 CHAPTER III. THE NERVOUS CONTROL OF BREATH- 
 ING . 43 
 
 Voluntary and reflex disturbances of breathing, 43. Experiments on man 
 
xii RESPIRATION 
 
 showing the non-existence of "vagus apnoea," 47. Afferent vagus excitations 
 coordinate the phases of breathing, 48. The depth and vigor of breathing de- 
 pend on the chemical stimuli to the respiratory center, 49. Effects of resistance 
 on the rhythm of breathing, 50. Artificial respiration and the vagus coordina- 
 tion of breathing, 50. Normal breathing and afferent nervous control, 53. 
 Evidence that the activity of the respiratory center depends on locally acting 
 chemical stimuli in the medulla oblongata, 53. Physiological significance of 
 this fact, 54. Fatigue of the respiratory center, 56. The breathing in "sol- 
 dier's heart" and allied conditions, 57. "Neurasthenia" and fatigue, 56. 
 Variations in individual susceptibility to fatigue of breathing, 57. 
 
 CHAPTER IV. THE BLOOD AS A CARRIER OF OXYGEN 59 
 
 General chemical properties of haemoglobin and oxyhaemoglobin, 59. Meth- 
 aemoglobin and its properties, 59. Action of ferricyanide in liberating oxygen 
 or CO from combination with haemoglobin, 60. Oxyhaemoglobin and CO 
 haemoglobin are molecular compounds, 61. The ferricyanide method of de- 
 termining the oxygen capacity of haemoglobin, 61. The oxygen capacity of 
 haemoglobin is exactly proportional to its coloring power, 61. The Gowers- 
 Haldane haemoglobinometer, 62. Normal variations in haemoglobin percentage 
 of blood, 63. Haemochromogen and its modifications, 64. Relation between 
 oxygen capacity and iron of haemoglobin, 64. Relation of haemochromogen 
 to haemoglobin, 66. Ferricyanide method for ordinary blood-gas determina- 
 tions, 66. Amount of available oxygen in human arterial blood, 67. Funda- 
 mental importance of the partial pressure of oxygen in the blood, 67. "Partial 
 pressures," "vapor pressures," "diffusion pressures," and "concentrations" of 
 substances in the living body, 67. Investigations of the laws of dissociation 
 of oxyhaemoglobin in blood, 68. Work of Paul Bert, Hiifner, Loewy and 
 Zuntz, Bohr, Barcroft, 70. Effects of salts, COz, and acids or alkalies, 72. 
 Physiological importance of the shape of the oxyhaemoglobin dissociation curve, 
 7 2.-^ Properties and dissociation curves of CO haemoglobin, 72. Nature of 
 alterations produced by CO2 on the dissociation curves of CO haemoglobin and 
 oxyhaemoglobin, 76. Relative affinities of oxygen and CO for haemoglobin, 
 74- Evidence of differences in the chemical structure of haemoglobin in differ- 
 ent individuals and species, 77. Use of haemoglobin for estimating partial 
 pressures of CO or oxygen, 79. Explanation of the peculiarities of the dissocia- 
 tion curve of oxyhaemoglobin in blood, 80. Equations for the dissociation 
 curves, 82. 
 
 CHAPTER V. THE BLOOD AS A CARRIER OF CARBON 
 
 DIOXIDE ........ 84 
 
 Amount of CO 2 in normal human and dog's arterial blood, 84. Amounts in 
 simple solution and chemical combination, 85. The CO 2 is combined with 
 alkali as bicarbonate, 84. Why the bicarbonate dissociates appreciably with a 
 small fall in the partial pressure of CO 2 in the blood, 85. Haemoglobin and 
 other proteins act as acids in the living body, and do not combine with CO 2 , 
 though they play a most essential part in its carriage, 88. The dissociation 
 curve for CO 2 of human blood, 89. Constancy of this curve for the same indi- 
 vidual, and relative constancy in different normal individuals, 89. Evidence 
 that oxygen has a chemical action in liberating CO 2 in the lungs, 92. The de- 
 
RESPIRATION xiii 
 
 oxygenation of the blood in the tissues helps the blood to combine with CO 2 and 
 thus partly prevents the pressure of COz from rising, 90. CC>2 may be given off 
 in the lungs although the CO2 pressure is lower in the venous blood than in the 
 alveolar air, 91. Approximate mathematical treatment of the dissociation curve 
 for CO2, 92. Effect of the CO2 in blood on the dissociation of oxyhaemoglobin 
 in the systemic blood, 94. The physiological buffers which prevent abrupt rise 
 or fall of CO2 pressure in the respiratory center, 96. Effects on the alveolar 
 CO2 pressure of holding the breath or forced breathing, 96. Abruptness of 
 rises or falls of oxygen pressure in the respiratory center, 100. This abrupt- 
 ness is the cause of periodic breathing when the alveolar oxygen pressure is low, 
 103. Artificial production of periodic breathing in healthy persons, 103. 
 Why shortage of oxygen and consequent periodic breathing do not occur nor- 
 mally, 104. Addendum, Discussion of some recent theories of the carriage of 
 CO2 by blood, 105. Interchange of acid between plasma and corpuscles, 106. 
 
 CHAPTER VI. THE EFFECTS OF WANT OF OXYGEN 108 
 
 Immediate dependence of the body for its oxygen supply on air, 108. Anox- 
 aemia produced by lowered pressure of oxygen in the air, 109. Effects on the 
 breathing, 109. These effects largely transitory, 109. Lowering of the thresh- 
 old of alveolar CO2 pressure, but alveolar CO2 pressure still regulates the 
 breathing, no. Variability of the effects in different individuals, in. 
 Death from anoxaemia caused by excessive removal of CO2 from the blood, 112. 
 Excess of COa in the air counteracts the effects of deficiency of oxygen, 112. 
 Mere increase of breathing does not diminish the anoxaemia, though it dimin- 
 ishes the cyanosis, 114. The peculiar symptoms produced by forced breathing 
 are apparently due mainly to anoxaemia, 115. Subsidiary effects of CO2 in re- 
 lieving anoxaemia, 117. Periodic breathing at high altitudes is caused by 
 anoxaemia, 117. Effects of anoxaemia on the frequency of breathing, 118. 
 Effects in causing fatigue of the breathing, 121. Effects of anoxaemia on the 
 circulation, 121. Increase in pulse rate is largely transitory, 123. Cyanosis 
 and anoxaemia not the same thing, 125. Effects on the nervous system, 125. 
 Insidious character of these effects, 125. Effects on muscular power, senses, 
 memory, and powers of judgment, 126. Personal experiences, 128. Moun- 
 tain sickness and conditions of its production, 128. Nervous after symptoms 
 following severe anoxaemia, 129. After effects on heart, 129. After effects 
 on respiratory center, 130. Adaptation to want of oxygen, 130. 
 
 CHAPTER VII. THE CAUSES OF ANOXAEMIA . .132 
 
 Defective saturation of arterial haemoglobin, 132. One cause of this is 
 defective distribution of air in the lungs, 133. Experimental proof and ex- 
 planation of this, 133. Effects of holding the breath, and explanation of the 
 anoxaemia produced, 141. Cause of difference between clinical Cheyne-Stokes 
 breathing and periodic breathing produced artificially in healthy persons, 141. 
 Significance of rapid breathing in cases of illness, 142. Danger of sudden 
 attacks of restricted and rapid breathing, 143. Causes of anoxaemia in em- 
 physema, bronchitis, and asthma, 145. Orthopnoea and its causes, 146. A 
 second cause of arterial anoxaemia is defective pressure of oxygen in the in- 
 spired air, 146. Immediate effects and after effects, 147. The percentage 
 saturation of the arterial haemoglobin is lower than corresponds to the oxygen 
 
xiv RESPIRATION 
 
 pressure of the mixed alveolar air, 148. With the same alveolar oxygen pres- 
 sure there is less anoxaemia at low atmospheric pressures than at normal atmos- 
 pheric pressure, 148. Anoxaemia due to hindered diffusion of oxygen into the 
 blood, 149. Poisoning by lung-irritant gases, 150. Arterial anoxaemia in 
 pneumonia, 150. Observations of Stadie, Harrop, and Meakins, 151. The 
 clinical administration of oxygen, 152. Description of apparatus for the pur- 
 pose, 154. Anoxaemia during muscular exertion, 156. Experiments of Briggs 
 on oxygen inhalation during muscular exertion, 157. Anoxaemia and 
 velocity of chemical reaction in the formation of oxyhaemoglobin, 158. 
 Anoxaemia due to defective oxygen-carrying power of the blood, 158. Evidence 
 that the symptoms of CO poisoning are due to anoxaemia, 160. CO is not 
 oxidized in the body, but passes in and out by the lungs, 160. Popular errors 
 as to the effects of CO poisoning and anoxaemia generally, 160. Relation be- 
 tween percentage of CO in air and percentage saturation of the blood with CO. 
 1 60. Relation between percentage saturation of the blood and symptoms, 161. 
 Causes of certain differences between the symptoms of CO poisoning and 
 those of anoxaemia produced in other ways, 162. Alteration of the dissociation 
 curve of oxyhaemoglobin in CO poisoning, 165. Acclimatization to CO poison- 
 ing, 1 66. Occurrence of NO haemoglobin in the body, 166. Methaemoglobin- 
 forming poisons, 166. Evidence that with these poisons death is due to anox- 
 aemia, 167. Recovery from methaemoglobin-forming poisons, 167. Hae- 
 molytic poisons, 168. Anaemia and anoxaemia, 168. Reasons why no anox- 
 aemia is present during rest in ordinary anaemia, 169. Anoxaemia due to de- 
 fective circulation, 169. Gum-saline injections in defective filling of the vessels 
 with blood, 170. 
 
 CHAPTER VIII. BLOOD REACTION AND BREATHING 171 
 
 Ordinary physiological indications of maintenance of a normal blood reaction, 
 *7i' Walter's experiments on acid poisoning and the defenses against it, 171. 
 Diabetic coma and acid poisoning, 173. "Titration alkalinity" and alkalinity 
 of the blood, 173. The "buffer substances" in the living body, 174. Modern 
 conceptions of alkalinity and acidity, 175. Osmotic pressure, molecular con- 
 centration, and molecular diffusion pressure, 176. lonization of molecules, 177. 
 lonization and reaction, 177. Electrometric measurement of reaction, 179. 
 Theories of acidosis and anoxaemia, 179. Hasselbalch's electrometric de- 
 terminations of relation of CO 2 pressure to reaction in blood, 182. Experiments 
 showing that variation of alveolar CO 2 pressure in the living body compensates 
 for variations in blood reaction which would otherwise occur, 183. Barcroft's 
 experiments on the Peak of Teneriffe, 183. Quantitative relation between varia- 
 tions of breathing and of blood reaction, 184. Extreme delicacy of regulation 
 of blood reaction, 185. Very small difference between the reactions of arterial 
 and venous blood, 185. Difference in reaction between oxygenated and fully 
 reduced normal blood, 186. Error in electrometric method, 188. Summary of 
 evidence as to the means by which blood reaction is regulated, 188. Dis- 
 turbance of blood reaction by anoxaemia, 189. Physiological evidence that the 
 blood becomes more alkaline, 189. Gradual, but incomplete, compensation for 
 this by the kidneys and liver, 192. This compensation mistaken for an 
 "acidosis," 192. Relief of the anoxaemia by the compensation, 193. Com- 
 pensatory blood changes brought about by exposure to excess of CO 2 , or by ex- 
 cessive removal by CO 2 from the body, 193. The amount of "alkaline reserve" 
 in the blood is no certain index of "acidosis" or "alkalosis," 194. Experiments 
 
RESPIRATION xv 
 
 on the urine excreted during forced breathing, 195. True acidosis caused by 
 excessive muscular exertion, 196. Disturbance of blood reaction in nephritis, 
 196. Ammonium chloride acidosis in man, 196. Remarks on indirect methods^ 
 used for measuring changes of reaction in the blood, 199. Method depending 
 on the dissociation curve of oxyhaemoglobin, 199. Method depending on ratio 
 of combined CO 2 to free CO 2 in blood, 200. Need for more delicate methods 
 than we possess at present, 202. Question as to the constancy of blood reaction 
 during normal life, 202. Action of drugs on the regulation of blood reaction, 
 204. Reasons why the alveolar CO2 pressure is not perfectly steady during 
 rest, 204. Effects of meats, 204. Effects of starvation and carbohydrate-free 
 diets, 205. The regulation of breathing in man during rest is practically 
 speaking regulation of blood reaction, 205. Addendum. Recent literature on 
 acidosis and alkalosis, 205. Definition of acidosis and alkalosis, 206. Ex- 
 treme delicacy and physiological importance of regulation of reaction in the 
 tissues, 207. 
 
 CHAPTER IX. GAS SECRETION IN THE LUNGS . . 208 
 
 Question as to active secretion of gas by the lung epithelium, 208. Oxygen 
 secretion by the swim bladder epithelium, 208. Function of the swim bladder, 
 208. Biot's discovery of oxygen secretion, 209. Experiments of Moreau, Bohr, 
 and Dreser, 210. Jager's discovery of the "oval" in the swim bladder, 211. 
 Histology of the swim bladder wall and "red body," 214. Probable function of 
 the "red body," 214. Gas secretion in Arcella. Experiments of Bles, 216. 
 Implications of secretion generally, 217. Ideas of Johannes Miiller on secre- 
 tion, 218. -Apparent gas secretion in Corethra larvae, 220. Ludwig and 
 Pfliiger on gas secretion by the lungs, 220. Experiments of Bohr and Fredericq, 
 221. Method and experiments of Krogh, 222. Carbon monoxide method of 
 measuring arterial oxygen pressure, 224. Fallacies in earlier measurements, 
 225. New experiments on animals. Conclusions, 226. New experiments on 
 men. Method, 229. Result that secretion is completely absent during rest 
 under normal conditions, but present under conditions producing want of oxy- 
 gen in the tissues, 233. Experiments after acclimatization on Pike's Peak, 236. 
 Evidence of constant active secretion, 237. Indirect evidence of oxygen 
 secretion, 238. Experiments of Briggs, 240. Experiments in a respiration 
 chamber at normal atmospheric pressure, 241. Acclimatization experiments in 
 a steel chamber, 242. Cause of difference between results by carbon monoxide 
 and aerotonometer methods, 243. Reason why the percentage of oxygen satura- 
 tion of the arterial blood is considerably less at high altitudes before acclimatiza- 
 tion than corresponds to the oxygen pressure of the alveolar air, 244. Bohr's 
 method of measuring the rate of diffusion of gases from the alveolar air into 
 the blood, 245. Experiments of A. and M. Krogh by this method, 246. 
 Paralysis of oxygen secretion under pathological conditions, 247. Direct evi- 
 dence that during hard muscular work at normal atmospheric pressure diffusion 
 of oxygen is quite insufficient to saturate the arterial blood with oxygen, 247. 
 Question of active excretion of COa by the lungs. Krogh's experiments, 247. 
 Reasons for suspecting that active secretion of COa may occur under certain 
 conditions, 248. Comparison of oxygen secretion by the lungs with glomerular 
 secretion by the kidneys, 250. Reply to some recent criticisms of the evidence 
 for oxygen secretions, 251. Addendum. Recent experiments of Barcroft and his 
 co-workers, 253. 
 
xvi RESPIRATION 
 
 CHAPTER X. BLOOD CIRCULATION AND BREATHING 257 
 
 Intimacy of connection between circulation and breathing, 257. The mosl 
 immediate need for circulation is the need for oxygen and for removal of COz, 
 257. The local circulation rates must be correlated in the main with these 
 needs, 258. Special value of experiments on man, 259. Experiments of Loewy 
 and von Schrotter with lung catheter, 260. Experiments of Krogh and Lind- 
 hard by the nitrous oxide method, 262. Yandell Henderson's experiments on 
 dogs, 263. Experiments on "heart-lung preparations," 264. New method in 
 which the whole of the lungs are used as an aerotonometer, 264. Results in 
 man during rest and work, 265. The circulation rate is rapid during rest, and 
 does not increase in direct proportion to work, 268. The oxygen consumption 
 per heartbeat and its significance, 269. The venous blood from different parts 
 of the body, 270. Significance of this as regards the mixed venous blood under 
 different conditions, 270. General conclusion as regards local regulation of 
 blood flow, 271. Yandell Henderson's experiments on local circulation and 
 COz pressure, 272. Evidence that excessive artificial respiration causes slowing 
 of the circulation and great local anoxaemia, 272. With moderate increase of 
 COz percentage in the inspired air the circulation does not increase with the 
 breathing, 273. But with great increase of COz percentage the circulation in- 
 creases, 274. Increase in oxygen pressure slows the circulation, 274. With 
 great deficiency of oxygen there is increase in the circulation, 275. Effects of 
 forced breathing and muscular exertion on venous blood pressure, 275. Gen- 
 eral conclusion as to regulation of local and general circulation, 276. Com- 
 parison of regulation of circulation with regulation of breathing, 277. Part 
 played by the heart in the circulation, 278. Regulation of heart's action, 278. 
 Coordination of contraction of muscular fibres of auricles and ventricles, 278. 
 Start, spread, and frequency of each contraction, 279. Regulation of filling of 
 ventricles, 279. Nervous regulation of frequency of heartbeat, 279. Regula- 
 tion of blood distribution, 281. Contractility of arteries, veins, and capillaries, 
 281. Vasomotor regulation of arterial and venous blood pressure, 283. 
 Abnormal defects in circulatory regulation, 284. Valvular defects and breath- 
 ing, 286. Nervous defects and breathing, 286. Loss of blood and its treat- 
 ment by gum-saline injections, 287. The condition of "shock," 288. Yandell 
 Henderson's investigations, 288. Shock from absorption of poisonous disin- 
 tegration products, 289. Regulation of blood volume, haemoglobin, and rate 
 of pulse and respiration in animals of different sizes, and after loss of blood or 
 transfusion, 290. Evidence that the haemoglobin percentage of the blood de- 
 pends on the oxygen pressure in tissue capillaries, 293. Chlorotic "anaemia" 
 and breathing, 297. Addendum. Further experiments on the circulation in 
 man, 298. 
 
 CHAPTER XI. AIR OF ABNORMAL COMPOSITION 300 
 
 Outside air of country and towns: effects of impurities, 300. Air of occupied 
 rooms. Common impurities and their effects, 302. Effects of temperature, 
 moisture, and movement of air, 303. General standard of air purity, 305. 
 Critical wet-bulb temperature, 305. The katathermometer, 306. Escape of 
 lighting gas and conditions determining their danger, 306. Importance of pro- 
 portion of CO in lighting gas, 310. Air of mines. Abnormal constituents 
 present, 311. Black damp: composition, sources, and properties, 311. Fire 
 damp: composition, sources, and properties, 313. Afterdamp from explosions 
 
RESPIRATION xvii 
 
 causes death by CO poisoning, 315. Causes and prevention of colliery ex- 
 plosions, 316. Composition of pure afterdamp and practical test for CO, 316. 
 Self-contained breathing apparatus for miners, 318. White damp and spon- 
 taneous heating of coal, 319. Smoke from fires and blasting: nitrous fumes, 
 319. Treatment of CO poisoning, 320. Wet-bulb temperature in mines, 321. 
 Effects of dust inhalation in mines, 322. Varying effects of different kinds of 
 dust. Miner's phthisis, 322. Physiology of dust excretion from the lungs, 323. 
 Air of wells. Barometric pressure and dangers of well sinkers, 325. Oxida- 
 tion processes in underground strata, 326. Air of railway tunnels, 326. Air 
 of sewers. Accidental impurities and their dangers, 327. Air of ships, 329. 
 Lung-irritant gas poisoning in warfare, and treatment, 329. 
 
 CHAPTER XII. EFFECTS OF HIGH ATMOSPHERIC 
 
 PRESSURES 334 
 
 Paul Bert's work on the physiological action of barometric pressure, 334. 
 The diver's equipment and the method of using it, 335. The diving bell and the 
 caisson, 336. Tunneling in compressed air, 337. Effects of air pressure on the 
 ears and voice, 338. Effects due to pressure of COz in diving, and their 
 avoidance, 339. Compressed air illness or "caisson disease," 340. Investiga- 
 tions of Paul Bert and others, 341. Medical recompression chambers, 343. 
 Theory of stage decompression and experiments on the subject, 345. Tables for 
 guidance of divers, 350. Treatment of compressed-air illness, 351. Diving 
 operations at a great depth off Honolulu, 351. Management of air locks in 
 tunnels, 353. Paul Bert's experiments on effects of increased oxygen pres- 
 sure, 355. Effects of oxygen in producing pneumonia, 356. 
 
 CHAPTER XIII. EFFECTS OF Low ATMOSPHERIC 
 
 PRESSURES 358 
 
 Occurrence of low atmospheric pressures at high altitudes. "Mountain sickness," 
 358. Summary of Paul Bert's fundamental experiments on the pressure effects 
 of gases, 358. His experiment on man in a steel chamber, 360. Reason why 
 a given lowering of alveolar oxygen pressure has less physiological effect at a 
 low atmospheric pressure than at ordinary atmospheric pressure, 362. Effect 
 of CO;j pressure in diminishing the anoxaemia of a low atmospheric pressure, 
 362. Mosso's "acapnia" theory, 363. Acclimatization to low atmospheric 
 pressures, 364. Effects of high altitudes in increasing the haemoglobin per- 
 centage of the blood, 364. Effect of increased atmospheric pressure in dimin- 
 ishing the haemoglobin percentage, 365. Beneficial effect of increased haemo- 
 globin in anoxaemia, 365. Increased breathing in acclimatized persons, 366. 
 Physiological effect of a mere increase of breathing, 367. The acclimatiza- 
 tion change is a compensation of alkalosis, 369. Alveolar COz pressure in 
 persons acclimatized at various altitudes, 370. Conclusions from the Duke of 
 Abruzzi's Himalayan Expedition, 372. Active secretion of oxygen in the lungs 
 after acclimatization, 373. Relation of physical training to power of oxygen 
 secretion, 373. History of high ascents in balloons, 375. High ascent by 
 Glaisher and Coxwell in 1862, 375. Fatal ascent of the Zenith in 1875, 376. 
 High ascent with use of oxygen in 1901, 378. Experiments of von Schrotter, 
 379. Recent high American aeroplane ascent, 379. Limits of height attainable 
 with use of ordinary oxygen apparatus, 379. Apparatus required for indefinitely 
 great heights, 380. 
 
xviii RESPIRATION 
 
 CHAPTER XIV. GENERAL CONCLUSIONS . . .382 
 
 The breathing and circulation are so regulated as to keep the diffusion pres- 
 sures of oxygen, and of hydrogen and hydroxyl ions, in the tissues normal, 382. 
 Breathing and circulation are responses to tissue activity, and do not pri- 
 marily determine it, 383. Claude Bernard and the regulation of internal en- 
 vironment, 383. Diffusion pressures and Bernard's "conditions of life/'sSs. 
 Diffusion pressure of water on the same footing as that of other blood constitu- 
 ents, 384. The blood constituents are in continuous active relation with the 
 living tissues, 385. Comparison of living tissue elements with dissociable 
 chemical molecules, 386. Conception of the living body as the seat of a system 
 of mutually dependent reversible reactions, 386. Defects of the mechanistic 
 and "hormone" theories of physiological inter-connection, 387. The dividing 
 line between biology and the physical sciences, 388. The fundamental con- 
 ception of biology, and the real work of the biological sciences, 389. This 
 work illustrated by the investigations detailed in previous chapters, 389. 
 Real nature of organic identity, 390. The existence of active maintenance of 
 organic identity is the foundation of medicine and surgery, as well as of physiolo- 
 gy and morphology, 391. Examination of the argument that the physical con- 
 ception of Nature is truer and more scientific than the biological, 392. The 
 previous question which is fatal to the physical conception, 394. Physical 
 reality a superficial sensuous appearance, 394. In describing biological phe- 
 nomena and putting her questions to Nature, biology must use her own working 
 hypothesis and not those of the physical sciences, 394. Nature as seen by the 
 biologist, 396. Supposed evolution from "inorganic" conditions, 396. Indi- 
 vidual life and life in association, 396. It is impossible to describe or define 
 conscious activity in either physical or biological terms, 397. Neither the 
 physical nor biological interpretation of Nature is, in the last resort, more than 
 a practical makeshift, 398. The rightful practical sphere of physiology does not 
 include distinctively conscious activity, 399. 
 
 APPENDIX ......... 400 
 
 A. Determination of oxygen capacity of blood haemoglobin by ferricyanide, 400. 
 B. Determination of oxygen capacity of blood haemoglobin by haemoglobin- 
 ometer, 404. C. Determination of oxygen and carbon dioxide in blood by ferri- 
 cyanide and acid, 407. D. Colorimetric determination of percentage saturation 
 of haemoglobin with CO, 418. E. Determination of blood volume in man 
 during life by CO, 424. 
 
CHAPTER I 
 Historical Introduction. 
 
 IN the history of physiological discovery the growth of knowledge 
 as to the physiology of breathing was comparatively late. Before 
 the middle of the seventeenth century hardly anything was known 
 about breathing except its muscular mechanism and the facts 
 that if the breathing of a man or higher animal is interrupted 
 for more than a very short time death ensues, and that the breath- 
 ing is increased by exertion and by some diseases. The discovery 
 by Harvey of the circulation threw no positive light on the physi- 
 ology of breathing, and it was still generally believed that the 
 main function of respiration is to cool the blood. Progress was 
 impossible without corresponding progress in chemistry, 
 
 The first beginnings of a better knowledge date from the work 
 at Oxford of Robert Boyle 1 and Mayow 2 a young doctor. Boyle 
 showed with the air pump that air is necessary to life, and Mayow 
 investigated and compared together the influences of niter in the 
 combustion of gunpowder, and of air in respiration and ordinary 
 combustion in air. He drew the conclusion that in all of these 
 processes a "nitro-aerial spirit" combines with "sulphur" (com- 
 bustible matter). As regards respiration he concluded that the 
 nitro-aerial spirit is present in limited proportion in air, and is 
 absorbed from the air in the lungs by the blood, carried by the 
 circulation to the brain, where it is separated off in the ventricles, 
 and thence passes down the supposed nerve- tubules to the muscles, 
 where it unites with "sulphur" and produces muscular contraction 
 by the resulting explosions. He explained the increased breathing 
 which accompanies muscular exertion as a necessary accompani- 
 ment of the increased consumption of the nitro-aerial spirit. 
 
 It will thus be seen that he had practically discovered oxygen r 
 in so far as the rudimentary chemical ideas which he had formed 
 permitted the discovery. He had also formed a sound physiological 
 conception of the relation between muscular work and increased 
 breathing. Mayow's conception of oxygen passing down the 
 
 1 Boyle, New ex-pertinents physico-mechanical, touching the Spring of the Air, 
 Oxford, 1666. Particularly Experiments XL and XLI, with the accompanying 
 "Digression containing some Doubts touching Respiration." 
 
 51 Mayow, Tractatus Quinque Meciico-physici, Oxford, 1673. In particular 
 Tractatus II, De Respiratione (26. Edition). 
 

 2 RESPIRATION 
 
 nerves was of course only a modification of the idea then current, 
 and elaborated by Descartes among others, that muscular con- 
 traction depends upon the "animal spirits" passing down the 
 supposed nerve tubules from the brain. This conception was ap- 
 parently confirmed by the effects of cutting or ligaturing nerves; 
 and Lower, 8 another Oxford physician, performed the striking 
 experiment of completely disturbing the action of the heart by a 
 ligature on the vagus nerve. He had stumbled upon inhibition 
 and misinterpreted it in favor of Mayow's theory. 
 
 About the same time another significant observation was made 
 by Hooke, 4 the Secretary of the Royal Society. He found that 
 when the chest of an animal was opened so that the lungs col- 
 lapsed, it could be revived and kept alive by artificial respiration, 
 and, if holes were pricked in the lungs so that air could pass 
 through them, the animal could still be kept alive if a stream of 
 air was continuously blown through the lungs, although they did 
 not move. 
 
 The foundations thus seemed to be laid of our present knowl- 
 edge of the physiology of breathing; but unfortunately the sig- 
 nificance of the discoveries made at Oxford was not appreciated, 
 and indeed the study of physiology and other branches of natural 
 science there was practically allowed to die out for the succeeding 
 two hundred years. 
 
 The next important step in connection with respiration was 
 the discovery, about the middle of the eighteenth century, by 
 Joseph Black of Edinburgh, that "fixed air" (carbon dioxide) 
 which he had found to be liberated by acids from mild alkalies 
 (carbonates) is given off by the lungs in respiration. Priestley 
 discovered soon afterward that what, in accordance with Stahl's 
 phlogiston theory, he called "dephlogisticated air" (oxygen) dis- 
 appears both in ordinary combustion and in animal respiration, 
 while it is produced by green plants in sunlight. Lavoisier then 
 followed up Black's and Priestley's work by showing that in 
 combustion what he for the first time called oxygen combines 
 with carbon and other substances, and that carbon dioxide is 
 produced by the combination of carbon and oxygen, while water 
 is produced by the combination of hydrogen and oxygen. He and 
 Laplace 5 also showed that the carbon dioxide produced by an 
 animal is nearly equivalent to the oxygen consumed, and that 
 
 "Lower, Tractatus de Corde, p. 86, 1669. 
 
 4 Hooke, Phil. Trans., II, p. 539, 1667. Hooke had been assistant to Willis 
 and Boyle at Oxford. 
 
 "Lavoisier and Laplace, Memoir es de I'Academie des Sciences, p. 337, 1780. 
 
RESPIRATION 3 
 
 the amount of heat formed by an animal is nearly equivalent to 
 that formed in combustion of carbon when an equal quantity of 
 oxygen is consumed in respiration and combustion. He thus made 
 it clear that in the living body, just as in combustion, oxygen 
 combines with carbon and other substances, producing carbon 
 dioxide and other oxidation products : also that this combination 
 is the source of animal heat. 
 
 He found in the course of experiments on man that during 
 muscular work the consumption of oxygen and output of carbon 
 dioxide is increased. Curiously enough, he expresses regret that 
 this should be so, as the laboring classes, who have least money 
 for buying food, consume more food than those who are better 
 off. 6 The essential connection between physiological work and 
 consumption of oxygen was still hidden from him, although, as 
 already seen, Mayow had fairly correct ideas on this subject. 
 It was not until 1845 that Mayer, 7 a German country doctor, 
 pointed out in connection with the general formulation of the 
 doctrine of conservation of energy, that in living animals, as in 
 steam engines, ordinary kinetic energy as well as heat has its 
 source in the potential energy liberated in the process of oxida- 
 tion. Oxidation is thus the ultimate source of the energy of animal 
 movements. Every exact experiment made since then on this 
 subject has confirmed Mayer's conclusion, and the increased 
 consumption of oxygen during muscular work became as intelli- 
 gible as it was on Mayow's crude theory. 
 
 The discoveries with regard to the chemistry of respiration 
 raised the further question as to what the exact nature of the 
 combustible material is, and where the combination of oxygen 
 with combustible matter occurs. As regards the first question it 
 was evident that since on an average the composition of the adult 
 living body remains constant, and the excreta, as compared with 
 the food taken, contain very little combustible material, the 
 material oxidized must correspond to the oxidizable matter of 
 the food. This material was classified by Prout as belonging almost 
 entirely to one or other of three groups of substances, known now 
 under the names of proteins, carbohydrates, and fats. Of these 
 the former alone contains nitrogen, which is excreted in the urine 
 in the form, mainly, of urea when the protein is oxidized. Only 
 water and carbon dioxide are formed in the oxidation of carbo- 
 
 8 Lavoisier and Sequin, Mem. de I'Acaci., p. 185, 1789. 
 
 7 Mayer, Die orgamsche Bewegung in threm Zusammenhange mit dem Stoft- 
 wechsel, Heilbronn, 1845. 
 
4 RESPIRATION 
 
 hydrates or fats, and by the ratios and amounts in which nitrogen 
 compounds and carbon dioxide are excreted and oxygen consumed 
 we can calculate how much protein, carbohydrate, and fat is 
 being consumed in the body. 
 
 As regards the second question there was for long much doubt. 
 It was, however, definitely shown by Magnus 8 in 1845 tnat mucn 
 gas is liberated from blood on exposing it to a vacuum, and that 
 less oxygen and more carbon dioxide are given off from venous 
 than from arterial blood. The mercurial blood gas pump was 
 then gradually perfected, mainly by Lothar Meyer, Ludwig, and 
 Pfliiger ; and it was gradually established that the oxygen which 
 disappears in the lungs is taken up by the blood almost entirely 
 in the form of a loose chemical compound with haemoglobin, the 
 colored albuminous substance in the red corpuscles. This compound 
 yields up part of its oxygen as the blood passes round the systemic 
 circulation, and returns to the lungs for a fresh charge, the charg- 
 ing being due to the higher partial pressure of oxygen in the 
 lungs, while the partial discharging in the systemic circulation 
 is due to the lower partial pressure there in consequence of con- 
 sumption of oxygen. The discharging is accompanied by a change 
 of color from scarlet to dark purple. Similarly carbon dioxide is 
 taken up mainly in the form of a loose chemical combination with 
 alkali, and discharged in the lungs as a consequence of the lower 
 partial pressure of the gas in the lungs. For a considerable time 
 there was much doubt as to how far the actual oxidation occurs in 
 the blood or in the tissue elements; but the investigations of 
 Pfliiger 9 about 1872 showed clearly that practically all the oxida- 
 tion occurs in the tissues. 
 
 So far I have discussed from an abstract physical and chemical 
 standpoint the main outlines of discovery relating to respiration. 
 It is now necessary to consider these discoveries more closely, and 
 from a physiological standpoint. For a long time the brilliance 
 of Lavoisier's discovery as to the relation between respiration and 
 animal heat carried physiologists to some extent off their balance, 
 as it came to be believed that heat production is a more or less 
 blind mechanical process under no direct organic control, and 
 presumably dependent simply upon the supply of oxygen and 
 oxidizable material. Thus Liebig, who was not only a great 
 chemist but also a great chemical physiologist, concluded that 
 every increase in the food consumed or the amount of oxygen 
 
 8 Magnus, Annalen der Physik, XL, 1838, and LXVI, 1845. 
 
 9 Pfliiger, Pjliiger's Archrv, VI, p. 43, 1872. 
 
RESPIRATION 5 
 
 introduced into the lungs must increase the rate of oxidation and 
 heat production. 10 This conclusion seemed to be confirmed when 
 he introduced his well-known method for the determination of 
 urea in urine and it was found that every increase in the amount 
 of nitrogenous food eaten was followed by a corresponding in- 
 crease in the amount of urea excreted, although during complete 
 starvation the excretion of urea was not diminished below a certain 
 minimum. He inferred that it is only the "vital force" which pro- 
 tects the body against indefinite oxidation, and that when more 
 food is introduced than is really required this protection is not 
 extended, so that the food material falls a prey to oxygen. In 
 assuming this influence of the "vital force" he was only applying 
 to the phenomena of physiological oxidation the ideas held by 
 the majority of contemporary physiologists. 
 
 When, however, the phenomena of physiological oxidation 
 came to be studied more closely by Bidder and Schmidt, Voit, and 
 other physiologists, it was found that although the excretion of 
 urea might fall greatly during starvation there was very little 
 fall in the consumption of oxygen. It thus became evident that any 
 diminution in the consumption of protein was accompanied by 
 increase in consumption of the fat and of any carbohydrate 
 remaining in the body. Further investigation of the ratios in which 
 protein, carbohydrate, and fat replaced one another in the oxida- 
 tions occurring in the body resulted in the striking discovery by 
 Rubner that within wide limits of variation in their supply to the 
 body they replace one another in proportion to the energy which 
 they liberate in their oxidation within the body. 11 Thus I gram of 
 fat furnishes as much energy as 2% grams of protein or carbohy- 
 drate, and I gram of fat from the reserve in the body takes the 
 place of 2*4 grams of protein or carbohydrate when the supply 
 of the latter in the food is cut off. The idea that the rate of oxida- 
 tion in the living body is determined by the rate of food supply 
 is thus erroneous. On the contrary the oxidation is regulated with 
 marvelous accuracy in accordance with its energy value in satis- 
 faction of what are commonly called the "energy requirements" 
 of the body. Rubner's discovery is one of the main physiological 
 foundations of scientific dietetics. 
 
 Just as the rate of physiological combustion, other things being 
 equal, is not determined in the higher organisms by the supply .of 
 food material, so it is not determined by the abundance of the 
 
 10 Liebig, Letters on Chemistry, Third English Edition, p. 314, 1855. 
 "Rubner, Zeitschr. f. Biologie, XIX, p. 313, 1883. 
 
6 RESPIRATION 
 
 oxygen supply. Lavoisier himself and afterwards Regnault and 
 Reiset found that a warm-blooded animal breathing pure oxygen 
 consumes no more oxygen than an animal breathing ordinary 
 air; and subsequent investigations have shown that the oxygen 
 percentage in air has to be reduced very low before the oxy- 
 gen consumption is diminished. Pfliiger also found that oxidation 
 in the tissues is within wide limits independent of the rate of supply 
 of oxygen through the blood circulation. We are thus again face 
 to face with "physiological requirements." 
 
 When temperature and heat production in the living body came 
 to be studied physiologically the first striking fact discovered 
 was that however much the external temperature might vary 
 within wide limits, the body temperature of warm-blooded ani- 
 mals remained practically the same during, health. Similarly, 
 although the heat production might be increased several times 
 by muscular exertion there was no material increase of body 
 temperature, and it became quite evident that the rise of tempera- 
 ture in fever is not due to increased heat production, but to dis- 
 turbance in the nervous regulation of heat discharge from the 
 body. Finally, when the influence of variations in external tem- 
 perature on heat production in the body was measured, it was 
 found by a succession of observers, including, besides Lavoisier, 12 
 Crawford in 1788, and Pfliiger and others in more recent times, 
 that, particularly in small animals, a lowering of external tem- 
 perature evokes through the influence of the nervous system a 
 rise in heat production, so that heat production becomes subservi- 
 ent to the maintenance of body temperature. This maintenance 
 is therefore one of the factors determining physiological energy 
 requirements. 
 
 When we inquire what determines the energy requirements of 
 the body as a whole we ' nd that the results of investigation point 
 us towards a numbei of associated conditions which we can 
 identify one by one by observation or experiment, but which ordi- 
 narily occur in conjunction with one another, and on an average 
 remain very constant. Thus the activity of the nervous system in 
 determining various forms of muscular and glandular activity 
 constitutes one of the chief factors. But the activities of the 
 nervous system are themselves subject to control in the form of 
 what we call on the one hand "fatigue," or on the other "exuber- 
 ance of spirits," finding its expression in man in games and what 
 appear at first sight to be mere "luxus" activities of all kinds. 
 
 u Pfluger, Pfluger's Archw, XII, p. 282, 1876. 
 
RESPIRATION 7 
 
 Hence apart from seasonal variations the daily nervous activities 
 are pretty constant in total amount. 
 
 Although the internal body temperature is actually very con- 
 stant, yet a very moderate actual rise or actual fall in body 
 temperature is sufficient to increase or diminish oxidation very 
 materially. In fever, for instance, the oxidation in the body is 
 greater than it would be without the rise of temperature but with 
 other conditions the same. The oxidation in fever is, however, 
 only a fraction of that during even very moderate exertion. 
 
 When we examine still more closely, and in the light of the facts 
 which are continuously becoming revealed by pathology and 
 pharmacology, we begin to realize that "energy requirements" 
 depend on an infinite multitude of associated "normal conditions." 
 An upset in the proportion of, say, calcium or potassium in the 
 blood, or in that of substances produced in minute amounts in 
 one or other of the "ductless glands" or supplied to the body 
 along with the other main constituents in ordinary food, will 
 dramatically end "energy requirements" by that mysterious phe- 
 nomenon which we call death* and which we are so familiar with 
 that we almost cease to speculate about its nature. 
 
 At first sight death may seem to become intelligible when we 
 find that in the higher animals its immediate cause is want of 
 oxygen in the tissues owing to interruption of the circulation or 
 breathing. But further examination shows us that death is no mere 
 stoppage of an engine owing to lack of air or fuel, but also total 
 ruin of what we took to be machinery. It is a mysterious dissolu- 
 tion in the association together of the infinitely complex group of 
 normals which constitute the life the <wrcs of an organism ; 
 and an examination of the fragments left has thrown no light on 
 why the association should have existed at all, or endured so long. 
 The outward form and internal arrangenent and composition of 
 the dead body tell their story of life to him who can interpret their 
 hieroglyphics ; but there is no life visible. The gulf between the 
 dead and the living is a gulf across which our present intellectual 
 vision does not reach, and we only deceive ourselves when we 
 sometimes imagine that it does. Thus when we ask what determines 
 those "energy requirements" which determine consumption of 
 oxygen and output of carbon dioxide in the living body, the only 
 answer we can at present elicit from experimental investigation 
 is that the energy requirements are one side of the <wris of the 
 organism. To those who object that the <f>v<ris is a mere name, 
 and that physiology must be simply physics and chemistry I can 
 
8 RESPIRATION 
 
 only reply, following the example of Hippocrates who protested 
 against the intrusion of abstract philosophical speculations into 
 medicine, that there can be no doubt about the existence of the 
 associated and persistent group of appearances which the word 
 <v<ris designates when applied to life. If we ignore this we reject 
 the one thing which gives us that grasp of biological phenomena 
 which enables us to predict them, and renders a scientific treatment 
 of biology and medicine possible. 
 
 The immediate subject of this book is the side of physiology 
 which concerns the means by which the supply of oxygen and 
 removal of carbon dioxide are so carried out and regulated that 
 physiological requirements are met. That this supply and removal 
 are through the lungs and blood has already been pointed out; 
 but the development of knowledge as to the means of regulation 
 must now be traced. Much difficulty arises, however, from the 
 fact that the problem itself was only recently realized with any 
 clearness. Respiration and circulation have been to a large extent 
 treated as if the requirements of the body were on the whole 
 constant. Actually, however, the consumption of oxygen and 
 production of carbon dioxide fluctuate greatly. A heavy exertion, 
 for instance, will increase tenfold the consumption of oxygen and 
 output of carbon dioxide for the whole body, and must certainly 
 increase in a far higher ratio the consumption and output of the 
 muscles actually at work. 
 
 It has of course been known from the earliest times that muscu- 
 lar activity causes great increase in the depth and frequency of 
 the breathing, and that rebreathing the same air has a similar 
 effect; but the very familiarity of these facts seems to have led 
 to a relative neglect of the problem of how the respiratory activi- 
 ties are regulated. An undue specialism has led to the investiga- 
 tion of each form of bodily activity as if it were something 
 separable from other bodily activities, and not a physiological 
 activity. Further confusion has arisen through the roughness of 
 many of the experiments made on animals, and corresponding 
 failure to detect the delicacy of physiological regulation. 
 
 In 181 1 it was discovered by Legallois that if a portion of tissue 
 definitely localized in the medulla oblongata is destroyed respira- 
 tion ceases and death ensues. 13 This part of the medulla has come 
 to be known as the respiratory center; and round the responses of 
 this "center" to various nervous and other stimuli the physiologi- 
 cal investigation of breathing has been focused. 
 
 "Legallois, Experiences sur la principe de la vie, Paris, 1812. 
 
RESPIRATION 9 
 
 It was found by Legallois and subsequent investigators that the 
 nervous connections both above and below the respiratory center 
 can be successively severed without preventing the rhythmic dis- 
 charges of inspiratory and expiratory impulses except in so far as 
 efferent nerves connected with the center are cut off from it. Thus 
 the rhythmic discharges of the center are not dependent on afferent 
 nervous impulses and continue regularly so long as normal arterial 
 blood is supplied to it. In this sense the action of the center is 
 automatic. On the other hand the mode of action of the center is 
 much affected by nervous stimuli. 
 
 In the first place its action is to a large extent under voluntary 
 control. Thus the breathing can easily be suspended for about a 
 minute, and in the actions of speaking, singing, etc., is greatly 
 interfered with. The rate and depth of breathing are also 
 under voluntary control, and may be much affected by emotion. 
 Consciously perceived stimuli of all kinds may also affect the 
 breathing particularly stimuli affecting the air passages. The 
 irregularity and variability of the breathing owing to all these 
 causes tended to direct the attention of physiologists away from 
 the central problem of how the breathing responds to fundamental 
 physiological requirements. 
 
 It was soon discovered that apart from consciously felt stimuli 
 the breathing is specially affected by afferent stimuli conducted 
 by the vagus nerve. Early last century it was noticed that when the 
 vagus nerves are severed the breathing becomes less frequent and 
 deeper; and on stimulating the vagi various marked effects, de- 
 pending on the strength of stimulus, were found to be produced 
 on the breathing. 
 
 In 1868 Hering and Breuer 14 made the striking discovery that 
 on mechanically interrupting, at the end of inspiration, the ex- 
 pulsion of air from the lungs the rhythm of respiratory effort is 
 interrupted for a time, until at last this interruption is broken by 
 an inspiratory effort, followed by alternating expiratory and in- 
 spiratory efforts showing that the center has renewed its rhythmic 
 activity. Similarly if at the end of expiration air is prevented from 
 entering the lungs there is an interruption before the center re- 
 turns to its normal rhythmic activity. These effects are completely 
 absent if the vagi have been divided. The slow rhythmic dis- 
 charges of the center go on quite independently of whether the 
 inflation or deflation of the lungs is prevented or not. 
 
 "Hering and Breuer, Sitzber. d. Wiener Akad., Math-natural. Cl. (2), LVII, 
 p. 672 and LVIII, p. 909, 1868. 
 
10 RESPIRATION 
 
 It was evident from these experiments, and from the marked 
 slowing and deepening of breathing after the vagi are cut, that 
 distention of the lungs stimulates the nerve endings of the vagi 
 in the lungs in such a way as to terminate inspiration and initiate 
 expiration, while deflation of the lungs produces a corresponding 
 stimulus acting so as to terminate expiration and initiate inspira- 
 tion. Thus inspiration seems to be the cause of expiration, and 
 expiration of inspiration. Hering described this as the "self-regu- 
 lation" of breathing. 
 
 Another series of observations relates to chemical stimulation 
 of the respiratory center. It was found that if air containing very 
 little oxygen is breathed, or a small volume of ordinary air is 
 repeatedly rebreathed, great panting ensues, followed by general 
 convulsions and final cessation of breathing. The same result was 
 found by Kiissmaul and Tenner to follow if the blood supply to 
 the brain is completely cut off, so that the blood remaining in the 
 vessels becomes venous. The respiratory center is thus first stimu- 
 lated to excessive action by imperfectly oxygenated or venous 
 blood, and later becomes exhausted and finally ceases to act. But 
 another most significant fact was definitely discovered by Rosen - 
 thai in i862. 15 If in an animal artificial respiration is pushed so 
 that the ventilation of the lungs is abnormally great the activity 
 of the respiratory center ceases entirely for a time, and this 
 condition he designated as apnoea. In most persons apnoea can be 
 produced easily by voluntarily forcing the breathing for a short 
 time. After a few deep and rapid breaths it will be noticed that 
 all natural tendency to breathe ceases for a time. 
 
 These observations suggested that ordinary breathing is de- 
 termined by the degree of arterialization of the blood supplying 
 the respiratory center. If the degree of arterialization is dimin- 
 ished the breathing is increased, and vice versa, so that the 
 respiratory center automatically maintains a normal degree of 
 arterialization. When the venous blood is arterialized in the lungs 
 two changes occur, as we have already seen. The blood takes up 
 oxygen, and also loses carbonic acid. It might be one or the other, 
 or else both, of these changes that determines the activity of the 
 respiratory center. The most immediately evident change in the 
 blood during its passage through the lungs is its change in color 
 from a bluish to a bright scarlet color, and this change, as already 
 seen, is due solely to its oxygenation and not to loss of carbonic 
 acid. We thus naturally tend to think of blue blood as venous and 
 
 "Resentful, Z>* At**mb*totg**g**, 1862. 
 
RESPIRATION II 
 
 scarlet as arterial ; and with the blood pump we can easily prove 
 that the scarlet blood contains more dissociable oxygen than the 
 blue. 
 
 Rosenthal came to the conclusion that it is solely or almost 
 solely in virtue of its varying oxygen content that the blood 
 stimulates the respiratory center or not. 18 Careful blood-gas de- 
 terminations showed that when apnoea had been produced by 
 forced ventilation of the lungs the arterial blood contained a little 
 more oxygen. On the other hand, when oxygenation was rendered 
 incomplete by letting an animal breathe air very poor in oxygen 
 there was an immediate great increase in the breathing, although 
 the discharge of carbonic acid was in no way interfered with. 
 Moreover, when air containing a very large excess of CO 2 was 
 breathed by an animal the rate of breathing remained normal. 
 Rosenthal also brought forward other evidence which appeared 
 to point in the same direction ; but the weak point in his argument 
 was the fact that there is no apnoea when pure oxygen is breathed, 
 although the arterial blood contains a good deal more oxygen 
 than usual. The truth is that he had been misled by the fact that a 
 very high percentage of CO 2 in the air breathed has a narcotic 
 effect, so that the breathing, which is in reality increased at first 
 by raising the percentage of CO 2 in the air of the lungs, quiets 
 down again when the percentage becomes very high. Pfliiger and 
 Dohmen 17 showed that both excess of CO 2 (provided that the CO 2 
 is not in too great excess) and want of oxygen excite the respira- 
 tory center. 
 
 A further fact, discovered originally by Traube, 18 but often 
 overlooked by subsequent investigators, was that apnoea could 
 be produced even by a gas such as nitrogen or hydrogen, in which 
 no oxygen was present. Thus if apnoea is due to "over-arterializa- 
 tion" of the arterial blood it can be produced by the simple re- 
 moval of CO 2 , whether or not the oxygen is also diminished, 
 although the artificial ventilation of the lungs must be much more 
 vigorous if apnoea is produced in the absence of oxygen. 
 
 Meanwhile another theory of apnoea was put forward, and 
 has led, as will be shown later, to the utmost confusion and com- 
 plete misinterpretation of the facts. When the lungs are distended 
 there is, as already mentioned, an interruption in the rhythm of 
 discharge from the respiratory center. The inspiratory muscles, 
 
 "Rosenthal in Hermann's Hanctbuch cter PkysioL, Vol. IV, 2, 1882. 
 
 17 Pfliiger, Pfliiger' s Archiv, I, p. 61, 1868. 
 
 13 Traube, Allgem. Med. Centralzeitung, 1862, No. 38, and 1863, No. 97. 
 
12 RESPIRATION 
 
 and specially the diaphragm, are, and remain till the interruption 
 is broken by an inspiratory effort, relaxed. This interruption of 
 inspiratory effort came to be interpreted as an apnoea, and appears 
 so if only inspiratory muscular movements are recorded, as, for 
 instance, with the method adopted in Hering's laboratory by 
 Head, 19 in which only the contractions of the diaphragm are re- 
 corded, or with other methods which do not record tonic expira- 
 tory effort. Hence it came to be assumed that there exists what is 
 called "vagus apnoea." The next step was to maintain that all 
 apnoea is in reality vagus apnoea, and this inference was supported 
 by the fact that "apnoea" can still be obtained when the arterial 
 blood is blue owing to air containing a very low percentage of 
 oxygen being breathed, and can also be produced (as Lorrain 
 Smith and I found) by air very rich in CO 2 . It was also affirmed 
 by Brown-Sequard that after the vagi are cut apnoea cannot be 
 produced, though this statement can easily be shown to be com- 
 pletely mistaken. With an efficient apparatus for increasing the 
 ventilation of the lungs apnoea can quite readily be produced 
 after section of the vagi. 
 
 On the other hand, increasingly clear evidence accumulated 
 that apnoea due to over- ventilation of the blood passing through 
 the lungs exists as a matter of fact. The most striking proof of 
 this was afforded by experiments in which Fredericq 20 crossed 
 the circulation of two animals by connecting the vessels in such a 
 way that the respiratory center of each animal was supplied with 
 arterial blood from the other animal. He then found that when 
 excessive artificial respiration was produced in one of the animals 
 apnoea was produced in the other, and when the artificial respira- 
 tion ceased hyperpnoea continued in the animal which had had 
 artificial respiration, since its respiratory center was now receiving 
 blood which was venous owing to the cessation of breathing in the 
 other animal. This hyperpnoea, on the other hand, maintained 
 the apnoea in the other animal, so that one of the animals re- 
 mained apnoeic while the other remained hyperpnoeic. 
 
 This experiment showed clearly the existence of a true "chemi- 
 cal" apnoea; but, as the existence of vagus apnoea was also con- 
 sidered to be firmly established, the existence of both forms of 
 apnoea came to be generally assumed. As regards vagus apnoea the 
 evidence was considered to show that when apnoea is produced by 
 distending the lungs with air or hydrogen it is vagus apnoea that 
 
 19 Head, Journ. of Physiol., X, i and 279, 1889. 
 M Fredericq, Arch, der Physiol., 17, p. 563, 1901. 
 
RESPIRATION 13 
 
 lasts on after the distention ceases, and from this supposed fact 
 the further inference was drawn that repeated distention of the 
 lungs produces a summed vagus effect resulting in vagus aprioea 
 after the distentions have ceased. Thus the same procedure that 
 causes chemical apnoea seemed to produce also vagus apnoea, 
 and the two kinds of apnoea could hardly be distinguished in 
 practice. Moreover hyperpnoea due to any chemical cause such as 
 want of oxygen or excess of CO 2 must apparently tend to be 
 prevented by the production of vagus apnoea due to repeated 
 distentions of the lungs. The two processes by which the breathing 
 appeared to be regulated acted, therefore, in opposite directions. 
 
 As regards the chemical stimuli acting on the respiratory center, 
 it remains to consider the further evidence as to the relative im- 
 portance of want of oxygen and excess of CO 2 ; also whether other 
 chemical stimuli act on the center. In 1885 Miescher 21 showed 
 by experiments on man that a given small increase in the per- 
 centage of CO 2 in air affects the breathing considerably, while a 
 corresponding diminution in the oxygen percentage has no such 
 effect. He was thus led to the conclusion that it is the CO 2 per- 
 centage in the air of the lungs that ordinarily determines the 
 chemical regulation of breathing, and not the oxygen percentage. 
 Thus CO 2 protects the body from want of oxygen so long as 
 ordinary air is breathed. It will be seen in the sequel how rela- 
 tively correct this general view of Miescher's was, although he 
 maintained the existence of vagus apnoea and thus shared in the 
 mistakes of his time. 
 
 In 1888 Geppert and Zuntz 22 published the results of a very 
 careful series of experiments on the effects of muscular work (pro- 
 duced by tetanizing the hind limbs of an animal after section of the 
 spinal cord) on respiration. After bringing forward new evidence 
 that it is the blood which carries the stimulus for increased breath- 
 ing to the respiratory center they showed that during the work 
 the proportion of CO 2 in the blood was greatly diminished, and 
 that there was also a slight increase in the oxygen percentage of 
 the blood. Hence, they argued, it is neither increase in CO 2 per- 
 centage nor diminution in oxygen percentage that causes the 
 hyperpnoea accompanying muscular exertion. They believed that 
 it is some acid substance produced in the muscles, and pointed out 
 that Walter had found that the breathing is much increased in 
 poisoning by acids. 
 
 21 Miescher, Arch. f. (Anat. uJ) Physiol., p. 355, 1885. 
 
 32 Geppert and Zuntz, Pfltiger's Archiv, XLII, pp. 195, 209, 1888. 
 
I 4 RESPIRATION 
 
 From the foregoing review of the knowledge existing up to 
 the beginning of the present century on the physiological regu- 
 lation of breathing it will be seen that the conclusions reached 
 were unsatisfactory in many ways, and to some extent contra- 
 dictory. On the one hand the nervous regulation through the vagi 
 and other nerves seemed to have no relation to the requirements 
 of the body for oxygen and for removal of CO 2 , and in fact to 
 act antagonistically to these requirements. On the other hand 
 the excitation of the breathing during muscular work seemed 
 also, from the results of Geppert and Zuntz, to have no definite 
 relation to increased requirements for oxygen and CO 2 . There 
 was also no definite quantitative information as to why in normal 
 breathing during rest the composition of the expired air is so 
 constant as it is. Without more exact and consistent physiological 
 knowledge it appeared to be very difficult to interpret the ab- 
 normal breathing so often met with in disease, or to know how to 
 set about investigating it. 
 
 From still another standpoint the existing knowledge was very 
 unsatisfactory to me personally. From a consideration of the 
 general characteristics which distinguish a living organism from 
 a machine I had become convinced that a living organism cannot 
 be correctly studied piece by piece separately as the parts of a 
 machine can be studied, the working of the whole machine being 
 deduced synthetically from the separate study of each of the parts. 
 A living organism is constantly showing itself to be a self-main- 
 taining whole, and each part must therefore always be behaving 
 as a part of such a self-maintaining whole. In the existing 
 knowledge of the physiology of breathing this characteristic could 
 not be clearly traced. The regulation of breathing did not, as 
 represented in the existing theories, appear to be determined in 
 accordance with the requirements of the body as a whole; and 
 for this reason I doubted the correctness of these theories, and 
 suspected that errors had arisen through the mistake of not study- 
 ing the breathing as one of the coordinated activities of the whole 
 body. In so far as the investigations detailed in succeeding chap- 
 ters originated with me, they were mainly inspired by the con- 
 siderations just mentioned; and, as will be seen in the sequel, the 
 same considerations have led to a reinvestigation and reinterpre- 
 tation of other physiological activities besides breathing. 
 
CHAPTER II 
 Carbon Dioxide and Regulation of Breathing. 
 
 MY attention was first directed to the regulation of breathing by 
 a series of experiments carried out by Lorrain Smith and myself 1 
 as to the question whether, as had shortly before been asserted by 
 Brown-Sequard and d'Arsonval as a result of a very definite and 
 apparently convincing series of experiments, a poisonous organic 
 substance is given off in expired air. The results of our experi- 
 ments, which were made partly on man and partly on animals, 
 were entirely negative, and left no doubt in our minds that the 
 apparent positive results described were due partly to undetected 
 air leaks which led to animals being asphyxiated, and partly to 
 other experimental errors. In the human experiments we used an 
 air-tight respiration chamber of about 70 cubic feet capacity, in 
 which the air became more and more vitiated by respiration. 
 
 The effects of the vitiated air on our breathing attracted our 
 attention specially. When the proportion of CO 2 in the air rose 
 to about 3 per cent, and the oxygen fell to about 1 7 per cent (there 
 being 20.94 per cent of oxygen and 0.03 per cent of CO 2 in pure 
 atmospheric air) the breathing began to be noticeably increased. 
 With further vitiation the increase in breathing became more and 
 more marked, until with about 6 per cent of CO 2 and 13 per cent 
 of oxygen the panting was very great, with much consequent 
 exhaustion. 
 
 When the experiment was repeated, with the difference that 
 the CO 2 was absorbed by means of soda lime, there was no notice- 
 able increase in the breathing before the oxygen fell below about 
 14 per cent. When, finally, the CO 2 was left in the air, but oxygen 
 was first added so that the oxygen remained abnormally high 
 throughout, the panting was just the same as when ordinary air 
 was used. In short experiments in which the same air was 
 rebreathed from a large bag till we could no longer stand the 
 experiment we found that we had to stop at about 10 per cent of 
 CO 2 , whether oxygen was added or not, and that the oxygen per- 
 centage made no difference to the distress produced. In these 
 experiments there was only about 8 to 9 per cent of oxygen in the 
 
 1 Haldane and Lorrain Smith, Journal of Pathology and Bacteriology, I, pp. 
 168 and 318, 1893. 
 
1 6 RESPIRATION 
 
 rebreathed air at the end of the experiment; but even this made 
 no difference to the breathing. When, on the other hand, a mixture 
 containing a greatly reduced oxygen percentage, without any 
 addition of CO 2 , was breathed, the breathing was increased sensi- 
 bly, as shown by graphic records, when the oxygen fell to about 
 12 per cent, and was greatly increased by lower percentages. 
 With extremely low percentages, such as 2 per cent, consciousness 
 was lost quite suddenly after about 50 seconds, before there was 
 time to notice any increase in the breathing. 
 
 It was evident from these experiments that when the same air 
 is rebreathed, or an insufficient proportion of fresh air is supplied, 
 the increased breathing produced is due simply to excess of CO 2 , 
 until, at least, the oxygen percentage becomes extremely low. It 
 appeared, therefore, that the variations in ordinary breathing in 
 response to variations in the respiratory exchange must be due 
 to the increased CO 2 produced, and not to the increased consump- 
 tion of oxygen. This conclusion was the same as that of Miescher, 
 and supported his views as to the regulation of respiration. 
 
 When more than about 10 per cent of CO 2 was breathed the 
 effect of the mixture was to produce stupefaction, which was very 
 marked with higher percentages. This effect was already well 
 known in animals, and CO 2 was one of the gases tried as an an- 
 aesthetic by Sir James Simpson before he adopted chloroform. 
 The effect of excess of CO 2 in producing ataxia, stupefaction, and 
 loss of consciousness has become very familiar to me in connection 
 with experiments with mine-rescue apparatus and diving appa- 
 ratus. These effects are readily produced in the presence of a large 
 excess of oxygen, and are therefore quite independent of the 
 effects of want of oxygen. The narcotic effect of a large excess of 
 CO 2 quiets down the respiration, and this effect in animals led 
 many previous observers to overlook almost entirely the ordinary 
 effects of CO 2 in stimulating the breathing. 
 
 During the next few years after our first experiments I was 
 engaged in the investigation of other problems connected with 
 general metabolism, respiration, and blood gases, but in 1903 
 returned to the regulation of breathing in a long series of experi- 
 ments carried out in conjunction with Dr. J. G. Priestley, who 
 was then a student at Oxford. 
 
 It seemed pretty evident that in order to reach clear ideas on 
 the regulation of breathing it was necessary to study very care- 
 fully the composition of the alveolar air which is in contact, 
 through the alveolar epithelium, with the blood passing through 
 
RESPIRATION 17 
 
 the lungs ; also that this could be best done on man. The composi- 
 tion of human alveolar air under different conditions had already 
 been calculated by Loewy 2 and Zuntz from the volume occupied 
 by a plaster cast of the respiratory passages in a dead body and 
 the average composition and volume of a breath of expired air. 
 The expired air is evidently a mixture of air from the alveoli with 
 the air which remains in the respiratory tubes at the end of inspira- 
 tion. This air is presumably but little altered by diffusion through 
 the walls of the respiratory tubes, and so far as respiratory ex- 
 change is concerned the volume of the lumen of these tubes must 
 constitute a "dead space" in breathing. The dead space is occupied 
 by alveolar air at the end of expiration, and by more or less pure 
 atmospheric air at the end of inspiration. 
 
 If we know the volume of the dead space, and the volume and 
 composition of the air expired at each breath, we can calculate 
 the average composition of the alveolar air. It is, however, im- 
 possible to estimate directly the volume of the dead space in a 
 particular individual with any accuracy, or to be sure that it 
 remains the same under different physiological conditions. The 
 bronchi and bronchioles are provided with a muscular coat by 
 means of which their lumen is capable of contracting or dilating. 
 Apart from this the air in the alveoli which are nearest to the end 
 of a bronchus will contain purer air during inspiration than during 
 expiration, and this introduces a further complication. 
 
 To get a reliable knowledge of the composition of alveolar air 
 it seemed desirable to make direct determinations. The method 
 introduced by Priestley and myself 3 is simply to make a sharp and 
 deep expiration through a piece of hose pipe about four feet long 
 and one inch in diameter, and provided with a plain glass mouth- 
 piece which is closed by the tongue at the end of the expiration 
 (Figure i). By means of a narrow bore glass tube filled with 
 mercury and introduced air-tight into the hose pipe near the 
 mouthpiece, a sample of the last part of the expired air is then 
 at once taken directly into the gas analysis apparatus as indicated 
 in Figure i, 4 or else into a vacuous sampling tube. 5 If the sample 
 is to be a normal one the breathing must be quite normal before 
 
 2 Loewy, Pfluger's Archiv, LVIII, p. 416, 1894. 
 
 8 Haldane and Priestley, Journ. of Pkysiol., XXXII, p. 225, 1905. 
 
 * For physiological work methods of air analysis which are both accurate and 
 rapid are required. A description of the methods which I introduced with this in 
 view will be found in my book, Methods of Air Analysis, London, Charles Griffin 
 & Co., Third Edition, 1920. 
 
 * If the sample is too large some pure air may be drawn in. 
 
18 
 
 RESPIRATION 
 
 the deep expiration ; and it requires some care to secure this. Under 
 normal resting conditions the depth of expiration needed in order 
 to give a reliable sample at the end of inspiration is at least 800 cc. 
 With less than this the sample is likely to be mixed with air of 
 
 Figure i. 
 Apparatus for obtaining and analysing alveolar air. 
 
 the apparent dead space; for though with normal breathing the 
 volume of the apparent dead space is far less than 800 cc., at least 
 three or four times its volume of alveolar air is needed in order 
 to flush it and the breathing tube out thoroughly. If more than 
 about 800 cc. are expired, the composition of the sample is the 
 same whatever the depth of the expiration, and we designated 
 air of this constant composition as "alveolar air" although, as 
 will be shown later, the composition of the air in the alveoli is by 
 no means such a simple matter as we thought. The following are 
 the averages of results which I obtained on this point when the 
 samples were taken just at the end of inspiration. 6 
 
 Vol. of air expired 
 through tube 
 
 190 cc. 
 
 335 
 5io 
 650 
 950 
 1350 
 
 Per cent of COz in sample 
 taken from tube 
 
 3-03 
 
 4-37 
 5-04 
 5-19 
 
 5-51 
 5.48 
 
 As soon as this method of sampling the alveolar air was applied 
 on ourselves and others it became evident that the alveolar CO 2 
 and O 2 percentage during rest under normal conditions are sur- 
 
 * Haldane, Amer. Journ. of Physiol., XXXVIII, p. 20, 1915. 
 
RESPIRATION 19 
 
 prisingly constant for each individual. As the depth of breathing 
 cannot be kept absolutely steady and the composition of the al- 
 veolar air varies slightly with inspiration and expiration it is 
 best to take at least two samples one just at the end of inspiration, 
 and another just at the end of expiration. The following tables 
 give the CO 2 percentages in samples of our normal resting 
 alveolar air, taken in the sitting position during rest at intervals 
 over about 20 months in 1903 to 1905. Since then we have made 
 many further determinations, but the percentages have remained 
 nearly the same. They are slightly lower or higher on some days 
 than on others, and other observers have noticed this in them- 
 selves. 
 
 
 J. 
 
 S. H. 
 
 
 Barometric 
 
 COi per cent, 
 
 CO* per cent. 
 
 COa per cent, 
 
 pressure in 
 
 end- of 
 
 end of 
 
 mean 
 
 mm. of Hg. 
 
 inspiration 
 
 expiration 
 
 
 759 
 
 5-33 
 
 5.76 
 
 5-545 
 
 747 
 
 5-47 
 
 5.69 
 
 5.56 
 
 748 
 
 5.56 
 
 5-70 
 
 5.63 
 
 748 
 
 5-59 
 
 5.87 
 
 573 
 
 748 
 
 5-38 
 
 5.60 
 
 5-49 
 
 748 
 
 5-33 
 
 5-94 
 
 5-40 
 
 749 
 
 5-8o 
 
 5-5i 
 
 5-87 
 
 749 
 
 5-66 
 
 5-59 
 
 5.585 
 
 765 
 
 5.63 
 
 5.83 
 
 5.6i 
 
 759 
 
 5-42 
 
 5-72 
 
 5-625 
 
 758 . 
 
 5-74 
 
 5-72 
 
 571 
 
 765 
 
 5-53 
 
 5-72 
 
 5-62 
 
 Mean 754 
 
 5-54 
 
 5-72 
 
 5-63 
 
 It will be seen that, as might be expected, the inspiratory 
 samples give on an average a somewhat lower result than the 
 expiratory ones. The average for one subject is 5.63 per cent and 
 for the other 6.28. The slight variations of individual results 
 from these averages are evidently not due merely to changes in 
 barometric pressure. 
 
 When ordinary air was breathed the oxygen percentage in the 
 alveolar air was nearly as steady as the CO 2 percentage. When, 
 however, the oxygen and CO 2 percentages in the inspired air 
 were varied it became quite evident that the breathing is regu- 
 
20 
 
 RESPIRATION 
 
 
 J. 
 
 G. P. 
 
 
 Barometric 
 
 COa per cent, 
 
 COz per cent, 
 
 COt per cent, 
 
 pressure in 
 
 end, of 
 
 end, of 
 
 mean 
 
 mm. of ffg. 
 
 inspiration 
 
 expiration 
 
 
 759 
 
 6.18 
 
 6-43 
 
 6.305 
 
 754 
 
 6. 5 I 
 
 6.63 
 
 6.57 
 
 747 
 
 6.10 
 
 6.70 
 
 6.40 
 
 753 
 
 6.81 
 
 6.86 
 
 6.835 
 
 758 
 
 5-95 
 
 6.74 
 
 6-35 
 
 758 
 
 5-82 
 
 6.23 
 
 6.025 
 
 758 
 
 5-93 
 
 6.21 
 
 6.07 
 
 754 
 
 6.12 
 
 6.33 
 
 6.215 
 
 754 
 
 6.26 
 
 6.20 
 
 6.23 
 
 754 
 
 6.23 
 
 6.05 
 
 6.14 
 
 75i 
 
 5-66 
 
 6.75 
 
 6.205 
 
 75i 
 
 5.98 
 
 5-99 
 
 5.985 
 
 762 
 
 6.37 
 
 6.29 
 
 6.33 
 
 762 
 
 6.24 
 
 6.09 
 
 6.165 
 
 765 
 
 6-39 
 
 6.43 
 
 6.41 
 
 Mean 756 
 
 6.17 
 
 6-39 
 
 6.28 
 
 lated so as to give a constant percentage of CO 2 and not of oxy- 
 gen. The following results were obtained with oxygen percentages 
 varied at intervals in the same subject. 
 
 OXYGEN 
 
 PERCENTAGE 
 
 CO2 
 
 PERCENTAGE 
 
 Inspired, at 
 
 r Alveolar air 
 
 Inspired, 
 
 air Alveolar air 
 
 80.24 
 
 72.21 
 
 0.20 
 
 5.84 
 
 63.67 
 
 57-57 
 
 0.14 
 
 5.41 
 
 20.93 
 
 14.50 
 
 0.03 
 
 5-54 
 
 16.03 
 
 10.39 
 
 O.O5 
 
 5-62 
 
 15.82 
 
 10.59 
 
 0.05 
 
 5.6o 
 
 15.63 
 
 10.60 
 
 0.07 
 
 5-45 
 
 12.85 
 
 8-34 
 
 0.06 
 
 5.37 
 
 12.78 
 
 7.80 
 
 0.07 
 
 5.28 
 
 11-33 
 
 8.96 
 
 O.IO 
 
 3.85 
 
 11.09 
 
 7.10 
 
 0.10 
 
 4-89 
 
 6.23 
 
 4.30 
 
 0.09 
 
 3-57 
 
 This table shows that increase in the oxygen percentage over 
 short periods had no noticeable influence on the alveolar CO 2 
 percentage, and that not until the oxygen percentage in the in- 
 
RESPIRATION 21 
 
 spired air was lowered to about 12 or 13 and the alveolar oxygen 
 percentage to about 8 was there any marked decrease in the CO 2 
 percentage. With a greater lowering of the oxygen percentage 
 than this, however, the breathing was so much increased as to 
 lower the CO 2 percentage considerably. 
 
 When the CO 2 percentage in the inspired air was increased, 
 on the other hand, the effect was strikingly different. Instead of 
 the alveolar CO 2 rising in any direct correspondence to the rise 
 in the inspired CO 2 , the increase in alveolar CO 2 was so slight as 
 to be hardly appreciable even with a rise of 2 or 3 per cent in the 
 CO 2 of the inspired air. This is evident from the following ex- 
 periments, made in the air-tight chamber. 
 
 SUBJECT CO 2 PER CENT CO 2 PER CENT IN RELATIVE 
 IN INSPIRED ALVEOLAR AIR RATES OF 
 AIR ALVEOLAR 
 VENTILATION 
 
 
 End of 
 
 End of 
 
 
 
 
 inspiration 
 
 expiration 
 
 Mean 
 
 
 J. S. H. 0.03 
 
 5-42 
 
 5-83 
 
 5.62 
 
 IOO 
 
 2.07 
 
 5.60 
 
 
 
 
 
 153 
 
 3-80 
 
 6.03 
 
 5-92 
 
 5-97 
 
 258 
 
 0.03 
 
 5-74 
 
 5-72 
 
 5-71 
 
 IOO 
 
 1.74 
 
 5-59 
 
 5.71 
 
 5.65 
 
 143 
 
 3.98 
 
 5-99 
 
 6.16 
 
 6.03 
 
 277 
 
 5.28 
 
 6.44 
 
 6.66 
 
 6-55 
 
 447 
 
 J. G. P. 0.03 
 
 6-85 
 
 6.28 
 
 6.31 
 
 IOO 
 
 5-29 
 
 6.92 
 
 6.86 
 
 6.89 
 
 392 
 
 6.66 
 
 7.62 
 
 7.72 
 
 7.67 
 
 622 
 
 7.66 
 
 8.34 
 
 8.56 
 
 8-45 
 
 795 
 
 The evident effect of adding CO 2 to the inspired air was so to 
 increase the breathing that, if the percentage added was not too 
 high, the CO 2 percentage in the alveolar air was kept nearly 
 constant. Of the delicacy of this reaction it is easy, from the fig- 
 ures, to form a fair estimate. With a moderate amount of hyperp- 
 noea, and provided that, as was actually the case, sufficient time 
 has elapsed to eliminate the influence of any temporary damming 
 back of CO 2 within the body, the discharge of CO 2 by the lungs 
 is about the same during hyperpnoea as during rest. Hence it 
 is possible to calculate how great a relative increase in the alve- 
 olar ventilation is brought about by a given increase in the alveolar 
 
22 RESPIRATION 
 
 CO 2 percentage. We found that about 0.23 per cent increase in 
 the alveolar CO 2 gives 100 per cent increase in the resting alveolar 
 ventilation. For instance with 4.16 per cent of CO 2 in the inspired 
 air, the alveolar CO 2 percentage would rise to about 6.06 per 
 cent, if it had been about 5.6 per cent when pure air was breathed. 
 As the difference between 4.16 and 6.06 is only a third of the 
 difference between o.o and 5.6, it follows that the alveolar ventila- 
 tion is thrice as great with the slightly raised alveolar CO 2 per- 
 centage. 
 
 A more precise measure of the effects of raising the alveolar 
 CO 2 percentage on the lung ventilation has more recently been 
 obtained by Campbell, Douglas, and Hobson, 7 who found that for 
 an increase of 10 liters per minute in the volume of air breathed 
 there was an increase of 0.28 per cent (or 2 mm. of mercury 
 pressure) in the alveolar CO 2 . An increase of 0.17 per cent was 
 sufficient to double the alveolar ventilation during complete rest 
 in a deck chair. 
 
 If an increase of 0.2 per cent in the alveolar CO 2 is sufficient 
 to double the alveolar ventilation it might be expected that a 
 decrease of 0.2 per cent would cause the breathing to cease. As 
 already mentioned, forced breathing or excessive artificial respira- 
 tion causes temporary cessation of natural breathing, or apnoea. 
 After forced breathing for about a minute the subsequent apnoea 
 commonly lasts for about 1^2 minutes in man. The alveolar CO 2 
 percentage is markedly diminished for a few seconds by even a 
 single extra deep breath of pure air, and correspondingly in- 
 creased by a breath of air containing more than 5 or 6 per cent of 
 CO 2 . It is easy to show, however, that the full effect of the dimin- 
 ished or increased percentage of CO 2 on the respiratory center is 
 not immediate. This is just what might be expected. The arterial 
 blood leaving the lungs at any moment is doubtless saturated with 
 CO 2 to a point corresponding with the existing percentage of CO 2 
 in the alveolar air; but when this blood reaches the tissues it 
 comes in contact with tissue and lymph saturated with CO 2 to the 
 normal extent, but possessing a considerable capacity for absorbing 
 more CO 2 . In consequence of this the tissues, including the res- 
 piratory center, take some time to get into equilibrium with the 
 new level of saturation with CO 2 in the arterial blood. Hence in 
 order to measure the real effect of any increase or diminution in 
 the alveolar CO 2 percentage, it is necessary to maintain this per- 
 centage constant for some time. When air containing an excess 
 
 'Campbell, Douglas, and Hobson, Journ. of Physiol., XLVIII, p. 303, 1914. 
 
RESPIRATION 23 
 
 of CO 2 is breathed, the alveolar CO 2 percentage naturally be- 
 comes constant after a few minutes ; but with forced breathing of 
 ordinary air it is not possible to maintain an alveolar CO 2 per- 
 centage which is below the normal by some required small amount. 
 
 To get over this difficulty we employed forced breathing with 
 air to which CO 2 had been added, and found that on successive 
 trials with increasing percentages of CO 2 in the inspired air 
 the duration of apnoea following forced breathing diminished 
 until, when there was more than about 4.7 per cent of CO 2 in the 
 inspired air, no apnoea at all was produced. It was thus evident 
 that a very small diminution in the alveolar CO 2 percentage 
 produces apnoea. The actual composition of the alveolar air at the 
 end of forced breathing in similar experiments was determined 
 later by Douglas and myself. 8 It was found that with more than 
 4.7 per cent of CO 2 in the inspired air no apnoea could be produced 
 by forced breathing, however hard, in a person whose normal 
 alveolar CO 2 percentage was about 5.6, and that apnoea was only 
 produced if the alveolar CO 2 was reduced by more than 0.2 per 
 cent below the normal. When, however, the CO 2 in the inspired 
 air was lower, so that the alveolar CO 2 percentage was reduced 
 by more than 0.2 per cent, apnoea was produced. 
 
 It is thus clear that the cause of the apnoea following forced 
 breathing is reduction in the CO 2 percentage in the alveolar air, 
 and that a reduction of as little as 0.2 per cent is sufficient to cause 
 apnoea. The astounding sensitiveness of the respiratory center to 
 CO 2 is thus clearly established in both an upward and a downward 
 direction. A mean increase or diminution of .01 per cent in the 
 alveolar CO 2 will evidently produce an increase or diminution of 
 5 per cent in the alveolar ventilation, or of about 400 cc. per minute 
 in the lung ventilation. 
 
 It may be useful to review briefly the sources of error in the 
 views current until recently with regard to the causes of the apnoea 
 produced by excessive ventilation of the lungs. One view was 
 that the excess of oxygen in the arterial blood causes the apnoea. 
 This theory had so little evidence to support it that it is very 
 surprising that it should have remained current so long. It is 
 true that during excessive artificial respiration the arterial blood 
 contains slightly more oxj^gen than usual; but there is a still 
 greater excess during the quiet normal breathing of pure oxygen, 
 which causes not the smallest sign of apnoea. Rosenthal 9 laid great 
 
 8 Campbell, Douglas, Haldane, and Hobson, Journ. of Physiol,, XLVI, p. 312. 
 - 
 Rosenthal in Hermann's Handbuch tier Physiologic, IV, 2, p. 266. 
 
24 RESPIRATION 
 
 stress on an experiment in which on slightly raising the pres- 
 sure in a spirometer from which an animal is breathing, the an- 
 imal stops breathing; and he attributed this to increase in the 
 partial pressure of the oxygen in the spirometer. The real cause 
 was quite evidently the distention of the animal's lungs by the 
 pressure, as in the experiments of Hering and Breuer. When a 
 man or animal has been rendered hyperpnoeic from want of oxy- 
 gen, and the hyperpnoea has reduced the normal percentage of 
 CO 2 in the alveolar air and blood, apnoea is produced by supply- 
 ing more oxygen; but this apnoea is of course dependent on de- 
 ficiency of CO 2 , and cannot, therefore, be cited in support of the 
 oxygen theory of ordinary apnoea. 
 
 The other erroneous theory that apnoea following forced 
 breathing is due to a summation of inhibitory vagus stimuli aris- 
 ing from distention of the lungs in the forced breathing 
 was based on two fallacies. The first was that intact vagi are 
 necessary for the production of apnoea by artificial respira- 
 tion. This is certainly not the case; for apnoea can be produced 
 quite promptly and easily after section of the vagi. It is necessary, 
 however, to make sure that the excessive artificial ventilation is 
 really effective in ventilating the lungs, since after section of the 
 vagi the natural breathing does not follow the rhythm of the 
 artificial respiration, and may thus partly annul the effects of 
 the latter. 
 
 The other fallacy connected with the vagus theory of ordinary 
 apnoea was that when air containing little or no oxygen is used 
 for artificial respiration an apnoea due to excessive aeration of 
 the blood is impossible. Advocates of the vagus theory wrongly 
 thought only of oxygen want in connection with aeration of the 
 blood. They thus attributed to vagus excitation any apnoea which 
 was produced in presence of defective oxygenation of the blood, 
 ignoring the fact that deficiency of CO 2 was present along with 
 defective oxygenation, and that this fact explained the observed 
 apnoea. Provided that the alveolar CO 2 percentage is sufficiently 
 reduced, apnoea can be produced readily in spite of great defi- 
 ciency of oxygen in the alveolar air. 
 
 The fact that apnoea is produced when forced breathing reduces 
 the alveolar CO 2 percentage by as little as 0.2 per cent (with the 
 alveolar oxygen percentage not abnormally low), and that if this 
 reduction is prevented no amount of excessive lung ventilation 
 will produce apnoea, affords, in conjunction with the other facts 
 already referred to, conclusive evidence that the apnoea following 
 
RESPIRATION 25 
 
 excessive lung ventilation is due to lowering of the alveolar CO 2 
 percentage, and not to either of the other causes to which the 
 apnoea has also been attributed. The vagus theory of the apnoea 
 caused by increased lung ventilation involved the very great 
 improbability that a special arrangement exists in the body for 
 bringing increased breathing to an end, regardless of whether a 
 continuance of the increased breathing is physiologically required 
 or not. It seemed almost incredible that such a theory could be 
 correct. 
 
 The ease with which apnoea due to reduction of CO 2 in the 
 alveolar air might be taken for an apnoea due to the after effect 
 of mere distention of the lungs is clearly shown by the stetho- 
 graphic tracings of human breathings reproduced in Figures 2 
 to 7. 10 Figure 2 shows apnoea as an after effect of inflation of the 
 lungs, while Figure 3 shows that when the inflation is made with 
 air containing 4.6 per cent of CO 2 , so as to prevent reduction of 
 the alveolar CO 2 percentage, no apnoea succeeds the period of 
 inflation. The apnoea appearing as an after effect in Figure 2 is 
 therefore due to reduction of the alveolar CO 2 in consequence of 
 the distention with pure air. 
 
 Figure 2. 
 
 Figure 3. 
 
 Effects of distention for 8 sees. Crosses show beginning 
 and end of distention. To read from left to right. In Fig. 2 
 pure air is used for distention ; in Fig. 3 air containing 
 4.62 per cent COz. 
 
 Figures 4, 5, and 6 illustrate the same point. In Figures 4 and 
 5 there is apnoea succeeding a short distention, but not immedi- 
 
 10 Christiansen and Haldane, Journ. of Physwl., XLVIII, p. 274, 1914. 
 
26 
 
 RESPIRATION 
 
 ately, since a few seconds were needed before the "apnoeic" blood 
 could affect the respiratory center. In Figure 6 the distention was 
 sufficiently prolonged for the "apnoeic" blood to affect the center 
 before the end of distention. The effect is therefore similar to 
 that in Figure 2. 
 
 Figure 4. 
 
 Figure 5. 
 
 Figure 6. 
 
 Effects of distention with pure air for increasing short 
 periods. Crosses show beginning and end of distention. To read 
 from left to right. Fig. 4 distention for i sec. ; Fig. 5 for 3 sees. ; 
 and Fig. 6 for 5 sees. 
 
 The regularity of ordinary breathing is constantly being inter- 
 fered with in various ways, as for instance during talking or 
 singing; and the breath can if necessary be held for about a minute 
 by voluntary effort. The readiness with which these interruptions 
 occur has given rise to the popular idea that the supply of air to 
 the lungs is to a large extent under voluntary control, and can be 
 increased or diminished by proper training. In reality the mean 
 
RESPIRATION 27 
 
 ventilation of the lungs is not affected by ordinary interruptions. 
 This is strikingly shown by experiments which we made on the 
 effects of voluntarily varying the frequency of breathing. 
 
 The frequency of breathing varies considerably among normal 
 individuals, or in the same individual at different times ; and it is 
 easy to vary the frequency while leaving the depth of breathing 
 to regulate itself in a natural manner. On making experiments of 
 this kind Priestley and I found the following percentages of CO 2 
 in the alveolar air : 
 
 ALVEOLAR CO 2 PERCENTAGE 
 
 
 RESPIRATIONS 
 
 End, of 
 
 End of 
 
 Mean 
 
 
 PER MINUTE 
 
 . inspiration 
 
 expiration 
 
 
 J. S. H. 
 
 9 
 
 5-59 
 
 5.87 
 
 573 
 
 
 19 
 
 5.56 
 
 570 
 
 5.63 
 
 J. S. H. 
 
 9 
 
 5-33 
 
 5-47. 
 
 540 
 
 
 20 
 
 5-44 
 
 5.60 
 
 5-52 
 
 J. G. P. 
 
 10.5 
 
 5-95 
 
 6.74 
 
 6-35 
 
 
 30 
 
 5.98 
 
 6.05 
 
 6.02 
 
 In a recent series, made on myself ten years later, 11 the fre- 
 quency was varied within much wider limits, with the following 
 results : 
 
 ALVEOLAR CO 2 PERCENTAGE 
 
 RESPIRATIONS 
 
 End of 
 
 End of 
 
 Mean 
 
 PER 
 
 MINUTE 
 
 inspiration 
 
 expiration 
 
 
 { 
 
 30 
 
 4 
 
 5-66 
 
 5-24 
 
 570 
 6.09 
 
 S-te 
 5-66 
 
 ( 
 
 24 
 
 5.48 
 
 5-49 
 
 5.48 
 
 ( 
 
 6 
 
 5-40 
 
 573 
 
 5.56 
 
 r 
 
 36 
 
 5.63 
 
 573 
 
 5-68 
 
 J 
 
 4 
 
 5.II 
 
 6.34 
 
 572 
 
 1 
 
 3 
 
 5-10 
 
 6.24 
 
 5.71 
 
 I 
 
 60 
 
 6.17 
 
 6.16 
 
 6.16 
 
 It will be seen that in spite of variations from 3 to 36 per minute 
 in the frequency of breathing the alveolar CO 2 percentage re- 
 
 "Haldane, Amer. Journ. of PhysioL, XXXVIII, p. 20, 1915. 
 
28 RESPIRATION 
 
 mained constant, since increased or diminished depth of breathing 
 compensated for diminished or increased frequency. The manner 
 in which this correspondence between depth and frequency is 
 brought about will be discussed in the next chapter. 
 
 During any considerable muscular exertion the discharge of 
 CO 2 from the lungs is enormously increased ; and in view of the 
 facts already described we should expect to find the breathing 
 similarly increased, with a rise in the alveolar CO 2 percentage 
 corresponding to the rise observed when the breathing is corre- 
 spondingly increased by breathing air containing an excess of 
 CO 2 . Priestley and I obtained the following mean results during 
 work on a somewhat primitive bicycle ergometer. 
 
 ALVEOLAR CO 2 
 
 PERCENTAGE 
 
 CALCULATED 
 
 End of 
 
 End of 
 
 Mean 
 
 RESPIRATORY 
 
 EXCHANGE 
 
 inspiration 
 
 expiration 
 
 
 J. S. H. Rest 
 
 I 
 
 5-54 
 
 570 
 
 5.62 
 
 Work 
 
 4-9 
 
 5-44 
 
 6.05 
 
 5-75 
 
 J. G. P. Rest 
 
 I 
 
 6.17 
 
 6-39 
 
 6.28 
 
 Work 
 
 3-8 
 
 6-45 
 
 6.98 
 
 6.72 
 
 Mean Rest 
 
 i 
 
 5.85 
 
 6.045 
 
 5.95 
 
 Work 
 
 4-3 
 
 5-945 
 
 6-545 
 
 6.235 
 
 In this series there was thus only a mean rise of 0.285 per cent 
 in the alveolar CO 2 , whereas we had expected to find a rise of 
 about 0.6. The correspondence was, however, in the right direc- 
 tion, and we endeavored, mistakenly as afterwards appeared, to 
 explain the lack of exact correspondence. 
 
 A more complete series was carried out later with much im- 
 proved apparatus by Douglas and myself, with Douglas as sub- 
 ject. 12 The accompanying table shows the data for volume of air 
 breathed, oxygen consumed, CO 2 given off, composition of ex- 
 pired air, and of alveolar air. In these experiments we used the 
 now well-known bag method of Douglas for determining the 
 respiratory exchange. 13 
 
 It will be seen from this table that with a CO 2 production in- 
 creased from 264 cc. per minute during rest standing to 1398 cc. 
 per minute during walking at 4 miles on grass the alveolar CO 2 
 percentage rose from 5.70 to 6.36, i.e., by 0.66 per cent. The vol- 
 ume of air breathed per minute was increased from 10.4 to 37.3, 
 
 u Douglas and Haldane, Journ. of Physiol., XLV, p. 235, 1912. 
 "Douglas, Journ. of Physwl., XLII, Proc. Physiol. Soc., p. xvii, 1911. 
 
RESPIRATION 
 
 29 
 
 tx O 
 COi per cent in alveolar ON tx 
 
 Tj" T^* ^* o co vO ^" O OO O 
 
 air m to 
 
 vO vo vo vo vO vo vo vo VO vo 
 
 ON Tf 
 
 COz per cent in expired air M . M . 
 
 10 ONOO 01 10 tx O 01 O ON 
 01 CO covo iovo iO txOO tx 
 
 CO co 
 
 Vol. of each breath in cc. 
 at 37 *- moist, and pre- tx 01 
 vailing barometric pres- *$ vo 
 sure. 
 
 vo M coiOO ^lO^-O 10 
 010l^-ioOOO l OOO-* 
 
 00 w 
 
 tXtxONOl TtOl 01 lOCOiO 
 
 Breaths per min. vo ^ 
 
 Ol ^rfvo ^OO txOOOO ON 
 
 Liters of air breathed per tx 
 
 COVO ONOO O fO 01 10 co ON 
 
 prevailing barometric pres- 2 
 
 VO 00 O * ON tx rj-vo M O 
 
 sure. 
 
 
 COz production in cc. per ONVO 
 min. at oC. and 760 mm. 
 
 M 01 tx 01 txOO M CO O vo 
 VO VO co 01 10 ON LOCO O CO 
 iovo txO\O fOOl txO co 
 
 M hi M M 01 01 
 
 Oz consumption in cc. per co 01 
 min. at oC. and 760 mm. * ^ 
 
 VOOO OVOOO O\o\o ^ "^ 
 
 VOtXONO HH LOTfO MlO 
 M M M M Ol 01 01 
 
 
 }H >-i >-i ^H >-( 
 
 $ S $ $ 
 
 C^ /'"N C3 /^*N rt x^v c^ ^"^ C^ ^* s ^ 
 
 
 r^ bjOr^ fcJDr^ tU) r^ tUO ^H tlJO 
 
 
 g 
 
 
 rt 
 
 ffj 
 
 fa 
 
 J, JLa 
 
 <8 ~ 
 
 1 a NCNXN 
 
 P^^2 
 
 ; W W coco^-^-^^uouo 
 
30 RESPIRATION 
 
 or by 26.9 liters. This corresponds very closely to the estimate by 
 Campbell, Douglas, and Hobson of an increase of 10 liters per 
 minute in the breathing for every .26 per cent of increased alveolar 
 CO 2 at normal barometric pressure. 
 
 When, however, the CO 2 production was increased still further, 
 the alveolar CO 2 percentage, instead of continuing to increase, 
 began to diminish, and was only 6.10 per cent with the maximum 
 CO 2 production (2386 cc.) and volume of air breathed (60.9 
 liters). Quite clearly, an additional factor or factors besides mere 
 increase in the alveolar CO 2 percentage was coming into play; 
 for with the higher rates of CO 2 production the lung ventilation 
 is not merely increasing in the same fixed proportion as before to 
 the increased production of CO 2 , but at a slightly higher rate. 
 What this additional factor is will be discussepl later; but mean- 
 while we may rest content with the broad fact that the increased 
 ventilation is almost in proportion to the increased production of 
 CO 2 , just as we should expect from the other facts already dis- 
 cussed with regard to the regulation of breathing. 
 
 It was shown by Paul Bert 14 that the physiological actions of 
 CO 2 , oxygen, and other gases present in the air breathed depend 
 on their partial pressure. It is only when the barometric pressure 
 is constant that their action depends on the percentage proportions 
 in which they are present in the air. The method of calculating 
 the partial pressure of the CO 2 in the alveolar air may be illus- 
 trated by an example. Let us suppose that the barometric pressure 
 is 760 mm., and that 5.6 per cent of CO 2 is found in the alveolar 
 air. In the first place allowance must be made for the aqueous 
 vapor present in the alveolar air, which in the living body must be 
 saturated with aqueous vapor at the body temperature. The pres- 
 sure exercised by this aqueous vapor is 47 mm. Hence the remain- 
 ing gas pressure is 760 47=713 mm. Of this pressure 5.6 per 
 cent is due to CO 2 (the results of the gas analysis being always in 
 terms of dry air) . Hence the pressure of CO 2 is 
 
 760 47 
 
 5-6 x = 39.9 mm., or 5.25 per cent of an atmosphere, 
 
 i oo 
 
 since 39.9 is 5.25 per cent of 760. 
 
 From Paul Bert's results it might be confidently predicted that 
 it is not the mere percentage but the pressure of CO 2 in the alve- 
 olar air which regulates the breathing, and our experiments left 
 no doubt on this point. On descending one of the deepest mines, 
 and ascending the highest hill in Great Britain, we found that the 
 
 "Paul Bert, La Pression barometrique, Paris, 1878. 
 
RESPIRATION 31 
 
 pressure of CO 2 in the alveolar air remained about constant, while 
 the percentage varied. A more conclusive experiment was made in 
 a large steel pressure chamber, employed at the Brompton Hospi- 
 tal, London, for the treatment of asthma. In this chamber the 
 only one then existing in England of the kind we compared 
 our alveolar air at normal atmospheric pressure, and at the highest 
 pressure which the chamber would stand. The mean results were 
 as follows : 
 
 Barometric pressure CO* per cent CO* pressure in 
 
 
 in mm. ffg. 
 
 in dry alveolar per cent of 
 
 
 
 atr 
 
 one atmosphere 
 
 J. G. P. 
 
 I26l 
 
 3-64 
 
 5.83 
 
 
 765 
 
 6.41 
 
 6.05 
 
 J. S. H. 
 
 1258 
 
 3.42 
 
 5.46 
 
 
 765 
 
 5.62 
 
 S-3I 
 
 Mean 
 
 1260 
 
 3.53 
 
 5.64 
 
 
 765 
 
 6.01 
 
 5-68 
 
 It is quite clear from these results that it is the pressure of CO 2 
 in the alveolar air, and not its mere percentage, which regulates 
 the breathing. It is also as evident from these experiments as 
 from those already mentioned in which the oxygen percentage 
 was varied, that the oxygen pressure in the alveolar air may be 
 increased very greatly without at the time affecting the regula- 
 tion of the CO 2 pressure. The actual alveolar oxygen pressure 
 was 13.0 per cent of an atmosphere in the observations at ordinary 
 pressure, and 26.8 per cent in those at the high pressure. 
 
 Still more striking results were obtained by Leonard Hill and 
 Greenwood, 15 and by Boycott 16 in steel chambers erected later 
 for the investigation of the effects of high atmospheric pressures. 
 Hill and Greenwood obtained the following results. 
 
 They considered at the time that their results showed that the 
 production of CO 2 remained unaltered during the experiments; 
 and it is evident that had the volume of air breathed and the mass 
 of CO 2 produced remained the same the results would have been 
 as they found. But the constancy of the partial pressure of CO 2 
 was certainly due, not to the cause which they suggested, but to 
 the fact that the breathing was regulated so as to keep the partial 
 pressure of CO 2 steady. 
 
 "Hill and Greenwood, Proc. Roy. Soc., 1906, B, LXXVII, p. 442, 1906. 
 18 Boycott and Haldane, Journ. of Physiol., XXXVII, p. 365, 1908. 
 
32 RESPIRATION 
 
 ATMOSPHERIC 
 PRESSURE IN ALVEOLAR CO 2 ALVEOLAR CO a PRESSURE 
 
 IN MM. HG. 
 
 PERCENTAGE 
 
 IN MM. HG. 
 
 
 Hill 
 
 Greenwood 
 
 Hill 
 
 Greenwood 
 
 760 
 
 4-7 
 
 5-3 
 
 33-5 
 
 37-8 
 
 4640 
 
 0-75 
 
 0.9 
 
 34-4 
 
 4L3 
 
 3860 
 
 0-95 
 
 1.0 
 
 36.2 
 
 38.1 
 
 3090 
 
 1.2 
 
 1-3 
 
 36.5 
 
 39.5 
 
 2310 
 
 1.8 
 
 1.8 
 
 40.7 
 
 40.7 
 
 1540 
 
 2.5 
 
 2.7 
 
 37-5 
 
 40.5 
 
 760 
 
 5-o 
 
 5-4 
 
 35-6 
 
 38.5 
 
 The results of Boycott and Haldane with Boycott as subject 
 are shown graphically in Figure 7. It will be seen that, provided 
 that the alveolar oxygen pressure was prevented from falling so 
 low as to cause want of oxygen, the alveolar CO 2 pressure re- 
 mained steady with variations of the barometric pressure from 
 300 to 2800 mm. and corresponding variations in the alveolar 
 CO 2 percentage from 15 to 1.5. 
 
 The daily variations of atmospheric pressure at any one place 
 are not sufficiently great to cause any considerable variations in 
 the alveolar CO 2 percentage, and there are other causes, discussed 
 below, which cause distinct variations in the alveolar CO 2 present. 
 Even, therefore, if we take into consideration the daily variations 
 of atmospheric pressure, the resting alveolar CO 2 pressure is not 
 quite constant at different times in the same individual, and varies 
 considerably in different individuals. 
 
 The differences in the alveolar CO 2 pressure in different indi- 
 viduals, and in different sexes and at different ages, were investi- 
 gated by Miss Fitz Gerald and myself. We obtained the following 
 results from a number of different persons, 17 living at Oxford. 
 
 ALVEOLAR CO a PRESSURES IN MM. OF MERCURY 
 
 
 Mean 
 
 Maximum 
 
 Minimum 
 
 Men 
 
 39-2 
 
 44-5 
 
 32.6 
 
 Women 
 
 36.3 
 
 41.0 
 
 30.4 
 
 Boys 
 
 37-2 
 
 42.1 
 
 30.6 
 
 Girls 
 
 35-2 
 
 40.1 
 
 31.2 
 
 17 Haldane and FitzGerald, Journ, of Physiol., XXXII, p. 491, 1905. 
 
RESPIRATION 
 
 33 
 
 The investigations of Priestley and myself brought out the 
 remarkable fact that the composition of the alveolar air is the 
 same no matter how deep the breath may be from the last portion 
 of which the sample is taken. According to descriptions commonly 
 
 3000 2600 
 
 200 
 
 2200 1800 1400 1000 
 air pressure mm Hg 
 
 Figure 7. 
 
 Effects of variation in barometric pressure on alveolar gas pres- 
 sures and percentage of CO 2 in A. E. B. The dotted lines show results 
 when oxygen was added to the air. 
 
 current of the anatomical relations of bronchioles to alveoli one 
 would have expected that the deeper parts of a breath, coming 
 from alveoli far from the bronchioles, would contain more CO 2 , 
 since these alveoli must get less fresh air than the alveoli near a 
 bronchiole. It was somewhat of a puzzle that this was not the case. 
 I was unaware of the anatomical investigations which had been 
 carried out ten years earlier by a distinguished American investi- 
 gator, W. S. Miller, who by using the laborious "reconstruction" 
 method had discovered the true anatomical arrangement. 18 Figure 
 8, modified from a colored plate in Miller's latest paper, shows 
 
 18 Miller, Journ. of Morphol., VIII, p. 165, 1893, and XXIV, p. 459, 1913. 
 
34 
 
 RESPIRATION 
 
 diagrammatically this arrangement. The finest ordinary bronchi- 
 oles divide up to form "respiratory bronchioles" with alveoli in 
 
 Figure 8. 
 
 Diagram showing arrangement of three lung lobules, with their 
 bronchiole, respiratory bronchioles, alveolar ducts, atria, and air- 
 sacs. (After colored plate by Miller, Journ. of Morphol. 24, p. 459, 
 1913.) 
 
 their walls, and the respiratory bronchioles branch into "alveolar 
 ducts" lined with ordinary alveoli, and each opening into from 
 
RESPIRATION 35 
 
 two to five distributing chambers which he named "atria," and 
 which are also lined with alveoli. From each atrium a number of 
 openings lead onwards into what he calls "air- sacs," which are 
 main cavities of which the walls are also constituted of alveoli or 
 air cells. By far the greater part of the alveoli belong to the air- 
 sac system, but a certain number belong to the respiratory bron- 
 chioles, alveolar ducts and atria; and the latter act partly as air 
 passages to the air sacs, and partly perform the same respiratory 
 functions as the air sacs themselves. 
 
 With this anatomical arrangement the whole of an air-sac sys- 
 tem is about equally well ventilated with fresh air, the only alveoli 
 which receive an undue supply of fresh air being those of the 
 respiratory bronchioles, alveolar ducts and atria. We can thus 
 understand why it is that the deeper parts of a very deep breath 
 have exactly the same composition as the middle parts. Evidently 
 however what Priestley and I called "alveolar air" is air-sac air. 
 
 The fact that the atria, etc., have partly a respiratory function, 
 and partly act as air passages to the air-sac system, enables us 
 also to clear up some otherwise unintelligible facts with regard 
 to the "dead space" in breathing. The dead space was first esti- 
 mated roughly by Loewy from the volume of a cast of the 
 respiratory passages, taken in a human lung after death. As this 
 method seemed uncertain, Priestley and I made determinations by 
 comparing the composition of a whole breath of expired air with 
 the composition of what we took to be the whole alveolar air. We 
 calculated the expired air as a mixture of this alveolar air with 
 fresh air occupying the dead space. In this way we found that 
 during rest the volume of the "effective dead space" is about 
 30 per cent of the volume of the average tidal air. For greater 
 certainty Douglas and I collected the whole of the expired air 
 over a certain period, and made the same calculation from the 
 average volume and composition of each breath, compared with 
 the composition of the alveolar air. 19 We then found that the 
 "effective dead space" is far greater during the hyperpnoea of 
 hard muscular work than during rest. As we were then still un- 
 aware of Miller's work we interpreted our observations as indicat- 
 ing that the bronchi or other respiratory passages become wider 
 during hyperpnoea, so as to enable air to enter the lungs more 
 easily. Any one who examines a section of lung must be struck at 
 once by the fact that the mucous membrane of the bronchi is 
 usually in folds, indicating that if the muscular coat relaxed the 
 
 19 Douglas and Haldane, Journ. of Physwl., XLV, p. 235, 1912. 
 
36 RESPIRATION 
 
 folds would open out and the lumen of the bronchi would greatly 
 increase. We thought it probable that such a relaxation occurs 
 during hyperpnoea, and that this explains the increase of the dead 
 space. 
 
 Using a method which Siebeck first introduced, Krogh and 
 Lindhard 20 then redetermined the dead space, and concluded that 
 it does not appreciably increase during hyperpnoea. Their method 
 was to take in a small measured breath of a hydrogen mixture: 
 they then made a deep expiration, which was measured, and 
 from the deeper part of which a sample of the alveolar air was 
 taken. From the percentage of hydrogen in the alveolar air, as 
 compared with the higher percentage in the whole expired air, the 
 volume of the dead space could be calculated on the assumption 
 that it was filled with the original hydrogen mixture. 
 
 The question was then independently reinvestigated about the 
 same time by Yandell Henderson, Chillingworth, and Whitney at 
 Yale, and myself at Oxford. We reached the same conclusion 
 namely that the apparent effective dead space is enormously in- 
 creased during hyperpnoea, as Douglas and I had found, but that 
 the increase is due simply to mechanical causes, and occurs 
 whether or not the respiratory center is excited by excess of CO 2 
 or other causes. Our papers appeared together in the American 
 Journal of Physiology. 2 ^ In their determinations Krogh and 
 Lindhard had inspired the same volume of the hydrogen mixture 
 whether there was air hunger at the time or not, and consequently 
 they got the same dead space; whereas our experiments were 
 made with the very deep breathing which is naturally associated 
 with air hunger, and consequently the dead space was increased. 
 
 Miller's investigations enable us to explain the great increase 
 of the "effective dead space" with deep inspirations. Considering 
 the relative thickness and stoutness of the bronchial walls it seems 
 very improbable that the bronchi, surrounded as they are by very 
 yielding lung tissue, could passively dilate appreciably owing to 
 a deeper inspiration, and this consideration led Douglas and 
 me to believe that they must dilate owing to a relaxation of 
 their muscular walls a theory negatived by the later experi- 
 ments. What dilate during deep breathing are evidently not the 
 bronchi but Miller's "alveolar ductules" and "atria," which serve 
 as air passages to the "air-sacs," and' which must expand along 
 
 20 Krogh and Lindhard, Journ. of Physwl., XLVIII, p. 30, 1913. 
 
 21 Yandell Henderson, Chillingworth, and Whitney; also Haldane, Amer. Journ. 
 of Physiol., XXXVIII, pp. i and 20, 1915. 
 
RESPIRATION 37 
 
 with the general expansion of the lungs. In addition, they are more 
 completely washed out by fresh air during inspiration. It also fol- 
 lows that the "effective or virtual dead space" is neither a definite 
 anatomical space nor a fixed dead space in any sense, but a value 
 dependent on several variable factors. These factors include the 
 rates at which CO 2 passes outwards and oxygen passes inwards 
 between the air and blood at different points in the alveolar sys- 
 tem. For this reason the "effective dead space" is different for 
 oxygen and CO 2 . The over-ventilation of the atria, etc., removes 
 from the blood circulating round them an extra proportion of 
 carbon dioxide, but cannot, for a reason which will be discussed 
 later, give to the blood any appreciable extra amount of oxygen. 
 During inspiration this extra proportion of CO 2 passes on to the 
 saccular alveoli, but not during expiration. The "respiratory 
 quotient," or ratio between the volume of carbon dioxide given 
 off and of oxygen absorbed, is thus abnormally high in the air 
 expired from the atria, etc., and as a consequence abnormally 
 low in the air sacs, so that the "effective dead space," as calculated 
 from deficiency of oxygen in the expired air, compared with that 
 in the "alveolar air," is greater than when the dead space is calcu- 
 lated from the relative CO 2 percentages. The respiratory quotient 
 for the "alveolar air" is also below the correct value as calculated 
 from the composition of the mixed expired air. 
 
 The following table, giving results on myself, shows the varia- 
 tions in the "effective dead space" with varying depth of breathing 
 as calculated both from CO 2 and from oxygen, and also the differ- 
 ences between the respiratory quotient as calculated from the 
 expired air and from the alveolar air. Using a slightly different 
 method, Henderson, Chillingworth, and Whitney got similar re- 
 sults. 
 
 It will be seen from this table how enormously the apparent 
 dead space varies with the depth of breathing and how much 
 greater the dead space calculated from the oxygen is than that 
 calculated from the CO 2 . A further point which comes out is that 
 with deep breathing the difference between the alveolar CO 2 
 percentages at the beginning and end of expiration is far less than 
 the difference between the oxygen percentages. This is mainly 
 because the extra CO 2 washed out of the alveolar ductules and 
 atria passes on into the saccular alveoli during inspiration. A 
 further point is that the true respiratory quotient is about a sixth 
 higher than the alveolar respiratory quotient. The fact that the 
 alveolar respiratory quotient is a good deal lower than the true 
 
38 
 
 RESPIRATION 
 
 M-- 
 
 o u 
 
 slS 
 
 S S S = 
 
 Sl g l 
 
 So 
 II 
 
 J3 U 
 I S 
 
 ON 
 
 ON N 
 
 ~ ON 
 
 -i vo ON 
 
 N M NH 
 
 t^ N !>>. VO I-M 00 
 
 N N M -l HH HH 
 
 ~ O 
 
 00 CO 
 
 vo N ON O 
 
 O O O O O O O 
 
 O N oo *- N u-v o 
 
 ^ O m oo t^. vo O 
 
 rn 4- 4- en ro ro rn 
 
 "^" l^ L^ -^- ON N uo 
 
 c OOQN t^-ONri-CO 
 
 O 00 ur> Ti- 00 U^ 
 
 Sc ^>ONO-I ON^-OONH 
 
 .: 
 
 r\ vo ^-*^ vo t^ *-^s ~^^ ON 
 
 u^i \j-\ u~v is-\ i^\ u~) vr> w<i 
 
 "^-ONQ ^ONONVOvo 
 O O l^ t^ l~^ -^> ^ o 
 
 c ^-O 
 
 WG vovo u^u^u-,o^u^^o 
 
 H..O ThN tnunvo u-NThoo 
 
 G 
 
 oou^u->-.i^ooOO 
 + l^ O i^oooooo N 
 S*S' oooo ONOOOOOO t-^oo 
 
 odddddod 
 
 OONTJ-NOONr^ 
 Q vovovof^t-^voi^oo* 
 
 ONVO ^t-"-" ONOO N ON 
 
 N lr\ N CO LO i-r\ N OO 
 
 4-4-4- 
 
 , o H J 
 < ac S 55 1^ 2 w 
 
 SSgisl! 
 
 Q W ^ < W 
 
 *<J- OO rr> m O en O 
 
 00 m CO vr\ rj- NN 
 
 ON TJ- TJ- vo vo vo rt- 
 
 N N - 
 
 vo oo t^ t~^ TJ- o 
 
 _ M N VO 
 
RESPIRATION . 39 
 
 quotient had been noticed by us before this in the work of the 
 Pike's Peak Expedition (to be referred to later), but had not been 
 explained. It is quite evident from the table that the composition 
 of the deep alveolar air cannot be even approximately calculated 
 from that of the expired air by assuming the existence of a con- 
 stant dead space. The latter assumption has caused great confusion 
 in recent years, particularly in the work of the Copenhagen School. 
 It was shown by Yandell Henderson and his coadjutors that 
 when air passes along an air passage the axial stream is much 
 faster than the peripheral stream, and that as a consequence of 
 this the air in the dead space is not pushed out bodily in front of 
 the alveolar air during expiration. Some of the tracheal and 
 bronchial air is at first left behind, and before pure alveolar air 
 issues at the nose or mouth the air passages have to be washed out 
 by three or four times their volume of alveolar air. This is illus- 
 
 ;r : '^ 
 
 Figure 9. 
 
 (a) Shows a "spike" of smoke moving through a glass tube, (b) 
 Shows the condition when the current is suddenly stopped and mixing 
 instantaneously occurs, (c) Shows clear air drawn in. 
 
 Figure 10. 
 
 Shows how a column of smoke crosses a bulb with little mixing 
 or sweeping out of the air within it. 
 
 trated by Figures 9 and 10, taken from their paper, and drawn 
 from experiments made with smoke. Both they and I found also 
 that a pause before expiration diminishes the volume of the ap- 
 parent dead space. This is easily understood, as the air in the 
 atria, etc., will during the pause come nearer in composition to 
 that of the saccular alveoli. With care in avoiding a pause I found 
 
40 RESPIRATION 
 
 that during rest with normal breathing it was necessary to expire 
 about 800 cc. of air before a reliable alveolar sample could be 
 obtained at the end of inspiration. If the breathing was deep and 
 slow much more air had to be expired. At the end of a normal 
 expiration, however, the air issuing from the mouth is practically 
 alveolar in composition. 
 
 The conclusion reached by Priestley, Douglas, and myself that 
 increased production of CO 2 , and consequent rise in the alveolar 
 CO 2 percentage, determines increased breathing during muscular 
 work was afterwards questioned by Krogh and Lindhard, 22 on 
 the ground that our determinations of the alveolar CO 2 percentage 
 were fallacious, and that the real alveolar CO 2 percentage during 
 muscular work is not only lower than we found, but also con- 
 siderably lower than during rest. Their argument is mainly 
 based on the assumptions, which have already been shown to be 
 wrong, that the "effective dead space" is not largely increased 
 during deep breathing, and that reliable samples of alveolar air 
 can be obtained at the end of a deep inspiration, without more 
 than a very shallow expiration to clear the extra dead space. This 
 part of their argument falls to the ground. They point out, how- 
 ever, what is a real source of slight error namely that a delay of 
 fully half a second occurs during the taking of an alveolar sample, 
 and that during this interval the alveolar CO 2 percentage must 
 rise appreciably. It was shown above that the difference in CO 2 
 percentage between samples of alveolar air taken at the beginning 
 and end of expiration during work corresponding to an increase 
 of 4.3 times in the CO 2 production was about 0.6 per cent. As an 
 expiration took nearly 2 seconds, there would be a rise of 0.15 per 
 cent in half a second, corresponding to the delay in taking the 
 alveolar sample. During rest, according to a similar calculation, 
 there would be a rise of 0.05 per cent. The net error in comparing 
 rest with work would thus be only about o. I per cent, a difference 
 too small to affect the conclusions materially. Owing to their 
 defective methods of estimating and directly determining the 
 alveolar CO 2 percentage at the beginning of expiration Krogh 
 and Lindhard enormously overestimated the error due to a delay 
 of half a second in obtaining a sample. The fact remains, however, 
 that when the work was pushed in the case of Douglas, and even 
 without pushing the work in my own case, the rise in alveolar CO 2 
 percentage was less than corresponded to the increase in breathing. 
 This significant fact will be discussed later. 
 
 22 Krogh and Lindhard, Journ. of Physiol., XLVIII, p. 30, 1913. 
 
RESPIRATION 41 
 
 It will be shown in Chapter IX that during rest under normal 
 conditions the gas pressures in the alveolar air and blood passing 
 through the alveoli come into exact equilibrium. Now it has just 
 been shown that in a very appreciable part of the lung alveoli 
 (those in the respiratory bronchioles, alveolar ducts, and atria) 
 the CO 2 pressure is lower, and the oxygen pressure higher, than 
 in the air-sac alveoli. We might therefore be led to infer that in 
 the mixed arterial blood the CO 2 pressure will be lower, and the 
 oxygen pressure higher, than in the blood from the air-sac alveoli, 
 and that in consequence of this the mixed arterial blood will have 
 -a lower CO 2 pressure than that of the deep alveolar air. Further 
 consideration shows, however, that this will not be the case. The 
 walls of the alveoli of respiratory bronchioles, etc., are in contact 
 on the one side with the air of air-passages, but on the other with 
 air in the air-sac alveoli. Hence the extra proportion of CO., 
 extracted from the blood in the air-passage alveoli is practically 
 taken from the air-sac alveoli, and this is why the apparent respira- 
 tory quotient of the air-sac alveoli is lower than the true respira- 
 tory quotient. We should be counting the lowering twice if we as- 
 sumed that in consequence of the extra discharge of CO 2 in the re- 
 spiratory bronchioles, etc., the CO 2 pressure of the arterial blood is 
 lower than corresponds to that of the air-sac alveoli. The same 
 argument applies also as regards the oxygen pressure of the air- 
 sac air, although under normal conditions hardly any extra oxy- 
 gen can pass into a given volume of blood in its passage through 
 the alveoli of respiratory bronchioles, etc. Hence the gas pressures 
 of the air-sac alveoli represent truly the mean gas pressures to 
 which the arterial blood is saturated in the various alveoli. This 
 is why the gas pressures of the deep alveolar air as determined by 
 the method which Priestley and I introduced are of so much 
 importance. 
 
 Krogh and Lindhard 23 still maintain that the mean gas pres- 
 sures to which the blood is equilibrated in passing through the 
 lungs is given, not by the composition of the deep alveolar air, 
 but by that of the alveolar air as calculated from a fixed, or almost 
 fixed, dead space. This involves the conclusion that during deep 
 breathing, including the deep breathing of muscular exertion, the 
 arterial CO 2 pressure is far lower than is shown by the direct 
 method of Priestley and myself. As, however, there is no cor- 
 responding apnoea, the whole theory of regulation of breathing in 
 accordance with the CO 2 pressure of the arterial blood must be 
 
 23 Krogh and Lindhard, Journ. of Pkysiol., LI, p. 59, 1917. 
 
42 RESPIRATION 
 
 abandoned if Krogh and Lindhard are correct. Their reasoning 
 is quite logical, but their premises are unsound. They have failed 
 to take into consideration trje anatomical relations of the air- 
 passage alveoli to the air-sac alveoli. 
 
 The fact that the mixed air from all the air sacs of the lungs is 
 the same in composition however much of this air is expelled in 
 taking a sample led us to assume almost unconsciously that the 
 composition of the air in practically all the air sacs is the same. 
 Nevertheless all that the experiments prove is that the average 
 composition of the air expelled from the air sacs is the same, while 
 in individual air sacs the composition may vary widely. 
 
 It is evident that in any particular air-sac system the mean 
 composition of the contained air will depend on the ratio between 
 the supply of fresh air and the flow of blood. If the supply of 
 fresh air is unusually small in relation to the supply of venous 
 blood there will be a lower percentage of oxygen and higher 
 percentage of carbon dioxide in the air of the air sac, and vice 
 versa. It seems probable that by some means at present unknown 
 to us a fair adjustment is maintained normally between air supply 
 and blood supply. For instance, the muscular walls of bronchioles 
 may be concerned in adjusting the air supply, or the arterioles 
 or capillaries may contract or dilate so as to adjust the blood 
 supply. In any case what seems to matter is the degree of arteri- 
 alization, not of the blood from individual air sacs, but of the 
 mixed arterial blood; and if the composition of the mixed air-sac 
 air served as a reliable index of the arterialization of the mixed 
 arterial blood we might dismiss as a matter of only academic 
 interest the question whether the air in individual air sacs varies 
 in composition. 
 
 It will be shown below that there can be little doubt that under 
 normal conditions the air in different air sacs varies appreciably 
 in composition, and that under abnormal conditions the variation 
 may be considerable. It will also be shown that the latter fact 
 is one of great importance in pathology and therapeutics. 
 
 Meanwhile it is clear from the experiments described in the 
 present chapter that under normal conditions, excluding heavy 
 work, the breathing in man is on an average regulated by the al- 
 veolar CO 2 pressure ; and a very slight increase or diminution in 
 the alveolar CO 2 pressure suffices to cause a very great increase or 
 diminution in the breathing. This conclusion has thrown a flood 
 of clear light on the physiology of breathing. 
 
CHAPTER III 
 The Nervous Control of Breathing. 
 
 IT is now necessary to discuss more closely the influence of nervous 
 control on breathing. The rhythmic activity of the respiratory 
 center is for short periods of time very completely under volun- 
 tary control a fact evidently connected with the very delicate 
 use of the lungs in phonation, as well as in other voluntary acts not 
 directly connected with "chemical" respiratory functions. Excita- 
 tion of various afferent nerves may also excite or inhibit inspira- 
 tion or expiration. Most of the effects thus produced appear to be 
 protective in various ways, or preparatory to some particular 
 effort, and they only disturb the main regulation of breathing 
 occasionally, just as voluntary interference does. In view of the 
 facts with regard to the control of breathing by chemical stimuli, 
 we might thus be led to the conclusion that the respiratory center, 
 when not interfered with by voluntary or other occasional nervous 
 disturbances, acts simply by producing rhythmic inspiratory and 
 expiratory discharges, determined in extent and frequency by 
 nothing but chemical stimuli dependent on the blood supply. 
 
 This simple conception is entirely inadequate, in view, more 
 particularly, of the facts discovered originally by Hering and 
 Breuer, and already referred to. These facts, apart from the 
 results of section of the vagi, can be observed very fully in man, 
 without the complications introduced by anaesthetics, and were 
 so studied in 1916 by Mavrogordato and myself. 1 We employed 
 a very simple arrangement which enabled us to breathe through 
 a wide-bored tap, and observe by a water manometer the pres- 
 sure between the mouth and the tap when the latter was closed, 
 the nostrils being closed by a clip. If the tap was closed at 
 the end of natural or forced inspiration or expiration, or in 
 any other phase of respiration, the phenomena could be studied. 
 By connecting the far end of the tap with a reservoir containing 
 pure air or air containing any required percentage of CO 2 , we 
 could observe the influence of hyperpnoea due to CO 2 , and by 
 suitable volume recorders connected with the far ends of the 
 reservoir and gauge the breathing and pressure could be recorded. 
 
 If expiration is interrupted by turning the tap, and all voluntary 
 
 1 Journ. of Physiol., L; Proc. Physiolog. Soc., p. xli, 1916. 
 
44 RESPIRATION 
 
 effort is suspended, the previous rhythm of the respiratory center 
 is interrupted by a prolonged expiratory phase, as indicated by 
 the gauge. The expiratory pressure is at first slight and constant, 
 but afterwards rises gradually and at an increasing rate, until, 
 if expiration is still prevented, there is at length an inspiratory 
 effort, as shown in Figure n. Similarly, if the breathing is ob- 
 
 RESPIRATION 
 
 /NTPAPULMONARY PRESSURE 
 
 RESPIRATION 
 
 I NTRAPULMONARY PRESSURE 
 
 Figure i i . 
 
 Effects of interrupting natural breathing. A. Respiration inter- 
 rupted during inspiration near end. B. Respiration interrupted 
 during expiration near end. Respirations inspiration up, expira- 
 tion down. Intrapulmonary pressure positive pressure down, neg- 
 ative pressure up. 
 
 structed during expiration there is a prolonged and increasing 
 inspiratory effort (Figure n). The initial inspiratory pressure 
 is somewhat greater than the initial expiratory pressure, and this 
 is in accordance with the opinion generally held that while ordi- 
 nary quiet inspiration is always an active process the correspond- 
 ing expiration is mainly passive. 
 
 With interruption at the end of an extra deep inflation or de- 
 flation of the lungs the phenomena are still more marked. If 
 apnoea has previously been caused by forced breathing, the initial 
 expiratory or inspiratory pressures are still produced as before, 
 but a long interval elapses before they begin to increase, and the 
 duration of the expiratory or inspiratory phase is much prolonged. 
 
RESPIRATION 45 
 
 If, on the other hand, the inflation or deflation was made during 
 the hyperpnoea caused by breathing air containing an excess of 
 CO 2 the expiratory or inspiratory pressures mount up at onre. 
 The mounting up of the initial pressure is thus dependent on the 
 accumulating chemical stimulus to the respiratory center. If the 
 breathing is interrupted, not just after, but before the completion 
 of inspiration or expiration, the inspiratory phase is continued if 
 inspiration has been interrupted, and the expiratory if expiration 
 has been interrupted, as shown in Figure II. 
 
 If, instead of interrupting the breathing by means of a tap or 
 other obstacle which cannot be overcome, the only interruption 
 is by a limited adverse pressure capable of being overcome by 
 the breathing, the apparent "apnoea" is terminated by an expira- 
 tion if the pressure is positive, or an inspiration if the pressure is 
 negative. This simply means that with a positive pressure the 
 expiration occurs at the moment when the expiratory effort has 
 increased sufficiently to overcome the adverse positive pressure, 
 and similarly with a negative pressure. This is illustrated by 
 Figures 12 and 13, which reproduce stethographic tracings ob- 
 tained in man. 2 The subject at first breathed quietly through the 
 limb of a wide-bore three-way tap open to the air. At the end of 
 an inspiration the tap was suddenly turned so that the mouth of 
 the subject was connected with the air of a bag under a pressure 
 of about 3 inches of water. The consequence of this was that the 
 
 Figure 12. 
 
 Figure 13. 
 
 Effects of prolonged distention of the lungs. To be read from left to right. Time 
 marker = seconds. Distention continued between the two crosses. In Fig. 12 pure 
 air was used for distention; in Fig. 13 air containing 7.3 per cent of COa and 
 8.2 per cent of oxygen. 
 
 lungs were suddenly distended with a large volume of air. It 
 will be seen that after about half a minute the apparent pause in 
 
 'Christiansen and Haldane, Journ. of Physiol., XLVIII, p. 272, 1914. 
 
46 RESPIRATION 
 
 the breathing was interrupted by an expiration, repeated after- 
 wards at gradually diminishing intervals. The diminution in 
 these intervals was evidently due to the fact that CO 2 was ac- 
 cumulating in the lungs; and this interpretation is confirmed by 
 Figure 13. 
 
 Figure 14 shows a corresponding effect with a negative pressure 
 applied, so as partially to deflate the lungs. In this case the ap- 
 parent pause was much shorter, as CO 2 began to accumulate very 
 rapidly, owing to the facts that not only had no fresh air been 
 introduced, but the volume of air in the lungs was diminished. 
 
 Figure 14. 
 
 Effects of partial deflation. Crosses show beginning and end of 
 deflation. To read from left to right. Time-marker = i second. 
 
 The supposed apnoeic pause produced by distention or inflation 
 of the lungs is simply a prolonged inspiratory or expiratory effort. 
 This effect is produced regardless of the chemical stimulus to the 
 center. Thus Lorrain Smith and I showed that it is even produced 
 when the lungs are distended with air containing 20 per cent of 
 CO 2 , though the prolongation is much curtailed in such a case. 3 
 
 It is thus clear that the continuance of an inspiratory or ex- 
 piratory discharge of the respiratory center depends on the extent 
 to which actual inspiration or expiration accompanies the dis- 
 charge. If the movements of inspiration or expiration are not 
 accomplished the ordinary respiratory rhythm is replaced by a 
 prolonged and increasingly powerful inspiratory or expiratory 
 discharge, tending to overcome the obstruction. The respiratory 
 center does not act independently of the lung movements, but 
 inspiratory or expiratory discharge of the center goes hand in' 
 hand with actual inspiration or expiration, as if the center were 
 one piece with the lungs. The term "vagus apnoea" is evidently 
 an entire misnomer, as prolonged inspiratory or expiratory effort 
 cannot be called apnoea. The tracings which apparently demon- 
 strate the existence of apnoea are only one-sided, and therefore 
 misleading, records. 
 
 Hering and Breuer found, as already mentioned in Chapter I, 
 that after section of both vagi the association of discharge of the 
 
 1 Haldane and Lorrain Smith, Journ. of Pathology, I, p. 168, 1892. 
 
RESPIRATION 47 
 
 center with the respiratory movements is annulled, so that infla- 
 tion or deflation of the lungs has no immediate influence on the 
 respiratory rhythm. Hence the afferent impulses through which 
 the discharges of the center are coordinated with the movements 
 of the lungs are conveyed by the vagi. After section, or better (so 
 as to avoid excitatory effects produced by actual section), freezing 
 of the vagi, the breathing, as has been known since early last 
 century, becomes deeper and less frequent, the inspirations in 
 particular taking on a dragging character which, until the work 
 of Schafer, referred to below, was entirely attributed to the ab- 
 sence of the normal inhibitory effect conveyed through the vagi 
 on distention of the lungs to a certain point. Nevertheless the 
 respirations continue to be rhythmic, and to respond in their depth 
 to the stimulus dependent on varying percentages of CO 2 in the 
 alveolar air. It was shown by Scott 4 however, that the control of 
 the alveolar CO 2 percentage when excess of CO 2 is present in the 
 air breathed becomes much less perfect, as the frequency of the 
 breathing cannot increase. 
 
 The analogy between the Hering-Breuer stimuli transmitted 
 through the vagi and what Sherrington has named the "proprio- 
 ceptive" stimuli participating in reflex or voluntary movements 
 of the limbs is evident; though the rhythmic discharges of the 
 respiratory center are dependent on stimuli, not from the surface 
 of the body, but from the blood acting on the center. 
 
 When, in addition to section of the vagi, the respiratory center 
 is also severed from its connections above the medulla oblongata, 
 the rhythmic discharges of the center become still less frequent, 
 and may be inadequate to prevent death from asphyxia. The 
 influence on the center of afferent stimuli from the respiratory 
 muscles has not yet been demonstrated directly; but the fact, 
 observed by Boothby and Shamoff, 5 that an animal in which the 
 pulmonary branches of the vagi have been severed without injury 
 to the recurrent laryngeal nerve recovers after a sufficient time 
 a normal control over respiration seems to point to the existence 
 of such stimuli. The same conclusion has been still more clearly 
 reached in a quite recent paper by Schafer, 6 who shows that the 
 slowed breathing after section of the vagi is largely due to ob- 
 struction caused by laryngeal paralysis. 
 
 We must now endeavor to correlate the facts relating to the 
 
 4 Scott, Journ. of Physwl., XXXVII, p. 301, 1908. 
 
 5 Boothby and Shamoff, Amer. Journ. of Physwl. , XXXVII, p. 418, 1913. 
 9 Schafer, Quart, Journ. of Exper. Physiol., XII, p. 231, 1919. 
 
48 RESPIRATION 
 
 Hering-Breuer phenomena with those relating to the governing 
 of the lung ventilation by the charge of CO 2 in the alveolar air 
 and arterial blood. It seems very clear that the immediate cause of 
 the arrest of inspiration during ordinary breathing is the disten- 
 tion of the lungs to a certain point, and a consequent inhibitory 
 stimulus transmitted up the vagi. The experiments of Head, 7 in 
 which the movements of a slip of the diaphragm, the most promi- 
 nent inspiratory muscle, were recorded, show that this inhibition 
 produced an instant relaxation of the diaphragm. If the vagi 
 have been frozen the relaxation is greatly delayed, and even 
 after the delay is at first very imperfect. The inhibition of inspira- 
 tion initiates an expiratory phase, which continues until, in its 
 turn, it also is cut short by deflation to a certain point, at which 
 the vagi transmit an influence which inhibits expiration and 
 initiates the inspiratory phase. It appears from Head's experi- 
 ments that if the vagi are frozen after the inspiratory or expira- 
 tory phase has been initiated, this phase still continues. If with 
 vagi intact the breathing is partially obstructed, inspiration or 
 expiration is continued till either act is complete. The influence 
 transmitted through the vagi initiates inspiration or expiration, 
 therefore; and the center persists in the inspiratory or expiratory 
 phase till the vagus gives the signal which terminates the phase 
 and initiates the complementary phase. The center behaves as if 
 it always remembered the last signal; and the analogy between 
 any act dependent on memory and the duration of the inspiratory 
 or expiratory phases of breathing is evident. We are equally 
 reminded of the "refractory period" in the phases of cardiac and 
 other muscular activity. 
 
 Where the "chemical" regulation of the respiratory center 
 exerts its preponderating influence is in determining the extent 
 to which inflation or deflation of the lungs must extend in order 
 that the Hering-Breuer stimuli should be effective, and also the 
 vigor and consequently the rapidity of the inspiratory and expira- 
 tory movements. Thus an increased CO 2 stimulus causes increased 
 depth of breathing, since a greater inflation or deflation of the 
 lungs is required before the stimulus of inflation or deflation 
 becomes effective. At the same time the movements of the chest 
 wall become more rapid, so that the frequency of breathing is 
 not diminished in consequence of the greater distances traveled 
 by the chest walls. The net result is thus ordinarily increase 
 in depth without diminution in frequency. But if the frequency 
 
 T Head, Journ. of Physiol., X, pp. i and 279, 1889. 
 
RESPIRATION 49 
 
 is diminished in consequence of voluntary or involuntary inter- 
 ference, the depth is correspondingly increased owing to a very 
 slightly increased CO 2 stimulus. This is the explanation of why 
 the mean alveolar CO 2 percentage remains so steady with varying 
 frequency of breathing. It is only, as a rule, when there is very 
 considerable increase in the breathing that there is any material 
 increase in the frequency; and during health the frequency is 
 hardly affected by moderate muscular exertions or moderate 
 stimulation by CO 2 in other ways. The frequency of breathing 
 is thus no measure of the amount of air breathed; but undue 
 frequency of breathing, as will be shown later, is a very important 
 abnormal sympton. 
 
 The response of the breathing to abnormal resistance has re- 
 cently been investigated by Davies, Priestley, and myself. 8 For 
 recording the depth and frequency of breathing we used the 
 recording "concertina" described in Chapter VII (Figure 43). 
 For a resistance to breathing we sometimes used partly closed 
 taps, the effects of which could be thrown in suddenly by closing 
 alternative inspiratory and expiratory air passages. In place of 
 the taps we also sometimes employed cotton wool resistances, as 
 with a cotton wool resistance the driving pressure varies directly 
 as the air flow, while with a tap the pressure varies as the square 
 of the air flow. The pressure was measured with a water ma- 
 nometer connected with the tubing between the mouth and the 
 resistance. 
 
 Figure 15. 
 
 Effects of resistance. In this and subsequent figures inspiration = upstroke. 
 Time marker = 10 seconds. To read from left to right. 
 
 It was found that when a resistance is thrown in the immediate 
 effect is a great slowing of the breathing. After the next breath 
 the respirations become deeper and less slow, and after several 
 breaths the breathing settles down to a rhythm in which the 
 respirations are deeper and correspondingly less frequent. With 
 a considerable resistance the frequency is often reduced to a fourth 
 of the normal rate, while the depth is almost correspondingly in- 
 
 8 Davies, Haldane, and Priestley, Journ. of PAysiol., LIII, p. 60, 1919. 
 
50 RESPIRATION 
 
 creased (Figure 15). The explanation of this is obvious from the 
 foregoing account of the physiology of the Hering-Breuer reflex. 
 When a resistance is thrown in deflation or inflation of the lungs 
 is slowed, but continues till the point is reached at which the 
 phase of respiration is reversed by the reflex. Meanwhile, how- 
 ever, CO 2 has begun to accumulate, so that the next respiration 
 is not only more vigorous but deeper ; and the final result is deeper 
 and less frequent respiration. 
 
 When there is no resistance to breathing the compensation of 
 diminished frequency by increased depth is almost perfect, as 
 shown by the experiments already quoted of Priestley and my- 
 self; but when the slowing is due to resistance the compensation 
 is less perfect, since the extra work performed by the respiratory 
 muscles implies a more powerful stimulus of CO 2 to the respira- 
 tory center. Accordingly the alveolar CO 2 percentage rises quite 
 considerably with resistance to breathing. The following table 
 shows the rises observed by Davies, Priestley, and myself with 
 varying resistances. 
 
 Just as, in the absence of resistance a very slight increase in the 
 alveolar CO 2 percentage, and consequent slight increase in the 
 chemical stimulus to the respiratory center, increases the depth 
 of breathing, so a slight diminution in alveolar CO 2 percentage 
 diminishes the depth. It was recently discovered independently by 
 Yandell Henderson in America and by Liljestrand, Wollin, and 
 Nilsson in Sweden that if apnoea is first produced and artificial 
 respiration then carried out by Schafer's or one of the other usual 
 methods the quantity of air which enters the chest at each artificial 
 inspiration is only about a third or less of what enters during 
 artificial respiration when the subject has simply suspended vol- 
 untarily his own breathing. With voluntary suspension of the 
 natural breathing, moreover, the volume of air which enters at 
 each artificial inspiration varies (roughly speaking) inversely 
 as the frequency of the artificial breathing, so that it is impossible 
 to produce a condition of true apnoea by increasing the frequency 
 of the artificial breathing. If, finally, the air artificially inspired 
 contains an excess of CO 2 , the volume introduced by the artificial 
 respiration increases just as it would with natural breathing. It is, 
 in fact, just as if the subject were himself breathing naturally all 
 the time, in spite of the undoubted fact that he has suspended his 
 natural breathing. 
 
 These phenomena are completely intelligible on the theory that 
 the limits within which inflation or deflation of the lungs inhibits 
 
RESPIRATION 51 
 
 SUBJECT ALVEOLAR CO 2 RESISTANCE IN 
 
 
 PERCENTAGE CM. OF H 2 O 
 
 
 During Inspira- 
 
 
 N ormal resistance tory Expiratory 
 
 Remarks 
 
 J. S. H. 5-40 5-34 4^ 1^2 
 
 Slight cotton-wool resistance. 
 
 
 Breathing slowed 
 
 J. G. P. 5-60 5-80 4^ ^A 
 
 Slight cotton-wool resistance. 
 
 
 Breathing slowed 
 
 H. W. D. 5.97 6.24 13 5 
 
 Heavier cotton-wool resistance. 
 
 
 Breathing slowed 
 
 5-99 5-93(?) 8 4 
 
 Lighter cotton-wool resistance. 
 
 
 Breathing slowed 
 
 6.61 
 
 Lighter cotton-wool resistance. 
 
 
 Breathing slowed 
 
 6.21 
 
 Lighter cotton-wool resistance. 
 
 
 Breathing slowed 
 
 6.26 
 
 Lighter cotton-wool resistance. 
 
 
 Breathing slowed 
 
 7.02 25 14 
 
 Heavy cotton-wool resistance. 
 
 
 Breathing slowed 
 
 J. S. H. 5.4 6.40 
 
 Heavy cotton-wool resistance. 
 
 
 Breathing quickened 
 
 6.60 
 
 Heavy cotton-wool resistance. 
 
 
 Breathing about 66 
 
 6.76 ? ? 
 
 Resistance lessened by partly 
 
 
 opening taps. Respirations 
 
 
 about 30 
 
 J. G. P. 5-37 5.76 ? ? 
 
 Tap resistance lessened by 
 
 
 partly opening taps. Respi- 
 
 
 rations about 4 
 
 J. S. H. 5.33 6.50 ? ? 
 
 Tap resistance lessened by 
 
 
 partly opening taps. Respi- 
 
 
 rations about 24 
 
 6.80 ? ? 
 
 Tap resistance increased. Res- 
 
 
 pirations about 40 
 
52 RESPIRATION 
 
 inspiration or expiration depend on the alveolar CO 2 percentage. 
 In apnoea a very slight amount of inflation or deflation is suffi- 
 cient to cause inhibition of inspiration or expiration. In conse- 
 quence of this the respiratory movements are nearly jammed in 
 a mean position during apnoea unless considerable force is ex- 
 erted, which is not the case with ordinary methods of artificial 
 respiration. With a normal stimulation of the respiratory center 
 by CO 2 and a normal respiratory frequency, the limits of inflation 
 or deflation at which the Hering-Breuer inhibition occurs are a 
 good deal wider, and with a diminished respiratory frequency, 
 or an increased percentage of CO 2 in the air inspired, the limits 
 are much wider still. Thus the respiratory center tends indirectly 
 to govern artificial respiration unless the latter is of a specially 
 vigorous kind. 
 
 That the center responds, even during apnoea, with tonic con- 
 traction of the diaphragm to deflation of the lungs, and with re- 
 laxation to inflation, was clearly shown by Head's experiments; 
 and the inspiratory or expiratory pressures produced by the 
 diaphragm and other respiratory muscles can easily be demon- 
 strated in man. The continued control of respiratory movements 
 during apnoea or voluntary suspension of the breathing, or during 
 voluntary variations in the frequency of breathing, is thus readily 
 intelligible. In voluntary forced breathing or in forcible artificial 
 respiration, this control is broken down. It must not, however, be 
 assumed that because the ordinary gentle methods of human 
 artificial respiration have such a small effect during ordinary 
 apnoea, the effect will be equally small where the suspension of 
 breathing has been caused by asphyxiation or the action of an 
 anaesthetic or other poison. In these cases the excitability of the 
 respiratory center to the Hering-Breuer stimuli is possibly as 
 much depressed as its excitability to CO 2 , in which case the 
 artificial respiration will not be insufficient. 
 
 The normal rate and depth of breathing in any individual is evi- 
 dently an expression of the normal balance between chemical and 
 nervous stimuli. The normal is fairly constant because the balance 
 is a stable one. It may, however, be greatly altered under abnormal 
 conditions, and it can easily be interfered with voluntarily. 
 
 It is evident from the foregoing discussion that we cannot 
 separate the nervous from the "chemical" control of breathing, 
 since each determines the other at every point. From too exclusive 
 a consideration of the nervous side of the control it has been sup- 
 posed, on the one hand, that the center is essentially automatic in 
 
RESPIRATION 53 
 
 its action, or that its alternate inspiratory and expiratory dis- 
 charges are, under normal resting conditions, determined simply 
 by alternating stimuli transmitted through the vagus nerves. On 
 the other hand a too exclusive consideration of the chemical side 
 leads to the erroneous impression that the discharges of the center 
 are, apart from occasional voluntary or other interferences, de- 
 termined in strength and duration solely by chemical stimuli. If, 
 finally, we attempt to determine, one by one, the "factors" in the 
 regulation of breathing, the sum of the supposed factors turns 
 out to be illusory, since no one of them is a constant quantity. The 
 evaluation of each factor depends on its varying relation to the 
 others. 
 
 The "respiratory center" is a small area situated in the medulla 
 oblongata. It has been found that when this area is destroyed, all 
 rhythmical respiratory movements cease, and that so long as this 
 area is intact and in connection with any efferent nerves supply- 
 ing respiratory muscles, discharges of the center through these 
 nerves continue, as shown by the rhythmical contractions of the 
 muscles, although all the other nervous connections upwards and 
 downwards have been severed. 
 
 It is also now clear that the activity of the center depends upon 
 the composition of the blood circulating through it, and not on 
 chemical stimuli acting elsewhere. If the circulation to the medulla 
 is interrupted by closure of all the four arteries supplying it, so 
 that its blood has time to become venous, violent hyperpnoea re- 
 sults, as Kiissmaul and Tenner showed about the middle of last 
 century; and the crossed circulation experiments of Fredericq, 
 already referred to, prove that either apnoea or hyperpnoea is 
 produced, according as the blood supplied to the central nervous 
 system is more aerated or less aerated in the lungs. 
 
 It has been suspected that although the stimuli dependent on 
 the composition of the blood act directly within the brain, nervous 
 end-organs situated elsewhere are also sensitive to these stimuli, 
 so that the corresponding nerves convey impulses which play an 
 important part in the regulation of breathing. It was, for instance, 
 believed by Traube that chemical stimuli are conveyed directly 
 from the lungs by the vagus nerve, and others have supposed 
 that stimuli to increased breathing are conveyed by direct nervous 
 paths from the muscles. This hypothesis was investigated with 
 great care by Geppert and Zuntz, 9 who severed all the nervous 
 connections between actively working muscles and the medulla, 
 
 'Geppert and Zuntz, Pfliiger's Archiv, XLII, pp. 195, 209, 1888. 
 
54 RESPIRATION 
 
 and found that the respiratory response to increased muscular 
 work was the same as before, but was entirely absent if the circu- 
 lation from the working muscles was interrupted. Similarly they 
 found that severance of the nervous connection between the lungs 
 and the center did not affect the response. Lorrain Smith and I 
 found, similarly, that when air containing about 20 per cent of 
 CO 2 was supplied to a rabbit there was no difference in the time 
 required for the onset of hyperpnoea after the vagi were cut. 
 
 No definite anatomical group of nerve cells has been defined at 
 the position occupied by the respiratory center; and the exact 
 meaning which ought to be attached to the expression "respira- 
 tory center" is still doubtful. It seems pretty clear, however, that 
 the center is at about the position which is sensitive to the chem- 
 ical respiratory stimuli. To judge from analogy the sensitive 
 elements are probably not the bodies of nerve cells, but end- 
 organs or arborizations. The central paths for the innervation of 
 inspiratory and expiratory movements must also be different, but 
 in what sense the center itself is double is still obscure. Its excita- 
 tion by chemical stimuli depends more upon the character of the 
 blood supplied to it than on substances generated by its own local 
 metabolism. Thus the temporary diminution of blood supply in 
 fainting does not produce the same prompt effect on the center as 
 changes in the arterial blood owing to imperfect aeration in the 
 lungs. In this respect the center is very well suited to fulfill the 
 function of taking a part in controlling the quality of the general 
 arterial blood supply of the body. The amount of arterial blood 
 supplied is controlled in other ways. 
 
 Like other parts of the central nervous system, the respiratory 
 center can easily be fatigued; and, as will be explained later, 
 fatigue of the respiratory center is of great importance in practical 
 medicine. Fatigue of respiration was recently studied by Davies, 
 Priestley, and myself, and its phenomena described in the paper 
 already referred to. The fatigue was produced by breathing 
 against a resistance, the breathing being also increased at the 
 same time, if necessar)', by muscular exertion. The resistance was 
 produced by cotton wool in the manner already described. 
 
 So long as the center is functioning normally it responds to the 
 resistance, in the manner indicated above, by producing a constant 
 slow and deep type of breathing. When, however, the resistance 
 is excessive and continued for some time, the breathing becomes 
 progressively shallower and more frequent. At the same time the 
 alveolar ventilation becomes less and less effective, until at last 
 
RESPIRATION 
 
 55 
 
 asphyxial symptoms begin to develop. Figure 16 is a tracing 
 which shows this change. Figure 1 7 shows a similar change pro- 
 duced, not by resistance alone, but by the combined effects of 
 resistance and the increased breathing due to muscular work. 
 
 IllJimHilliiiiih. 
 
 Ron 
 
 iRo// 
 
 Figure 16. 
 Effects of heavy resistance. To read from left to right. 
 
 fR.dwr.Jf 
 
 Figure 17. 
 Effects of resistance and gentle work. To read from left to right. 
 
 It will be shown later that even a slight deficiency in the oxy- 
 genation of the arterial blood favors greatly the development of 
 fatigue symptoms in the respiratory center. But addition of oxy- 
 gen to the air does not prevent the development of fatigue due 
 simply to great extra work thrown on the respiratory center. When 
 the breathing is quite free, and the oxygenation of the blood 
 normal, fatigue does not at all readily show itself, and greatly 
 increased breathing goes on in a normal manner over long periods. 
 During muscular exertion, however, as will be shown later, the 
 oxygenation of the blood may become impaired, in which case 
 fatigue of the breathing may easily show itself, so that the subject 
 becomes in a literal sense "short of breath," since each breath 
 is short. 
 
56 RESPIRATION 
 
 During the war cases were very common of what, according as 
 one nervous symptom or another was most prominent, was desig- 
 nated as "chronic gas poisoning," "soldier's heart," "disordered 
 action of the heart," "neurasthenia," etc. In these cases "shortness 
 of breath" on exertion was a common and prominent symptom. 
 Their breathing was investigated by Meakins, Priestley, and my- 
 self 10 and we found a marked deviation from normality in its reg- 
 ulation. In many of these persons the frequency of the breathing 
 was very abnormally increased during rest, and in nearly all there 
 was on exertion a quite abnormal increase of frequency, with a 
 corresponding reduction of the normal increase of depth. The 
 symptoms were thus the same as those of fatigue of the respira- 
 tory center, and on extra exertion these patients were liable to 
 lose consciousness with asphyxial symptoms, just as in ordinary 
 overfatigue of the center. Another prominent symptom was that 
 the patients were unable to hold a deep breath for anything like 
 a normal period, even if they were given oxygen to help. Many of 
 them were also subject, particularly at night, to attacks of rapid 
 shallow breathing with a sense of impending suffocation. 
 
 The condition of the breathing in these patients was evidently 
 such as would be produced by an abnormal increase in the readi- 
 ness with which the Hering-Breuer reflex is elicited, and we 
 therefore described the respiratory condition as one of "reflex re- 
 striction" in the depth of breathing. At the time we were not 
 aware of the symptoms of fatigue of the respiratory center. In the 
 condition of fatigue the shallow and rapid breathing is just what 
 would result from an increase in the strength of the Hering- 
 Breuer reflex, and a similar apparent exaggeration of this reflex 
 is present, as already seen in connection with the results of artificial 
 respiration, in the condition of apnoea. In view, therefore, of all 
 the facts relating to the respiratory movements in fatigue, apnoea, 
 and neurasthenia, it seems probable that the apparent increased 
 strength in the Hering-Breuer reflex is due to a diminution in 
 the persistency of the individual inspiratory and expiratory dis- 
 charges from the center, rather than to any real increase in the 
 inhibitory Hering-Breuer discharges up the vagus nerves. It is 
 thus only the weakness of the center that enables the Hering- 
 Breuer reflex to gain the upper hand. 
 
 If we apply the same general conception to the other exag- 
 
 10 Haldane, Meakins, and Priestley, Reports of the Chemical Warfare Meciical 
 Committee, No. 5, Reflex Restrictions of Breathing, 1918, and No. n, Chronic 
 Cases of Gas Poisoning, 1918; also Journ. of Physiol., LII, p. 433. 
 
RESPIRATION 57 
 
 gerated reflexes and general failure of nervous coordination in 
 "neurasthenia," fatigue, and "shock," we seem to render these 
 conditions more intelligible. Thus the great general nervous ir- 
 ritability, exaggeration of circulatory reflexes, tendency to sweat- 
 ing, and occasional instability of temperature, as observed in 
 "neurasthenia/' are probably analogous to the exaggerated re- 
 flex restriction in the depth of breathing and the inability to hold 
 a breath. All these symptoms seem to be due to what Hughlings 
 Jackson called "release of control." 
 
 In the causation of military neurasthenia the nervous over- 
 strain of war, and the shocks to the nervous system in connection 
 with various incidents of warfare and gross bodily injuries had 
 evidently played a prominent part; but it was equally evident 
 that infections of different sorts were also in part responsible for 
 the condition, the nervous system being apparently weakened by 
 toxic influences. In the same way ordinary fatigue of the respira- 
 tory center or other parts of the nervous system may be due not 
 merely to extra work, but also partly to want of oxygen (as will 
 be shown later) , or to other chemical influences. Neurasthenia may 
 thus be regarded as only a more lasting and persistent form of 
 ordinary fatigue or exhaustion. It will be shown later that a very 
 important secondary effect of the shallow breathing characteris- 
 tic of neurasthenia or fatigue of the respiratory center is im- 
 perfect oxygenation of the blood. 
 
 The readiness with which a given resistance to breathing pro- 
 duces signs of fatigue of the breathing varies greatly in different 
 individuals. In some persons a comparatively small resistance 
 suffices to produce shallow breathing and rapid exhaustion of the 
 respiratory center, though in other quite healthy persons a very 
 considerable resistance is needed. Men with symptoms of neu- 
 rasthenia are, as might be expected, particularly sensitive to re- 
 sistance. This matter is, of course, important in connection with 
 the design of respirators, etc. A respirator causing any consider- 
 able resistance may easily disable a man for muscular exertion. 
 
 The threshold alveolar CO 2 pressure at which the respiratory 
 center begins to be excited may be altered by various abnormal 
 conditions which will be discussed further in later chapters. The 
 threshold may be lowered by want of oxygen or by the presence 
 in the blood of an abnormally low proportion of available alkali, 
 or by certain drugs, including, as Yandell Henderson 11 has pointed 
 
 11 Yandell Henderson and Scarbrough, Amer. Journ. of Physiol., XXVI, p. 
 279, 1910. 
 
58 RESPIRATION 
 
 out, ether in low concentrations, or by massive afferent nervous 
 stimuli. On the other hand the threshold is raised by such anaes- 
 thetics as chloroform, morphia, or chloral; and under their influ- 
 ence the alveolar CO 2 pressure is raised 12 and the breathing is 
 commonly so much diminished that the arterial blood becomes 
 markedly blue. These facts are of great importance in connection 
 with the use of anaesthetics. Henderson showed also that morphia 
 affects the chemical more than the afferent threshold of the res- 
 piratory center. Rise of body temperature has a marked effect in 
 lowering the threshold. 13 
 
 u Collingwood and Buswell, Journ. of Physiol. (Proc. Physiol. Soc.), XXXV, 
 p. xxxiv, and XXXVI, p. xxi, 1907. 
 
 13 Haldane, Journ. of Hygiene, V, p. 503, 1905; see also Haggard, Journ. of Biol. 
 Chem. XLIV, p. 131, 1920. 
 
CHAPTER IV 
 The Blood as a Carrier of Oxygen. 
 
 THE evidence has already been referred to that nearly all the 
 available oxygen in the blood is present in the form of a chemical 
 compound with the haemoglobin of the red corpuscles, and that 
 this compound has the remarkable property of dissociating with 
 fall in the partial pressure of oxygen, at the same time changing 
 its color from bright scarlet to a dark purple. It dissociates com- 
 pletely when the oxygen pressure is reduced to zero, and the 
 readiness with which the dissociation occurs is dependent on 
 temperature and other conditions which will be discussed below. 
 It is contained in the corpuscles to the extent of about 30 per cent 
 of their weight, and on liberation from them it can be crystallized 
 out with comparative ease by the help of cold and of substances 
 which diminish its solubility. There is considerable variation in 
 the form of the crystals obtained from the blood of different 
 animals. 
 
 To what extent, and in what directions, the elementary composi- 
 tion of haemoglobin varies is not yet definitely known; but the 
 haemoglobin of birds has been found to contain phosphorus, while 
 none is present in the haemoglobin of mammals. Iron is always 
 present. A given amount of blood, whether or not the corpuscles 
 have been dissolved and the haemoglobin liberated and diluted, 
 takes up, on saturation with air at room temperature, a perfectly 
 fixed and definite amount of oxygen in chemical combination. No 
 further measurable quantity is taken up, except in simple physical 
 solution, on saturation with oxygen. An exactly equal volume of 
 carbon monoxide or nitric oxide is taken up in combination in 
 presence of either of these gases. There is no shadow of doubt that 
 the combination is a chemical one, though some extraordinary 
 attempts, based on ignorance of well-ascertained facts, have re- 
 cently been made to explain the combinations of oxygen and CO 2 
 in blood as due to adsorption. 
 
 Haemoglobin not only enters into dissociable chemical combina- 
 tions with oxygen, carbon monoxide and nitric oxide, but also in 
 presence of various oxidizing agents, such as ferricyanides or 
 chlorates, or very weak acids, etc., when oxygen is also present, 
 passes into a modification called by Hoppe Seyler methaemoglobin. 
 
60 RESPIRATION 
 
 This substance, which crystallizes in a similar form to oxyhaemo- 
 globin but has a dull brown color in acid solution and a brownish 
 red color in alkaline solution, was found by Hiifner to take up in 
 its formation from haemoglobin just as much oxygen as oxy- 
 haemoglobin ; but the oxygen is not given off in a vacuum. On the 
 other hand it yields its oxygen much more rapidly to a reducing 
 agent than oxyhaemoglobin or free oxygen does, and is thus an 
 oxidizing agent of some activity. Thus if a drop of ammonium 
 sulphide solution is mixed with a solution of methaemoglobin in 
 the absence of free oxygen the methaemoglobin is instantly re- 
 duced to haemoglobin, as shown by the change of color and spec- 
 trum. But if free oxygen is present the color and spectrum of 
 oxyhaemoglobin appear, since the ammonium sulphide combines 
 far more slowly with free oxygen, or with the combined oxygen 
 of oxyhaemoglobin, so that the haemoglobin formed instantly 
 from the methaemoglobin is able to combine with the free oxygen 
 and remain for a long time as oxyhaemoglobin. 
 
 While investigating the action of poisons which form met- 
 haemoglobin in the living body I noticed that when ferricyanide 
 and certain other reagents act on oxyhaemoglobin to form methae- 
 moglobin fine bubbles are liberated, and on further investigation 
 the liberated gas was found to be oxygen. 1 I then measured ac- 
 curately the liberated oxygen, and found that the volume of oxy- 
 gen liberated by ferricyanide from blood agrees exactly with the 
 volume liberated by the mercurial pump from combination in the 
 blood. Ferricyanide also liberates carbon monoxide from its com- 
 bination with haemoglobin, and the volume liberated corresponds 
 with the volume of oxygen liberated by a corresponding quantity 
 of oxyhaemoglobin. The following figures were obtained. 
 
 Combined gas in cc. liberated, from 
 the haemoglobin of 100 cc. of blood 
 and measured dry at QC and 760 mm. 
 
 By blood pump alone from blood saturated with air 18.18 
 
 By ferricyanide from blood saturated with air 18.20 
 
 By ferricyanide from blood saturated with CO 18.07 
 
 From their behavior, it appears that oxyhaemoglobin and CO- 
 haemoglobin are molecular compounds in which the molecules of 
 
 1 Haldane, Journ. of Physiol., XXII, p. 298, 1898. 
 
RESPIRATION 6l 
 
 gas are directly combined as such with the molecules of haemo- 
 globin, just as molecules of water are combined with molecules of 
 a salt or other substance to form hydrate molecules. In methaemu- 
 globin, on the other hand, the atoms in the molecules of oxygen 
 which enter into combination are separately combined just as in 
 ordinary chemical compounds containing oxygen. When the 
 oxidation of haemoglobin to methaemoglobin occurs the new 
 molecule formed loses its capacity for forming the molecular 
 compounds oxyhaemoglobin and carboxyhaemoglobin. In conse- 
 quence of this the molecular oxygen and carbon monoxide are 
 liberated from oxyhaemoglobin or carboxyhaemoglobin by the 
 action of ferricyanide, and can be measured with the greatest ac- 
 curacy in the gaseous form by a simple method which I described 
 in 1900 (see Appendix). 2 
 
 The ferricyanide method afforded a ready means of measuring 
 directly the gas combined in the molecular form with haemo- 
 globin, and for this purpose replaced the complicated procedure 
 and involved calculations required when the mercurial pump 
 was used. One of the first discoveries made with the new method 
 was that the coloring power of haemoglobin or any one of its 
 molecular compounds with gases varies exactly as its capacity for 
 combining with gas. Hence the "oxygen capacity" of the haemo- 
 globin in blood in other words its power of fulfilling its physio- 
 logical function of carrying oxygen can be measured easily by 
 means of a reliable colorimetric method. 3 The following table 
 (p. 62) shows the results we obtained on this point. 
 
 That oxygen capacity and depth of color run parallel also in 
 various anaemias and other pathological conditions was shown 
 by Morawitz ; 4 and Douglas 5 showed that even during the rapid 
 regeneration of haemoglobin after loss of blood this also holds. 
 
 At the time when the ferricyanide method was introduced there 
 existed several well-known forms of "haemoglobinometer." Of 
 these the apparatus of the late Sir William Gowers was by far the 
 most convenient. In his method 20 cubic millimeters of blood, 
 obtained from a prick of the skin, are introduced into a small 
 graduated tube and diluted with water until the depth of color 
 is the same as that of a standard solution of picrocarmine in an- 
 other similar tube. The depth of color of the picrocarmine solution 
 
 a Haldane, Journ. of Phystol., XXV, p. 295, 1900. 
 
 * Haldane and Lorrain Smith, Journ. of Physiol., XXV, p. 331, 1900. 
 
 4 Morawitz and Rohmer, Deutsch. Arch. f. kUn. Med., XCIII, p. 223, 1908. 
 
 5 Douglas, Journ. of Physiol., XXXIX, p. 453, 1910. 
 
62 
 
 RESPIRATION 
 
 is that of normal human blood diluted to i/iooth; and the 
 graduated tube gives the strength of color of the blood under 
 examination in terms of this normal standard. One defect of the 
 method was that the picrocarmine standard is not permanent, 
 and another that the color of the picrocarmine solution is not the 
 
 OXYGEN CAPACITY 
 PER 100 CC. 
 
 Ferricyanide Colortmetric 
 method method 
 
 PERCENTAGE 
 DIFFERENCE IN 
 RESULT BY 
 COLORIMETRIC 
 METHOD 
 
 Ox blood 18.51 
 
 18.42 
 
 0.5 
 
 15.05 
 
 15 
 
 33 
 
 + 1.9 
 
 20.29 
 
 19.85 
 
 2.2 
 
 15.04 
 
 15 
 
 17 
 
 +0.9 
 
 Horse blood 18.37 
 
 18.39 
 
 +0.1 
 
 Ox blood 19.75 
 19.90 
 
 $ 20.00 
 
 +0.9 
 
 18.94 
 
 18.94 
 
 +0.0 
 
 Rabbit's blood 14.62 
 
 \ 
 
 
 
 
 
 
 O. I 
 
 14-55. 
 
 ) 
 
 
 
 Sheep's blood 17.44 
 
 17- 
 
 30 
 
 0.8 
 
 17.44 
 
 
 
 
 Ox blood 21.50 
 
 21.42 
 
 0.4 
 
 21.55 
 
 
 
 
 16.16 
 
 16.06 
 
 0.6 
 
 Human blood 2 1 .08 
 
 21.27 
 
 +0.9 
 
 Mean 18.07 
 
 18 
 
 .06 
 
 0.055 
 
 same spectrally as that of the blood solution. As a consequence of 
 this both the depth and the quality of the tints of the two solutions 
 are differently affected by variations in the quality of the light at 
 the time of using the instrument. Thus if the tints agree at one time 
 of day they may be different at another; and in ordinary artificial 
 light the results given are totally different from the results by 
 daylight. Moreover, in consequence of individual differences in 
 vision, a color match for one person is not the same as that for 
 another person, even in the same light. To remedy these defects I 
 substituted for the picrocarmine a one per cent solution of blood 
 of the average oxygen capacity of the blood of adult men (18.5 
 
RESPIRATION 63 
 
 cc. of oxygen per 100 cc. of blood) , and introduced other improve- 
 ments. 6 
 
 In the presence of free oxygen haemoglobin is a very unstable 
 substance, and soon decomposes, owing to the action of bacteria, 
 etc. ; but in the absence of oxygen the color of haemoglobin is per- 
 fectly stable, and this is also the case for carboxyhaemoglobin. The 
 standard solution was therefore saturated with carbon monoxide 
 in the absence of oxygen, and in this form is permanent. The blood 
 under examination is also saturated with carbon monoxide by 
 contact with coal gas or a little carbon monoxide. The two solu- 
 tions are thus spectrally the same. With these improvements the 
 Gowers haemoglobinometer became an extremely accurate instru- 
 ment for ascertaining the oxygen capacity of blood, and the ac- 
 curacy of any particular instrument could be controlled at once 
 by the ferricyanide method. Certain ever-recurring criticisms of 
 the instrument are almost entirely based on want of acquaintance 
 with the physiological principles of colorimetric methods, or of 
 the chemical facts on which the method is based. A detailed de- 
 scription of the method will be found in the Appendix. 
 
 The percentage oxygen capacity (or haemoglobin percentage) 
 in the blood varies quite appreciably from hour to hour and day 
 to day, according as the total volume of the blood varies from ad- 
 dition or withdrawal of liquid. There are also variations associ- 
 ated with age and sex ; and pathological variations may be very 
 marked and significant. As regards age and sex I found the follow- 
 ing average relative figures for the percentage oxygen capacity 
 of the blood. 
 
 Men 18.5 
 
 Women 16.5 
 
 Children 16.1 
 
 It has been known for long that when an oxyhaemoglobin 
 solution is overheated or treated with various simple reagents the 
 oxyhaemoglobin decomposes into a coagulated protein and a 
 deeply-colored brown substance soluble in alcohol and certain 
 other solvents, and known as haematin. The haematin contains 8.7 
 per cent of iron, and the coagulated protein is free from iron. To 
 the haematin the formula C3 4 H 34 N 6 O 5 Fe has been assigned. By 
 the action of reducing agents the haematin loses oxygen and 
 changes to a purple color, with a corresponding change of spec- 
 trum, described by Stokes at the same time as he described the 
 
 8 Haldane, Journ. of Physiol., XXVI, p. 497. 
 
64 RESPIRATION 
 
 spectra of oxyhaemoglobin and haemoglobin. To this reduced 
 haematin Hoppe Seyler gave the very suitable name haemo- 
 chromogen, as he believed it to be the parent substance of the color 
 of haemoglobin and its varied derivatives. Thus we can regard 
 haemoglobin as a compound of haemochromogen with a protein, 
 also haematin as an oxygen compound of haemochromogen, while 
 compounds of haemochromogen with carbon monoxide and nitric 
 oxide are also known. 
 
 This conception is confirmed by the fact that the oxygen ca- 
 pacity of haemoglobin varies as its coloring power, and by another 
 still more recently established fact. Till a few years ago it still 
 seemed very doubtful whether there is a fixed and definite rela- 
 tionship between the iron in haemoglobin and its oxygen capacity; 
 and Bohr 7 thought that he had obtained evidence of the existence 
 of marked variations in the relation between iron and oxygen 
 capacity ; and that this relation differs in arterial and venous 
 blood. The doubts on this subject turned round the reliability of 
 the methods of determining iron. But in 1912 Peters, using a new 
 and very reliable method of iron determination, found that there 
 is a fixed and simple relationship between the oxygen capacity and 
 iron, one molecule of combined oxygen corresponding to one 
 atom of iron. 8 
 
 Still other considerations point in the same direction. When we 
 examine the colors and spectra of the various direct derivatives 
 of haemoglobin and haemochromogen a striking general cor- 
 respondence emerges. Methaemoglobin and haematin have very 
 similar colors and spectra, which differ in a more or less similar 
 manner in acid or alkaline solutions, and give a similar red color 
 and corresponding spectrum on addition of hydrocyanic acid. 
 With carbon monoxide haemochromogen gives the same color 
 and spectrum and takes up the same volume of carbon monoxide 
 as haemoglobin. With the nitric oxide compounds there appears 
 also to be a correspondence. Thus I found that the red color of 
 raw salted meat is due to the presence of NO-haemoglobin, 
 formed by the action on haemoglobin of the reduction product of 
 the niter which is mixed with the salt; and the color is still red 
 after the meat is cooked and the NO-haemoglobin broken up to 
 yield a haemochromogen compound on heating. NO-haemoglobin 
 is also found post mortem in poisoning by nitrites. Between 
 haemoglobin and haemochromogen there is also more or less of 
 
 7 Bohr, Nagel's Handbuch der Physiologie, I, p. 95, 1905. 
 
 8 Peters, Journ. of Physiol., XLIV, p. 131, 1912. 
 
RESPIRATION 65 
 
 correspondence; but oxyhaemochromogen, the molecular oxygen 
 compound of haemochromogen, is missing, and it seems that 
 haematin is so readily formed by haemochromogen in the pres- 
 ence of oxygen that oxyhaemochromogen cannot exist. Figure 18 
 shows the positions of the absorption bands in the spectra of NO- 
 haemoglobin and NO-haemochromogen. 
 
 C D E b F 
 
 Figure 18. 
 
 i. Nitric oxide haemoglobin. 2. Oxyhaemoglobin. 3. Carbonic 
 oxide. haemoglobin. 4. Nitric oxide haemochromogen. 5. Obtained 
 by action of nitrous acid on haematin. 
 
 If haemochromogen has been formed from haemoglobin by 
 the action of acids or caustic alkali and heat, a substance possess- 
 ing the spectrum and properties of natural haemoglobin is gradu- 
 ally re-formed on neutralization. 9 As proteins are greatly altered 
 in properties by heating with alkali it would seem from this ob- 
 servation that there may be a number of different haemoglobins, 
 in which, though the haemochromogen part of the molecule is 
 the same in all, the protein part varies. As will be shown later, 
 there is evidence that not only in different species, but also in 
 different individuals of the same species, the protein part of the 
 haemoglobin molecule varies, thus producing slight variations in 
 the properties of the haemoglobin as a carrier of gases, although 
 there is no variation in the oxygen capacity per unit weight of 
 iron present. The haemochromogen part of the molecule seems, 
 on the other hand, to be constant in all the different sorts of haemo- 
 globin, and this brings about the identity of the relations between 
 oxygen capacity, coloring power, and percentage of iron in all 
 the different varieties of haemoglobin, although as regards other 
 properties haemoglobins from different sources vary distinctly. 
 
 9 See Menzies, Journ. of Phystol., XVII, p. 415, 1895, and XLIX, p. 452, 1915- 
 
66 RESPIRATION 
 
 The original ferricyanide method for determining the oxygen 
 capacity of haemoglobin was very accurate, but required a good 
 deal of blood, and was also slow on account of the time needed 
 for exact equalization of temperature and gas pressures. Mr. 
 Barcroft was then beginning his important series of investiga- 
 tions on the metabolism of the salivary glands and other organs. 
 As he required a blood-gas method suitable for very small volumes 
 of blood he asked me whether the ferricyanide method could be 
 adapted for the purpose, and I designed an apparatus which we 
 jointly tested and described, and which turned out so successfully 
 that, in one form or another, it has now almost displaced the 
 mercurial blood pump. 10 In this apparatus the oxygen combined 
 in the haemoglobin of the very small quantity of blood required 
 is liberated by ferricyanide, and the CO 2 by acid. The amount of 
 gas liberated in either case is determined, not from the increase 
 in volume which its liberation causes, but from the increase of 
 pressure when the total volume of gas is kept rigorously constant. 
 I adopted this principle as the result of much previous experience 
 in the measurement of small differences in gas volumes. Certain 
 causes of difficulty are eliminated by the pressure method, and by 
 the adoption, as in the original ferricyanide method, of a control 
 arrangement by which the effects of changes in temperature and 
 barometric pressure during the experiment are eliminated. Vari- 
 ous improvements in the technique of collecting and sampling 
 blood drawn directly from blood vessels were also introduced by 
 Mr. Barcroft. 
 
 This apparatus has been modified in various ways by different 
 investigators, and some of the modifications are improvements. 
 Others, however, seem to me to be the reverse. In the Appendix 
 there is a description of a new and much more exact method in 
 which the volumes of oxygen and CO 2 are measured directly. 
 
 Besides the oxygen chemically combined with haemoglobin, 
 the blood contains a certain small amount of oxygen in simple 
 solution. In accordance with Henry's law of solution of gases in 
 liquids this amount varies with the partial pressure of oxygen in 
 the atmosphere with which the blood is saturated, which in the 
 case of arterial blood in the living body is (with certain reserva- 
 tions discussed in Chapters VII and VIII), the alveolar air. The 
 amount of oxygen in free solution can be measured directly when 
 the haemoglobin is by one means or another put out of action in 
 respect to its power of entering into molecular combination with 
 
 10 Barcroft and Haldane, Journ. of Physiol., XXVIII, p. 232, 1902. 
 
RESPIRATION 67 
 
 oxygen. Bohr found that at body temperature 2.2 cc. of oxygen 
 (measured at o and 760 mm.) go into simple solution in 100 cc. 
 of blood when the partial pressure of oxygen is one atmosphere, 11 
 and this is about 8 per cent less than dissolves in water. In the 
 alveolar air the partial pressure of oxygen is only about 13 per 
 cent of an atmosphere, and in the mixed arterial blood about 1 1 
 per cent, or 84 mm., of mercury. Hence the amount of free oxygen 
 dissolved in the 100 cc. of the arterial blood of a man is only about 
 0.24 cc. (measured at oC. and 760 mm. pressure) whereas 
 about 17.4 cc. are present in combination with haemoglobin, as 
 will be shown below. It is evident, however, that the amount in 
 free solution is of great importance; it depends upon the partial 
 pressure of oxygen in the atmosphere with which the blood is in 
 equilibrium; and, as already pointed out, Paul Bert found that 
 the physiological action of oxygen and of any other gas depends 
 upon its partial pressure in this atmosphere. 
 
 From the standpoint of physical chemistry the "partial pres- 
 sure" of a gas in solution is simply the vapor pressure of the dis- 
 solved gas, i.e., its tendency to pass out of the solvent at any free 
 surface, or the gas pressure which will just balance this tendency 
 so that the amount of gas in solution neither increases nor de- 
 creases. But the vapor pressure of a substance in solution, or of the 
 solvent itself, varies directly, as I showed in a recent paper, 12 
 with the diffusion pressure of the substance in solution. Hence 
 vapor pressure is a direct index of diffusion pressure; and this is 
 the reason why the partial pressure of a gas in solution is of so 
 great importance. It is owing to differences in diffusion pressure 
 that water or substances dissolved in it tend, independently of 
 active "secretory" processes, to pass in one direction or another 
 in the living body or outside it. For instance, when water passes 
 through a semi-permeable membrane into a solution of sugar or 
 salt, this is because the diffusion pressure of the pure water is 
 greater than that of the diluted water in the sugar or salt solution. 
 Van't Hoff's brilliant discovery that there is a connection between 
 the fundamental "gas laws" and the phenomena of osmotic pres- 
 sure was unfortunately marred by his failure to interpret either 
 the connection or the experimental facts correctly. As a conse- 
 quence, osmotic pressure and diffusion pressure were either com- 
 pletely misinterpreted or confused with one another. There seems 
 now to be no doubt that it is the diffusion pressures, and not the 
 
 11 Bohr, Nagel's Hancibuch der Physiol., I, p. 62, 1905. 
 19 Haldane, Bio-Chemical Journal, XII, p. 464, 1918. 
 
68 RESPIRATION 
 
 mere concentration of substances in the body, that are of physio- 
 logical importance. To illustrate this distinction, the concentra- 
 tion of water in blood is much less than in a two per cent solution 
 of sodium chloride ; but the diffusion pressure of water in the blood 
 is much greater than in the salt solution. Hence water will pass 
 from the blood into salt solution. Similarly carbonic acid probably 
 passes by diffusion from the muscular substance into the blood 
 although the concentration of free carbonic acid in the muscle is 
 less than in the blood. 
 
 Paul Bert's conclusion that it is the partial pressure of a gas 
 which is of importance as regards its physiological action can thus 
 be extended to every other substance present in the living body, 
 not excepting water. The partial pressure of a dissolved gas is of 
 decisive importance because the gaseous partial pressure, or vapor 
 pressure, is an index of the diffusion pressure of a substance in 
 solution ; but where the gaseous partial pressure is so low that it 
 cannot be measured, we must have recourse to other indices of the 
 diffusion pressure. 
 
 It has been shown how important are the gas pressures in al- 
 veolar air. But the gas pressures of the blood in the systemic 
 capillaries are of still more fundamental importance. It is clear 
 that in order to understand how the oxygen pressure of the blood 
 is regulated we must know the connection between dissociation of 
 the oxyhaemoglobin of blood and fall in oxygen pressure. In other 
 words we must know what is called the dissociation curve of oxy- 
 haemoglobin in blood. 
 
 The history of the growth of knowledge on this subject is some- 
 what curious. Paul Bert 13 made some rough determinations with 
 the pump of the amounts of oxygen in dogs' blood saturated with 
 air in which the oxygen pressure was varied. His results indicated 
 that in presence of oxygen reduced to a pressure of about 20 mm. 
 the blood at body temperature had lost half its oxygen. In a living 
 animal breathing air with an oxygen pressure of about 55 mm. (the 
 alveolar oxygen pressure being unknown) the blood had also lost 
 half its oxygen. When the blood was at a temperature below that 
 of the body the oxygen was dissociated much less readily. 
 
 The subject was taken up again by Hiifner, who used a solution 
 of oxyhaemoglobin crystals in dilute sodium carbonate solution. 
 As the result, partly of experiments, and partly of calculation, he 
 published in 1890 a very symmetrical curve, according to which 
 oxyhaemoglobin does not lose half its oxygen till the oxygen pres- 
 
 18 Paul Bert, La Pression Barometrique, p. 694, 1878. 
 
RESPIRATION 
 
 69 
 
 sure is reduced to 2.6 mm. 14 This curve was totally at variance 
 with Paul Bert's results, and made it very difficult to understand 
 the effects on animals breathing air with a low oxygen pressure. 
 In 1904 Loewy and Zuntz 15 published further experiments with 
 defibrinated blood giving results much nearer to those of Paul 
 Bert. Meanwhile the subject was taken up by Bohr, 16 who not only 
 confirmed Paul Bert in the main, but for the first time showed that 
 the dissociation curve for blood or haemoglobin solutions has a 
 very peculiar shape, with a double bend (Figure 19) , and that the 
 100 
 
 10 20 30 40 50 60 70 80 90 100 HO 120 130 44O 150 
 
 Figure 19. 
 
 Curves representing the percentage saturation of haemoglobin 
 with oxygen at different partial pressures of oxygen and CO2. 
 Dog's blood at 38C. Ordinates = percentage saturation with 
 oxygen ; abscissae = partial pressures of oxygen in millimeters 
 of mercury. (Bohr, Hasselbalch, and Krogh.) 
 
 curve for a haemoglobin solution differs considerably from that 
 for blood. For this reason he inferred that the haemoglobin in 
 blood ("haemochrome") differs chemically from crystallized 
 haemoglobin. Bohr, Hasselbalch and Krogh 17 then made the 
 important discovery that the dissociation curve of haemoglobin or 
 "haemochrome" is greatly influenced by the partial pressure of 
 the CO 2 present (Figure 19), the CO 2 helping to expel oxygen 
 from its combination, so that, as the blood takes up CO 2 in its 
 passage through the capillaries, oxygen is liberated from the oxy- 
 haemoglobin more readily than would otherwise be the case. 
 
 "Hiifner, Arch. f. (Anat. u.) Physwl., p. i, 1890. 
 
 "Loewy and Zuntz, Arch. f. (Anat. u.) Physwl., p. 166, 1904. 
 
 "Bohr, Centralbl. f. Physwl., 17, p. 688, 1904. 
 
 17 Skand. Arch. f. Physwl., 16, p. 602, 1904. 
 
70 RESPIRATION 
 
 The investigation was now taken up by Barcroft and his pupils, 
 who have made a number of important advances during the last 
 few years with the help of one form or another of the ferricyanide 
 apparatus. 18 
 
 They found that the form taken by the dissociation curve of 
 oxyhaemoglobin is greatly influenced by the salts present in the 
 red blood corpuscles, or in a solution of their oxyhaemoglobin. 19 
 When all the salts were removed by dialysis the curve became a 
 rectangular hyperbola, 20 as in the curve published by Htifner. If 
 the reversible reaction between oxygen and haemoglobin is rep- 
 resented by the uncomplicated equation Hb + O 2 *^HbO 2 , the 
 curve would, in accordance with the well-known law of Guldberg 
 and Waage, be a rectangular hyperbola. This is the case when 
 salts are absent and the solution is neutral, as in the dialysed solu- 
 tion. When, however, salts are present, the form of the curve is 
 altered towards the characteristic form given by blood, and the 
 nature and extent of the alteration was found to depend on the 
 nature and concentration of the salts. Thus when dialysed dogs' 
 haemoglobin was dissolved in a salt solution of the same composi- 
 tion and concentration as in human blood corpuscles the dissocia- 
 tion curve obtained was similar to that of human blood. 
 
 These discoveries rendered it unnecessary to assume with Bohr 
 and others that there is any essential chemical difference between 
 the haemoglobin present in blood corpuscles and in a solution of 
 crystallized haemoglobin. At the same time they furnished a key 
 to the explanation of the apparently divergent observations as to 
 the dissociation curve of oxyhaemoglobin. Barcroft and Orbeli 21 
 found that not only does CO 2 shift the curve in the direction dis- 
 covered by Bohr and his pupils, but that other acids added in 
 such small quantities as not to decompose the haemoglobin have a 
 similar effect, while alkalies have the opposite effect. As will be 
 explained later Barcroft and his associates concluded that this 
 alteration affords a very sensitive measure of any alteration in the 
 reaction, or hydrogen ion concentration of the blood; and they 
 have used it for this purpose. 
 
 The form of the dissociation curve of the oxyhaemoglobin in 
 human blood at body temperature and with a constant pressure of 
 
 18 A summary of these investigations is given in Barcroft's book, The Respira- 
 tory Function of the Blood, 1914. 
 
 19 Barcroft and Camis, Journ. of Physiol., XXXIX, p. 118, 1909. 
 
 20 Barcroft and Roberts, Ibid,., XXXIX, p. 143, 1909. 
 
 21 Barcroft and Orbeli, Journ. of Physiol., XLI, p. 353, 1910. Barcroft, Ibid., 
 XLII, p. 44, 191 1. 
 
RESPIRATION 
 
 40 mm. of CO 2 , as in average human alveolar air, was worked out 
 by Barcroft, and his results for one individual (Douglas) were 
 
 MOO 
 
 18 19 20 21 22 23 24 25 26 27 28 23 
 _O_ 
 
 *80 
 
 I 2 34567 8 9 IO II 12 13 14 13 16 17 
 
 PRESSURE OF OXYGEN IN PERCENTAGE OF ONE ATMOSPHERE . 
 
 Figure 20. 
 
 Dissociation curves of oxyhaemoglobin in presence of 40 mm. pressure of 
 CO2 at 38 (i per cent of an atmosphere =7.60 mm. pressure). 
 
 O Blood of C. G. D., using ammonia in blood-gas apparatus. 
 
 Blood of C. G. D., using NaaCOs in blood-gas apparatus. 
 D Blood of J. S. H., using ammonia in blood-gas apparatus. 
 
 Blood of J. S. H., using Na 2 CO 3 in blood-gas apparatus. 
 
 X Mixed blood of six mice, using ammonia in blood-gas apparatus. 
 
 approximately confirmed by Douglas and myself, working with 
 a different apparatus. Figure 20 shows the curves given by the 
 blood of Douglas and myself in a very exact series of observa- 
 tions, with the individual observations marked. Our curves as will 
 be seen are sensibly the same; but Barcroft has found that the 
 curves of different individuals may vary very distinctly. With the 
 blood of Douglas and myself, for instance, half-saturation of 
 the haemoglobin with oxygen occurs at an oxygen pressure of 4.0 
 per cent of an atmosphere or 30.4 mm. With that of other individ- 
 uals, and the same pressure (40 mm.) of CO 2 , half -saturation 
 may, according to Barcroft, occur at as low an oxygen pressure as 
 24 mm. 
 
 22 
 
 23 Barcroft, The Respiratory Function of the Blood, p. 218, 1913. 
 
72 RESPIRATION 
 
 On examining the dissociation curve it will be seen that the 
 steepest part of the curve is in the middle. In the case of oxy- 
 haemoglobin dissociating in the living body as the blood passes 
 through the capillaries, and in doing so taking up CO 2 , this part 
 of the curve is still steeper, for the reason given by Bohr and his 
 pupils. It is clear that with this form of curve the oxygen pressure 
 in the capillaries must tend, after the first fifth of the oxygen has 
 been given off, to remain comparatively steady during the giving 
 off of the next three-fifths : for at this stage a large amount of 
 oxygen is given off from the oxyhaemoglobin with a compara- 
 tively small fall in the oxygen pressure. In this way the oxygen 
 supply to the tissues is maintained at a far higher and also much 
 steadier pressure than if the curve were a rectangular hyperbola. 
 As will be seen later, a man would die on the spot of asphyxia if 
 the oxygen dissociation curve of his blood were suddenly altered 
 so as to assume the form which Hiifner supposed it to have in the 
 living body. The salts of the red corpuscles and the particular 
 hydrogen ion concentration of the blood are of essential impor- 
 tance in connection with the oxygen supply of the tissues. 
 
 Haemoglobin, as already mentioned, forms specially colored 
 dissociable compounds, not only with oxygen, but also with carbon 
 monoxide and nitric oxide, and the compound with CO is of 
 special physiological interest, apart from its practical importance 
 in connection with the frequency of CO poisoning. As compared 
 with the oxygen compound the CO compound, which was dis- 
 covered by Claude Bernard, 23 is characterized by its relative 
 stability, which is so great that at one time it was supposed that 
 CO-haemoglobin is not dissociable. 
 
 Blood of which the haemoglobin is saturated with CO has a 
 scarlet color similar to that of blood saturated with oxygen ; but if 
 the CO-haemoglobin is highly diluted, or examined in a very thin 
 layer, its color is pink, as compared with the yellow color of diluted 
 oxyhaemoglobin. By taking advantage of this fact one can easily 
 recognize the presence of CO-haemoglobin in blood. This test, as 
 I have often pointed out, is far more delicate than the older 
 spectroscopic test, but requires daylight or some similar light. By 
 adding carmine solution to diluted normal blood one can exactly 
 match the color of the diluted blood containing CO, 24 and by 
 using a suitable carmine solution I found it possible to estimate 
 
 33 Claude Bernard, Compt. Rend., XLVIII, p. 393, 1858. 
 
 "A detailed description of this method in its latest form will be found in 
 the Appendix. 
 
RESPIRATION 
 
 73 
 
 with great accuracy the percentage saturation of haemoglobin 
 with CO. 
 
 With the help of this method Douglas and I worked out "dis- 
 sociation curves for the CO-haemoglobin of human blood at 
 38 C in the absence, of course, of oxygen, but in the presence 
 of varying partial pressure of CO 2 - 25 The results are shown in 
 Figure 21. 
 
 too 
 
 90 
 80 
 70 
 
 50 
 40 
 30 
 
 10 
 
 4 
 
 c* 
 
 o 0</ 
 
 O05 -010 -015 -O20 -025 -0*30 -035 '040 -O45 -0>5O 
 
 PRESSURE OF CO IN PERCENTAGE OF ONE ATMOSPHERE. 
 
 Figure 21. 
 
 Dissociation curves of CO haemoglobin in absence of oxygen, at 38 and 
 with various pressures of CO 2 . O Blood of C. G. D. Blood of J. S. H. 
 
 These curves, like the curve for the oxyhaemoglobin of human 
 blood in Figure 20 are drawn free-hand. On comparing them we 
 found that, allowing for possible small errors due to insufficient 
 determinations, they are all the same curve when the scale on 
 which the abscissae of each are plotted is altered by a suitable 
 
 26 Douglas, J. S. Haldane, and J. B. S. Haldane, Sourn. of Physiol., XLIV, p. 
 275, 1912. 
 
74 RESPIRATION 
 
 constant. It thus appears that the effect of substituting CO for O 2 , 
 or of varying the partial pressure of CO 2 , is only to alter a simple 
 constant in the equation to the curve. In other words it is only 
 the affinity of haemoglobin for the gas saturating it which alters. 
 With respect to the oxyhaemoglobin curve the same conclusion 
 was reached by Barcroft and Poulton, 26 who found that variations 
 in the partial pressure of CO 2 had, within wide limits, the same 
 effects on the dissociation curve of oxyhaemoglobin, as on that of 
 CO-haemoglobin. In the case of Bancroft's blood it requires a 
 little over twice as high a partial pressure of oxygen to produce 
 half-saturation of the haemoglobin in presence of 40 mm. pres- 
 sure of CO 2 as when CO 2 is absent; just as in the blood of Douglas 
 it takes a little over twice as high a partial pressure of CO. Bar- 
 croft and Means 27 have, however, also shown that in the case of a 
 salt-free or nearly salt-free solution of haemoglobin the effect of 
 CO 2 is not merely to alter the affinity of oxygen for haemoglobin, 
 but also to alter the mathematical form of the curve, just as salts 
 do. Hence it is only in the case of whole blood that the affinity 
 alone is altered ; and probably we should find that it is only within 
 definite limits of variations in the hydrogen ion concentration of 
 whole blood that the mathematical form of the dissociation curve 
 is sensibly unaltered. 
 
 When blood or haemoglobin solution is exposed to a mixture 
 of CO and air the haemoglobin becomes partly saturated with CO 
 and for the rest with O 2 . I found many years ago that with a dilute 
 solution of blood the curve representing the percentage saturation 
 of the haemoglobin with CO when increasing percentages of CO 
 are added to the air in the saturating vessel is a rectangular hyper- 
 bola. 28 Figure 22 shows curves obtained by Douglas and myself 
 with undiluted blood at body temperature from two persons and 
 two mice. 29 
 
 It will be seen that in each case the curve is a rectangular 
 hyperbola, corresponding to the simple reversible reaction HbO 2 
 + CO^HbCO + O 2 . Thus for my own blood the proportions of 
 HbCO to HbO 2 are I : I with .07 per cent of CO, 2 : 1 with 2 x .07 
 per cent of CO, 3 : 1 with 3 x .07 per cent of CO, etc. For each kind 
 
 *" Barcroft and Poulton, Journ. of Physwl., XLVI, Proc. Physwl. Soc., p. iv, 
 1913- 
 
 87 Barcroft and Means, Journ. of Physwl., XLVII, Proc. Physwl. Soc., p. 
 xxvii, 1914. 
 
 "Haldane, Journ. of Physiol., Vol. XVIII, p. 449, 1895. 
 
 28 Journ. of Physiol., Vol. XLIV, p. 278, 1912. 
 
RESPIRATION 
 
 75 
 
 of blood the curve remains exactly the same when the blood is 
 diluted, or rendered less or more alkaline, or when neutral salts 
 are added. This is of course quite different from what happens 
 with the simple dissociation curves of oxyhaemoglobin and CO- 
 haemoglobin. 
 
 15 -20 -25 
 
 PERCENTAGE OFCO. 
 
 Figure 22. 
 
 Dissociation curves of CO haemoglobin in presence of air (20.9 per cent 
 O 2 ) at temperature of 38. I. Blood of J. S. H. II. Blood of C. G. D. III. 
 Blood of mouse A. IV. Blood of mouse B. The crosses indicate points deter- 
 mined in the presence of 40 mm. pressure of added COa. 
 
 When the percentage of CO in the air is kept constant and the 
 percentage of oxygen is varied the curve is again a complete rec- 
 tangular hyperbola, as shown in Figure 23, provided that the per- 
 centage of CO is sufficient to saturate the haemoglobin completely 
 in the absence of O 2 , as in the upper curve. 
 
7 6 
 
 RESPIRATION 
 
 It is thus evident that when we have determined the percentage 
 saturations of a sample of haemoglobin with CO and O 2 in a solu- 
 tion saturated with a gas mixture containing CO and O 2 at known 
 concentrations or partial pressures, what we have really de- 
 termined is the relative affinities of the haemoglobin for CO and 
 100 
 
 90 
 
 580 
 
 60 
 
 
 Q- 10 
 
 \ 
 
 in 
 
 10 
 
 20 
 
 30 40 ,50 60 70 
 PERCENTAGE OF OXYGEN. 
 
 80 90 
 
 100 
 
 Figure 23. 
 
 Dissociation curves of CO-haemoglobin in presence of constant percentage 
 of CO and varying percentage of oxygen, at atmospheric pressure. I. Blood of 
 J. S. H. : CO = 0.0945 per cent. Blood of mouse C : CO = 0.090 per cent. 
 III. Blood of mouse D : CO = 0.0635 per cent. 
 
 O 2 (without allowing, however, for the slight difference in solu- 
 bility between the two gases). In my own blood the haemoglobin 
 is equally divided between CO and O 2 when the partial pressures 
 of CO and O 2 are as .07 to 20.9 i.e., as I to 299. Hence the 
 affinity of the haemoglobin for CO is 299 times its affinity for O 2 . 
 For the haemoglobin of Douglas the corresponding figure is 246. 
 For his haemoglobin we can also compare the affinities for CO and 
 
RESPIRATION 77 
 
 O 2 in another way. In presence of 40 mm. of CO 2 his blood be- 
 comes half -saturated with CO (in the absence of oxygen) at a 
 pressure of .017 per cent of an atmosphere of CO, as shown in 
 Figure 21, and half -saturated with O 2 (in the absence of CO) at 
 a pressure of 4.0 per cent of an atmosphere, as shown in Figure 
 20. These pressures are in the ratio of I 1235, which is nearly the 
 same ratio as when the relative affinities are estimated by the pre- 
 vious method. 
 
 As already seen, we may be able to account for varying dis- 
 sociation curves of the oxyhaemoglobin in whole blood by the 
 varying composition and concentration of the salts contained in 
 the red corpuscles, and by varying alkalinity; but we cannot so 
 account for the varying relative affinities of different specimens 
 of haemoglobin for CO and O 2 , since the curves in Figure 22 are 
 not affected by varying concentration of salts or degrees of alka- 
 linity. There seems to be no escape from the conclusion that in 
 different individuals of the same species, as well as in different 
 species, the haemoglobin molecules are different. Whether the 
 haemoglobin in each individual is made up of homogeneous mole- 
 cules, or is a mixture in some definite proportion of two or more 
 different kinds of haemoglobin, we do not as yet know. What 
 seems pretty certain, however, is that each individual has a specific 
 kind of haemoglobin just as he has a specific shape of nose. At 
 whatever time we have investigated my own and Dr. Douglas's 
 haemoglobin their specific differential characters have appeared 
 to be sensibly the same. It seems pretty certain that, since the ratio 
 of oxygen capacity to both the coloring power and amount of 
 iron in haemoglobin is constant, the difference in the haemoglobin 
 molecule in different kinds of blood is due to the protein and not 
 the haemochromogen fraction of the molecule; but as yet there 
 are no data to indicate more specifically the nature of the differ- 
 ence. It is of considerable biological significance to have found, 
 however, that, looking at living organisms from a purely chemical 
 standpoint, individual differences express themselves, not merely 
 in the relative amounts of the different molecules which can be 
 separated from different parts of the body, but also in their chemi- 
 cal constitution. 
 
 Since the dissociation curve of CO-haemoglobin in presence of 
 a constant pressure of oxygen and varying pressure of CO, or in 
 presence of a constant pressure of CO and varying pressure of 
 oxygen, is a rectangular hyperbola, provided that the gases are 
 present at sufficient pressure to saturate the haemoglobin, it is 
 
78 RESPIRATION 
 
 clear that provided we know the relative affinities of the two gases 
 for the haemoglobin, and the pressure at which one is present, we 
 can tell from an observation of the percentage saturation of the 
 haemoglobin the pressure of the other. Hence we can use haemo- 
 globin solutions for determining small percentages of CO in air. 
 All that is necessary is to introduce a little blood solution into a 
 small bottle of the air, shake till the solution takes up no more CO, 
 and then determine colorimetrically the percentage saturation of 
 the haemoglobin with CO, and calculate the percentage of CO 
 present. 30 Still more important in physiological work is the con- 
 verse determination of the oxygen pressure by observation of the 
 percentage saturation of haemoglobin exposed to a constant pres- 
 sure of CO. By this means, as we shall see later, it is possible to 
 measure the partial pressure of oxygen in the arterial blood within 
 the living body and so decide the question whether active secretion 
 of oxygen inwards occurs in the lungs. 
 
 Douglas and I found that when the combined pressure of O 2 
 and CO are insufficient to saturate the haemoglobin the dissocia- 
 tion curve of CO-haemoglobin in presence of a constant pressure 
 of CO and diminishing pressure of O 2 begins to diverge from the 
 rectangular hyperbola which it would otherwise have followed, 
 and then proceeds to trace out the peculiar hump shown on the 
 lower two curves in Figure 23, and in greater detail in Figure 24. 
 We thus have what seems at first sight a most anomalous fact, 
 namely that although all other facts show that increase in the 
 pressure of oxygen tends to keep out CO more and more from 
 combination with haemoglobin, yet at very low pressure of oxygen 
 and CO the reverse is the case, and increase of oxygen pressure 
 helps the CO to combine with haemoglobin. There can be no doubt 
 that the converse is also the case namely that at low pressures of 
 CO the presence of the CO helps the oxygen to combine with the 
 haemoglobin. This explains a very anomalous fact noticed by 
 Lorrain Smith and myself many years ago 31 namely that the 
 presence of a small percentage of CO helps animals to resist the 
 effect of a very low oxygen pressure, or at any rate does not make 
 them worse. We had expected that a given percentage of CO 
 would become more and more poisonous the more the oxygen 
 pressure was diminished, and this was the case within certain 
 limits ; but we were then quite at a loss to understand why with 
 very low oxygen pressures the CO seemed to do no harm. 
 
 ""Haldane, Methods of Air Analysis, p. 119, 19 1 g. 
 
 "Haldane and Lorrain Smith, Journ. of Physiol., XXII, p. 246, 1897- 
 
RESPIRATION 
 
 79 
 
 The explanation of the anomalous hump in the curves on Fig- 
 ures 23 and 24 is in reality easy enough in view of the peculiar 
 double-bended form of the simple dissociation curves of oxy- 
 haemoglobin and CO-haemoglobin in whole blood. When CO is 
 present at a pressure insufficient to saturate the blood, and the 
 
 100 
 
 90 
 
 2eo 
 5 
 
 Iro 
 3 
 
 O 
 
 ?60 
 
 50 
 
 I- 
 u. 
 
 z 40 
 o 
 
 j 30 
 
 ^ 20 
 I '0 
 
 E 
 
 t/ 
 
 y\ 
 
 4 6 3 10 12 14 16 18 2O 22 
 
 PRESSURE OF OXYGEN IN PERCENTAGE OF ONE ATMOSPHERE. 
 
 Figure 24. 
 
 Dissociation curves of CO-haemoglobin in blood at 38 and in presence of 
 40 mm. CO 2 , with constant pressure of CO and varying pressures of oxygen. 
 
 oxygen pressure is gradually raised from zero, the two gases to- 
 gether will trace out curves representing the total saturation of 
 the haemoglobin, as shown in the thin lines on Figure 24. These 
 curves are calculated on the theory that the proportion of oxy- 
 haemoglobin to CO-haemoglobin is exactly what is required in 
 view of the known relative affinities of oxygen and CO for the 
 haemoglobin of the blood used. As, however, the thin curves start 
 at the steep part of the joint curve a very small addition of oxygen 
 
8o RESPIRATION 
 
 will produce such a large effect that not only will a large amount 
 of oxygen go into combination, but also an increased proportion 
 of CO. The thick lines show the curve for CO-haemoglobin as 
 calculated on this hypothesis, and the dots show the actual ob- 
 servations. There is in reality perfect agreement with the theory 
 that oxygen and CO combine with haemoglobin in exact propor- 
 tion to their relative affinities for haemoglobin and their partial 
 pressures, just as in the upper curve of Figure 23. The great sig- 
 nificance of this in connection with the explanation of CO poison- 
 ing will be referred to later. 
 
 It remains to discuss the explanation of the various dissocia- 
 tion curves to which reference has been made. We have seen above 
 that Barcroft and his pupils found that when a solution of oxy- 
 haemoglobin is freed, or approximately freed, from salts it gives 
 a dissociation curve which is a simple rectangular hyperbola, in 
 accordance with the simple reaction 
 
 Hb + O 2 ^HbO 2 . 
 
 A. V. Hill pointed out in 1910 that the varying values obtained 
 for the osmotic pressure of haemoglobin solutions in presence of 
 salts indicates that the molecules are more or less aggregated to- 
 gether owing to the influence of the salts ; and he showed that this 
 fact was capable of explaining the deviation from a rectangular 
 hyperbola of the dissociation curve. Thus if, in consequence of 
 the aggregation, the reaction were 
 
 Hb 2 + 2O 2 ^Hb 2 O 4 , 
 
 the curve would no longer be a rectangular hyperbola but would 
 approximate to that given for oxyhaemoglobin in presence of a 
 certain proportion of salts. By assuming a suitable proportion of 
 aggregation of the haemoglobin molecules as Hb 2 , Hb 3 , etc., we 
 can therefore construct equations which will give the actual dis- 
 sociation curves. He also gave a general form of equation to meet 
 the varying cases. In this equation there are two constants, which 
 must be suitably chosen. 
 
 The subject was also taken up by Douglas, J. B. S. Haldane, 
 and myself. We adopted Hill's aggregation theory, but in a dif- 
 ferent form. It seemed to us that the aggregation in protein solu- 
 tions is a phenomenon of the same general nature as precipitation, 
 the precipitate being, however, only formed in very small parti- 
 cles consisting of only two, three, or at any rate a few molecules. 
 
RESPIRATION 8 1 
 
 On this view the aggregated haemoglobin molecules have their 
 molecular affinities saturated, and therefore go out of the^e^ 
 action between oxygen or CO and haemoglobin, thus following 
 the general principle that corpora non agunt nisi soluta. The only 
 reaction taking place between the haemoglobin and oxygen is thus 
 the first one mentioned above. To explain the actual form of the 
 dissociation curve for blood or salt solutions we assumed that the 
 degree of aggregation depends on the concentration of the haemo- 
 globin or oxyhaemoglobin in the solution, in accordance with the 
 reactions 
 
 Hb 
 
 Hb + Hb 2 <HHb 3 etc. 
 HbO 2 + HbO 2 ^Hb 2 O 4 
 HbO 2 + Hb 2 O 4 *Hb 3 O 6 etc. 
 
 Thus reduced haemoglobin and oxyhaemoglobin molecules ag- 
 gregate separately; and if we assume that reduced haemoglobin 
 aggregates more readily than oxyhaemoglobin we can explain at 
 once the distortion of the curve from the primary rectangular 
 hyperbola obtained by Barcroft. For as the oxyhaemoglobin be- 
 comes reduced the aggregation of the reduced haemoglobin mole- 
 cules must increase more rapidly than the aggregation of the 
 oxyhaemoglobin diminishes. Hence at what would, but for the 
 aggregation, be half-saturation, there are fewer free reduced 
 haemoglobin molecules and more free oxyhaemoglobin molecules 
 than would be the case if the oxyhaemoglobin molecules aggre- 
 gated as readily as the reduced haemoglobin molecules. Hence 
 the actual saturation will be much less than half, and not just half, 
 as would be the case if the tendency to aggregation were the same 
 for the two kinds of molecules. The actual dissociation curve will 
 also have the double bend which is characteristic of it. We also 
 assumed that the saturated molecules of HbCO have just as much 
 tendency to aggregate with one another and with the saturated 
 molecules of HbO 2 as have the molecules of HbO 2 . For this reason 
 the dissociation curve of HbCO in blood in presence of oxygen 
 must be a rectangular hyperbola, as is actually the case, though its 
 dissociation curve in the absence of oxygen has the same form as 
 the dissociation curve of HbO 2 . 
 
 By making certain assumptions (for a statement of which I 
 must refer to our original paper) J. B. S. Haldane found that the 
 following equation to the curve for human blood in Figure 20 
 
82 RESPIRATION 
 
 resulted, and fitted the experimentally determined curve very 
 closely. 32 
 
 1.65 (985) 
 
 (I-S) (I+2S) 
 
 where p = pressure in percentages of one atmosphere, and 
 5 = fractional saturation of the haemoglobin with oxy- 
 gen. 
 
 Thus if 5 be 50 per cent = = y 2 , p will be 4.0, as we actu- 
 ally found to be the case. To express the result in millimeters of 
 mercury pressure, p must of course be multiplied by 7.6, and would 
 thus become, in the above example, 30.4. 
 
 As explained above, the simple dissociation curves for oxy- 
 haemoglobin or CO-haemoglobin in normal human blood 33 are, 
 so far as our present knowledge goes, the same, when allowance 
 is made for the differing affinities of the two gases for haemo- 
 globin. The above equation may therefore be generalized in the 
 form 
 
 1.65 (985) 
 
 pa = 
 
 (i S) (i+25) 
 
 taking a as representing the affinity of the gas for haemoglobin 
 as compared with the affinity represented in the curve on Figure 
 20, giving half-saturation with a gas pressure of 4.0 per cent of an 
 atmosphere. Thus for the fourth curve on Figure 21 (dissociation 
 curve of CO-haemoglobin in the blood of Douglas, in presence of 
 42 mm. CO 2 pressure), at half-saturation pa = 4.0. Hence as p 
 was .017, a was 235, or the affinity of the haemoglobin for the CO 
 (determined without taking into account the solubilities of CO and 
 O 2 ) was 235 times its affinity for oxygen in the standard curve of 
 Figure 20. This is a convenient and easily intelligible method of 
 putting the results. 
 
 82 In working out this equation it was assumed that (as found by Barcroft and 
 Roberts for dogs' haemoglobin) a dialysed solution of the haemoglobin of Douglas 
 and myself becomes half -saturated with oxygen at 38 C and a pressure of 1.6 per 
 cent of an atmosphere of oxygen, and that in human blood saturated with oxygen 
 2/3 of the oxyhaemoglobin is aggregated, and in completely reduced blood 8/9 of 
 the reduced haemoglobin. The curve of the dialysed solution would give the 
 
 i-S 
 
 equation p = 
 
 i. 60 
 
 38 For abnormal human blood the curves are probably different, as will be 
 pointed out in Chapter VIII. 
 
RESPIRATION 83 
 
 The corresponding equation worked out by Hill is 
 
 y_ 
 
 ioo zz 
 
 where x = oxygen pressure in mm. of mercury, 
 
 y =i percentage saturation of the haemoglobin, 
 K = a constant varying for different curves. 
 
 For the blood of Douglas (which was the first to be investi- 
 gated completely by Barcroft, and which was also investigated by 
 ourselves) the value of K was . 000196. 34 
 
 Hill's equation gives curves almost identical with ours, and as 
 he had kindly communicated it to us by letter we should certainly 
 have adopted it had we seen how the theory on which it is based 
 could be brought into definite relation with the particular rec- 
 tangular hyperbola given by dialysed haemoglobin, or reconciled 
 with the fact that the dissociation curve of CO -haemoglobin in 
 presence of a constant oxygen pressure is a rectangular hyperbola. 
 Hill soon afterwards offered a possible explanation as regards the 
 latter point. 35 It seems to me that this explanation is improbable, 
 but so also, it must be confessed, are certain assumptions connected 
 with the deduction of our own equation. At present the data are 
 lacking for a decision as to whether either theory is correct, al- 
 though both equations are for all practical purposes satisfactory. 
 I cannot see, however, how to escape the conclusion that there is 
 more aggregation among the unsaturated than among the satu- 
 rated molecules of haemoglobin. It is evident that far more data 
 are needed to enable us to understand the dissociation of oxy- 
 haemoglobin in blood. 
 
 With the help of the chemical facts described in the present 
 chapter we might proceed at once to the discussion of a number of 
 physiological and pathological problems; but such a discussion 
 would be incomplete and misleading in the absence of the facts 
 relating to the carriage of CO 2 by the blood, and this subject will 
 therefore be considered in the next chapter. 
 
 34 The value of K as calculated from our own results (Fig. 20) is, for the blood 
 of both Douglas and myself, outside the normal limits given by Barcroft and rep- 
 resented graphically in Figure 109, page 226, of his book The Respiratory Function 
 of the Blood, The cause of this discrepancy is not yet clear. 
 
 "A. V. Hill, Bio-Chemical Journal, VII, p. 471, 1913. 
 
CHAPTER V 
 The Blood as a Carrier of Carbon Dioxide. 
 
 WE must now turn to the consideration of the blood as a carrier of 
 CO 2 . Mammalian arterial blood has usually been found to contain 
 about 40 or 50 volumes of CO 2 per 100 volumes of blood, while 
 venous blood from the right side of the heart contains several 
 volumes more. The following average results obtained with the 
 mercurial pump by Schoeffer 1 illustrate the difference between 
 venous and arterial dogs' blood, although much doubt must exist 
 as to whether the circulation and respiration were at normal rest- 
 ing values when the samples were taken. Much more reliable data 
 will be given for man in Chapter X. 
 
 OXYGEN CO 2 
 
 Arterial blood 19.2 39.5 
 
 Venous blood from right heart 11.9 45.3 
 
 Difference 7.3 5.8 
 
 In man, as will be shown below, normal arterial blood contains 
 during rest about 53 volumes per cent of CO 2 if the blood is satu- 
 rated with CO 2 at the pressure (about 40 mm.) existing in average 
 alveolar air of adult men. As 100 volumes of blood, according to 
 Bohr's 2 calculation, take up in simple solution about 51 volumes 
 of CO 2 in presence of a pressure of one atmosphere of CO 2 at body 
 
 temperature, they can only take up ~- x 51 = 2.7 volumes at the 
 
 760 
 
 normal alveolar pressure of 40 millimeters or 5.3 per cent of an 
 atmosphere. Hence only 2.7 volumes per cent of the CO 2 are in 
 simple solution, the other 50.3 volumes being in chemical combi- 
 nation. As will be shown below, the difference between the partial 
 pressures of CO 2 in human arterial and venous blood during rest 
 is only about 6 mm. or 0.8 per cent of an atmosphere. Hence the 
 physically dissolved CO 2 given off in the lungs is only 0.4 volumes 
 
 1 Schoeffer, Sitz. ber. d. Wiener Acad., math. nat. cl., XLI, p. 589, 1860. 
 8 Bohr, Nagel's Handbuch der Physiol., II, p. 63, 1905. 
 
RESPIRATION 85 
 
 per cent, while actually about 4 volumes per cent are given off. It 
 is evident, therefore, that the giving off of CO 2 in the lungs is 
 almost entirely dependent on the dissociation of its chemical 
 combinations in the blood. 
 
 In what form is CO 2 chemically combined in the blood? We 
 cannot answer this question in the same comparatively simple 
 and definite manner as in the case of the combination of oxygen 
 with blood. CO 2 dissolved in water has acid properties, and by the 
 addition of other stronger acids to blood the dissociable chemical 
 combinations with CO 2 are entirely broken up and CO 2 liberated. 
 It is thus quite evidently as an acid (i.e., as H 2 CO 3 ) that CO 2 
 enters into combination with blood. On analysis blood is found to 
 contain an excess of alkali (for the greater part soda) not com- 
 bined with mineral acids. In other words hydrochloric, phosphoric, 
 and small amounts of sulphuric, acids are present in blood, but 
 not in sufficient amounts to saturate the alkali. Hence CO 2 is ap- 
 parently free to combine with the excess of alkali, forming, since 
 an excess of free CO 2 is present, bicarbonates. As Zuntz 3 pointed 
 out, if blood were nothing but a solution of the well-recognized 
 acids and bases present in it, we could account for the quantity of 
 CO 2 which it is capable of combining with chemically. Zuntz cal- 
 culated that the excess of alkali present in the blood is equivalent 
 to at least a 0.2 per cent solution of soda. This could take up as 
 bicarbonate as much CO 2 as blood can take up in combination. 
 Nevertheless the properties of such a solution in respect to the 
 carriage of CO 2 would not approach to those of blood: for the 
 soda would remain completely saturated as bicarbonate when ex- 
 posed to the CO 2 in the alveolar air, and there would not be any 
 appreciable dissociation, so that the solution would be no better 
 than distilled water as a physiological carrier of CO 2 . This point 
 has been rendered specially clear by Bohr, who investigated the 
 dissociation curve for CO 2 of a dilute sodium bicarbonate solution. 
 
 To reach an insight into the actual behavior of blood as a car- 
 rier of CO 2 we have to take into consideration another factor. 
 Proteins have the very peculiar property of being able to act either 
 as weak alkalies towards acids or as weak acids towards alkalies. 
 This is shown, for instance, by the familiar fact that an ordinary 
 indicator such as litmus ceases to give a sharp end-point when a 
 protein is present, and that not only neutral but even slightly acid 
 
 'Zuntz, Hermann's Handbuch d,er Physiol., IV, 2, p. 65, 1882. To Zuntz's 
 admirably clear and thorough discussion of the subject I am greatly indebted. This 
 discussion is far ahead of most of what has appeared in later textbooks and papers. 
 
86 RESPIRATION 
 
 protein solutions will combine with CO 2 . A considerable excess 
 of acid or alkali must be added to a neutral protein solution before 
 a marked acid or alkaline reaction is reached. The protein acts as 
 a "weak," or very slightly ionized, acid, such as carbonic acid, and 
 likewise acts as a correspondingly weak alkali, since the protein 
 molecule possesses both acid and alkaline affinities. It is thus, like 
 carbonic acid, or any other weak acid, or weak alkali, a buffer 
 substance, which prevents any abrupt change from acid to alkaline 
 reaction or vice versa. Not until all the CO 2 combined in a solu- 
 tion of carbonate has been liberated by acid is there a sudden de- 
 velopment of acid reaction, or so long as any free CO 2 is present 
 of strong alkaline reaction. The CO 2 acts as a buffer substance on 
 the alkaline side only, whereas protein is capable of acting on 
 either side of the neutral point. In the living body, however, blood 
 is always a little alkaline, so that the combination of CO 2 with 
 proteins does not come into account. 
 
 We can now see a reason why blood should act towards CO 2 as 
 it does in the living body and in the vacuum pump. The total 
 alkali in the blood is combined, partly with strong acids, such as 
 HC1, partly with carbonic acid, and partly with protein com- 
 pounds ; partly also, perhaps, with other substances capable of 
 acting, like the proteins, as very weak acids. In the living body, 
 however, free carbonic acid is always present, and the mass in- 
 fluence of the free carbonic acid prevents part of the protein from 
 combining with alkali, while the protein in a similar manner 
 keeps out the carbonic acid. We have thus a chemical system 
 which is disturbed at once by any variation in the concentration 
 of free carbonic acid present, i.e., by any variation in the partial 
 pressure at which the blood is saturated with CO 2 . When the 
 pressure of CO 2 falls, more of the proteins are at once enabled to 
 take the place previously occupied by the carbonic acid in the 
 chemical combinations which constitute the system ; and vice versa 
 with a rise of CO 2 pressure. In the vacuum pump the CO 2 pressure 
 is reduced to zero, since, although the total pressure in the vacuum 
 chamber of the pump is, owing to aqueous vapor, always above 
 zero, the CO 2 is carried off in the stream of aqueous vapor passing 
 away. To recover the whole of this CO 2 in the same gaseous form, 
 however, a perfect and dry vacuum in the receiving chambers of 
 the pump is needed. Since the CO 2 pressure is zero the whole of the 
 CO 2 in combination is expelled by the mass influence of protein act- 
 ing as an acid. Pfliiger showed that even when a moderate amount 
 of sodium carbonate is added to blood, the additional CO 2 in the 
 
RESPIRATION 87 
 
 carbonate is expelled in the vacuum pump, and can be recovered 
 in the gaseous form with the help of the perfect vacuum of the 
 Pfliiger blood pump. 4 This can now be easily understood in terms 
 of the theory just stated. The blood must be either boiled or 
 shaken : otherwise the disengagement of CO 2 is excessively slow. 
 
 When serum alone, and not whole blood, is exposed to the 
 vacuum of the pump, most of the CO 2 can be pumped out, but not 
 quite all. It is necessary to add some acid in order to obtain the 
 whole of the CO 2 at any rate within any reasonable time. The 
 proteins of the serum are not present in sufficient amount to effect 
 the dissociation of the whole of the sodium carbonate, but the ex- 
 pulsion is easy when the haemoglobin of the corpuscles is added. 
 Both haemoglobin and serum proteins act towards sodium carbon- 
 ate as acids, and it was shown by Sertoli 3 that much of the CO 2 
 can be expelled in the pump from sodium carbonate solution if 
 serum proteins are first added. 
 
 Bohr found that haemoglobin solutions, even if they are first 
 rendered slightly acid, will combine with considerable amounts 
 of CO 2 , and he was thus led to what seems to me to be the er- 
 roneous conclusion that haemoglobin has a specific power, apart 
 from its alternative acid or basic properties, of combining with 
 CO 2 . Equally erroneous, as Priestley 6 has recently shown, is a 
 similar conclusion which was put forward on spectroscopic 
 grounds. 
 
 We have already seen what predominant physiological im- 
 portance is attached to the pressure of CO 2 in the arterial blood, 
 and with what exactitude this pressure is regulated. We should 
 therefore expect to find that the pressure of CO 2 in the tissues of 
 the body generally is of the same importance and subject to simi- 
 lar regulation. To understand this regulation it is of primary 
 importance that we should know the laws of dissociation of CO 2 
 from its combination in blood. Until quite recently our knowledge 
 on this subject was very limited, although Bohr 7 had constructed 
 a tentative dissociation curve from observations partly by Jacquet 
 and partly by himself, on samples of blood from the ox and dog. 
 
 The matter was taken up a short time ago by Christiansen, 
 Douglas and myself 8 with the help of the new method of blood- 
 
 4 Pfliiger, Ueber die Kohlensdure des Blutes, p. 6, 1864. 
 
 5 Sertoli, Hoppe-Seyler's Med.-Chem. Unters., Ill, p. 356, 1868. 
 
 9 Priestley, Journ. of Physiol., LIII, Proc. Physiol. Soc., p. LVIII, 1920. 
 
 7 Bohr, Nagel's Handbook der Physiol., II, p. 106, 1905. 
 
 8 Christiansen, Douglas, and Haldane, Journ. of Physiol., XLVIII, p. 244 
 1914. 
 
88 
 
 RESPIRATION 
 
 gas determination mentioned in Chapter IV. Warned by previous 
 failures of physiologists to recognize the exactitude of normal 
 physiological regulations, we used defibrinated human blood, of 
 which fresh samples could be obtained at any time from the same 
 individual under normal conditions. At the outset we wasted much 
 time, however, through failing to realize that it was necessary to 
 have the blood fresh for each experiment, as blood outside the 
 body undergoes slow changes which diminish its capacity for 
 carrying CO 2 . 
 
 Figure 25 shows the results obtained with my own blood. 
 
 ABSORBED by 100 VOLUMES of BLOOD 
 
 n -^ Cn O o 0) ^ 
 3 O O O O C 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 .o-' 
 
 ^-^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^^ 
 
 ^ 
 
 -- '^ r 
 
 
 -* 
 
 *-- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ~~2> 
 
 ^ 
 
 
 
 - 
 
 , 
 
 ^-^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 ** 
 
 
 - c ^_ 
 
 ^ 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ** 
 
 ^U 
 
 
 ^* 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ,' 
 
 ^ 
 
 
 ^ 
 
 -" 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 I* 
 
 ^f 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 /r 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 / 
 
 r 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c/ 
 
 '/ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 Vj 
 
 
 /. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 >->0 
 
 1 
 
 f 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ; 
 
 J/ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 ^ JO 
 
 // 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Q 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 20 30 40 5O 60 7O 80 90 100 HO 12 
 
 PRESSURE of CO 2 in MM. Hy. 
 
 Figure 25. 
 
 Lower curve absorption of CO2 by blood of J. S. H. in presence of air and C(>2. 
 Upper curve absorption of CC>2 by blood of J. S. H. in presence of hydrogen and 
 C0 2 . 
 
 Attention may first be directed to the lower curve, showing the 
 amounts of CO 2 taken up in the presence of air and varying pres- 
 sures of CO 2 . The first, and by far the most striking, point to be 
 noted is that, although the different determinations were made on 
 different days covering a period of about six months, they all lie 
 on one curve. The samples were taken at different times of the day 
 
RESPIRATION 89 
 
 during ordinary laboratory work. In regulating the temperature 
 of the bath containing the saturator, analyzing the samples of air 
 from the saturator, observing the barometric pressure, measuring 
 the sample of blood (of which about I cc. was used for each anal- 
 ysis), and determining the CO 2 by means of the blood-gas ap- 
 paratus (we used Brodie's modification of the original apparatus), 
 it was impossible to avoid combined errors of I or 2 per cent of 
 the quantities to be measured, so that we could not say how exact 
 Nature's regulation of the curve is. At any rate it was so exact 
 for my blood that the most exact existing chemical methods did 
 not show any deviations from the curve, any more than they 
 could show deviations from the oxyhaemoglobin or CO-haemo- 
 globin dissociation curves. Marked temporary deviations could, 
 however, be produced by severe muscular exertion ; and probably 
 very distinct deviations may occur after meals. 
 
 With the blood of other persons the results were only slightly 
 different. Thus the curves, so far as ascertained, for the blood of 
 Miss Christiansen and Dr. Douglas were slightly below, and 
 otherwise parallel to mine under normal conditions. The blood of 
 most persons seemed to take up about 50 volumes of CO 2 per 100 
 volumes of blood at 40 millimeters pressure of CO 2 ; but under 
 abnormal conditions, as will be shown below, there are great 
 temporary variations from this standard, corresponding to the 
 great variations observed under the unfavorable conditions in 
 experiments on animals. 
 
 More than 50 years ago it was suspected by Ludwig that oxygen 
 may have some influence in turning out CO 2 from the venous 
 blood which conies to the lungs. The experiments made to ascer- 
 tain whether oxygen helps to turn out CO 2 from blood gave, how- 
 ever, only a negative result, and more recent work by Bohr, 
 Hasselbalch, and Krogh 9 led to similar negative conclusions. We 
 had been making experiments to investigate the rise of alveolar 
 CO 2 pressure when the breath was held, or when a small quantity 
 of air was rebreathed. One result of these experiments was to show 
 that if the alveolar oxygen pressure fell much below normal the 
 percentage of CO 2 in the alveolar or rebreathed air was always, 
 without exception, lower after any definite interval of time, than 
 was the case under the same conditions but with the alveolar oxy- 
 gen percentage high. This brought us back to Ludwig's old ques- 
 tion, which with the new blood-gas method we could investigate 
 
 "Bohr, Hasselbalch, and Krogh, Skand. Arch. f. Physiol., XVI, p. 411, 1904. 
 
90 RESPIRATION 
 
 far more easily and exactly than when nothing but the blood 
 pump and the old methods of gas analysis were available. 
 
 The first pair of experiments showed us that Ludwig's old 
 suspicion was correct, and that at the same pressure of CO 2 blood 
 takes up considerably more CO 2 in the absence than in the presence 
 of oxygen. The upper curve in Figure 25 is the absorption curve 
 for my own blood in the absence of oxygen, and shows that at the 
 physiologically important part of the curve the blood takes up 
 from 5 to 6 volumes per cent more of CO 2 if oxygen is absent. We 
 found that the excess of CO 2 taken up runs parallel, not to the 
 partial pressure of oxygen, but to the extent to which the oxy- 
 haemoglobin of the blood is dissociated. Saturation of the haemo- 
 globin with CO had just the same effect on the curve as saturation 
 with oxygen. The effect may be due to saturated haemoglobin 
 being a less alkaline substance than reduced haemoglobin, but is 
 more probably dependent on the molecules of reduced haemo- 
 globin having a much greater tendency to aggregate than those 
 of saturated haemoglobin. The reasons for this assumption with 
 regard to aggregation were given at the end of last chapter. 
 The aggregated haemoglobin molecules would presumably have 
 less mass influence in keeping out the CO 2 from combination with 
 alkali than the unaggregated molecules. 
 
 Let us now see what physiological deductions can be drawn 
 from the absorption curves in Figure 25. Human blood contains 
 about 1 8 volumes per cent of oxygen, and if all this oxygen were 
 used up in the tissues about 15 volumes of CO 2 would be formed. 
 But during the using up of the oxygen the absorption curve for 
 CO 2 starting from 40 mm. would pass from the lower to the upper 
 curve of Figure 25, following upwards the thick line shown in 
 Figure 26. 
 
 Hence the CO 2 pressure, instead of rising to 80 mm., as would be 
 the case if the lower curve were followed, would only rise to 62 
 mm. Actually, as will be shown later, not more than about a fifth 
 of the oxygen is used up during rest, so the pressure of CO 2 in the 
 mixed venous blood rises only about 5 or 6 mm. This makes it 
 far more easy to understand why the pressure of CO 2 in the arte- 
 rial blood should be so exactly regulated as it is. If it had been the 
 case that the resting CO 2 pressure in the systemic capillaries were 
 far above the arterial CO 2 pressure, the necessity for such exact 
 regulation of the arterial CO 2 pressure would have been hard to 
 understand. 
 
 While the venous blood is being aerated in the lungs, the ab- 
 
RESPIRATION 
 
 sorption curve for CO 2 will follow the thick line downwards. It 
 will be seen that, if we assume the resting excess pressure of CO 2 
 in the venous blood, the quantity of CO 2 given off when the CO 2 
 pressure in the lung capillaries falls to that of the alveolar air 
 will be about 55 per cent greater than if no oxygenation had oc- 
 curred. If, on the other hand, we assume a certain excess charge 
 75 
 
 50 
 
 40 
 
 /A 
 
 80 
 
 90 
 
 30 40 50 60 70 
 
 PRESSURE of CO 2 in MM. Hq. 
 
 Figure 26. 
 
 Upper curve absorption of CO2 by blood of J. S. H. in pres- 
 ence of hydrogen and CC>2. 
 
 Middle curve absorption of COz by blood of J. S. H. in pres- 
 ence of hydrogen and CC>2. 
 
 Lower curve absorption of CO2 in blood of ox and dog in pres- 
 ence of air and CO 2 (Bohr's data). 
 
 Thick line A B represents the absorption of CO2 by the blood 
 of J. S. H. within the body. 
 
 of CO 2 in volumes per cent in the venous blood, the discharge of 
 CO 2 will ordinarily be about 55 per cent greater than if no oxy- 
 genation had occurred. 
 
 We can also see that under abnormal conditions, such as may 
 easily occur when the breathing is suspended or reduced in 
 
92 RESPIRATION 
 
 amount, as after forced breathing, or during excessive artificial 
 respiration, or other respiratory disturbances, CO 2 may easily be 
 given off by the lungs when there is no excess of venous over 
 alveolar CO 2 pressure, or even when the venous CO 2 pressure is 
 considerably lower than that of the alveolar air. For when the 
 blood reaches the lungs the process of oxygenation so reduces the 
 capacity of the blood for CO 2 that its CO 2 pressure is raised 
 above that of the air, and diffusion results. If the respiratory 
 quotient has fallen temporarily to a third or less of its normal 
 value, the thick line of Figure 26 will become vertical in the living 
 body, or incline to the left instead of to the right. It is merely 
 necessary to suspend the breathing for a very short time in order 
 to realize this condition. Only if air containing a large excess of 
 CO 2 is breathed, will CO 2 be absorbed backwards, and the thick 
 line pass downwards as well as to the left. 
 
 Tlje discovery that oxygenation of the haemoglobin helps to 
 turn out CO 2 from blood gives us the key to the proper interpreta- 
 tion of the fact that, as was found by ourselves in human experi- 
 ments, and earlier by Werigo, 10 and by Bohr and Halberstadt, 11 
 more CO 2 is given off into the air of the lungs when oxygen is 
 present. Thus in Halberstadt's experiments it was found that if 
 one lung was ventilated with air, and the other with hydrogen, 
 the lung ventilated with air gave off nearly 50 per cent more CO 2 
 than the lung ventilated with hydrogen. This result is precisely 
 what would be expected in view of the facts just described; but 
 as Bohr was misled by the apparent results of his experiments 
 with blood outside the body, he wrongly attributed Halberstadt's 
 and Werigo's results to the supposed fact that in presence of air 
 there is a large formation of CO 2 in the lungs, owing to a process 
 of oxidation occurring there. As will be shown later, hardly any 
 formation of CO 2 occurs in the lungs. 
 
 In a quite recent paper Parsons 12 has investigated mathemati- 
 cally the form of the absorption curve of blood for CO 2 on the 
 theory that the blood is a chemical system consisting of carbonic 
 acid and what may be regarded as one other free acid (consisting 
 of the proteins present) with a fixed concentration of available 
 alkali distributed between them. This fixed concentration he 
 estimated from blood-ash analyses and in other ways, to be about 
 4.5 x io~ 2 N. He found that the form of the curve given by calcula- 
 
 10 Werigo, Pfliiger's Archiv., LI, p. 321, 1892. 
 
 11 Bohr, Nagel's HancLb. der Physiol., I, p. 208. 
 13 Parsons, Journ. of PAysiol., LIII, p. 42, 1919. 
 
RESPIRATION 
 
 93 
 
 tion corresponded satisfactorily with the curves which both we 
 and he had obtained experimentally for human blood. This is 
 illustrated in Figure 27, reproduced from his paper. We had not 
 attempted to calculate the form of the curve, as several proteins 
 are involved in the chemical system; but by the simplifying as- 
 sumption which he made Parsons overcame this difficulty. 
 
 Pco, 
 
 10 20 30 40 50 60 7O 8O 9O IOO 110 120mm. 
 
 Figure 27. 
 
 Comparison between the theoretical curve and experimental results for 
 completely reduced blood of Haldane. 
 
 In the previous chapter we have seen that, other things being 
 equal, a rise of CO 2 pressure shifts the dissociation curve of oxy- 
 haemoglobin to the right if the curve is represented as in Figure 
 19 or 28. In the living body the pressure of CO 2 is constantly 
 rising as the blood becomes more and more venous in its passage 
 through the systemic capillaries. The data embodied in Figure 25 
 gave us the means of calculating this rise, and it will be seen that 
 it is much less than previously existing knowledge would have 
 led us to believe. Figure 27 shows the oxygen dissociation curve 
 of my own blood in the living body, calculated from Figure 26, on 
 the assumption that the shifting of the curve to the right is pro- 
 portional to the increase of CO 2 pressure in the blood as it passes 
 along the systemic capillaries. 
 
 Bohr believed that the shifting of the dissociation curve to the 
 right by the influence of increasing CO 2 pressure in the systemic 
 
94 
 
 RESPIRATION 
 
 capillaries is an important factor in facilitating the unloading of 
 oxygen from the blood ; and this line of argument has been further 
 elaborated by Barcroft. The actual shifting is, however, very 
 small under normal conditions, and of much less physiological 
 importance than the effect of the shifting of the CO 2 absorption 
 curve in consequence of reduction of oxyhaemoglobin. 
 
 100 
 
 h 
 
 S 80 
 
 1 
 
 C 70 
 60 
 
 ^40 
 
 ;r 30 
 5 20 
 
 *,. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^^ 
 
 ^*^ 
 
 -^ 
 
 = 
 
 _^ BB 
 
 
 
 ^OK. 
 
 ?=. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^. 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^i 
 
 * 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 y 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 , 
 
 7 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ' / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 > / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 7 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / f 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 '/ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 '/ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 X 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Pressure of oxyyen in Percentage of one atmosphere. 
 
 Figure 28. 
 
 The thick line shows the dissociation curve of oxyhaemoglobin in the blood of 
 J. S. H. and C. G. D. in the presence of 40 mm. pressure of COa. The thin line 
 represents the dissociation curve of oxyhaemoglobin in the blood of J. S. H. and 
 C. G. D. within the body. 
 
 We are now in a position to interpret much more completely 
 the facts concerning the regulation of breathing by small varia- 
 tions in the alveolar CO 2 pressure. How very small the mean 
 variations are, we have already seen. On the other hand the 
 breathing is constantly being interrupted or interfered with in 
 one way or another during ordinary occupations, such as speak- 
 ing or singing, and the breath can be held for a few seconds with- 
 out any noticeable air hunger being produced. During these 
 interferences the alveolar CO 2 pressure must be constantly rising 
 and falling on either side of the normal limit, but the physiological 
 effect seems almost nil, and to popular imagination it seems as 
 if the breathing, instead of being regulated so rigorously as was 
 shown to be the case in the second chapter, is hardly regulated at 
 
RESPIRATION 95 
 
 all. We are also familiar with instructions to increase the breathing 
 so as to "improve the oxygenation of the blood" and with quack 
 advertisements based on the same idea. How does it come about 
 that although the regulation is so exact on the average, yet 
 temporary deviations from this average exactitude do not cause 
 any discomfort? How is it, also, that when the production of CO 2 
 is suddenly increased to perhaps ten times the normal, as on a 
 sudden muscular exertion, yet the breathing responds gradually 
 and easily to the new conditions? 
 
 The answer to this question is that there are physiological buf- 
 fers between the stimulus of increased production of CO 2 , or 
 increase in the alveolar CO 2 pressure, and stimulation of the 
 respiratory center, and that if it were not so the respiratory center 
 would work in a jerky, irregular, and extremely inconvenient 
 manner. The first of these buffers is the large volume of air always 
 present in the lungs. Thus in my own case the mean volume of air 
 in the lungs at the end of inspiration during rest is 3650 cc., 
 measured dry at oC, including about 3000 cc. of saccular al- 
 veolar air containing about 5.6 per cent of CO 2 . Let us assume 
 that the breath is held at the end of inspiration during rest, and 
 consider what happens. About 250 cc. of CO 2 would be normally 
 given off per minute, or 20 cc. in 5 seconds; and if the latter 
 quantity were given off with the breath held the mean CO 2 pres- 
 sure in the lung air would rise by 0.6 per cent in 5 seconds. But, 
 as will be shown later, about 700 cc. of blood will pass through the 
 lungs in 5 seconds, and as the arterial blood will be more highly 
 saturated with CO 2 if the alveolar CO 2 percentage rises, some of 
 the CO 2 which would ordinarily have been given off will be 
 dammed back in the blood. Figure 25 shows that for every rise of 
 2.5 mm. or .36 per cent in the alveolar CO 2 pressure the blood will 
 take up, or hold back, i volume per cent of CO 2 . Hence the actual 
 rise in the mean CO 2 pressure within the lungs cannot be more 
 than about 0.4 per cent in the 5 seconds during which the breath 
 is held. The net result is that about two-thirds of the CO 2 which 
 the suspension of the breathing prevents from escaping from the 
 body is temporarily accommodated in the lung air, which thus 
 acts as a first buffer for preventing too sudden a change in the 
 arterial CO 2 pressure. 
 
 A second buffer is provided by the tissues and lymph in and 
 around the respiratory center itself. So far as we know the re- 
 action in all parts of the body is slightly alkaline, just as in the 
 blood; and the tissues and lymph have, like the blood, a con- 
 
9 6 
 
 RESPIRATION 
 
 siderable capacity for absorbing CO 2 . Hence it will take some 
 time for the blood to saturate the tissues and lymph up, or de- 
 saturate them down, to a new CO 2 pressure. Here we have a 
 second, and very powerful, buffer action, tending to smooth out 
 the influence on the respiratory center and other tissues of all 
 variations of short duration in the CO 2 pressure of the arterial 
 blood, and also to prolong the influence of variations of longer 
 duration. 
 
 This subject was investigated by Douglas and myself. 13 The 
 following table shows the results we obtained on determining the 
 alveolar CO 2 pressure at various times after holding the breath. 
 In order to throw out disturbing effects due to the action of oxy- 
 gen want on the respiratory center, some of the experiments were 
 made after a few normal breaths of oxygen had been taken, so that 
 there should be plenty of oxygen in the lungs up to the end of the 
 stoppage of respiration. 
 
 PRESSURE IN 
 
 MM. OF 
 
 HG. IN ALVEOLAR AIR 
 
 
 C0 2 
 
 2 
 
 At end of period of holding breath for 30" 
 
 49.2 
 
 62.6 
 
 At fifth expiration following 
 
 29.1 
 
 
 
 At ninth expiration following 
 
 31-5 
 
 
 
 At twelfth expiration following 
 
 32.0 
 
 
 
 At twentieth expiration following 
 
 33-8 
 
 
 
 At thirtieth expiration following 
 
 37-0 
 
 
 
 At fortieth expiration following 
 
 38.8 
 
 
 
 At fifth expiration after holding 40" 
 
 28.4 
 
 117. 
 
 At eighth expiration following 
 
 29.4 
 
 
 At end of holding breath for 130" after oxygen 
 
 61.9 
 
 274. 
 
 At sixth expiration afterwards 
 
 24.8 
 
 
 At twentieth expiration afterwards 
 
 33-3 
 
 
 
 At fortieth expiration afterwards 
 
 31-2 
 
 
 
 Normal average 
 
 39-75 
 
 105. 
 
 Figure 29 is a stethographic tracing of the respirations during 
 an experiment, and shows that the breathing returns gradually 
 to normal after the hyperpnoea following the stoppage. 
 
 The table is extremely instructive, and shows very clearly what 
 a long period of increased breathing, with the alveolar CO 2 pres- 
 sure distinctly below normal, is required in order to compensate 
 
 13 Douglas and Haldane, Journ. of Physwl., XXXVIII, p. 420, 1919. 
 
I 
 
 5 ^ 
 
 ss 
 
 m .s 
 
 bi ft 
 .SvO 
 
 p 
 t" 
 
 :* 
 
 Is 
 
 .-8. 
 
 i 
 
 a e- 
 
98 RESPIRATION 
 
 for the cumulative action of the stoppage of breathing. After the 
 long stoppage of 130 seconds the breathing and alveolar CO 2 
 pressure had not nearly returned to normal, even after the fortieth 
 breath following the stoppage. 
 
 Figure 30 shows the converse experiment. Forced breathing 
 was continued two minutes so as to wash out CO 2 from the lungs, 
 arterial blood, and respiratory center; and oxygen had been taken 
 into the lungs, so as to cut out the effects of want of oxygen. The 
 apnoea lasted 4% minutes, and an alveolar sample (the taking of 
 which is recorded on the tracing and somewhat disturbs it) was 
 obtained as soon as the slightest inclination to breathe was noticed. 
 It will be seen that the CO 2 percentage in this sample was 7.12 
 per cent (51.5 mm. of CO 2 pressure) a value far above the normal 
 40 mm. required to excite the center under normal conditions. 
 Separate experiments showed that by the end of two minutes of 
 forced breathing the alveolar CO 2 pressure had fallen to about 
 13 mm. and during the apnoea rose to normal again at the end of 
 2j^ minutes. During the last 2 minutes the alveolar CO 2 pressure 
 was above normal ; but sufficient CO 2 had not accumulated in the 
 tissues of the respiratory center to stimulate it, till the alveolar 
 CO 2 pressure had gradually risen to 51.5 mm. At this point the 
 center, which had now just reached its normal CO 2 pressure, began 
 to work quietly and smoothly, reducing the alveolar CO 2 pressure 
 to normal, and picking up the normal regulating activity. The 
 breathing cannot indicate a gradual return of the CO 2 pressure 
 in the center to normal, corresponding to the gradual return in 
 Figure 29, since, as is shown by the experiments described in 
 Chapter II, complete apnoea results from a fall of 0.2 per cent or 
 1.5 mm. of the CO 2 pressure in the respiratory center. 
 
 The apnoea following forced breathing can be temporarily 
 interrupted by sending a block of blood highly charged with CO 2 
 to the respiratory center. The effect of this is shown in Figure 31. 
 As soon as the breathing and the "apnoeic" venous blood return- 
 ing to the lungs have removed the extra CO 2 introduced into the 
 lungs the apnoea returns again. 
 
 The washing out of CO 2 from the body during forced breath- 
 ing, and its gradual reaccumulation during the next ten or twenty 
 minutes, were strikingly illustrated in some experiments carried 
 out by Boothby. 14 Thus in an experiment on myself he found that 
 during i*/2 minutes of forced breathing I had removed about 
 1,400 cc. extra of CO 2 from the body. During the subsequent ap- 
 
 14 Boothby, Journ. of Physwl., XLV, p. 328, 1912. 
 
RESPIRATION 
 
 99 
 
 noea of 2 minutes about 600 cc. of CO 2 were regained, and about 
 200 cc. more during two minutes of periodic breathing which 
 followed. The remainder was regained during the following six 
 or eight minutes. In this latter period the alveolar CO 2 pressure 
 
 Figure 31. 
 
 Effect of a breath of air containing 9.0 per cent of CO 2 during apnoea follow- 
 ing forced breathing. Crosses show inspiration and expiration of breath. After 
 an interval there are three deep, and two shallow, breaths, followed by a long 
 apnoeic interval, after which the usual periodic breathing begins. To read 
 from left to right. Time-marker = i second. 
 
 was practically normal, but the respiratory quotient very low, in 
 correspondence with the very high respiratory quotient during 
 the forced breathing. 
 
 What approximately happens to the CO 2 pressure in the al- 
 veolar air and respiratory center is represented in Figures 32 and 
 
 -so 
 
 -40 
 
 -30 
 
 -10 
 
 Time in Minutes 
 
 Figure 32. 
 
 Approximate variations in CO 2 pressure of arterial and venous blood dur- 
 ing and after forced breathing of oxygen for two minutes. 
 
 33. The pressure of CO 2 in the respiratory center is assumed to be 
 about equal to that of the mixed venous blood, though it is prob- 
 ably lower. 
 
100 
 
 RESPIRATION 
 
 The very powerful steadying influence on the CO 2 pressure of 
 the capacity of the tissues for taking up CO 2 is evident from these 
 figures. In consequence of this influence, and in a much less degree 
 that of the reserve of air in the lungs, variations of short duration 
 in the alveolar CO 2 pressure hardly count, although even the 
 slightest variations of a more prolonged character count a great 
 deal. 
 
 2 4 
 
 T/ne in Minutes 
 
 Figure 33. 
 
 Approximate variations in COa pressure of arterial 
 and venous blood during and after holding the breath 
 for 130 seconds with oxygen. 
 
 On examining Figure 32 it will be seen that, although the 
 venous CO 2 pressure is below that of the alveolar air during most 
 of the apnoea, CO 2 is being given off all the time into the alveolar 
 air. This is due to the effect of oxygenation in decreasing the 
 capacity for CO 2 and thus raising its pressure in the blood. This 
 effect is explained by the fact that the thick line of Figure 26 will 
 be inclined to the left, as very little CO 2 is being given off by the 
 tissues, impoverished as they are of CO 2 by the forced breathing. 
 
 In order to realize how important the steadying influence just 
 mentioned is, we have only to turn to what happens when want 
 of oxygen, instead of CO 2 , is exciting the center. Oxygen is no 
 
fil 
 
 18 
 
 T3 00 
 O >n 
 "C in 
 
 I! 
 
 o> H 1 
 
 pill 
 
 - 
 
 S8 
 
 *J 
 
 "sl 
 
 g ,s 
 
 Us 1 
 
 " s^ 
 
 o S S 
 
102 
 
 RESPIRATION 
 
 more soluble in the tissues and lymph than in water. They have 
 thus practically no power of storing free oxygen. In the course of 
 our investigations on the effects of want of oxygen it became 
 evident that the center works very jerkily when excited by want 
 of oxygen, and the subject was studied in further detail by Doug- 
 las and myself. 15 We found that the effects of regulation of the 
 center by oxygen want could be observed very conveniently at 
 the end of the apnoea caused by forced breathing of ordinary air. 
 When apnoea is produced by forced breathing of air for about 
 two minutes, the oxygen percentage in the lungs runs down very 
 low before the pressure of CO 2 in the respiratory center has nearly 
 risen to its normal value. In some subjects there is an alarming 
 
 fr-4 
 
 VK. Mf 
 
 mv.'i\Mi'.',lr, 
 
 Figure 35. 
 
 Variations in alveolar gas pressures after forced breathing 
 for two minutes. Thin line = oxygen pressure, thick line = CO 2 
 pressure. Double line = normal alveolar CO 2 pressure. The 
 actual breathing is indicated at the lower part of the figure. 
 
 appearance of blueness in the face before any desire to breathe is 
 felt. Ultimately, however, the stimulus of oxygen want (together 
 with the subliminal CO 2 stimulus) suffices to start the breathing. 
 But the first four or five breaths greatly raise the alveolar oxygen 
 percentage and thus quiet the center down again, so that apnoea 
 again follows, which is again followed by breathing and subse- 
 quent apnoea, this periodic rising and dying away of the breath- 
 ing going on for about five minutes, as shown in Figure 34, though 
 not all subjects react alike. 
 
 Figure 35 shows the variations of the alveolar oxygen and CO 2 
 
 "Douglas and Haldane, Journ. of PAysiol., XXXVIII, p. 401, 1909. 
 
RESPIRATION 103 
 
 pressure, as determined in samples of alveolar air. Reference to 
 Figure 31 shows that at no time during the periodic breathing is 
 the CO 2 pressure in the respiratory center more than just suf- 
 ficient to excite the center by itself. 
 
 It is very easy to see what has been happening. The oxygen 
 want caused by the partially reduced blood coming from the 
 lungs at the end of the apnoea has, along with the CO 2 present, 
 sufficed to excite the center; but this oxygen want is at once re- 
 lieved by the breaths which follow, since the oxygen pressure in 
 the lungs is raised beyond the exciting point. The result is a 
 prompt return of the apnoea, till the oxygen in the alveolar air 
 again returns to the stimulating point. The respiratory governor 
 is "hunting" just as the governor of a steam engine or turbine 
 hunts if there is no heavy flywheel or other steadying influence. 
 The chief flywheel of the respiratory center is the great storage 
 
 Figure 36. 
 
 Breathing through soda lime and long tube. Sample of alveolar 
 air at the end of a dyspnoeic period, Oz= 8.70 per cent, CO2= 5.48 
 per cent. 
 
 capacity in the tissues for CO 2 . There is no such storage capacity 
 in connection with oxygen, so the flywheel has disappeared. 
 
 When slight oxygen want, and not merely excess of CO 2 , is 
 exciting the center, the breathing very readily becomes periodic. 
 To realize this condition in a permanent manner we only had to 
 breathe in and out through a tin of soda lime with a piece of hose 
 pipe of variable length attached on the far side, so as to give a 
 suitable dead space. By this means the alveolar oxygen pressure 
 can be reduced to any required extent. Figure 36 shows the effect 
 of such an arrangement. This effect is at once knocked out if oxy- 
 gen is breathed. 
 
104 RESPIRATION 
 
 Some years ago it was discovered by Pembrey and Allen 16 that 
 the well-known pathological form of periodic breathing named 
 after Drs. Cheyne and Stokes, who described it (though it was 
 previously described by John Hunter), is abolished by giving the 
 patient pure oxygen to breathe. This observation indicates with 
 great certainty that ordinary pathological Cheyne-Stokes breath- 
 ing is caused also by want of oxygen participating in the excita- 
 tion of the center. Pathological periodic breathing and that of 
 hibernating animals will be discussed later. 
 
 The normal pressure of oxygen in the alveolar air is about 
 100 mm. or 13.1 per cent of an atmosphere. On looking at the 
 dissociation curve of oxyhaemoglobin in human blood (Figure 
 20) it will be seen that a fall of 4 per cent of an atmosphere, or 
 30 mm., makes very little difference to the saturation of the haemo- 
 globin. Nor has such a fall any appreciable influence on the rest- 
 ing breathing at the time. It is thus evident that, although there 
 is no appreciable store of readily available oxygen in the liquids 
 of the body outside the red corpuscles and certain muscles which 
 contain a little haemoglobin, there is a store of oxygen, available 
 without any inconvenience, in the air of the lungs. If the breathing 
 is temporarily stopped during some occupation this store is drawn 
 on. Thus if the breath is held for half a minute the oxygen runs 
 down by about 4 per cent in the alveolar air during rest ; but under 
 normal conditions it is quite impossible to hold the breath long 
 enough to imperil seriously the oxygen supply to the tissues. In 
 spite of the gradual manner in which, as we have just seen, CO 2 
 acts on the respiratory center, there is never, except under very 
 artificial conditions, any considerable oxygen want. The com- 
 paratively large volume of air which is always in the lungs gives 
 sufficient oxygen storage to guard against the temporary want of 
 oxygen. Were this amount of air much less the danger would be 
 always present, and, as we shall see later, this danger or incon- 
 venience is present at high altitudes, when the mass of oxygen in 
 the lungs is greatly diminished. At a high altitude one cannot 
 hold the breath for more than a few seconds without feeling an 
 imperative desire to breathe, and such operations as shaving, or 
 reading a barometer, are thus rendered troublesome. Nature sees 
 to it that ordinary mortals who live under a pressure of about one 
 atmosphere carry about sufficient oxygen in their lungs to pre- 
 vent oxygen want; and there seems to be some evidence that 
 
 16 Pembrey and Allen, Journ. of Physiol., XXXII, Proc. Physiol. Soc., p. xviii, 
 1905 ; also Medico- Chirurg. Trans., XI, p. 49, 1907. 
 
RESPIRATION 105 
 
 persons who inhabit very high parts of the earth develop a greatly 
 increased chest capacity. 
 
 Addendum. The account given in this chapter of the manner in 
 which CO 2 is carried by the blood represents what I have taught 
 for many years, and is largely based, as mentioned above, on the 
 teaching of Pfliiger and Zuntz. A very different view of the 
 subject has recently been presented by Buckmaster, Bayliss, and 
 others. According to this view the extra CO 2 taken up in the 
 venous blood is combined, not with alkali, but with haemoglobin, 
 and may also be in part adsorbed by haemoglobin and other 
 proteins. As evidence that haemoglobin and other proteins do 
 not play the part of weak acids in expelling CO 2 from its combi- 
 nation with alkali, Buckmaster cites experiments in which he 
 found that, contrary to Pfliiger's statement, blood or haemo- 
 globin is not capable of expelling CO 2 from a weak carbonate 
 solution in the vacuum of a blood pump at body temperature. 1 
 It seems to me that these experiments were fallacious because the 
 blood was neither boiled nor shaken. Boiling, shaking, or bubbling 
 is necessary to remove the CO 2 . When Pfliiger's experiment was 
 repeated in a simple form by Adolph in my laboratory the ex- 
 pulsion of CO 2 from sodium carbonate by blood was found to 
 occur quite readily. 2 As already mentioned, Buckmaster's con- 
 tention that haemoglobin gives a characteristic spectrum with 
 CO 2 was also found to be incorrect. 
 
 The supposition that an extra amount of gas is adsorbed by the 
 proteins of blood has no basis. The careful experiments of Bohr 
 and other previous observers show clearly that apart from chemi- 
 cal combination blood takes up, not more, but considerably less, 
 gas than an equal volume of water. The only apparent exception 
 to this rule was the fact that oxygenated blood (but not reduced 
 blood) yields slightly more nitrogen than the quantity calculated 
 from its estimated solubility. The existence of this small surplus 
 was confirmed by Buckmaster and Gardner. 3 The apparent surplus 
 is almost certainly due to what is a rather common source of 
 slight error in gas analysis. When the gas pumped off from oxy- 
 genated blood is analyzed, the first step is to bring the gas into 
 contact with potash solution to absorb the CO 2 . When this is ab- 
 sorbed a gas mixture consisting almost wholly of oxygen is left 
 in contact with the potash solution. But the latter is saturated 
 
 1 Buckmaster, Journ. of Phystol., LI, p. 105, 1917. 
 
 'Adolph, Journ. of Physiol., LIV, Proc. Physiol. Soc. p. XXXIV, 1920. 
 
 8 Buckmaster and Gardner, Journ. of Physwl., XLIII, p. 401, 1912. 
 
106 RESPIRATION 
 
 with air, and as a consequence nitrogen diffuses from the potash 
 solution into the gas mixture, while oxygen diffuses into the 
 potash solution. The consequence is that the residue of nitrogen 
 found in the gas after the oxygen has been absorbed is greater 
 than was originally present in the gas. This source of error is 
 absent if little or no oxygen is present in the gas pumped off from 
 the blood. We can thus explain why no extra nitrogen has been 
 found in reduced blood. 
 
 Bayliss 4 contends that the bicarbonate and the plasma proteins 
 present in blood play no part in the physiological carriage of 
 CO 2 between the tissues and the lungs, and that haemoglobin is 
 alone concerned in the carriage, since it does not, under actual 
 physiological conditions, compete as an acid with CO 2 for the 
 alkali available in the blood. The experiments cited in support of 
 this conclusion seem to me quite unconvincing; and if it were cor- 
 rect we should expect to find that blood saturated at the alveolar 
 partial pressure with CO 2 would contain more combined CO 2 than 
 a solution of bicarbonate of the same strength in titratable alkali 
 as the blood. Actually, the blood, especially at body temperature, 
 contains far less combined CO 2 . It seems quite impossible to 
 reconcile Bayliss' theory with this fact ; and I cannot see how any 
 other theory than that given in the first part of this chapter is 
 capable of interpreting the facts as a whole. It may be that a small 
 amount of CO 2 is combined with free haemoglobin ; but it seems 
 evident that under physiological conditions haemoglobin and 
 other proteins act, for all practical purposes, simply as weak acids. 
 It is in virtue of this action, and the more powerful action of 
 oxyhaemoglobin than reduced haemoglobin as an acid, that blood 
 functions so efficiently as a physiological carrier of CO 2 . Campbell 
 and Poulton, who entirely disagree, and on substantially the same 
 grounds as I do, with the conclusions of Buckmaster and Bayliss, 
 have recently shown that an artificial mixture of dialysed cor- 
 puscles and dilute sodium bicarbonate solution takes up, within 
 physiological limits of CO 2 pressure, much less CO 2 than the 
 bicarbonate alone holds. 5 
 
 For the sake of simplicity I did not discuss separately the 
 action of plasma and corpuscles in combining with CO 2 ; but much 
 attention has been given recently to this subject. Zuntz 6 pointed 
 out that when plasma or serum is separated from blood collected 
 
 4 Bayliss, Journ. of Physiol., LIII, p. 162, 1919. 
 
 5 Campbell and Poulton, Journ. of Physiol., LIV, p. 157, 1920. 
 
 6 Zuntz, Hermann's H ' andbuch der Physiol., IV, 2, p. 77, 1882. 
 
RESPIRATION 107 
 
 as it flows from a vessel, the corpuscles are capable of taking up 
 from pure CO 2 > more combined CO 2 than an equal volume _pf 
 the plasma. If, on the other hand, the blood is artificially saturated 
 with pure CO 2 , or air containing a high percentage of CO 2 , and 
 then separated into plasma and corpuscles, the plasma contains 
 more combined CO 2 than the corpuscles. He concluded that alkali 
 previously combined with haemoglobin in the corpuscles combines 
 with CO 2 when a high concentration of the latter is present, and 
 passes out as bicarbonate into the plasma. Further investigation 
 of this phenomenon by Giirber 7 showed that alkali does not pass 
 out of the corpuscles, but acid passes in, leaving the corresponding 
 alkali behind in the plasma. The walls of the corpuscles seem, 
 therefore, as Hamburger 8 in particular has pointed out, to be 
 practically impermeable to sodium and potassium ions, but per- 
 meable to chlorine and other anions. Hence the proportions of 
 alkali to chlorine, etc., in the plasma depend upon the corpuscles, 
 and are regulated by them according as the pressure of CO 2 in 
 the blood rises or falls. Yandell Henderson and Haggard, who 
 have quite recently investigated this phenomenon closely from 
 the physiological standpoint, point out what striking effects this 
 regulating action may produce. 9 During forced breathing, for 
 instance, the weakly combined alkali of the plasma may be con- 
 siderably diminished, although the total weakly combined alkali 
 in the blood need not necessarily be altered. 
 
 The relation of the corpuscles to the available alkali in the 
 plasma suggests at once the question whether there is not a similar 
 relation as regards other tissue elements. Henderson and Haggard 
 showed that with vigorous and continued artificial respiration 
 the available alkali in the whole blood, and not merely in the 
 plasma, diminishes greatly, and that this diminution is accompa- 
 nied by signs of irretrievable damage to the body. This suggests 
 excessive draining of acid from the tissue elements with the 
 result that the whole body suffers, although the alkalinity of the 
 blood itself is partly prevented from falling. The matter will, 
 however, be discussed further in Chapter VIII. 
 
 7 Giirber, Sitz-der. d. physik-med. Gesellsch. zu Wurzburg, p. 28, 1895. 
 
 Anionenwanderungen in serum und Blut unter den einfluss von CO 2, saure und 
 alkali. Biochem. Zett. Vol. 86, p. 309-324, 1918. 
 
 9 Haggard and Henderson, Journ. of Biol. Chem., XLV, p. 199, 1920. 
 
CHAPTER VI 
 The Effects of Want of Oxygen. 
 
 IN the higher organisms, as Paul Bert first pointed out, the im- 
 mediate cause of death of the body as a whole is practically always 
 want of oxygen, owing to failure of the circulation or breathing. 
 This fact arises from the circumstance that the body has hardly 
 any internal storage capacity for oxygen, but depends from 
 moment to moment for its supply from the air. We can deprive 
 the body for long periods of its external supplies of food or water, 
 or we can prevent for some time the excretion of urinary products 
 or even of carbon dioxide, but we cannot interfere with the supply 
 of oxygen to the blood without producing at once the most threat- 
 ening symptoms. Almost the only appreciable storage capacity 
 for surplus oxygen is in the lungs. In virtue of this small store 
 breathing can be prevented for about 1*4 minutes in a man at 
 rest and previously breathing normally before urgent symptoms 
 of oxygen want appear ; but if the oxygen in the lungs and blood 
 is rapidly washed out by breathing pure nitrogen, nitrous oxide, 
 or other gas free from oxygen, loss of consciousness occurs almost 
 at once. Lorrain Smith and I found that even with quiet breathing 
 of pure hydrogen, so that some time was needed to wash out the 
 lungs, sudden and complete loss of consciousness was produced 
 within 50 seconds. 
 
 Even when the oxygen supply, though not cut off, is insuffi- 
 ciently free, the ill effects develop rapidly, and may very soon 
 become serious. Hence few things are of more importance in 
 practical medicine than the causes and effects of want of oxygen. 
 
 Want of oxygen in the systemic circulation may be produced 
 either by deficiency in the available oxygen in the arterial blood, 
 or by abnormal slowing of the circulation, so that too much of the 
 available oxygen is used up in the systemic capillaries. It will be 
 convenient to consider first the effects of want of oxygen or "an- 
 oxaemia," and afterwards discuss the various ways in which it 
 may be produced. 
 
 The effects of anoxaemia can be observed most conveniently in 
 persons breathing air from which part of the oxygen has been 
 removed without the addition of any other gas producing by 
 itself a physiological effect; or in persons breathing pure air at 
 
RESPIRATION 109 
 
 reduced atmospheric pressure. In either case the partial pressure 
 of the oxygen breathed is reduced, and the haemoglobin tends to 
 become imperfectly saturated with oxygen in the lungs in cor- 
 respondence with the dissociation curve for the oxygen in human 
 blood (Figure 20). 
 
 The effects on the breathing have already been touched upon 
 in Chapter II, but must now be discussed fully. In most persons 
 the percentage of oxygen in the air breathed, or the barometric 
 pressure, must be reduced by about a third before any evident 
 effect on the breathing is produced at the time; and this effect 
 differs according as the reduction is produced rapidly or slowly. 
 With a greater reduction the contrast in this latter respect is still 
 more marked. With rapid reduction there is at first a quite notice- 
 able increase in the depth, and also in the frequency, of breathing. 
 In the course of several minutes, however, the increase diminishes 
 markedly. This phenomenon and the causes of it were described 
 and investigated by Poulton and myself. 1 We found that the in- 
 creased breathing causes, as could be anticipated, a distinct fall 
 in the alveolar CO 2 pressure. As a consequence, more CO 2 than 
 usual is washed out of the blood, and the respiratory quotient, or 
 ratio of the volume of CO 2 given off to that of oxygen absorbed, 
 is increased. Thus it increased from the normal of about 0.8 to as 
 much as 2.8 when there was sudden and considerable oxygen de- 
 ficiency. Soon, however, the extra discharge of CO 2 from the 
 blood began to cease and there was only a slight further fall in 
 the alveolar CO 2 pressure. Part passu the breathing quieted down 
 so as, in spite of the diminished discharge of CO 2 , to maintain a 
 certain level of alveolar CO 2 pressure, this level being of course 
 below the normal level. At the same time the alveolar oxygen 
 pressure dropped, since the lung ventilation had diminished while 
 the rate of absorption of oxygen remained undiminished. The 
 drop in alveolar oxygen pressure tended, of course, to increase the 
 symptoms of want of oxygen and thus prolong the period of in- 
 creased breathing; but finally a balance was struck, for the time 
 at any rate. When the deficiency of oxygen was produced quite 
 gradually the initial marked increase of breathing was not notice- 
 able, as the extra CO 2 was washed out gradually. 
 
 By further experiments, we found that the new and lower 
 level of alveolar CO 2 pressure had become the regulating level 
 for the atmosphere breathed. That is to say, a small increase above 
 this level caused a great increase in the breathing, while a small 
 
 *Haldane and Poulton, Journ. of Physwl., XXXVII, p. 390, 1908. 
 
1 10 RESPIRATION 
 
 diminution caused apnoea, just as when pure air is breathed. It 
 was evident, therefore, that the CO 2 pressure, though at a lower 
 level, was controlling the breathing still. The primary marked 
 increase in the breathing was due to the alveolar CO 2 pressure 
 and the CO 2 pressure in the whole of the body being above the 
 new level, and the quieting down of the breathing was due to 
 the gradual washing out of CO 2 from the whole body till the at- 
 tainment of the new normal level, which was itself determined 
 by the alveolar oxygen pressure. 
 
 A fuller discussion of these facts, and of the ultimate physio- 
 logical response to long-continued slight anoxaemia, must be 
 postponed to Chapter VII, but meanwhile it is evident that they 
 throw a new light on the physiology of breathing. Hitherto we 
 have considered the amount of lung ventilation as if it were de- 
 termined solely by a certain excess of partial pressure of CO 2 in 
 the arterial blood; but now we see that the excess is something 
 variable and dependent, for one thing, on the pressure of oxygen 
 in the arterial blood, just as the action of the Hering-Breuer re- 
 flex depends, not merely on the amount of distention or collapse 
 of the lungs, but also on the pressure of CO 2 in the arterial blood. 
 Similarly the action of want of oxygen on the breathing depends 
 on the CO 2 pressure. On how many other factors which together 
 make up "normal conditions" the action of CO 2 or want of oxy- 
 gen on the respiratory center depends we do not know. We always 
 find normal conditions in a healthy organism, and we are there- 
 fore apt to overlook their unknown complexity. If we represented 
 the relation between arterial CO 2 pressure, oxygen pressure, and 
 lung ventilation in the form of an equation, this equation would 
 only be valid under conditions otherwise normal. In other words 
 an unknown constant C would have to be set down in the equation. 
 
 That this constant exists during life in other words that living 
 organisms maintain fundamental normals of structure and ac- 
 tivity representing the <ris of Hippocrates is one basis of 
 biological science. Apart from this basis physiology would be a 
 mere chaos of unconnected "bio-physical" and "bio-chemical" 
 fragments. 
 
 The effect produced on the breathing by a given reduction in 
 the oxygen pressure of the inspired air or alveolar air varies con- 
 siderably in different individuals. Some respond much more 
 readily by increased breathing than others, and for this reason 
 seem to be better protected against the other and more serious ef- 
 fects of want of oxygen, since the increased breathing raises the al- 
 
RESPIRATION III 
 
 veolar oxygen percentage. In some persons a lowering by as little 
 as 5 per cent in the oxygen percentage of the inspired air will 
 sensibly increase the breathing, but in most persons a lowering of 
 at least 7 per cent (i.e., from 20.94 to 14) is needed to produce a 
 measurable effect, while in others very little effect is produced 
 before consciousness is lost from want of oxygen. It is thus for 
 many persons peculiarly dangerous to pass into an atmosphere in 
 which the oxygen percentage is very low, or to ascend to a very 
 great height in a balloon, since increased breathing may give 
 very little warning, particularly if the change is gradual, so that 
 the extra CO 2 is blown off gradually. 
 
 It was discovered in 1908 by Yandell Henderson 2 that when 
 effective artificial respiration in an animal has been pushed to 
 excess for some time, so that the pressure of CO 2 in the blood and 
 tissues is very greatly reduced, there is not only a prolonged 
 succeeding apnoea, but the animal dies of want of oxygen without 
 attempting to draw a single breath. The artificial respiration must 
 be performed somewhat forcibly, by means of a suction and ex- 
 haust pump ; and the reason for this will be evident from what has 
 already been said in Chapter III as to the control of the chest- 
 movements by the Hering-Breuer reflex during artificial respira- 
 tion produced by ordinary means. 
 
 This important experiment shows that when the CO 2 pressure 
 is reduced below a certain point in the respiratory center the latter 
 ceases to respond to even the extremest stimulus of want of oxy- 
 gen. The apnoea produced in the ordinary way by voluntary 
 forced breathing is terminated, as shown in Chapter V, by the 
 combined stimulus of CO 2 and want of oxygen, and in some 
 persons the oxyhaemoglobin in the arterial blood runs down so 
 low that the lips and face become alarmingly blue before breath- 
 ing begins. In the case of Poulton, for instance, his face presented 
 such an alarming appearance when he demonstrated our experi- 
 ments at a meeting of the Physiological Society that one or two 
 members of the Society could hardly be restrained from applying 
 artificial respiration on the spot. In my own case, and that of many 
 others, the blueness is much less marked, although, as already 
 shown, the termination of the apnoea is quite clearly due to want 
 of oxygen, and not merely to accumulation of CO 2 . 
 
 It is evident from the foregoing account that the respiratory 
 response to the stimulus of uncomplicated oxygen want is a com- 
 plex one. The anoxaemia tends to increase the breathing, but the 
 
 'Yandell Henderson, Amer. Journ. of Physiol., XXI, p. 142, 1908. 
 
112 RESPIRATION 
 
 increased breathing, by washing out CO 2 , checks this increase 
 very quickly, so that the net result for the time is only a small 
 increase. Where the anoxaemia is only slight this net increase will 
 be practically inappreciable, and this, as will be shown in Chap- 
 ter VIII, is due, not to the fact that there is no appreciable anox- 
 aemia, but to the masking of the natural response to anoxaemia 
 by the opposite response to the washing out of CO 2 . After a suffi- 
 cient interval of time the former response, as we shall see, becomes 
 unmasked by the compensation of the latter response, so that in 
 the long run there is a very definite response of the breathing to 
 even a very small fall in the oxygen pressure of the inspired air. 
 
 When diminution in the oxygen pressure of the inspired air is 
 accompanied by a corresponding increase in the pressure of carbon 
 dioxide, it is evident that within wide limits the pressure of oxy- 
 gen in the alveolar air will remain almost normal, since the in- 
 creased breathing due to the extra carbon dioxide will so raise 
 the alveolar oxygen pressure as to compensate approximately for 
 the oxygen deficiency in the inspired air. There will thus be no 
 appreciable anoxaemia, and consequently the oxygen deficiency 
 in the inspired air will produce no effect at all, although a similar 
 deficiency in the absence of the excess of CO 2 would produce a 
 marked effect. For instance, by adding CO 2 to the inspired air we 
 can easily compensate within wide limits for the deficient oxygen 
 pressure which affects airmen at high altitudes. This is not be- 
 cause, as Mosso 3 imagined, the effects of high altitude are due 
 primarily to excessive loss of CO 2 ("acapnia"), but because the 
 oxygen pressure, as well as that of CO 2 , is kept approximately 
 constant by the increased breathing due to the CO 2 . When, how- 
 ever, the conditions are such that the extra breathing due to ex- 
 cess of CO 2 does not prevent the alveolar oxygen pressure from 
 falling very low, the stimulus of anoxaemia is added co that of 
 CO 2 , and an enormously greater effect is produced on the breath- 
 ing than by the CO 2 stimulus alone. This extra effect, as was 
 recently shown by Meakins, Priestley, and myself 4 is due to in- 
 crease in the frequency of the breathing; and increased frequency, 
 provided the depth of breathing is sufficient, is, for a reason which 
 will appear in the next chapter, particularly effective in prevent- 
 ing anoxaemia. 
 
 A further complication in the effects of anoxaemia and forced 
 breathing on the respiratory center and the body as a whole is 
 
 8 Mosso, Life of Man on the High Alps, Chapter XXII, London, 1898. 
 4 Haldane, Meakins, and Priestley, Journ, of Physiol., LII, p. 420, 19 19. 
 
RESPIRATION 113 
 
 introduced by the fact that, as Bohr discovered (see Chapter V 
 and Figure 19), deficiency of carbon dioxide causes haemoglobin 
 to hold on more tightly to oxygen. The consequence of this is, that 
 when increased breathing lowers the pressure of CO 2 in the al- 
 veolar air and in the body as a whole, on the one hand the haemo- 
 globin of the blood passing through the lungs is more highly 
 saturated with oxygen than it otherwise would be; on the other 
 hand the blood holds this oxygen so firmly that the oxygen pres- 
 sure in the tissues falls lower than it otherwise would. There may 
 thus be considerable anoxaemia though the blood is almost as red 
 as usual, and the existence of this anoxaemia is only revealed by 
 the immediate physiological effects of raising the alveolar oxygen 
 pressure. 5 
 
 On reducing, in a steel chamber, the atmospheric pressure to 
 half an atmosphere there is a quite appreciable permanent increase 
 in the breathing, and consequent drop in alveolar CO 2 pressure 
 caused by anoxaemia, but, in my own case at any rate, no very 
 striking blueness of the lips, although at the time the alveolar 
 oxygen pressure is only about 34 mm. This pressure would only 
 be sufficient to saturate the oxyhaemoglobin of the blood to the 
 extent of 57 per cent if the pressure of CO 2 were that of normal 
 alveolar air (see Figure 20). Blood with this percentage satura- 
 tion would be very strikingly blue. Owing, however, to the dimin- 
 ished pressure of CO 2 , the saturation is much higher, and this 
 accounts for the color of the lips being nearly normal. The exist- 
 ence of considerable anoxaemia was, however, revealed at once 
 by the effects of adding oxygen to the inspired air : for vision and 
 hearing were at once strikingly improved and the breathing di- 
 minished. The degree of blueness of the lips is thus only a rough 
 index of anoxaemia when anoxaemia is taken in its physiological 
 meaning, as diminution in the oxygen pressure, rather than merely 
 of the oxygen content, of the blood. It is the diminution in the 
 amount of free oxygen, whether or not the amount of reserve oxy- 
 gen combined with the haemoglobin is also diminished, which is 
 functionally important. 
 
 Thus the benefit produced by diminished pressure of CO 2 (as, 
 for example, during forced breathing) in increasing the percent- 
 age saturation of the haemoglobin in the arterial blood is neu- 
 tralized by the disadvantage in the tissues owing to the same cause. 
 The venous blood may, in fact, be as red as usual, although the 
 venous oxygen pressure is abnormally low : for the saturation of 
 
 * Haldane, British Medical Journal, July 19, 1919. 
 
H 4 RESPIRATION 
 
 the arterial blood with oxygen can be only very slightly increased 
 by the lowering of alveolar CO 2 pressure. The oxygen pressure 
 of the venous blood must in consequence be lowered, so that anox- 
 aemia might be produced without any diminution, and even with a 
 slight increase, in the saturation of the haemoglobin of the venous 
 blood. On the other hand if the haemoglobin of the arterial blood, 
 with normal alveolar CO 2 pressure, were only half-saturated, a 
 lowering of the alveolar CO 2 pressure would considerably in- 
 crease the saturation of the haemoglobin in both arterial and 
 venous blood, but without sensible alteration of the venous oxygen 
 pressure. Only in the practically impossible case of the saturation 
 of the arterial haemoglobin being much below half would there 
 be any rise in the venous oxygen pressure. Practically speaking, 
 therefore, the Bohr effect, the increased oxygen content in blood, 
 due to lowering of alveolar CO 2 pressure, is never of service in 
 increasing the real oxygen supply to the tissues, and is sometimes 
 of great disservice, although it always tends to make the venous 
 blood less blue, and so diminishes cyanosis. On the other hand the 
 corresponding effect due to raising of alveolar CO 2 pressure will 
 practically never diminish the oxygen supply to the tissues, and 
 will usually increase it, though the venous blood will always be 
 more blue. 
 
 With forced breathing of normal air there is, as mentioned in 
 Chapter I, a slight increase in the oxygen present in the arterial 
 blood. This is due, partly to the Bohr effect and partly to the effect 
 of the increased alveolar oxygen pressure. Hence the saturation 
 of the haemoglobin is increased from about 95 to 100 per cent. 
 There is also a small increase in the free oxygen dissolved in the 
 arterial blood. On the other hand the amount of CO 2 and its 
 partial pressure are enormously reduced in the arterial blood, and 
 to a less extent the venous blood, since the circulation rate, as 
 will be shown in Chapter X, is much diminished. The net result 
 must be a considerable fall in the oxygen pressure in the tis- 
 sues. Now it is well known that forced breathing produces a 
 train of symptoms which, if the forced breathing is pushed, 
 tend towards unconsciousness, so that forced breathing has 
 even been used by dentists as a means of producing partial 
 anaesthesia. In many respects these symptoms are similar to those 
 of anoxaemia, except for the absence of spontaneous increased 
 breathing. It was discovered by Hill and Flack 6 that when the 
 forced breathing is with oxygen instead of with air the symptoms 
 
 8 Hill and Flack, Journ. of Physiol., XL, p. 347, 1910. 
 
RESPIRATION 115 
 
 are greatly diminished. The most natural explanation of this is 
 that the oxygen, by increasing largely the amount of free oxygen 
 in the blood, diminishes the anoxaemia, since an oxygen supply 
 which is not dependent on the Bohr effect is added to the ordinary 
 oxygen supply from oxyhaemoglobin. Probably, therefore, the 
 symptoms referred to are mainly produced by anoxaemia caused 
 by the Bohr effect. The subject will be discussed further in Chap- 
 ter X. 
 
 It is a very interesting fact that in many persons forced breath- 
 ing does not produce apnoea at all, although in such persons the 
 breathing is regulated in accordance with the alveolar CO 2 pres- 
 sure, just as in other persons. This fact was investigated by Dr. 
 Boothby some years ago while he was working with me. 7 He 
 found that at the end of continuous forced breathing for one or 
 two minutes there was in himself not only no sign of apnoea, but, 
 on the contrary, increased natural breathing for a short time. This 
 soon passed away, but at no time was there any apnoea, though 
 the excretion of CO 2 in the expired air was much diminished for 
 a considerable period. The cause of this absence of apnoea is not 
 yet clear. It seemed possible that the stimulus of anoxaemia from 
 the Bohr effect might, in persons who do not become apnoeic, 
 account for the absence of apnoea ; but even after forced breathing 
 of oxygen the apnoea was absent in one of these persons whom I 
 tested. His power of voluntarily holding a deep breath was 
 markedly increased by forced breathing of air, but natural apnoea 
 did not occur. 
 
 Owing, apparently, to the existence of the Bohr effect, the in- 
 fluence of CO 2 in relieving the general symptoms of anoxaemia 
 is not due merely to increased breathing and consequent rise in 
 the alveolar oxygen pressure. Lorrain Smith and I observed that 
 animals in a semi-comatose state from the anoxaemia of carbon 
 monoxide poisoning were revived by substituting expired air for 
 pure air without alteration of the percentage of carbon monoxide. 
 With the expired air mixture there could be no rise in the alveolar 
 oxygen pressure, and there was no alteration in the percentage 
 saturation of the blood with carbon monoxide. A still more strik- 
 ing effect is produced by simply adding CO 2 to the air inspired 
 during CO poisoning. At the time we could not understand this 
 effect, as Bohr's discovery had not yet been made. But this dis- 
 covery furnishes an explanation of why a rise in the alveolar CO 2 
 pressure, without alteration of the alveolar oxygen pressure, 
 
 7 Boothby, Journ. of PhysioL, XLV, p. 328, 
 
u6 
 
 RESPIRATION 
 
 should relieve the symptoms in CO poisoning : for the increased 
 CO 2 pressure will enable the oxygen to come off more easily from 
 the oxyhaemoglobin present in the blood, and will thus tend to 
 relieve the anoxaemia. The circulation rate will also be increased, 
 as will appear in Chapter X. There would seem to be a considerable 
 future scope for the therapeutic use of CO 2 in anoxaemic condi- 
 
 
 |^- -vtlWA^^wvA/UlA/v^A/^vA/iJt^^ J^\J\[lAAs*^sJ\jll\jy\~-~l^ 
 
 2 OFF 
 
 Figure 37. 
 
 Tracing i. (Stethograph) Douglas, July 12. Evening of arrival on Pike's Peak. Natural 
 periodic breathing. 
 
 Tracing 2. Haldane, July 12. Evening of arrival. Natural periodic breathing with more 
 sharply defined periods after making six forced breaths. 
 
 Tracing 3. July 16, Haldane. Natural periodic breathing abolished by administration of 
 oxygen. Reappearance of periodic breathing after withdrawing the oxygen. 
 
 tions of all kinds, whether or not these conditions are due to im- 
 perfect oxygenation of the arterial blood. 
 
 Even when simple anoxaemia is so extreme that consciousness 
 is on the point of being lost, the breathing in man, except at first, 
 is hardly more than doubled, as shown by the fact that the alveolar 
 
RESPIRATION 
 
 117 
 
 CO 2 pressure is only reduced to about half. During heavy muscu- 
 lar exertion, on the other hand, the breathing may easily be in- 
 creased to ten or fifteen times its normal amount. The relatively 
 slight increase in the amount of air breathed during very serious 
 anoxaemia is frequently lost sight of in the interpretation of 
 clinical symptoms. There is nearly always a considerable increase 
 in the frequency of breathing, but the depth of breathing is 
 usually only slightly increased, and may be diminished, as will 
 be explained more fully below. In the very dangerous pure anox- 
 aemia of high altitudes or CO poisoning, increase in the breathing 
 is not a prominent symptom. 
 
 It has been known for long that at high altitudes the breathing 
 is very apt to be periodic. This phenomenon was fully observed 
 on Monte Rosa by Mosso, 8 who, however, had completely failed 
 to realize the significance of Paul Bert's researches on the effects 
 of gases, and thus failed to interpret correctly the cause of the 
 periodic breathing. The periodic breathing is usually not con- 
 tinuous, but can easily be started by disturbing the ordinary 
 rhythm of breathing, as by taking a few long breaths, or holding 
 the breath. It is also very apt to occur at night. It is distinguished 
 from ordinary clinical Cheyne- Stokes breathing by the shortness 
 
 INTERVALS OF 5 SECONDS 
 
 Figure 38. 
 
 Henderson, August 13. Quantitative record of the respiration during periodic 
 breathing. Inspiration upwards. 
 
 of the periods. There are usually groups of only about three to 
 six breaths, followed by a pause, and this periodic sequence con- 
 tinues almost indefinitely (Figure 37). Sometimes the middle 
 breath of the group is deepest, sometimes the last breath (Figure 
 38) or sometimes the breaths are about equal in depth. Some- 
 times the periodicity only shows itself by periodically recurring 
 single deep breaths. 
 
 The general explanation of this periodic breathing has already 
 been given in Chapter V. That this explanation is the correct one 
 is shown by the fact that, as is seen in Figure 37, on adding oxy- 
 
 8 Mosso, Life of Man on the High Alps, Chapter III, London, 1898. 
 
n8 
 
 RESPIRATION 
 
 gen to the inspired air the periodicity disappears. This experiment 
 was carried out repeatedly by Douglas, Yandell Henderson, 
 Schneider, and myself, on Pike's Peak, and never failed. 9 Mosso 
 had attempted to carry it out, but got a negative result owing to a 
 defective mode of administering the oxygen. 
 
 As already seen periodic breathing can easily be produced at 
 ordinary barometric pressure by suitable means. As the barometric 
 pressure is reduced the periodic breathing is produced more and 
 more readily, and is more and more persistent, just as might be 
 expected; and the same is true if, instead of a reduction of bar- 
 ometric pressure, there is a reduction in the oxygen percentage of 
 the inspired air. This form of periodic breathing has no pathologi- 
 cal significance, and occurs during perfect health. 
 
 The special characters of the increased breathing caused by 
 
 Figure 39. 
 
 (a) Rebreathing Concertina filled with oxygen CO* accumulating, 
 
 (b) Rebreathing Concertina filled with air CO 2 accumulating. 
 Time-marker = 2 seconds. Arrow shows point where lips were distinctly blue. 
 
 anoxaemia were recently studied by Meakins, Priestley, and 
 myself. 10 The differences between increased breathing caused by 
 excess of CO 2 and that caused by anoxaemia, or by anoxaemia 
 accompanied by excess of CO 2 , are very striking. Speaking gen- 
 erally, the effect of excess of CO 2 is mainly to increase the depth 
 of breathing, and only a moderate increase of frequency is pro- 
 duced. On the other hand anoxaemia produces a marked increase 
 in frequency and only a moderate increase in depth. But when the 
 
 * Douglas, Haldane, Yandell Henderson, and Schneider, Philos. Trans. Royal 
 Society, B. 203, p. 231. 
 
 "Meakins, Haldane, and Priestley, Journ. of Physiol., LII, p. 420, 1919. 
 
RESPIRATION 
 
 119 
 
 effects of excess of CO 2 and anoxaemia are combined there is 
 great increase of both depth and frequency, so that far more air is 
 breathed than when either excess of CO 2 alone, or anoxaemia 
 alone, is the stimulus. In my own case, for instance, when the 
 breathing was pushed, in short experiments, to as much as seemed 
 bearable, 131 liters per minute, with a depth of 1.98 liters and a 
 frequency of 66 per minute, were breathed when the effects of 
 excess of CO 2 and anoxaemia were combined ; and only 8 1 liters, 
 with a depth of 2.69 and a frequency of 30, when the only stimulus 
 was excess of CO 2 . 
 
 Figure 39 shows quantitatively the effects of rebreathing a 
 small volume (about 2 liters) of air or oxygen from the recording 
 concertina already described. It will be seen that the increase in 
 frequency was much less when the effects of anoxaemia were cut 
 out by the oxygen. 
 
 Figure 40 A shows the effect on the same subject of similar re- 
 breathing when the accumulation of CO 2 was prevented by inter- 
 
 iiiiiii.ini 
 
 Figure 40. 
 
 Rebreathing through soda-lime from concertina. Time-marker = 2 seconds, 
 (a) Subject Cpl. M. (b) Subject J. S. H. 
 
 posing a layer of soda lime. It will be seen that the frequency 
 increases, but not the depth. Figure 40 B shows the effects on 
 another subject, whose respiratory center responds much more 
 readily to the effects of anoxaemia. In this case depth as well as 
 frequency are considerably increased. It must, however, be borne 
 in mind that, in short experiments such as these, the increased 
 breathing, as already explained, is mainly due to the necessity of 
 
-E 
 
 * 
 
 
 
 o 
 
 t~t 
 
 W> 
 
 I 
 
 E-8 
 
RESPIRATION 121 
 
 removing from the body the large amount of preformed CO 2 
 which has become superfluous owing to the effect of anoxaemia 
 in lowering the threshold of CO 2 pressure. 
 
 Figures 41 and 42 show the effects of anoxaemia combined with 
 those of the slight resistance associated with the recording ap- 
 paratus. The effects are complicated owing to the fact that with a 
 certain degree of anoxaemia, varying greatly for different indi- 
 viduals, periodic breathing is produced readily, as shown in some 
 of the tracings. Periodic breathing, or else very shallow breathing, 
 is also produced invariably after the anoxaemia, as shown in all 
 the tracings. This is of course due to the fact that so much CO 2 
 has been removed from the body by the hyperpnoea of anoxaemia, 
 just as it is removed by forced breathing. 
 
 In Figure 41 B and Figure 42, A, C, and D, it will be seen that 
 after an initial increase in depth the breathing became progres- 
 sively shallower and more frequent just as in fatigue due to ex- 
 cessive resistance; and after a time asphyxial symptoms were 
 usually impending owing to the ineffectiveness of the shallow 
 breaths. When the experiments were made we had not investigated 
 the effects of fatigue caused by resistance, and there is now no 
 doubt that the slight resistance due to the apparatus, combined 
 with the effects of anoxaemia on the respiratory center, accounted 
 for the specially rapid failure of breathing shown in the figures. 
 When the breathing is quite free, as in a steel chamber at low 
 pressures, failure of the respiratory center does not occur nearly 
 so readily, but the difference is only one of degree; and failure 
 of the respiratory center, as shown by shallow and frequent res- 
 pirations, is the inevitable result of serious arterial anoxaemia. 
 With the increasing shallowness of the breaths the arterial an- 
 oxaemia increases, owing to causes discussed in Chapter VII. 
 This increases the failure of the respiratory center; and unless 
 relief comes the inevitable result of the vicious circle thus pro- 
 duced is death. 
 
 We must now turn to the other symptoms and signs of want of 
 oxygen, beginning with the circulatory symptoms. Unfortunately 
 we cannot as yet measure the volume of blood circulated per 
 minute in the same easy way in which we can measure the volume 
 of air breathed. Our knowledge of the effects of want of oxygen 
 on the circulation is thus imperfect as yet. It will be discussed 
 more fully in Chaper X. When moderate symptoms of anoxaemia 
 are produced experimentally, as in a steel chamber at reduced 
 atmospheric pressure, or when air deficient in oxygen is breathed, 
 
RESPIRATION 
 
 123 
 
 there is at first an increase of the frequency, and apparently also 
 in the strength, of the heartbeats. This indicates an increase in the 
 circulation rate. But just as in the case of the respirations, the 
 frequency and vigor of the pulse soon fall again, though the fre- 
 quency remains above normal, just as does the frequency of res- 
 piration. Thus the pulse may rise to about 120 at first, and then 
 fall after a few minutes to about 90, and remain steady. With 
 greater anoxaemia the increase in rate is more marked. The great 
 temporary increase in blood pressure with acute anoxaemia in 
 animals is also a well-known phenomenon. 
 
 At first sight it might seem that a great increase in both res- 
 pirations and circulation would be the natural physiological 
 response to anoxaemia, since the increased respiration will raise 
 the alveolar oxygen pressure and the increased circulation rate 
 will increase the amount of oxygen left in the red corpuscles of 
 the blood passing through the capillaries. But, as already seen, 
 the increased respiration lowers the pressure of CO 2 in the respira- 
 tory center and tissues, and this lowering rapidly reduces the in- 
 creased breathing to within relatively narrow limits. A similar 
 lowering of CO 2 pressure in the tissues must also be produced by 
 increased circulation rate ; and the f alling-off in the initial increase 
 of pulse rate is probably at bottom due to the same cause as the 
 falling-off in the initial depth and frequency of breathing. With 
 further increase in the anoxaemia the heartbeats, like the respira- 
 tions, become more and more feeble. A fuller discussion of the 
 relatively little that is at present known definitely as to the physio- 
 logical regulation of the circulation will be found in Chapter X. 
 It is of course evident that the physiology (not the mere physics) 
 of the circulation is intimately related to that of the breathing. 
 
 As a sign of anoxaemia, the appearance of the lips, tongue, and 
 face is of much importance, but requires careful interpretation. 
 The bluish color or cyanosis seen in the lips and skin during 
 ordinary anoxaemia is, of course, due to the fact that in the blood 
 passing through the capillaries the proportion of oxyhaemoglobin 
 to haemoglobin is abnormally low. A somewhat similar color may 
 be produced by the action of poisons which produce methaemo- 
 globin and other colored decomposition products in the blood; 
 and this condition, which is of course quite exceptional, and can 
 quite easily be distinguished, will be referred to in Chapter VII. 
 Cyanosis may either be due to general or local slowing of the 
 circulation, or to the fact that the arterial blood is imperfectly 
 oxygenated, and the latter cause, as will be shown in Chapter VII, 
 
124 RESPIRATION 
 
 is far more common than was, till recently, supposed. Portions of 
 the skin may be blue from local slowing of the circulation due to 
 cold and other causes; but abnormal blueness of the lips and 
 tongue points to either imperfect oxygenation of the arterial 
 blood or general slowing of circulation. According as there is 
 much or little blood in the capillaries the color is full or unsatu- 
 rated. Thus in extreme cyanosis the lips may be either almost 
 black, or only leaden gray ; and in slight cyanosis the color may 
 be either a full or a pale purplish red. 
 
 Ordinary cyanosis of one kind or another is commonly seen in 
 patients who, though suffering from some chronic ailment, are 
 not particularly ill. Hence the significance of cyanosis under other 
 conditions is apt to be overlooked unless all the symptoms and 
 other circumstances are taken into account. It must, in the first 
 place, be pointed out that the degree of cyanosis is no direct 
 measure of the degree of physiological anoxaemia. The latter is 
 due to a lowering in the partial pressure of oxygen in the blood of 
 the capillaries, while the former is due to a diminution in the ratio 
 of oxyhaemoglobin to haemoglobin. Under ordinary conditions 
 the latter effect is an index, though, owing to the form of the 
 dissociation curve of oxyhaemoglobin (Figure 20), not a direct 
 measure, of the former effect. When, however, the matter is com- 
 plicated by an alteration in the Bohr effect of CO 2 pressure on the 
 dissociation of oxyhaemoglobin, the relationship between oxygen 
 pressure and dissociation of oxyhaemoglobin is at once altered. If, 
 for instance, the pressure of CO 2 in the arterial blood is reduced 
 by increased breathing, there may be much less cyanosis for a 
 given degree of physiological anoxaemia than when the CO 2 
 pressure in the blood is normal. Thus there is no fixed relationship 
 between cyanosis and physiological anoxaemia; and this fact is 
 of great importance in the clinical interpretation of cyanosis. 
 Moreover, as Barcroft showed, the Bohr effect is due to the action 
 of CO 2 as an acid. Hence, owing to the adjustments which, as will 
 be shown in Chapter IX, occur in the living body when time is 
 given, the CO 2 pressure in the alveolar air may be no guide as to 
 how far the Bohr effect is disturbing the ordinary relations be- 
 tween cyanosis and true anoxaemia. The word "anoxaemia" 
 should evidently be taken as signifying a condition in which the 
 free oxygen in the systemic capillary blood is abnormally dimin- 
 ished ; and this of course, in accordance with Henry's law, comes 
 to the same thing as diminution in the oxygen pressure. 
 
 The symptoms produced in the nervous system generally by 
 
RESPIRATION 125 
 
 anoxaemia must now be described. A knowledge of them is of 
 great importance in practical medicine. If a pure anoxaemia is 
 produced very suddenly, as by breathing pure nitrogen, hydro- 
 gen, methane, or nitrous oxide, loss of consciousness occurs quite 
 suddenly and with no previous warning symptoms. Thus a miner 
 who puts his head into a cavity in the roof full of pure, or nearly 
 pure, methane drops suddenly as if he had been felled ; and when 
 he recovers after breathing pure air for a few seconds he some- 
 times even imagines that he has been knocked down by another 
 man, and acts accordingly. If the anoxaemia is produced with 
 only moderate rapidity the marked temporary disturbances, al- 
 ready referred to, in the breathing and circulation give, as a rule, 
 some warning of what is coming. But when the onset is gradual 
 there is little or no preliminary discomfort, and for this reason 
 the onset of pure anoxaemia is very insidious, and the condition 
 is, therefore, in practice a dangerous one, as is well seen in CO 
 poisoning, or in ascents to very high altitudes in balloons or aero- 
 planes, or in many clinical cases. Thus although CO is not very 
 poisonous as compared with other gaseous poisons, it is responsible 
 for a far larger number of deaths than any other gaseous poison 
 not used in warfare. 
 
 As the slow onset of anoxaemia advances, the senses and intellect 
 become dulled without the person being aware of it; and if the 
 anoxaemia is suddenly relieved by means of oxygen or ordinary 
 air, the corresponding sudden increase in powers of vision, hear- 
 ing, etc., is an intense surprise. The power of memory is affected 
 early, and is finally almost annulled, so that persons who have ap- 
 parently never lost consciousness can nevertheless remember noth- 
 ing of what has occurred. Powers of sane judgment are much 
 impaired, and anoxaemic persons become subject more or less to 
 irrational fixed ideas, and to uncontrolled emotional outbursts. 
 Muscular coordination is also affected, so that a man cannot walk 
 straight or write steadily. With further increase in the anoxaemia, 
 power over the limbs is lost; the legs first being paralyzed, then 
 the arms, and finally the head. The senses are lost one by one, hear- 
 ing being apparently the last to go. The sense of painful impres- 
 sions on the skin seems to be lost early. Thus miners suffering from 
 CO poisoning, but not to the point of losing consciousness, are 
 often burnt by their lamps or candles without their being aware of 
 the burn at the time. 
 
 In many respects the symptoms of anoxaemia resemble those 
 of drunkenness, and a man suffering from anoxaemia cannot be 
 
126 RESPIRATION 
 
 held responsible for his actions. Without reason he may begin to 
 laugh, shout, sing, burst into tears, or become dangerously violent. 
 He is, however, always quite confident that he himself is perfectly 
 sane and reasonable, though he may notice, for instance, that he 
 cannot walk or write properly, cannot remember what has just 
 happened x and cannot properly interpret his visual impressions. 
 When unable even to stand, owing to experimental CO poisoning 
 or to anoxaemia produced by low pressures in a steel chamber, I 
 have always been quite confident in my own sanity, and it was 
 only afterwards that I realized that I could not have been in a 
 sane state of mind. 
 
 A recent experience of this kind was in a steel chamber in which 
 Dr. Kellas, who is an experienced climber in the Himalayas and 
 has exceptional powers of resisting anoxaemia, was with me. 11 
 We had reduced the pressure to 320 mm., and as I could no longer 
 write or make any observations I handed him the notebook. He 
 afterwards told me that I remained sitting, but always answered 
 his questions quite deliberately and confidently, and insisted on 
 his keeping the pressure at 320 mm. This went on for an hour and 
 a quarter, of which time I could afterwards remember absolutely 
 nothing. At last Dr. Kellas obtained my assent to raising the pres- 
 sure to 350 mm., after which I took up a mirror to look at my lips, 
 though Dr. Kellas observed that for some time I looked at the 
 back instead of the front of the mirror. I had, however, begun to 
 realize that we had been far longer at the low pressure than we 
 had intended, and agreed to a rise to 450 mm. On reaching this 
 pressure my mind had cleared and I noticed a return of feeling 
 and power in my legs. After coming out I could vaguely remember 
 taking up the mirror, but nothing before that, after handing over 
 the notebook. We had no intention of staying at so low a pressure 
 that it was impossible for me to take notes, and my persistence 
 was quite irrational. Dr. Kellas was much bluer than I was during 
 the stay at 320 mm., but could still write quite well, watch the ba- 
 rometer, and manage the regulating tap ; but whether he was quite 
 normal mentally seemed rather doubtful. Perhaps he shared to 
 some extent my irrational desire to continue the experiment: 
 otherwise I think he would have noticed how abnormal my condi- 
 tion was. We were both at the time unacclimatized to low pres- 
 sures. 
 
 This personal experience illustrates some of the peculiar 
 dangers associated with atmospheres which produce anoxaemia, 
 
 11 Haldane, Kellas, and Kennaway, Journ. of Physiol., LIII, p. 181, 1919. 
 
RESPIRATION 127 
 
 whether in virtue of defective oxygen pressure or of the presence 
 of poisonous proportions of CO. In the first place it is evident that 
 a man may advance for some distance into such an atmosphere 
 before he begins to be seriously affected; for the temporary 
 marked increase in the breathing may, when the oxygen pressure 
 is defective, at first prevent an appreciable fall in the alveolar 
 oxygen pressure. This must, for instance, happen while a balloon 
 or aeroplane is rising rapidly, or while a miner is advancing with 
 an electric lamp into an atmosphere very highly charged with 
 fire-damp. When the breathing begins to quiet down again the 
 effects of the atmosphere will develop fully and it may then be too 
 late to turn. At 320 mm., for instance, I was at first quite capable 
 of making observations and taking notes, including a note of the 
 increased breathing and its subsequent quieting down. 
 
 Another, and often still more serious, danger arises from the 
 gradual and insensible failure of judgment. A man suffering from 
 anoxaemia will usually go on, and insist in going on, with what 
 he set out to do. An airman will very probably continue to ascend, 
 oblivious to danger ; and a miner engaged in rescue or exploration 
 work, or in dealing with an underground fire, will insist in going 
 on when he is suffering from the anoxaemia of CO poisoning, 
 and will often fight any one who tries to make him desist. 
 
 All these considerations apply equally to clinical cases of anox- 
 aemia ; and for this reason the condition is quite commonly never 
 recognized till too late. The early recognition of clinical anox- 
 aemia is a matter of great importance. 
 
 Besides the immediate symptoms of anoxaemia there are a 
 number of delayed symptoms or after effects. They depend partly 
 on the length, and partly on the severity, of the exposure. A 
 short exposure, even with loss of consciousness, produces no 
 serious after symptoms ; but occasionally a man's behavior is very 
 abnormal for a few minutes after recovery. One of my ac- 
 quaintances has twice knocked persons down on waking up from 
 a short loss of consciousness caused by anoxaemia; and my own 
 behavior was distinctly abnormal just after coming out from the 
 steel chamber in the experiment alluded to above. Similar ab- 
 normalities after slight CO poisoning have often come under my 
 observation. Thus a well-known inspector of mines, on returning 
 to the surface after he had been affected by CO from an under- 
 ground fire, first shook hands very cordially with all the by- 
 
128 RESPIRATION 
 
 standers. A doctor who was present then offered him an arm ; but 
 this the inspector regarded as an insult, with the result that he 
 took off his coat and challenged the doctor to a fight. 
 
 The best-known delayed effect of slight anoxaemia is the train 
 of symptoms originally called "mountain sickness." This is a 
 condition in the typical form of which there is nausea, vomiting, 
 headache, sometimes diarrhoea, and always great depression. The 
 symptoms appear, as a rule, some hours after the beginning of 
 the exposure, and may not appear at all till after the exposure is 
 over. In CO poisoning it is usually after the exposure, and often 
 after the CO has practically disappeared from the blood, that 
 these symptoms begin. The duration of exposure required for their 
 production depends upon the degree of anoxaemia. Thus the 
 higher a mountain is, or the greater the altitude at which an air- 
 man has been flying, the shorter is the exposure required. On 
 Pike's Peak, at 14,100 feet (barometer about 458 mm.) the usual 
 stay (an hour or two) of visitors by train is too short to produce 
 mountain sickness, though the ordinary immediate symptoms of 
 anoxaemia are usually very evident, and even very great cyanosis 
 and fainting are observed occasionally. A stay of several hours 
 is usually required to induce mountain sickness, which usually 
 begins about 8 to 12 hours after the beginning of the exposure. 
 Thus the symptoms may only develop after the return downwards. 
 
 With a sufficient period of exposure mountain sickness may 
 develop at much lower altitudes than that of Pike's Peak. It is 
 often observed at even 7,000 or 8,000 feet, where the degree of 
 anoxaemia is not sufficient to produce any noticeable immediate 
 effect on the breathing. Similarly a percentage of CO which pro- 
 duces no noticeable immediate effect will, with sufficiently long 
 exposure, cause headache, nausea, etc. These facts are of the 
 greatest significance in clinical medicine, for it is now evident that 
 even a very slight degree of continued anoxaemia is of much 
 importance to the patient. Mountain sickness and the effects of 
 CO poisoning are not isolated phenomena unrelated to the rest of 
 physiology and pathology, but symptoms of anoxaemia, which is 
 in reality one of the commonest conditions during illness. At 
 present we can only conjecture as to the nature of the slight 
 temporary pathological changes of which the mountain sickness 
 symptoms are the manifestations. 
 
 With severe and prolonged exposure to want of oxygen the 
 nervous after symptoms are of an extremely formidable nature, 
 
RESPIRATION 129 
 
 and often end in death. 12 For a reason which will be explained 
 in a later chapter they are most commonly met with after CO 
 poisoning, and whatever their origin they are often grossly mis- 
 interpreted. The patient does not recover at once on removal of 
 the oxygen want, as in short exposures. In cases of CO poisoning 
 consciousness may not be recovered, although within an hour or 
 two after removal to fresh air most of the CO has already disap- 
 peared from the blood. It is exactly the same with men who have 
 remained unconscious for, perhaps, several hours in air very poor 
 in oxygen. Or if consciousness has been partially recovered the 
 patient may lapse again into unconsciousness. During gradual 
 recovery there is usually a very marked spastic condition of the 
 muscles, and occasional epileptiform seizures, and there may be 
 various partial paralyses and other nervous symptoms. Sometimes 
 the patient lingers on for weeks in a comatose condition with 
 spastic muscles and occasional opisthotonos. The body tempera- 
 ture is unstable, and every function of the central nervous system 
 seems to be more or less affected. Gross hemorrhages in the brain 
 have been described, and Mott has found small multiple hemor- 
 rhages. The symptoms are, however, evidently due in the main 
 to widespread injury to the nerve cells themselves during the ex- 
 posure. Loss of memory, mental incapacity, and even definite 
 mania may follow the exposure; but whatever the nature of the 
 symptoms may be, they nearly always pass off gradually if the 
 patient survives the first few days. One interesting nervous after 
 effect occasionally observed is what appears from the symptoms 
 to be peripheral neuritis. 
 
 The heart may also suffer severely in prolonged exposure to 
 want of oxygen ; and if the exposure has been accompanied by 
 much muscular exertion, as in efforts to escape or to rescue other 
 men, the after symptoms may be mainly cardiac. In these cases 
 the pulse is feeble and irregular, the heart dilated, with a blowing 
 systolic murmur; and any muscular exertion produces collapse. It 
 may be a considerable time before the heart fully recovers. 
 
 Probably every other organ and tissue in the body feels the 
 after effects of severe exposure to want of oxygen. The patient 
 often enough dies of pneumonia. Acute nephritis and gangrene of 
 extremities have been noticed as sequelae to the acute broncho- 
 pneumonia and oedema of the lungs in chlorine poisoning. As 
 
 " An interesting description of these symptoms by Dr. Shaw Little will be 
 found in Appendix B to my Report on the Causes of Death in Colliery Explosions, 
 Parliamentary Paper C. 8112, 1896. 
 
1 30 RESPIRATION 
 
 the patients have been exposed to very grave oxygen want in 
 consequence of the lung condition, it seems probable that the af- 
 fections just mentioned are after effects of the oxygen want, 
 aggravated by the after effects on the heart, and often complicated 
 by secondary infections. 
 
 With anoxaemia, as already explained, the respiratory center 
 becomes very easily susceptible of fatigue, as manifested by di- 
 minishing depth of the breathing. It is now well known that in 
 the resuscitation of persons who have been nearly asphyxiated by 
 drowning, asphyxiating atmospheres, etc., the most effective 
 remedy is artificial respiration. This is because the respiratory 
 center has completely or almost completely failed or become 
 "fatigued," and the patient would die if this condition were not 
 compensated for by artificial respiration. Respiration seems almost 
 always to fail before the heart fails. The respiratory center may 
 also take a long time to recover sufficiently to be able, without 
 artificial aid, to keep the patient alive. For this reason it may be 
 necessary to prolong the artificial respiration for hours. 
 
 Diminishing depth with increasing rate of respiration is always 
 a sign of the onset of fatigue of the breathing; and when the 
 depth continues to diminish without compensation from increased 
 rate the condition rapidly becomes dangerous, as will be shown 
 in Chapter VII, since secondary anoxaemia develops. In a person 
 dying quietly the diminishing depth can be observed until the 
 resulting anoxaemia ends in death. The immediate cause of death 
 seems to be failure of the respiratory center. When death from 
 anoxaemia occurs at very high altitudes (as, for instance, in the 
 case referred to in Chapter XII, of the balloonists, Tissandier and 
 Croce Spinelli) it is evidently failure of the respiratory center 
 which precipitates the anoxaemia, thus making the conditions 
 so very dangerous ; and the same remark applies to asphyxiation 
 in atmospheres containing a low percentage of oxygen in mines, 
 wells, etc. In CO poisoning, as will be explained in Chapter VII, 
 there is not so much danger from this cause, so that extreme anox- 
 aemia may exist for a long time without death occurring. 
 
 After the respiratory center has been over-fatigued in conse- 
 quence of anoxaemia, the effects may not pass off for a very long 
 period. The breathing on exertion, or even during rest, is ab- 
 normally shallow ; and the peculiar group of symptoms observed 
 in the neurasthenic condition so familiar during the war, and al- 
 ready referred' to in Chapter III, is observed. This condition may 
 
RESPIRATION 131 
 
 remain for months after severe anoxaemia, and is often mistaken 
 for organic heart injury. 
 
 In considering the effects of anoxaemia a factor comes in which 
 must always be borne in mind namely that of adaptation or ac- 
 climatization. This may act in two different ways. In the first 
 place adaptation may bring it about that the anoxaemia which 
 would, without adaptation, exist is greatly diminished. This form 
 of adaptation is very clearly seen in persons living at great alti- 
 tudes, and will be discussed in detail in later chapters. In the 
 second place the tissues may adapt themselves to a lower partial 
 pressure of oxygen. About this second form of adaptation our 
 knowledge is at present very imperfect; but it seems to me 
 that clinical evidence points strongly to its existence. Perhaps the 
 clearest evidence is afforded by cases of congenital heart defect, 
 in which part of the venous blood passes direct to the left side of 
 the heart without first passing through the lungs. In these cases 
 of "Morbus coeruleus" the arterial blood is always more or less 
 blue, and becomes extremely blue on muscular exertion, so that 
 one can always recognize this condition in persons walking in the 
 street. The remarkable point, however, is that in spite of the an- 
 oxaemic condition of the arterial blood these persons may get on 
 quite well, and be able to walk at a good pace. On account of the 
 large increase in their haemoglobin percentage, they have plenty 
 of oxygen in their blood, but at a low partial pressure. It seems 
 hardly possible to doubt, therefore, that their tissues have become 
 adapted to the low partial pressure of oxygen; and the same 
 adaptation probably exists to a considerable extent in many 
 chronic cases of valvular heart disease, emphysema, etc. 
 
 The fact that cyanosis may exist without harm in chronic cases 
 of disease has certainly contributed greatly to the general failure 
 to recognize the gravity of anoxaemia in persons not adapted. 
 Adaptation is a process which always requires time, and the time 
 factor must, therefore, be taken into account in judging of the 
 physiological effects of anoxaemia. 
 
CHAPTER VII 
 The Causes of Anoxaemia. 
 
 IN the previous chapter anoxaemia has been defined as the condi- 
 tion in which the partial pressure of oxygen, or, what comes to 
 practically the same thing, the amount of free oxygen, in the 
 systemic capillaries generally, is abnormally low. The causes of 
 this condition must now be examined. 
 
 The first and most important cause of anoxaemia is defective 
 saturation of the arterial haemoglobin with oxygen. This may, as 
 we shall see, arise from several causes ; but the most obvious of 
 these is defective partial pressure of oxygen in the alveolar air. 
 It will be shown in Chapter IX that during rest under normal 
 conditions oxygen passes into the blood through the alveolar 
 epithelium by a process of simple diffusion, and that the oxygen 
 pressure in the arterialized blood leaving each alveolus is exactly 
 that of the air in the alveolus. For the purposes of the present 
 discussion we may provisionally assume that this is always the 
 case during rest, so long as the lungs and the inspired air are 
 normal, although modifications in this assumption must be intro- 
 duced later. 
 
 In the light of this assumption and of our knowledge of the 
 dissociation curve of oxyhaemoglobin, it might seem at first that 
 we are justified in assuming that the oxygen pressure of mixed 
 arterial blood is simply that of mixed alveolar air as ordinarily 
 obtained for analysis by the methods already described. In favor 
 of this assumption is the now well-ascertained fact that the 
 breathing is regulated under ordinary conditions in close ac- 
 cordance with the pressure of CO 2 in the mixed alveolar air, as 
 explained in Chapter II. Variations in average alveolar CO 2 pres- 
 sure are thus a direct measure of variations in the CO 2 pressure 
 of the arterial blood ; and it was natural to assume, as was done 
 by myself and others till lately, that variations in alveolar oxygen 
 pressure must also be a measure of variations in the oxygen pres- 
 sure of the arterial blood. One known difficulty in this assumption 
 lay in the fact that the arterial oxygen pressure, as measured in 
 animals by the aerotonometer (Chapter IX) is nearly always 
 lower, and sometimes considerably lower, than the alveolar oxy- 
 gen pressure; but various explanations of this difficulty had been 
 adopted by myself and others. 
 
RESPIRATION 
 
 133 
 
 A new and important light was thrown on the whole subject in 
 the course of a study by Meakins, Priestley, and myself of the 
 "neurasthenia" produced by gassing and other causes during the 
 war. 1 As mentioned in Chapter III, the breathing in these patients 
 is abnormally frequent and shallow, particularly on exertion. It 
 was also found that addition of oxygen to their inspired air was of 
 considerable service during any ordinary exertion, and that in 
 some of them the lips became blue on exertion unless oxygen was 
 given. As there was no sign of anything seriously abnormal in 
 their lungs, we were led to suspect that the shallow breathing was 
 somehow causing anoxaemia. This led us to make experiments 
 
 Figure 43. 
 "Concertina" apparatus for continuous record of respiration. 
 
 on the effects of shallow breathing in normal persons, and for this 
 purpose we devised the apparatus 2 shown in Figure 43. The 
 subject inspires through the mouthpiece and inspiratory valve 
 from the recording "concertina." The bottom of this moves up- 
 wards with inspiration, and records the movement by means of 
 an inked pen on the drum. The bottom comes down on a movable 
 stop, and by moving this upwards the maximum capacity of the 
 concertina can be reduced to whatever is desired. During expira- 
 tion the expired air passes out by the rubber expiratory valve. 
 At the same time the expiratory pressure is communicated to a 
 
 1 Haldane, Meakins, and Priestley, Journ. of Physiol., LII, p. 433, 1919. 
 
 2 Made by Messrs. Siebe Gorman & Co., 187 Westminster Bridge Road, London. 
 
RESPIRATION 135 
 
 tambour the movement of which, as shown, closes a circuit from 
 an accumulator or from the lighting circuit through a rheostat. 
 This circuit passes through an electromagnet which instantly 
 lifts a valve and admits air freely into the concertina, which at 
 once refills itself. At the end of expiration the circuit is instantly 
 broken and the valve closes, so that only the volume of air 
 contained in the concertina can be inspired at the next inspiration. 
 In this way the amount of air taken in per breath can be limited, 
 and a continuous record is at the same time obtained of the depth 
 and frequency of respiration. With the concertina fully open 
 ordinary records of the breathing are obtained, and any gaseous 
 mixture can be supplied through a glass cylinder which incloses 
 the electromagnet and valve. The advantage of this method is 
 that it is capable, not merely of permitting a study of shallow 
 breathing, but also of giving a continuous quantitative record of 
 any sort of breathing. The old stethographic method of recording 
 the breathing is apt to be misleading, since it does not give a 
 quantitative record. 
 
 When the depth of inspiration is limited by means of this ap- 
 paratus the natural impulse, at first, is to continue the inspiratory 
 effort at the end of each inspiration, as the Hering-Breuer reflex 
 has not given the signal for expiration. With a little practice, 
 however, the breathing goes quite easily, and the frequency in- 
 creases in proportion as the depth is diminished. When the depth 
 is greatly limited the breathing becomes very frequent 100 or 
 more a minute. 
 
 On observing the breathing when the depth was gradually more 
 and more limited, we found that the breathing became periodic 
 very readily. As already explained, periodic breathing is a symp- 
 tom of anoxaemia, and this fact led us to try the effect of adding 
 a little oxygen to the inspired air. This promptly abolished the 
 periodic breathing, as shown in Figure 44. There could thus be 
 no doubt that the periodic breathing was due to imperfect oxy- 
 genation of the arterial blood. In some persons, such as myself, 
 the periodic breathing was produced much more readily, and in a 
 more striking degree, than in other persons. This, as already 
 mentioned in Chapter VI, is due to individual differences in the 
 response of the respiratory center to anoxaemia. 
 
 We at first thought that the anoxaemia must be due to the fresh 
 air not penetrating properly to the deep (air-sac) alveoli when 
 the breathing was shallow ; but on examining samples of the deep 
 alveolar air during a prolonged experiment, we were disappointed 
 
136 
 
 RESPIRATION 
 
 to find that in the deepest alveolar air the oxygen percentage, so 
 far from being lower, was actually higher than usual. There was 
 thus hyperpnoea from want of oxygen, and yet the deep alveolar 
 air contained more oxygen than usual. The breathing was, how- 
 ever, very inefficient and therefore greatly increased in amount, 
 as the dead space told much more than with normal breathing, 
 so that the percentage of CO 2 in the expired air was very low. 
 
 On turning the matter over, we bethought ourselves of some 
 anatomical observations collected by Professor Arthur Keith in 
 "Further Advances in Physiology," edited by Professor Leonard 
 Hill, 1909. He showed in this essay that during inspiration the 
 lungs do not expand equally and simultaneously at all parts, 
 but open out part by part, somewhat like the opening of a lady's 
 fan. The parts nearest the moving chest walls (for instance the 
 diaphragm) expand first, and other parts follow. It follows from 
 this, that in shallow breathing the lungs will be very unevenly 
 ventilated. Only certain parts will expand properly, and on ac- 
 
 6.6 6.0 
 
 PRESSURE OF CO^ 
 PERCENT OF AN ATMOSPHERE 
 S3 4.6 4.0 3.3 2.6 
 
 20 /.3 0.7 
 
 024 
 
 68/0/2 14 16 Id 20~Z2 24 26 26 30 32 54 36 38 4042 44 3 
 PERCENT OF AN ATMOSPHERE ^ 
 
 OXYGEN PRESSURE 
 
 Figure 45. 
 Dissociation curves of blood for COa and oxygen. 
 
 count of the increased frequency of breathing they will receive 
 much more than their proper share of fresh air, while the other 
 parts which do not expand will receive much less. 
 
 The consequence of this will be that the venous blood passing 
 through the unexpanded parts of the lungs will be very im- 
 perfectly arterialized, whereas in the expanded parts the blood 
 will be more arterialized than usual. The mixed arterial blood will 
 thus be a mixture of over-arterialized and under-arterialized 
 
RESPIRATION 
 
 137 
 
 blood. To see what the results of this mixture will be, we must 
 refer to the respective dissociation curves for the oxygen and the 
 CO 2 present in blood, taking also into account the action of oxygen 
 in expelling CO 2 from venous blood, as shown on the curve in 
 Figure 26. For convenience' sake the two relevant curves are 
 plotted together in Figure 45, taken from our paper. It will at 
 once be seen that the over-ventilation in some parts of the lungs 
 will wash out CO 2 from the blood in the same proportion as the 
 under-ventilation fails to wash it out. The mixed arterial blood 
 will thus be normal as regards its content of CO 2 if the total al- 
 veolar ventilation is normal. On the other hand, the over- ventila- 
 tion will hardly increase at all the charge of oxygen in the blood 
 from the over- ventilated alveoli, since this blood is on the flat part 
 of the curve with the alveolar oxygen pressure at perhaps 16 or 
 1 8 per cent of an atmosphere. The under-ventilation, on the other 
 hand, will leave the venous blood nearly venous and on the steep 
 part of the oxyhaemoglobin curve, with a large deficiency of oxy- 
 gen. The mixed arterial blood will, therefore, be seriously deficient 
 in oxygen, and symptoms of anoxaemia will consequently be pro- 
 duced. As one of these symptoms is increase in the breathing, there 
 will be some compensation, and the CO 2 percentage of the mixed 
 alveolar air will fall somewhat, there being a corresponding rise 
 also in the oxygen percentage, as was actually found in our ex- 
 periments. 
 
 There is thus a very complete explanation of our experimental 
 results, and also of the symptoms of anoxaemia in the neurasthenic 
 cases ; but clearly it is necessary to modify radically the idea that 
 the alveolar oxygen pressure gives the oxygen pressure of the 
 mixed arterial blood. We have no guarantee that even during 
 quite normal breathing the distribution of air in the individual 
 lung alveoli corresponds exactly with the distribution of blood to 
 them. Unless this correspondence is exact some alveoli will re- 
 ceive more air in proportion to their blood supply than others, 
 and as a consequence the mixed arterial blood will be a mixture 
 of more and less fully arterialized blood, with some of the con- 
 sequences first discussed. It is probable indeed that in some way 
 or other the air supply is proportioned to the blood supply, 
 whether by regulation through the muscular coats of the bron- 
 chioles or regulation of the blood distribution ; but it is also certain 
 that this proportioning is only an approximation. The fact that 
 in animals the aerotonometer gives a lower arterial oxygen pres- 
 sure than the alveolar oxygen pressure (Chapter IX) is most 
 
I 3 8 RESPIRATION 
 
 naturally explained on the theory that the proportioning is only 
 approximate, and there are various other facts which point in the 
 same direction. 
 
 One of these facts is as follows. When the breathing is suddenly 
 interrupted voluntarily the breath can be held for a certain time 
 usually about 40 seconds if only an ordinary breath is inspired 
 before the interruption. Leonard Hill and Flack 3 discovered, how- 
 ever, that if the lungs are filled with oxygen first the breath can 
 be held for two or three times longer ; also that the alveolar CO 2 
 percentage is considerably higher at the breaking point. On the 
 other hand, when the same air was rebreathed continuously from 
 a small bag filled at the start with a breath of alveolar air, the 
 alveolar CO 2 percentage went as high as when the breath was 
 held with oxygen, though not so high as when oxygen was re- 
 breathed from the bag. The following table, illustrating these 
 results, is taken from Hill and Flack's paper. 
 
 It was difficult, at the time, to interpret these results satisfac- 
 torily, since the alveolar oxygen percentages, when the breath was 
 held after breathing ordinary air, did not seem to be low enough to 
 stimulate the breathing appreciably. In order to obtain still more 
 definite information Douglas and I repeated the observations, 
 but in such a way as to have great variations in the alveolar oxy- 
 gen percentage. 4 We then found that the beneficial effects of in- 
 creasing the alveolar oxygen percentage were still evident, though 
 to a diminishing extent, till 1 7 per cent of oxygen was present in 
 the alveolar air. Oxygen in excess of this made no difference. But 
 1 7 per cent is 3 per cent more than what is present in normal al- 
 veolar air; and, as we have already seen, there are no effects on 
 the breathing from want of oxygen when ordinary air is breathed 
 by normal persons, or even when the oxygen percentage of the 
 alveolar runs down to 10 or even 8 per cent. The results were 
 therefore very mysterious at the time, and we were compelled to 
 adopt the improbable hypothesis that holding the breath has some 
 considerable effect on the circulation in the brain, leading to 
 anoxaemia of the respiratory center. There is, however, no reason 
 whatever to expect such an effect. 
 
 The experiments on shallow breathing have furnished the 
 solution to this mystery. It is evident that the relation between 
 blood supply and ventilation in individual groups of alveoli is 
 not an even one. In some alveoli the oxygen runs down and CO 2 
 
 'Leonard Hill and Flack, Journ. of Physiol., XXXVII, p. 77, 1908. 
 4 Douglas and Haldane, Journ. of Physiol., XXXVIII, p. 425, 1909. 
 
o o o 
 
 CO vo T}- 
 
 oq oo upoq 
 
 ^t- rf X 
 ~ ^ d> 
 
 co 
 
 IX O \O IH Os 
 
 tx tx M q w 
 
 od o o* d 6 
 
 O ID 
 
 o 01 
 
 .* 3> 
 
 s^-2 
 
 r 1 
 
 ^6 
 u 
 
 ON M 
 co M 
 
 00 
 
 00 
 
 <N O VO I C^ 
 N 00 ON 00 
 
 00 tx tx tx 
 
 to O *O O 
 
 CO m ON 
 
 q\ 
 
 1/5 Tt- M 
 CO TJ- CO 
 
 IH 
 
 vo OO 10 M tx 
 
 q up w \ o o 
 tx txod od 06 
 
 10 o 
 
 f o 
 
 to <N 00 COOO 
 
 v o <N w 
 
 K 
 
 ** 
 
 v 8 
 
 W tX N M O 
 
 coco oq co tx 
 vd <o vd tx tx 
 
 a 
 
140 
 
 RESPIRATION 
 
 accumulates faster than in others. Hence in some the blood is 
 less perfectly oxygenated; and if the breath is held for a time 
 this imperfect oxygenation becomes more and more marked till 
 at last the mixed arterial blood is very considerably short of oxy- 
 gen, just as when the breathing is very shallow. Hence the oxygen 
 percentage of the mixed alveolar air becomes altogether deceptive 
 as an index of the degree of oxygenation of the mixed arterial 
 blood, although the CO 2 percentage remains, for the reasons al- 
 ready given, a reliable index of the degree of saturation of arterial 
 blood with CO 2 . The results of these experiments on holding the 
 breath are thus very valuable as furnishing evidence that, even 
 with normal or increased inspirations, the relation between blood 
 supply and air supply varies considerably in different alveoli. 
 
 That the arterial blood does actually become imperfectly oxy- 
 genated when the breath is held has been quite recently demon- 
 strated by Meakins and Davies. 5 They found that on holding a 
 deep breath of air for 40 seconds, the haemoglobin of the arterial 
 
 -?r 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r 
 
 * 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 >* 
 
 
 ~^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 %H 
 
 u 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 & 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 a 30 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 40 60 80 100 120 M 
 
 '6O 180 200 220 240 260 280 300 320 540 36O 380 400 420 440 
 PRESSURE OF O z iN MM.H<J. 
 
 Figure 46. 
 Alveolar CO 2 during breath holding after inhalation of oxygen. 
 
 blood drawn from the radial artery was only 83.8 per cent satu- 
 rated with oxygen, although the mixed alveolar air contained 
 13.4 per cent of oxygen. Had air of this composition been dis- 
 tributed evenly throughout the alveoli the haemoglobin would 
 have been 97 per cent saturated with oxygen. 
 
 A further series of experiments which Douglas and I performed 
 is very instructive in this connection. As already mentioned in 
 Chapter V, the alveolar CO 2 percentage rises high above its 
 normal value before the end of an apnoea after forced breathing 
 with extra oxygen. We observed how high the alveolar CO 2 pres- 
 sure went when there were varying pressures of oxygen in the 
 
 'Meakins and Davies, Journ. of Pathol. and, Bacter., XXIII, p. 451,1920. 
 
RESPIRATION I 4 I 
 
 mixed alveolar air at the end of the apnoea produced by two 
 minutes of forced breathing, and the results are plotted in Figure 
 46. 6 It will be seen that the CO 2 pressure (and of course also the 
 length of the apnoea) rises with the alveolar oxygen pressure 
 until the latter reaches about 120 mm. (corresponding to about 
 17 per cent of oxygen in the dry alveolar air), beyond which a 
 further rise in alveolar oxygen pressure has no effect. In this case 
 the oxygen pressure in all the lung alveoli would be at a more or 
 less equal high level at the beginning of an apnoea, but would 
 fall at unequal rates in the different alveoli. Accordingly at the 
 end of apnoea the mixed arterial blood would' be getting venous 
 unless the average alveolar air contained more than 17 per cent 
 of oxygen. And yet as little as 8 per cent would be enough to 
 prevent this effect if the air was evenly distributed in relation to 
 the blood supply of the alveoli, or if respiratory movements pre- 
 vented anything more than comparatively slight variations in 
 the oxygen percentages in different alveoli. 
 
 Judging from aerotonometer experiments on normal animals, 
 and from direct determinations on human arterial blood, the hae- 
 moglobin of average human arterial blood is only about 94 to 96 
 per cent saturated with oxygen about 2 per cent less than if the 
 whole arterial blood was saturated to the oxygen pressure of the 
 mixed alveolar air. A very accurate series of determinations 
 described by Meakins and Davies in the paper just quoted showed 
 that in different healthy persons the saturation varies from 94 to 
 96 per cent. The slight variations seem to be due to the variations 
 which Barcroft described in the oxyhaemoglobin dissociation 
 curves of different individuals. 
 
 The periodic breathing produced by shallow breathing differs 
 strikingly from the periodic breathing produced by anoxaemia 
 in normal persons. As will be seen from Figure 44, the periods 
 are much longer, and in this respect bear a striking resemblance 
 to ordinary clinical Cheyne-Stokes breathing. The reason why 
 the periods are longer is evident enough : for the shallow breath- 
 ing is very ineffective in raising the oxygen percentage in the 
 badly ventilated parts of the lungs and so relieving the anox- 
 aemia. The relief thus comes slowly. The breathing, therefore, 
 "waxes and wanes" gradually, as in clinical Cheyne-Stokes 
 breathing. In hibernating animals similar breathing is often ob- 
 served and can be explained in the same way, as, owing to the 
 small production of CO 2 , the breathing is very shallow. 
 
 9 Douglas and Haldane, Journ. of Physiol., XXXVIII, p. 401, 1909. 
 
I 4 2 RESPIRATION 
 
 Ordinary clinical Cheyne- Stokes breathing is evidently a symp- 
 tom of anoxaemia due often to the shallow breathing which char- 
 acterizes a failing respiratory center. This failure may be that of 
 approaching death, since the anoxaemia itself tends to hasten the 
 failure of the center, as already explained in Chapter VI. There 
 is thus a vicious circle which, unless broken in some way, must end 
 in death from anoxaemia, just as in the case of an airman at a 
 dangerously high altitude. The color of the lips, in conjunction 
 with the diminishing depth of the breathing, points clearly to 
 what is happening. 
 
 It is now evident that the anoxaemia so often present in disease, 
 but so seldom recognized as such, is due in a large number of 
 cases to the shallow breathing characteristic of a damaged or 
 "fatigued" respiratory center, whatever the original cause of the 
 damage or fatigue may be. It is also evident that frequency of 
 breathing has assumed a significance which it did not previously 
 possess, since frequency is very often an index of shallowness of 
 breathing, damage to the respiratory center, and consequently 
 impending danger from anoxaemia. The frequent and shallow 
 breathing in surgical shock, or in various forms of influenza and 
 pneumonic conditions, or as it may occur in many other forms of 
 disease, is a symptom of which the possible deadly import will be 
 evident enough to those who have read the preceding chapter in 
 connection with what has just been said. In this connection I 
 should like also to emphasize the fact that, as fully explained in 
 the last chapter, it is unsafe to judge of the degree of anoxaemia 
 by the degree of cyanosis. The anoxaemia is, and must be, ac- 
 companied by alkalosis, so that the oxyhaemoglobin holds on more 
 tightly to its oxygen, and this alkalosis may become extreme with 
 very shallow and rapid breathing. 
 
 Chronic fatigue or failure of the respiratory center is seen in 
 neurasthenia and various other forms of disease; but failure of 
 the respiratory center may also occur in acute and sudden attacks, 
 which are often associated, either primarily or secondarily, with 
 anginal pain. The patient may feel that he cannot expand his 
 chest to breathe, just as if it were mechanically constricted; and 
 he rapidly develops asphyxial symptoms, with very frequent and 
 shallow breathing. In reality, apparently, he is in the grip of the 
 Hering-Breuer reflex, which, as explained in Chapter III, assumes 
 exaggerated influence, owing to the failure of the respiratory 
 center. These attacks, though they usually pass off, are sometimes 
 very dangerous; and many sudden deaths appear to be due to 
 
RESPIRATION 
 
 143 
 
 them. They are specially liable to occur at night. The rapid 
 breathing is apt to produce the impression in a physician that it is 
 the heart and not the breathing, that has failed ; and this impres- 
 sion may be apparently confirmed by the presence of secondary 
 anginal pain. In all doubtful cases, the effects of properly admin- 
 istering oxygen will decide the diagnosis. If the immediate cause 
 of the symptoms is failure of the respiratory center, the effects 
 of the oxygen are rapid and prompt, and have been so in cases 
 which have chanced to come under my own observation. 
 
 It is evident that anoxaemia caused by irregular distribution 
 of oxygen among the lung alveoli may be due to a variety of 
 causes. One of these is emphysema ; for the emphysematous parts 
 of a lung will naturally be supplied with far more than the proper 
 proportion of air to suit their greatly diminished respiratory 
 surface, while the other parts will receive correspondingly less 
 air. The arterial blood will thus be a mixture of over-arterialized 
 I 2 
 
 llllllllllltllllllllllllllllllllllllllM 
 
 Figure 47. 
 
 Subject J. S. H. Rebreathing in and out of 50 liter cylinder. Time marker = 
 2 seconds, i. Sitting. 2, Lying. 
 
 and under-arterialized blood, with resulting anoxaemia, which 
 may or may not be compensated by one or other of the processes 
 to be described in succeeding chapters. 
 
 Another cause of the same general character is bronchitis or 
 asthma. The irregular partial blocking or muscular constriction 
 of the bronchi and bronchioles in these conditions must lead to 
 
144 
 
 RESPIRATION 
 
 irregular distribution of fresh air to the alveoli, even though the 
 average distribution, as shown by the volume of air breathed, is 
 greatly increased. Hence the mixed arterial blood will be deficient 
 in oxygen, and grave anoxaemia may develop. Here, also, the 
 effects of oxygen administration will decide the diagnosis of the 
 condition. 
 
 We found that the recumbent position greatly favors the de- 
 velopment of periodic breathing, and therefore of anoxaemia. 
 We also found that when a normal person assumes the recumbent 
 position, the usual result is that the breathing becomes slower and 
 deeper. In my own case, for instance, the frequency diminishes 
 from about 15 in the sitting or upright position to 7 or 8, while 
 the depth correspondingly increases, so as to keep the alveolar 
 CO 2 pressure nearly the same (see Figure 47). The cause of this 
 phenomenon is not altogether clear, but is probably the increased 
 resistance thrown on the diaphragm in the recumbent position, 
 as the weight of the liver and other abdominal organs assists the 
 descent of the diaphragm in the upright position. Rontgen ray 
 
 Figure 48. 
 
 Subject J. G. P. i. Breathing restricted by concertina lying. 2. Breathing 
 restricted to same extent sitting. 3. Breathing further restricted sitting. Oxy- 
 gen given. Curves read left to right. Inspiration upstroke. Time marker = 
 seconds. 
 
 photographs which we took to show the position of the diaphragm 
 favored this explanation ; and, assuming it to be correct, the effect 
 of the recumbent position may well be similar to the slowing 
 effect produced by resistance as shown in Chapter III. Whatever 
 
RESPIRATION 
 
 145 
 
 the cause of the natural increased depth may be, it is evident that 
 in the recumbent position the tendency to irregular distribution 
 of fresh air in the lung alveoli with any given depth of breathing 
 is much increased, so that anoxaemia from this cause, as shown 
 in normal persons by periodic breathing, is much more readily 
 produced. In my own case periodic breathing is rapidly produced 
 in the recumbent position when the breathing is kept at over 20 
 per minute by artificially limiting the depth by means of our 
 apparatus, whereas in the upright position there is no such effect. 
 The effect of the recumbent position is shown in Figure 48. 
 
 We have thus a simple explanation of a phenomenon which 
 has been familiar to physicians since early times, but which has 
 hitherto never been satisfactorily explained. When patients are 
 
 42 
 
 41 
 40 
 39 
 
 10 
 
 fc 37 
 | 36 
 
 ! 
 
 I 34 
 1.33 
 3*32 
 
 V. 
 
 I 31 
 *30 
 
 29 
 
 28 
 
 800 700 600 500 400 
 Barometric Pressure m mm of mercury 
 
 Figure 49. 
 
 Effects of diminished barometric pressure on the 
 alveolar gas-pressures. The thick lines show the 
 alveolar CC>2 pressure, and the thin lines the alveolar 
 O 2 pressure. The dotted lines refer to the experiment 
 in which oxygen was added to the air. 
 
 short of breath during illness they are often very uncomfortable 
 in the recumbent position, and may become dangerously worse 
 if not propped up in bed or in a chair. This condition is known as 
 
 30Q 
 
I 4 6 RESPIRATION 
 
 orthopnoea, and its causation now seems evident. With a failing 
 respiratory center, and consequent abnormal shallowness of respi- 
 ration, anoxaemia is the natural result of the recumbent position; 
 and the prevention of this anoxaemia by keeping the patient in a 
 sitting position becomes an important part of treatment unless 
 the same object is attained by oxygen administration. 
 
 Defective distribution of air in the lung alveoli is, of course, 
 only one of the causes of defective oxygenation of the arterial 
 blood; but I have dealt with this cause first, not only because it 
 is of very great importance in medicine, but because an under- 
 standing of it is essential to the understanding of other causes of 
 defective oxygenation. 
 
 A second and hitherto much better known cause of defective 
 oxygenation of the arterial blood is a deficiency in the partial 
 pressure of oxygen in the inspired air, and consequent fall in the 
 alveolar oxygen pressure. As shown in Chapter II, it usually re- 
 quires a fall in oxygen percentage from the normal of 20.9 to 
 about 14 per cent, or a third, before any evident effect on the 
 breathing is produced at the time by the oxygen deficiency. Simi- 
 larly a fall of about a third in barometric pressure (corresponding 
 to about 1 1,500 feet above sea level) is required. Figure 49, from 
 a paper by Boycott and myself, 7 shows that until the barometric 
 pressure in a steel chamber falls by about a third, the normal 
 alveolar CO 2 pressure is very little disturbed. The alveolar CO 2 
 percentage simply goes up as the barometric pressure goes down, 
 but the pressure of CO 2 remains almost the same in the alveolar 
 air. In the same investigation we found that even when the bar- 
 ometric pressure was reduced to 300 mm. the alveolar CO 2 pres- 
 sure remained the same, provided that any excessive fall in the 
 oxygen pressure of the inspired air was prevented by adding oxy- 
 gen to the air of the chamber. There is thus no trace of foundation 
 for Mosso's contention 8 that the diminished mechanical pressure 
 of the air produces by itself a diminished saturation of the blood 
 with CO 2 . 
 
 Since the alveolar air, with the breathing normal, contains about 
 a third less oxygen than the inspired air, it follows that when the 
 oxygen percentage or partial pressure in the inspired air is reduced 
 by a third the alveolar oxygen percentage will be reduced to about 
 half i.e., from about 13 per cent of an atmosphere to about 6.5 
 per cent. On comparing this with the dissociation curve of oxy- 
 
 T Boycott and Haldane, Journ. of Physiol., XXXVII, p. 355, 1908. 
 8 Mosso, Life of Man on the High Alps, London, p. 287, 1898. 
 
RESPIRATION 
 
 147 
 
 haemoglobin it will be seen that such a diminution corresponds 
 to a saturation of about 80 per cent of the haemoglobin with oxy- 
 gen, and that any further diminution will cause a rapid fall in the 
 saturation. The air produces at the time no noticeable discomfort, 
 and the breathing is not sensibly affected, although the lips are 
 slightly bluish. The natural conclusion is that a diminution of 
 about 15 per cent in the saturation of the haemoglobin, or a dimi- 
 nution to half in the arterial oxygen pressure, is of no physio- 
 logical importance, even though the lips are rather dull in color. 
 This wholly mistaken idea is, however, rudely shaken by the 
 effects of remaining for a sufficient time in the atmosphere : for 
 the observer will be almost certainly prostrated by an attack of 
 mountain sickness which he is never likely to forget afterwards. 
 
 If, now, in order to escape mountain sickness, the pressure of 
 oxygen in the inspired air is only diminished by one-seventh (cor- 
 responding to a height of 4,500 feet; or an oxygen percentage of 
 17 at ordinary atmospheric pressure), there will be no appreciable 
 blueness, and the corresponding saturation on the oxyhaemoglobin 
 dissociation curve will be only 3.5 per cent below that for normal 
 alveolar air. Nevertheless there will, if sufficient time is given, be 
 quite appreciable physiological responses, which will be discussed 
 in succeeding chapters. The truth is that in the long run the body 
 responds in a fairly delicate manner to quite small diminutions 
 in the oxygen pressure of the inspired air. 
 
 Let us now look at the matter in the light of the new knowledge 
 as to the somewhat imperfect manner in which air is distributed 
 in the alveoli. In the course of our investigation on military 
 neurasthenia, we placed several of the patients in a steel chamber 
 and observed the effects of diminished pressure. A very slight dim- 
 inution, corresponding to only about 5,000 feet, was sufficient to 
 produce in them urgent respiratory and other symptoms, although 
 they were doing no work. Even in normal persons the dissociation 
 curve of oxyhaemoglobin and composition of the mixed alveolar 
 air are, as was shown above, no certain guides to the percentage 
 saturation of the haemoglobin, or oxygen pressure in the mixed 
 arterial blood. As a matter of fact the blueness of the lips seen in 
 persons freshly exposed to very low atmospheric pressure seems 
 to be often much greater than would correspond to the oxygen 
 pressure in their alveolar air when due allowance is made for the 
 Bohr effect of lowered alveolar CO 2 pressure. We may thus be 
 quite sure that at diminished atmospheric pressure the saturation 
 
148 RESPIRATION 
 
 of the mixed arterial blood with oxygen is or may be distinctly 
 lower than corresponds to the oxygen pressure of the alveolar air. 
 
 Poulton and I found that when a small quantity of air about 
 6 liters was rebreathed continuously up to the verge of loss of 
 consciousness, the CO 2 being completely absorbed by soda lime, 
 the inspired air contained only 4.8 per cent of oxygen, and the 
 alveolar air 3.7 per cent. There was very great hyperpnoea; for 
 the preformed CO 2 had not had time to escape in the manner 
 already referred to in Chapter VI. The respiratory quotient of 
 the alveolar air was as high as 2.8. The experiment was then 
 repeated with a large volume of air, and under such conditions 
 that the oxygen percentage only fell very slowly. The lowest per- 
 centage of oxygen that could now be reached in the inspired air 
 without great confusion of mind was about 9.4, with about 4.6 
 per cent (or 33 mm.) in the alveolar air. There was no noticeable 
 hyperpnoea, and the respiratory quotient was normal. The al- 
 veolar CO 2 percentage was only reduced from the normal of 5.7 
 per cent to 4.6, indicating that the alveolar ventilation was only 
 increased by about a fourth. 
 
 From these experiments we may conclude that air containing 
 less than 9.5 per cent of oxygen would ordinarily cause disable- 
 ment within half an hour. At a barometric pressure of 368 mm., or 
 a little less than half an atmosphere, corresponding to about 
 20,500 feet above sea level, there would be a corresponding drop 
 in the alveolar oxygen pressure; but judging from my own ob- 
 servations the physiological effects are very distinctly less severe. 
 This is probably due to the fact that in rarefied air the diffusion of 
 oxygen within the lung alveoli is much more free than at atmos- 
 pheric pressure. 9 As a rule no very serious symptoms are ex- 
 perienced at the time till the barometric pressure has fallen to 
 about 350 mm. (corresponding to 21,500 feet) ; but in this respect 
 different individuals vary considerably. It must also be borne in 
 mind that nervous symptoms of anoxaemia begin to appear at 
 altitudes not nearly so great. At 320 mm. (about 24,000 feet) 
 most persons, including myself, are soon very seriously affected 
 in the manner described in Chapter VI, unless they are acclima- 
 tized. 
 
 Another cause of imperfect oxygenation of the arterial blood 
 is that there may not be sufficient time for the required quantity 
 of oxygen to pass into the blood through the alveolar epithelium. 
 This cause of anoxaemia came into prominence in connection with 
 
 * Haldane, Kellas, and Kennaway, Journ. of Physiol., LIU, p. 195, 1915. 
 
RESPIRATION 
 
 149 
 
 the effects of lung-irritant poison gas during the war. It was evi- 
 dent from the first cases which I saw in April, 1915, that there was 
 acute anoxaemia due to imperfect oxygenation of the arterial 
 blood. There were the ordinary chlorine symptoms of acute bron- 
 chitis, alveolar inflammation, and oedema of the lungs. The faces 
 of the patients were deeply cyanosed, in spite of considerably in- 
 creased breathing of adequate depth. At first it was suspected that 
 the cyanosis was due to "toxaemia," causing the formation in the 
 blood of methaemoglobin or some similar dark-colored decompo- 
 sition product; but on diluting a drop of the blood, saturating 
 with CO, and comparing the solution with the tint of similarly 
 treated normal blood, I found that there was no abnormal pigment 
 present, so that the blue color was due simply to anoxaemia. That 
 this anoxaemia was, in the main at least, due to delay in the pas- 
 sage of oxygen into the arterial blood was then confirmed by the 
 fact that on administering oxygen the blue color changed to red, 
 and the patients improved in other respects. It was evident that 
 with the greatly increased partial pressure of oxygen in the al- 
 veolar air, the oxygen was able to pass into the blood at a sufficient 
 rate to saturate or nearly saturate the blood, and thus maintain 
 life. The delayed passage was probably due mainly to the fact 
 that the alveolar walls were swollen and oedematous, so that they 
 did not allow oxygen to pass inwards at a normal rate. As will be 
 pointed out in Chapter IX, this condition was produced experi- 
 mentally in animals by Lorrain Smith. The distribution of air in 
 the lung alveoli was doubtless also gravely interfered with by the 
 bronchitis and emphysema caused by the actions of chlorine, 
 though at the time I was ignorant of the importance of this cause. 
 To judge by the increased breathing there was also much dis- 
 turbance in the excretion of CO 2 by the lungs; and the great dis- 
 tention of the veins and other signs in the chlorine cases pointed 
 in this direction also. 
 
 In the cases of poisoning by phosgene and other lung irritants 
 used later, the symptoms of irritation of the air passages were 
 much less prominent. The general symptoms corresponded more 
 closely with those of pure anoxaemia. This was particularly true 
 in the earlier seen, or less severe, cases, when there was no evi- 
 dent oedema of the lungs. Thus, at first, the symptoms of acute 
 anoxaemia were shown only on muscular exertion sufficient to 
 cause a greatly increased need for oxygen ; and some of the men 
 who were apparently at the time only slightly affected lost con- 
 sciousness or died as a result of muscular exertion. Others suf- 
 
150 RESPIRATION 
 
 fered only from general malaise or symptoms similar to those of 
 mountain sickness, and apparently due to slight anoxaemia. In 
 the graver cases the anoxaemia was usually unaccompanied by 
 distention of the lips and veins with blood, and the cyanosis was 
 thus of the leaden or gray type, just as in cases of slowly advancing 
 anoxaemia from other causes. In death from gradual CO poison- 
 ing, for instance, there is no extra distention of the lips or veins 
 with blood, although, of course, the lips are not gray but light 
 pink. Death, in the phosgene cases and probably in others, seems 
 to have been finally due to failure of the respiratory center, the 
 breathing becoming more and more shallow till the resulting 
 increase in the anoxaemia ended in death. Orthopnoea was a very 
 common symptom so long as the men were conscious. 
 
 In favorable cases of ordinary croupous pneumonia the lips 
 remain of a good color, and there are no evident signs of anoxae- 
 mia ; but the breathing is rapid, and correspondingly shallow. 
 The danger of anoxaemia is therefore not far off. At Cripple 
 Creek (at an altitude of about 10,000 feet) I was told that cases of 
 commencing pneumonia were at once put on the train and sent 
 down to the prairie level, as it had been found that they had a 
 very poor chance if treated locally. This indicates the danger 
 from anoxaemia, and led us, in the Report of the Pike's Peak 
 Expedition, to advocate the use of chambers containing air en- 
 riched with oxygen for treating pneumonia. The fact that there 
 is often no cyanosis in spite of very extensive lung consolidation 
 seems to show that the pulmonary circulation has practically 
 ceased in the consolidated areas. The blood supply of these areas 
 may be solely through the bronchial arteries, the high-pressure 
 supply from which joins the pulmonary circulation. This inference 
 has recently been confirmed by Gross, 10 who found by means of 
 X-ray photographs of lungs injected with an injection mass 
 opaque to X-rays, that the pulmonary vessels are nearly blocked 
 off in the consolidated parts in pneumonia. In the unaffected parts 
 of the lungs, the oxygen seems to penetrate the alveolar walls 
 readily enough in pneumonia. Where anoxaemia becomes danger- 
 ous in croupous or disseminated pneumonia it seems usually to be 
 failure of the respiratory center and consequent shallow breathing 
 that is mainly responsible for the anoxaemia. 
 
 The fact that in pneumonias of all kinds the arterial blood is 
 commonly more or less imperfectly saturated with oxygen has 
 
 Gross, Canadian Med. Assoc. Jourtt., p. 632, 1919. 
 
RESPIRATION 151 
 
 quite recently been shown directly by Stadie, 11 who examined 
 samples of arterial blood drawn usually from the radial artery by 
 means of a syringe. In normal persons he found an average of 
 95 per cent saturation of the haemoglobin with oxygen ; and this 
 is about what might be expected in view of what has been said 
 above. In cases of pneumonia the saturation varied from 95 to 
 42 per cent; and as a rule the cases where the saturation was 
 below 76 per cent ended fatally. Cardiac cases were soon after- 
 wards investigated by Harrop, 12 who found that in many of them 
 there was imperfect saturation of the arterial blood. This was 
 almost certainly due, frequently, to partial failure of the respira- 
 tory center and consequent shallow breathing. 
 
 The significance of these analyses will be evident from what 
 has been said in the previous and present chapters ; and the danger 
 to a patient of permitting any serious arterial anoxaemia to con- 
 tinue when it can be prevented is, I hope, already evident. 
 
 As anoxaemia is such a common and often dangerous condition, 
 and can frequently be combated by the addition of oxygen to the 
 inspired air, it will be in place to refer here to clinical methods 
 of administering oxygen. In the first place it is necessary to have 
 clear ideas as to the objects aimed at, in administering oxygen. If 
 the oxygen is only given to enable a patient to surmount some quite 
 temporary crisis due to anoxaemia produced, it may be, by one 
 of the sudden angina-like attacks of reflex restriction of breath- 
 ing referred above a very simple method of administration will 
 suffice. A small cylinder of oxygen furnished with an india-rubber 
 tube by means of which a stream of oxygen may be directed into 
 the patient's open mouth will suffice; and such an arrangement 
 would probably often be useful in certain cases, as the oxygen 
 could be given promptly by a competent nurse at any time. 
 
 In the great majority of cases, however, the cause of the an- 
 oxaemia is one which may last for a considerable time, so that 
 the administration of oxygen, in order to be useful, must be 
 continued. In this connection it should be clearly realized that 
 the object of the oxygen administration is not simply palliative, 
 but curative. By preventing the anoxaemia we not only avert 
 temporarily a cause of danger or damage to the patient ; but give 
 the body an interval for recovery from the original cause, what- 
 ever it may be, of the anoxaemia, or for adaptation. We also break 
 a vicious circle: for if the anoxaemia is allowed to continue, it 
 
 11 Stadie, Journ. of Exper. Med., XXX, p. 215, 1919. 
 a Harrop, Journ. of Exper. Med., XXX, p. 241, 1919. 
 
152 RESPIRATION 
 
 will itself make the patient worse, or tend to prevent the recovery 
 which would otherwise naturally occur. We are not dealing with 
 a machine, but with a living organism; and a living organism 
 always tends to return to the normal if the opportunity is given. 
 
 Oxygen is still often given by methods which are either quite 
 ineffective or extremely wasteful. One method is to place a funnel 
 over the patient's face, and allow some quite indefinite amount of 
 oxygen to pass into the funnel. By this method the patient re- 
 breathes a good deal of expired air, but may hardly get any of 
 the oxygen, as the latter, being heavier, runs out below. A far 
 better method is to insert a rubber catheter or other soft tube into 
 the patient's mouth or nose, and pass a stream of oxygen through 
 the tube. Another good method, when pure oxygen has to be given, 
 is to allow the oxygen to pass at a sufficient rate into a rubber bag 
 connected with the inspiratory valve of an anaesthetic mask placed 
 over the patient's mouth and nose. The patient inhales from the 
 bag, and exhales to the outside through the expiratory valve in 
 the mask. 
 
 In ordinary cases the patient does not require pure oxygen, but 
 only a sufficient addition to the air of oxygen to prevent the an- 
 oxaemia. In any case it would be very undesirable to continue the 
 administration of pure oxygen for more than a limited time, as 
 pure, or nearly pure, oxygen has a slow irritant action on the 
 lungs, as will be shown in Chapter XII. If the mask is left open 
 to the air, so that the patient can breathe as much air as he likes, 
 and a stream of oxygen is allowed to pass into the mask directly, 
 the oxygen which passes in during expiration is of course wasted. 
 
 It became evident during the war that an efficient apparatus 
 for the continuous administration of oxygen with maximum econo- 
 my in oxygen was greatly needed, particularly in the treatment 
 of acute cases of poisoning by lung-irritant gas. I therefore de- 
 vised an apparatus so arranged that by a simple device the patient 
 inspired through a face piece the whole of the added oxygen, 
 without waste during expiration, while the proportion of oxygen 
 could easily be cut down or increased, according as was needful. 
 The original form of this apparatus was described in the British 
 Medical Journal, February 10, 1917, page 181, after it had already 
 been supplied extensively to the army in France. Its use there for 
 gas cases was initiated, and' the management of it carefully inves- 
 tigated, by Lieutenant Colonel C. G. Douglas of Oxford. Other 
 well-known medical officers also made very valuable observations 
 on the effects of oxygen inhalation. The results, particularly in 
 
RESPIRATION 
 
 153 
 
 gas cases, were strikingly successful; and practically continuous 
 administration could easily be carried out over the two or three 
 days during which there was danger from anoxaemia. Patients 
 can sleep comfortably during the administration. 
 
 The apparatus was afterwards simplified, with the special object 
 of making it both easy for a nurse to handle, and available for 
 front line and stretcher work, including treatment of "shock" 
 cases. Figure 50 shows the arrangement of the apparatus. It con- 
 sists of : ( I ) an oxygen cylinder provided with an easily worked 
 
 f/ICf 
 
 p/ece 
 
 Figure 50. 
 Apparatus for administering oxygen. 
 
 and efficient main valve; (2) a pressure gauge showing how much 
 oxygen is in the cylinder; (3) a reducing valve which reduces 
 the pressure to a small amount which remains constant till the 
 cylinder is exhausted; (4) a graduated tap indicating the flow of 
 oxygen in liters per minute; (5) thick- walled rubber tubing con- 
 veying the oxygen to the patient and a light rubber bag; (6) a 
 face piece with a minimum of dead space, and provided with 
 elastic straps and a pneumatic cushion which can be taken off for 
 disinfection. 
 
 The patient can inspire and expire freely through an opening 
 in which there is a rubber flap to cause a very slight resistance. 
 During expiration the oxygen collects in the bag, and is sucked 
 into the face piece at the beginning of inspiration. From the move- 
 ments of the bag it can be seen at any time whether the patient 
 is receiving the oxygen. To put the apparatus in action the main 
 valve is opened freely, and the tap is adjusted to give 2 liters a 
 minute or whatever greater or less amount suffices. With a de- 
 
154 RESPIRATION 
 
 livery of 2 liters a minute a 40- foot cylinder would last nearly 
 ten hours. 
 
 The effects of continuous oxygen inhalation with this apparatus 
 on the arterial blood in pneumonia and bronchitis have quite 
 recently been investigated by Meakins. 13 He found that with 2 
 liters a minute the percentage saturation of the haemoglobin in a 
 pneumonia case with almost complete consolidation of one lung 
 rose from 82 per cent to 91 per cent, but went back on stopping the 
 oxygen to 84 per cent, slight cyanosis returning also. On then 
 giving 3 liters a minute, the saturation rose to 97 per cent, which 
 is 2 per cent above the normal value for healthy persons. In a 
 bronchitis case with slight cyanosis and orthopnoea, the satura- 
 tion rose from 88.6 to 97.0 per cent on giving 2 liters a minute, 
 and the cyanosis and orthopnoea disappeared. In a normal man 
 the saturation rose from 95.6 to 98.1 on giving 2 liters a minute. 
 
 The plan of treating patients in an air-tight chamber contain- 
 ing a high percentage of oxygen was introduced towards the end 
 of the war at Cambridge under Barcroft's direction; 14 and a 
 similar chamber was erected at Stoke-on-Trent. Favorable results 
 were obtained in chronic cases of gas poisoning, as might be 
 anticipated in view of the disturbed nervous control of breathing, 
 already described in Chapters III and VII. It now seems evident 
 that the administration of air enriched with oxygen is likely to 
 be successfully introduced in the treatment of various illnesses 
 in which arterial anoxaemia is present. 
 
 During considerable muscular exertion the rate at which oxy- 
 gen has to penetrate from the alveoli into the blood is enormously 
 increased. Hence it is during muscular work that we should ex- 
 pect to find any signs of anoxaemia in healthy persons breath- 
 ing normal air at normal atmospheric pressure. That a certain 
 amount of anoxaemia is commonly produced can be shown 
 indirectly in various ways. In the first place the alveolar CO 2 
 pressure, particularly in some persons, does not rise during mus- 
 cular exertion in the proportion that would be expected if the 
 increased breathing were simply due to the increased production 
 of CO 2 and consequent rise in the alveolar CO 2 pressure. Thus in 
 the experiments of Priestley and myself, my own alveolar CO 2 
 pressure rose only by .13 per cent, in place of an expected rise of 
 
 11 Meakins, Brit. Med. Journ., March 5, 1920. A number of further cases have 
 still more recently been recorded by Meakins, Journ. of Pathol. and, Bacter., XXIV, 
 p. 79, 1921. 
 
 14 Barcroft, Dufton, and Hunt, Quarterly Journ. of Medicine, XIII, p. 179, 
 1920. 
 
RESPIRATION 
 
 155 
 
 about .8 per cent, if the increased breathing had been due to CO 2 
 alone; while in the case of Priestley (who was in much better 
 physical training than I was) the rise was .44 per cent in place of 
 an expected rise of about .56. I have since then frequently found 
 that my alveolar CO 2 pressure does not rise appreciably with 
 muscular exertion, and falls if the exertion is very great; though 
 in younger men there is almost always a marked rise, as in the 
 experiments on Douglas, mentioned in Chapter II. The absence 
 of a rise in me when ordinary air is breathed is not due to the 
 formation of lactic acid referred to in Chaper VIII. I found in 
 1917, however, that there is a well-marked rise when a little oxy- 
 gen is added to the inspired air. The failure of my alveolar CO 2 
 to rise was therefore due apparently to slight anoxaemia during 
 muscular exertion. 
 
 It has for long been well known to engineers that men perform 
 hard physical work more easily when they are working in com- 
 pressed air. This was very evident, for instance, during the work 
 on the Blackwall tunnel under the Thames, which I visited about 
 25 years ago. At the existing air pressure the alveolar oxygen 
 pressure would have 3^/2 times its normal value. In breathing 
 nearly pure oxygen while wearing a mine rescue apparatus, I 
 share the very common experience, that in spite of the weight of 
 the apparatus, heavy exertion, such as walking very fast, is much 
 easier. On the other hand, even a very moderate increase in alti- 
 tude increases considerably the panting on exertion. 
 
 Some years ago Hill and Flack 15 published a number of ob- 
 servations on the apparent effects of oxygen before and after 
 muscular exertion. Many of their observations were concerned 
 with very striking effects, already referred to, of oxygen in pro- 
 longing the time during which the breath can be held. They 
 showed that this effect is just as marked when exertion is per- 
 formed with the breath held as during rest. They also found that 
 oxygen given during the distress immediately following severe 
 exertion has a distinct effect in raising the blood pressure, improv- 
 ing the pulse, and alleviating the distress. This indicates that a 
 raised partial pressure of oxygen in the alveolar air increases the 
 oxygenation of the blood, and that part of the distress caused by 
 severe muscular work is caused by deficient oxygenation of the 
 arterial blood. I am unable to agree, however, with their further 
 conclusion that when oxygen is breathed a large amount of free 
 
 15 Hill and Flack, Journ. of PhyswL, XXXVIII, Pro. PhyswL Soc,, p. xxviii, 
 1909 ; and XL, p. 347, 1910. 
 
156 RESPIRATION 
 
 oxygen is stored in the blood and tissues, and that for this reason 
 a man who has breathed oxygen for a time has a distinct physio- 
 logical advantage as regards performance of work over a man 
 who has simply breathed air. Douglas and I found 16 that if oxy- 
 gen is breathed quietly before an exertion there is no physiological 
 advantage if the breath is not held. The extra oxygen in the lungs 
 is quickly washed out by the breathing, and there is nothing to 
 indicate the existence of any other extra store of oxygen in the 
 body. If, however, the breathing is forced before the exertion, 
 there is considerable advantage whether air or oxygen is breathed 
 during the forced breathing ; and this advantage is due simply to 
 washing out of CO 2 . As will be shown in Chapter XII, the tis- 
 sues and venous blood cannot become highly saturated with oxy- 
 gen when this gas is simply breathed at ordinary atmospheric 
 pressure; and if oxygen had any appreciable effect apart from 
 that due to the actual presence of an increased percentage of oxy- 
 gen in the lungs the result would be very unintelligible. 
 
 A clear and striking light has been thrown on this subject by 
 some recent experiments by Dr. Henry Briggs. 17 He found that 
 when equal work is done on a Martin's ergometer the percentage 
 of CO 2 in the expired air is, in persons not in good physical train- 
 ing, considerably higher when air rich in oxygen is breathed than 
 when ordinary air is breathed. In persons in the best physical 
 training, on the other hand, there is practically no difference until 
 the work done is very excessive. The following table is from his 
 
 PERCENTAGE CO a IN EXPIRED AIR 
 
 Work in foot pounds 
 
 Subject A 
 
 Subject 
 
 B 
 
 per minute 
 
 Breathing 
 
 Breathing 
 
 Breathing 
 
 Breathing 
 
 Pedaling with brake off 
 
 air 
 
 oxygen 
 
 air 
 
 oxygen 
 
 
 3-9 
 
 4-1 
 
 4-4 
 
 4-5 
 
 3,000 
 
 4-65 
 
 5.25 
 
 5-3 
 
 5.45 
 
 6,000 
 
 4-7 
 
 5-8 
 
 6.2 
 
 6.2 
 
 9,000 
 
 4-3 
 
 5-8 
 
 6.1 
 
 6.3 
 
 10,000 
 
 4.1 
 
 5-7 
 
 6.0 
 
 6.2 
 
 12,000 
 
 
 
 5-6 
 
 6.0 
 
 " Douglas and Haldane, Journ. of Physiol., XXXIX, Proc. Physiol. Soc., p. i, 
 1909. 
 
 17 Briggs, Henry Fitness and breathing during exertion, /. Physiology, Vol. 53, 
 1919-1920, Proc. Physiological Soc., p. 38-40. 
 
RESPIRATION 
 
 157 
 
 paper. Subject A was out of training, and Subject B in good 
 training. 
 
 The reason why anoxaemia is absent in persons who are in 
 good training will be discussed in Chapter IX. 
 
 There can be little doubt, in view of all the evidence adduced 
 above, that muscular work produces some degree of anoxaemia 
 in untrained persons, and that the anoxaemia increases with the 
 work. The anoxaemia can hardly be due to any other cause than 
 
 LITERS GAS INSPIRED PER MINUTE 
 
 Subject A 
 Breathing Breathing 
 
 Subject B 
 Breathing Breathing 
 
 air 
 
 oxygen 
 
 atr 
 
 oxygen 
 
 12 
 
 25 
 
 13 
 22.5 
 
 14 
 20 
 
 II 
 
 18 
 
 40 
 
 33 
 
 27 
 
 27 
 
 54 
 57 
 
 43 
 46 
 
 37 
 40-5 
 50 
 
 37 
 40.5 
 48 
 
 that the blood is passing through the lungs so quickly that suffi- 
 cient oxygen to saturate the haemoglobin has not time to pass in 
 through the alveolar epithelium, just as occurs to a far greater 
 extent even during rest in a case of phosgene poisoning. 
 
 Another possible explanation might perhaps suggest itself, and 
 seems, indeed, to be suggested in Chapter XI of Mr. Barcroft's 
 book, "The Respiratory Functions of the Blood." This is that the 
 velocity of the chemical reaction, which occurs when haemoglobin 
 comes into contact with oxygen at a certain partial pressure of 
 oxygen, is so low that there is not time for the change to complete 
 itself in the lungs during muscular exertion. The rate at which 
 haemoglobin takes up oxygen, or oxyhaemoglobin gives it off, 
 in presence of a certain partial pressure of oxygen is so extremely 
 rapid that at present we have no means of measuring it. We can 
 form some conception of what must be the velocity if we consider 
 what is happening in the circulation of a small warm-blooded 
 animal, such as a mouse or bird. As was shown by Dr. Florence 
 Buchanan 18 the pulse rate of such an animal is, even during rest, 
 
 "Buchanan, Journ. of Physiol,, XXXVII, Proc. Physiol. Soc., p. bcxix, 1908; 
 and XXXVIII, Proc. Physiol. Soc., p. Ixii, 1909. 
 
158 RESPIRATION 
 
 about 700 to 800 a minute. A volume of blood equal to the whole 
 of that in the animal will pass round the circulation in one or two 
 seconds during exertion, so that any portion of blood will only be 
 present for an instant in the pulmonary capillaries in each round 
 of the circulation. Yet the time is sufficient for the chemical change 
 to occur in the blood, and doubtless far more than sufficient, since 
 we have to allow also for the time needed for the passage of oxy- 
 gen through the layer of living tissue separating the air from the 
 blood. In man the time available is much greater, so that the 
 absolute velocity of the chemical change does not come into con- 
 sideration at all, though of course the relative rates at which oxy- 
 gen is chemically associated with or dissociated from haemoglobin 
 at varying partial pressures of oxygen and varying temperatures, 
 determine the corresponding dissociation curves as experimentally 
 determined. 
 
 A further group of causes of anoxaemia depends not on defec- 
 tive saturation in the lungs, but on defect in the charge of available 
 oxygen carried by the arterial blood, so that, with the existing 
 rate of circulation, the oxygen pressure in the systemic capillaries 
 falls too low. Of this group, carbon monoxide anoxaemia will be 
 considered first. 
 
 The laws of combination of carbon monoxide with haemoglobin 
 have already been discussed in Chapter IV. My own interest in 
 carbon monoxide arose out of my connection with coal mining, as 
 it had become evident to me that carbon monoxide poisoning was a 
 common occurrence, and I wished to understand it as thoroughly 
 as possible. When Claude Bernard discovered the combination of 
 CO with haemoglobin he attributed death from CO poisoning to 
 the anoxaemia resulting from the fact that CO displaces the oxy- 
 gen of oxyhaemoglobin. CO was, however, very generally be- 
 lieved to have other physiological actions than those of anoxaemia, 
 and my first experiments were made with a view to clearing this 
 matter up. 
 
 To put the matter to the test, I devised the following experi- 
 ment 19 (Figure 51). A mouse was dropped into a thick glass 
 measuring vessel filled with pure oxygen, and the pressure of 
 oxygen in this cylinder was then raised to two atmospheres by 
 connecting it with an oxygen cylinder in the manner shown. The 
 oxygen was then clamped off and another clamp opened, through 
 which the oxygen was directed into the top of another measuring 
 vessel full of water, and the water driven over into a third measur- 
 
 "Haldane, Journ. of Physiol., XVII, p. 201, 1905. 
 
RESPIRATION 
 
 159 
 
 ing vessel filled with pure carbon monoxide, so arranged that the 
 gas was driven into the vessel containing the mouse. The animal 
 was now in a mixture consisting of two parts of oxygen and one 
 of carbon monoxide, at a total pressure of two atmospheres of 
 oxygen and one of carbon monoxide. It could also be killed by 
 drowning in this atmosphere if water was forced over. 
 
 My calculation was that in the presence of two atmospheres of 
 oxygen the animal would have in simple solution sufficient oxy- 
 gen in its arterial blood to supply the oxygen requirements of its 
 tissues, at any rate during rest; and that it would thus be inde- 
 pendent of the oxygen supply shut off through the action of the 
 
 Figure 51. 
 Apparatus for exposing mouse to atmosphere of oxygen and CO. 
 
 CO, with which the haemoglobin would be almost completely 
 saturated. If, however, the CO had any toxic action apart from 
 its action in producing anoxaemia this action would certainly 
 manifest itself at once, since the partial pressure of the CO was 
 100 per cent of an atmosphere, whereas in CO poisoning as ordi- 
 narily met with in non-fatal cases, the partial pressure of CO 
 
160 RESPIRATION 
 
 is not more than about 0.2 per cent of an atmosphere. The amount 
 of free oxygen which would go into solution in blood at the body 
 temperature with an atmospheric pressure of two atmospheres is 
 4.2 volumes per 100 cc. of blood, which is just about as much as is 
 ordinarily taken from the blood as it passes through the tissues 
 (see Chapter X). 
 
 The mouse remained quite normal and seemingly unconcerned, 
 except that when it exerted itself in climbing up the jar it seemed 
 to become more easily tired than usual. Thus CO has no appreci- 
 able physiological action except that of producing anoxaemia. 
 It is, physiologically speaking, an indifferent gas, like nitrogen, 
 hydrogen, or methane, and, like these gases, only acts physio- 
 logically by cutting off the supply of oxygen. Its only specific 
 physiological action, so far as I am aware, is that it has a slight 
 garlic-like odor. It is not an "odorless gas" except to those who 
 are afraid even to smell it on account of the mythical properties 
 commonly attributed to it. Animals which have no haemoglobin 
 pay no more attention to CO than to nitrogen. I kept a cockroach 
 for a fortnight in an atmosphere consisting of 80 per cent of CO 
 and 20 per cent of oxygen, and it remained perfectly well. CO is 
 not oxidized or otherwise decomposed in the living body of any 
 animal. 20 It passes in by the lungs and passes out far more 
 rapidly than is generally supposed by the lungs, without there 
 being the smallest loss. For this and other reasons it is a most 
 valuable physiological reagent. 
 
 The popular idea that CO remains for long in the blood is based 
 simply on failure to realize the nature of the symptoms which fol- 
 low severe or long-continued anoxaemia. In the light of present 
 knowledge it is childish to suppose that as soon as anoxaemia is 
 relieved a patient will recover, or that anoxaemia is in itself a 
 trifling matter if life is not immediately imperiled. If there were 
 only one clinical lesson derived from a perusal of this book, I hope 
 it would be that anoxaemia is a very serious condition, the con- 
 tinuance of which ought to be prevented if at all possible. 
 
 The properties of CO as a poison can now in the main be under- 
 stood in the light of preceding chapters. As the molecular affinity 
 of haemoglobin for CO is enormously more powerful than its 
 affinity for oxygen, it is evident that a very small proportion of 
 CO in the air is capable of saturating the blood to a noticeable 
 extent. The proportion of oxygen in dry alveolar air is about 14 
 
 20 For experiments and references on this subject see Haldane, Journ. of PAysiol., 
 XXV, p. 225, 1899, and M. Krogh, Pfliiger's Archiv, 162, p. 94, 1915. 
 
RESPIRATION 161 
 
 per cent, and the affinity of haemoglobin for CO (in my own case 
 at least) is about 300 times its affinity for oxygen. It follows that, 
 if we assume for the moment that the oxygen pressure of the blood 
 is that of the normal alveolar air, the blood will gradually become 
 
 half-saturated with CO if air containing = .047 per cent of 
 
 300 
 
 CO is breathed continuously for a sufficient time. If the per- 
 centage is .0235 per cent, the final saturation will only be one 
 third ; and if the percentage is .012 the saturation will be a fourth ; 
 and so on. If pure air were again breathed the CO would be ex- 
 pelled from the body through the unbalanced action of the al- 
 veolar oxygen pressure in expelling CO from its combination. The 
 rates of absorption and of elimination of the CO can also be 
 calculated on the same principles from the mean percentage of 
 CO in the alveolar air, allowing for the fact that as the haemo- 
 globin approaches the balancing saturation the rate of absorption 
 will gradually fall off; and similarly the rate of elimination will 
 gradually fall off as the blood loses CO. As will be shown in 
 Chapter IX, however, this theoretical course of events is pro- 
 foundly modified by active secretion of oxygen inwards by the 
 lung epithelium. 
 
 It is evident also that in air abnormally poor in oxygen a given 
 percentage of CO will become more poisonous, and in air ab- 
 normally rich in oxygen less poisonous. This I verified experi- 
 mentally on animals. It remained to ascertain in man what effects 
 corresponded to a given saturation of the haemoglobin ; and this I 
 ascertained by experiments on myself, 21 using for the purpose the 
 carmine titration method referred to in Chapter IV, and fully 
 described in its latest form in the Appendix. 
 
 I found in these experiments that no particular effect was ob- 
 served until the haemoglobin was about 20 per cent saturated. At 
 about this saturation an extra exertion, such as running upstairs, 
 produced a very slight feeling of dizziness and some extra palpita- 
 tion and hyperpnoea. At about 30 per cent saturation very slight 
 symptoms, such as slight increase of pulse rate, deeper breathing, 
 and slight palpitations, became observable during rest, and run- 
 ning upstairs was followed in about half a minute by dizziness, 
 dimness of vision, and abnormally increased breathing and pulse 
 rate. At 40 per cent saturation these symptoms were more marked, 
 and exertions had to be made with caution for fear of fainting. 
 At 50 per cent saturation there was no real discomfort during 
 
 21 Haldane, Journ. of PhyswL, XVIII, p. 430, 1895. 
 
1 62 RESPIRATION 
 
 rest, but the breathing and pulse rate were distinctly increased, 
 vision and hearing impaired, and intelligence probably greatly 
 impaired. It was also hardly possible to rise from the chair with- 
 out assistance. Writing was very bad, and spelling uncertain. 
 Movements were very uncertain, and it was difficult to recognize 
 objects distinctly or estimate their distance correctly, so that things 
 a long way off were grasped at in vain. Attempts to go any distance 
 caused failure of the legs and collapse on the floor. In one experi- 
 ment the saturation reached 56 per cent. It was then hardly 
 possible to stand, and impossible to walk. After each of these 
 experiments the saturation of the blood fell rapidly when fresh 
 air was breathed ; and within three hours the saturation had fallen 
 below 20 per cent. 
 
 Shortly after these experiments, I examined the bodies of a 
 large number of men who had been killed in colliery explosions, 
 and found that nearly all had died of CO poisoning. The satura- 
 tion of the haemoglobin with CO was usually about 80 per cent, 
 but in some cases not more than 60 per cent. In fatal cases of 
 poisoning by lighting gas Lorrain Smith found similar satura- 
 tions. 
 
 The general similarity between the symptoms of CO poisoning 
 and those of anoxaemia produced in other ways is evident; and 
 the after-symptoms appear to be identical with those of mountain 
 sickness and related disorders. There is, however, a difference 
 between the symptoms of CO poisoning and those of anoxaemia 
 produced by imperfect oxygenation of the arterial haemoglobin. 
 This difference lies in the fact that in CO poisoning fainting, or a 
 tendency to fainting, is much more prominent than respiratory 
 distress. A man at a high altitude pants excessively on exertion, 
 but does not easily faint. A man suffering from CO poisoning 
 faints very readily on exertion, and the tendency to dizziness and 
 collapse is far more prominent than the hyperpnoea. The fainting 
 on exertion is evidently due to the fact that from lack of the mass 
 of oxygen needed the heart cannot compensate by sufficiently 
 increased output of blood for the greatly increased flow of blood 
 through the working muscles. The blood pressure therefore falls, 
 with the result that the circulation to the brain is diminished and 
 anoxaemia then causes loss of consciousness. But why does this 
 occur so much more readily in CO poisoning? The fact that it does 
 so indicates that relatively speaking the respiratory center is less 
 affected in the anoxaemia of CO poisoning, in which the mass of 
 oxygen in the blood is reduced but the pressure of oxygen in the 
 
RESPIRATION 163 
 
 arterial blood remains normal. That is to say, with a degree of 
 anoxaemia which would not seriously affect the heart in anoxae- 
 mia from imperfect oxygenation of the available haemoglobin 
 there will be marked response to anoxaemia in the respiratory 
 center, but not in CO poisoning. This points clearly to the very 
 important conclusion that it is practically speaking to the oxygen 
 pressure of the arterial blood that the respiratory center responds. 
 The blood which bathes the receptive end-organs (or whatever 
 else is sensitive to the respiratory chemical stimuli) of the respira- 
 tory center must therefore be blood which has lost very little of its 
 arterial charge of oxygen. 
 
 There are other facts pointing in the same direction. Thus in 
 fainting or dizziness from fall of blood pressure there is no im- 
 mediate panting, although the anoxaemia which immediately 
 results in the cerebrum is sufficient to cause loss or impairment of 
 consciousness. The arterial blood, however, remains normal as 
 regards its pressures of oxygen and CO 2 during fainting; and in 
 accordance with the conclusion just reached, the breathing is not 
 stimulated till the stagnation of blood in the respiratory center is 
 very marked. 
 
 It is to be kept in mind that at a moderate altitude the pressure 
 of oxygen in the arterial blood is diminished far more than the 
 mass of the oxygen, as expressed by the percentage saturation of 
 the haemoglobin. With CO it is the mass of oxygen which is 
 diminished in the blood, while the pressure may be normal. 
 
 It also seems a priori probable that the respiratory center should 
 be continuously sampling and controlling the gas pressures of the 
 arterial blood. For it has to act for the whole body. Its function 
 is evidently, not to keep normal the gas pressures in the capillaries 
 of one particular part of the body, such as the medulla oblongata, 
 but to keep normal the arterial blood upon which every part of 
 the body draws in accordance with varying local requirements. 
 It keeps the gas pressures normal just as the heart keeps the blood 
 pressure normal, so that every part of the body can always indent 
 for arterial blood of standard quality and sufficient quantity. 
 
 A further peculiarity of CO poisoning is that quite commonly 
 consciousness is lost for long periods in the poisonous atmosphere 
 without death occurring. Thus cases of CO poisoning afford strik- 
 ing opportunities of studying the effects of prolonged general 
 anoxaemia of the brain and every other organ in the body. The 
 reason why death does not occur more readily seems to be that, 
 although the amount of oxygen transported by the blood is dimin- 
 
164 
 
 RESPIRATION 
 
 ished, the oxygen pressure in the arterial blood remains normal, 
 and as a consequence the respiratory center does not rapidly fail 
 in the same manner as it does when the arterial oxygen pressure is 
 very low, as explained in Chapter VI. This characteristic seems to 
 be common to all forms of anoxaemia in which the arterial oxygen 
 pressure remains about normal, including anoxaemia due simply 
 to a failing heart. 
 
 If the action of CO were simply to diminish the oxygen- 
 carrying power of the haemoglobin, without modification of the 
 properties of the remaining haemoglobin, the symptoms of CO 
 poisoning would be very difficult to understand in the light of 
 other knowledge. Thus a person whose blood is half-saturated 
 
 20 
 
 Pressure of 2 in Mm. of \\g. 
 
 30 40 50 6.0 70 
 
 80 
 
 90 
 
 g 
 
 
 
 &4%526272& 91 
 
 Pressure of 0, in % of an atmosphere 
 
 Figure 52. 
 
 Curve I, o per cent saturation with CO; II, 10 per cent; III, 25 per cent; 
 IV, 50 per cent; V, 75 per cent. 
 
 with CO is practically helpless, as we have just seen ; but a person 
 whose haemoglobin percentage is simply diminished to half by 
 anaemia may be going about his work as usual. Miners may be 
 doing their ordinary work though their haemoglobin percentage 
 is reduced to half or less by ankylostomiasis ; and women may be 
 going about their duties with their haemoglobin percentage re- 
 
RESPIRATION 165 
 
 duced to a third by chlorosis. Even in the extremest "anaemia," 
 with the haemoglobin below 20 per cent of its normal value, and 
 the lips of extremest pallor, the patient is perfectly conscious, 
 though hardly capable of any muscular exertion. 
 
 The key to this seeming paradox is furnished by the discovery 22 
 that the oxyhaemoglobin left in the arterial blood when it is 
 partially saturated with CO has its dissociation curve altered in 
 such a way that the haemoglobin holds on more tightly to the oxy- 
 gen. The oxygen still present as oxyhaemoglobin is therefore less 
 easily available, so that the oxygen pressure in the tissues must 
 fall lower in order to get off the combined oxygen. With a given 
 amount of available oxygen in the blood the physiological anox- 
 aemia is thus increased. Figure 52, from a paper by J. B. S. Hal- 
 dane, 23 shows the alterations in the dissociation curves of the 
 oxyhaemoglobin with varying percentage saturations of the blood 
 with CO. It will be seen, for instance, that with 50 per cent satu- 
 ration of the blood with CO the oxygen pressure must fall to less 
 than half the usual value, and with 75 per cent saturation to less 
 than a third, in order to dissociate half the oxygen present in the 
 arterial blood as oxyhaemoglobin. No wonder, therefore, that the 
 symptoms of CO poisoning are much more severe than those of a 
 corresponding simple deficiency of haemoglobin in the blood. It 
 will be seen also that the shape of the dissociation curve is com- 
 pletely altered. The characteristic double bend (which, as already 
 seen, is of such vital physiological importance) in the oxyhaemo- 
 globin curve tends to disappear altogether, so that an enormous 
 fall in oxygen pressure is needed to make the bulk of the oxygen 
 in the oxyhaemoglobin dissociate. 
 
 In the investigations which Lorrain Smith and I made on the 
 effects of continuously breathing a definite percentage of CO all 
 the experiments were made on ourselves, and in a series which 
 was more or less continuous from day to day. From the results of 
 these experiments we estimated that it required about .05 per cent 
 of CO in the air to produce the 30 per cent saturation of the blood 
 which was necessary for any very noticeable symptoms of CO 
 poisoning. In isolated experiments made later, however, we found 
 the CO much more poisonous, so that it only required about .02 
 per cent to produce the required saturation. In the original ex- 
 periments we had become "acclimatized" without knowing it. The 
 
 22 Douglas, J. S. Haldane, and J. B. S. Haldane, Journ. of Physwl., XLIV, 
 p. 293, 1912. 
 
 83 J. B. S. Haldane, Journ. of Physiol., XLV, Proc. Physiol. Soc., p. xxii, 1912. 
 
1 66 RESPIRATION 
 
 great significance of this "acclimatization" will be discussed in 
 succeeding chapters. 
 
 The other gas, besides CO, which enters into molecular com- 
 bination with haemoglobin is nitric oxide. But as free nitric oxide 
 combines at once with the oxygen in air to form yellow "nitrous 
 fumes," and these are intensely irritant and produce very danger- 
 ous inflammation, nitric oxide poisoning in the same sense as CO 
 poisoning is impossible. Sir Humphrey Davy nearly killed him- 
 self when he attempted to breathe nitric oxide (NO) at the time 
 when he discovered the effects of nitrous oxide, or "laughing gas" 
 (N 2 O). NO haemoglobin is, however, formed to some extent in 
 the living body during poisoning by nitrites, as was discovered by 
 Makgill, Mavrogordato, and myself ; 24 and some time after death 
 from nitrite poisoning the whole of the haemoglobin becomes 
 combined with NO. Hence the body is red, just as in a fatal case of 
 CO poisoning, so that the case might easily be mistaken for CO 
 poisoning on mere spectroscopic examination of the blood. The 
 condition can be distinguished at once by the fact that the blood 
 and tissues remain red on boiling, just as in the case already al- 
 luded to of salted meat. 
 
 Another cause of an anoxaemia analogous to that of CO poison- 
 ing is present in the case of the action of poisons which produce 
 methaemoglobin in the living body. The first of these to be dis- 
 covered was chlorate of potash, which in former times, before the 
 dangerous properties of chlorates were realized, used to be ad- 
 ministered freely as an oxidizing agent, and has even been recom- 
 mended as an antidote for the anoxaemia of high altitudes. The 
 discovery that in a fatal case of diphtheria treated with chlorate 
 of potash the blood contained much methaemoglobin drew atten- 
 tion to the possible dangers from anoxaemia in poisoning by any 
 of the numerous substances which are capable of producing me- 
 thaemoglobin in the living body. 
 
 The possibilities of anoxaemia being produced were investi- 
 gated by Makgill, Mavrogordato, and myself. As ferricyanide 
 does not penetrate the walls of the red corpuscles, and chlorates 
 do not do so in the animals we were using, we used chiefly nitrites 
 for the experiments; and we did so for the reason, partly, that 
 nitrites have other important physiological actions besides that of 
 producing methaemoglobin (in reality a mixture of methaemo- 
 globin with a certain proportion of NO haemoglobin). Having 
 discovered the dose required to produce death we then, as soon 
 
 34 Makgill, Mavrogordato, and Haldane, Journ. of Physiol., XXI, p. 160, 1897. 
 
RESPIRATION 
 
 167 
 
 as serious symptoms began to develop after administration of the 
 dose, placed the animals in compressed oxygen. The result was 
 that the serious symptoms disappeared and the animals recovered. 
 If, however, they were removed into ordinary air, they died at 
 once with anoxaemic convulsions. When kept in the oxygen for a 
 sufficient time, however, they completely recovered and could be 
 returned to ordinary air. Oxygen at ordinary atmospheric pres- 
 sure was often sufficient to save the animals. 
 
 Having worked out a method for estimating colorimetrically 
 the proportional extent to which the haemoglobin was altered by 
 the poison, we then found that the dangerous symptoms depended, 
 just as in CO poisoning, on the extent of the alteration. It was 
 thus evident that the cause of death, and of the dangerous symp- 
 toms, was anoxaemia, just as in CO poisoning. We also found that 
 the methaemoglobin and NO haemoglobin soon disappeared, leav- 
 ing the blood quite normal, if death was averted. The methaemo- 
 
 IOO 
 
 70 
 
 ^ 
 
 60 
 
 50 
 
 I 
 
 ^ 40 
 1 
 
 20 
 10 
 
 Mours offer srr/'ec{t'o/i 
 
 Figure 53. 
 Methaemoglobin due to sodium nitrate. 
 
 globin was simply reduced back again, just as on the addition of 
 a reducing agent to a methaemoglobin solution outside the body. 
 It was also evident that the reduction process was constantly going 
 on and tending to neutralize the poison even while the relatively 
 large amounts of it were still present in the blood. In proportion 
 
1 68 RESPIRATION 
 
 as the poison was destroyed or excreted the reduction process got 
 the upper hand. There are, therefore, reducing agents of some 
 kind or another within the corpuscles. Figure 53 shows the per- 
 centage conversion to methaemoglobin in the blood of a rabbit at 
 intervals after a non-poisonous dose of sodium nitrite. It will be 
 seen that after four hours the blood had completely recovered. 
 
 The action of methaemoglobin-forming poisons is rendered 
 evident at once by the marked cyanosis which they produce. The 
 methaemoglobin has a dark color, and the arterial blood becomes 
 therefore of a chocolate or coffee color. This form of cyanosis may 
 become very marked indeed without serious real symptoms of an- 
 oxaemia being present. Thus in acute poisoning by dinitrobenzol 
 (an ingredient of certain explosives) a man may become very blue 
 in the face and yet be going about as usual, although he presents 
 a most alarming appearance. 
 
 Many of the poisons which produce methaemoglobin cause, in 
 addition, radical decomposition in the haemoglobin, and even 
 breaking up of the red corpuscles. This is, for instance, the case, 
 to a large extent, with dinitrobenzol, so that there are other colored 
 decomposition products present as well as methaemoglobin; and 
 for the present it is not possible to specify their nature. Their pres- 
 ence, or that of methaemoglobin, can, however, be detected at once 
 on diluting a drop of the blood till the color begins to become 
 yellowish, then saturating with coal gas or CO, and comparing 
 the tint with that of normal blood diluted to a corresponding ex- 
 tent and similarly saturated. If any colored decomposition prod- 
 ucts are present the normal blood solution will be pinker, as the 
 CO does not combine to give a pink color with these foreign 
 substances. 
 
 When a poison causes solution of the red corpuscles (haemo- 
 lysis), or decomposes the haemoglobin beyond the methaemo- 
 globin stage, the haemoglobin is lost to the body, and "anaemia" 
 is one result of this, as well as jaundice. Thus chronic poisoning 
 by dinitrobenzol and similarly acting substances causes very seri- 
 ous anaemia. This also results from chronic poisoning by arsenu- 
 retted hydrogen, which has the peculiar action of injuring the 
 walls of the red corpuscles and so causing haemolysis, with re- 
 sulting haemoglobinuria, jaundice, and often nephritis. We are 
 thus brought to the consideration .of the anoxaemia caused by 
 anaemia, the word "anaemia" being taken to mean simply a 
 diminution in the percentage of haemoglobin in a given volume 
 of blood, whether the blood volume itself is diminished, or normal, 
 
RESPIRATION 169 
 
 or increased. As a matter of fact the blood volume is usually much 
 increased in "anaemia," as was first shown by Lorrain Smith. 25 
 
 It was found by Miss FitzGerald that in ordinary cases of 
 anaemia there is no appreciable diminution in the alveolar CO 2 
 pressure. 26 As will be shown more fully in Chapter VIII, a chronic 
 arterial anoxaemia, however slight, invariably lowers the alveolar 
 CO 2 pressure if time is given, and if the anoxaemia continues 
 during rest. The absence of a lowered alveolar CO 2 pressure in 
 cases of anaemia is thus clear evidence of the absence of anox- 
 aemia, in spite of greatly diminished oxygen-carrying capacity of 
 the blood. It is evident, therefore, that the circulation rate is much 
 increased in anaemia and this inference is confirmed by the ab- 
 sence of cyanosis. A little consideration will show that this in- 
 creased circulation rate, while it serves to maintain the normal 
 oxygen pressure of the blood in the systemic capillaries, will prob- 
 ably not reduce too much the pressure of CO 2 in the tissues. The 
 CO 2 conveying power of the blood in the living body depends, as 
 shown in Chapter V, on the concentration of haemoglobin present 
 in the blood, and this concentration is greatly reduced in anaemia. 
 Diminution in the actual CO 2 -conveying power of the blood in 
 the living body will therefore advance pari passu with the diminu- 
 tion of the oxygen-carrying power. Thus (as shown in Chapter 
 X) an increased circulation rate is brought about by the combined 
 stimulus of diminished oxygen pressure and increased CO 2 pres- 
 sure. This is not so in the case of anoxaemia from defective satura- 
 tion of the haemoglobin in the lungs ; nor, for the special reason 
 given above, in the anoxaemia of CO poisoning. The reason why 
 imperfect saturation of the arterial blood causes such serious 
 anoxaemia in the cerebrum and tissues elsewhere, while anaemia 
 causes so little anoxaemia (during rest) unless it is very extreme, 
 is probably bound up with this difference as regards effects on 
 CO 2 pressure in the tissues. The matter will, however, be discussed 
 more fully in Chapter X. 
 
 The last cause of anoxaemia to be considered is that due prima- 
 rily to defective circulation ; and it will be referred to very briefly 
 here, as the relation of circulation to respiration will be discussed 
 in Chapter X. When the blood pressure is very defective owing to 
 failure of heart action or failing supply of venous blood to the 
 heart, the inevitable result is failure in the general circulation rate, 
 and failure also in the proper distribution of blood within the body. 
 
 25 Lorrain Smith, Trans. Path. Soc. Lond., LII, p. 315. 
 86 Journ. of Pathol. and Bacter., XIV, p. 328, 1910. 
 
I ;o RESPIRATION 
 
 This must result in anoxaemia in the tissues, together with an 
 undue rise in their CO 2 pressure. But owing to the combination of 
 these two conditions the fall in oxygen pressure and rise in CO 2 
 pressure will both be moderate until the slowing of circulation is 
 excessive : for the oxygen will fall along the steep part of the 
 dotted curve in Figure 21, while the CO 2 pressure will rise along 
 the thick line in Figure 26. This means that a great diminution in 
 the charge of oxygen in the haemoglobin, and consequently a very 
 considerable cyanosis, will be possible with a comparative small 
 fall in the oxygen pressure or rise in the CO 2 pressure. Hence 
 cyanosis due to slowing of the circulation is not in itself such a 
 serious indication as cyanosis due to failing saturation of the blood 
 with oxygen, although of course indicative of possible more 
 serious failure of the circulation. 
 
 When fall of arterial blood pressure is. due to defective filling 
 of the large veins leading to the heart, benefit may be expected 
 from the intravenous injection of suitable saline solution, as this 
 will tend to fill up the veins, and to bring about adequate filling 
 of the heart. A simple salt solution tends, however, to leak out 
 again very quickly from the circulation. To remedy this defect 
 Bayliss 27 has introduced the plan of adding gum to the salt solu- 
 tion, the gum fulfilling the same function in preventing leakage 
 as the proteins normally present in blood plasma. This procedure 
 has proved very successful, and avoids the risks and practical 
 difficulties associated with transfusion of blood or liquids con- 
 taining proteins. For the reasons already pointed out, the dilution 
 of the blood by the saline injection does not cause anoxaemia. 
 
 As will be pointed out in Chapter X, failure in the venous return 
 to the heart may be due to deficient pressure of CO 2 in the systemic 
 capillaries, owing to excessive washing out of CO 2 in the lungs ; 
 and this excessive washing out may be secondary to arterial anox- 
 aemia. Arterial anoxaemia and deficiency of CO 2 may also be the 
 cause of failure of the heart muscle. It is probable, therefore, that 
 in many cases the vicious circle may be more effectively broken by 
 administration of oxygen or even CO 2 than by injection of gum- 
 saline solution or transfusion of blood; but in other cases injection 
 or transfusion would quite clearly be required. 
 
 "Bayliss, Intravenous Injection in Wound, Shock, 1918. 
 
CHAPTER VIII 
 Blood Reaction and Breathing. 
 
 IT has been known for long that the reaction of blood to litmus 
 paper is always slightly alkaline, while the living tissues are also 
 alkaline, though they change to acid in dying. Knowledge as to the 
 connection between the blood reaction and normal breathing is, 
 however, mostly of very recent origin ; and the same may be said 
 of knowledge as to the extreme exactitude with which the reaction 
 of the blood is regulated, and the physiological importance of the 
 very slightest deviation, from the normal reaction of the blood 
 and tissues. 
 
 That the reaction within the body is physiologically regulated 
 was originally indicated, not only by the reaction of the blood to 
 litmus and other indicators being always the same, but also by the 
 fact that on administration of sufficient doses of sodium bicarbon- 
 ate or other alkalies the urine, which is normally acid in man, 
 becomes alkaline. The same effect is produced by a vegetable diet, 
 which contains a large amount of organic acids combined with 
 alkali. The acids are mostly oxidized with formation of CO 2 within 
 the body, thus leaving alkaline carbonates, so that the excess of 
 alkali must be, and actually is, excreted in order that the reaction 
 within the body may remain normal. In herbivorous animals the 
 urine is always alkaline. On the other hand, in carnivorous ani- 
 mals, and in man with his usual mixed diet, the urine is acid. This 
 is because there is an excess of non-volatile acid formed within the 
 body by the oxidation of. the sulphur, phosphorus, etc., in the 
 food constituents and this excess is partly, at least, got rid of by 
 the kidneys, and the normal alkalinity of the blood and tissues 
 thus preserved. 
 
 More than forty years ago an important series of investigations 
 bearing on the physiology of the blood reaction was carried out 
 under Schmiedeberg's direction at Strassburg. The effect on rab- 
 bits of the administration of large doses of dilute hydrochloric acid 
 was investigated by Walter, 1 and it was found, as one result, that 
 the breathing of the animals was very greatly increased, becoming 
 extremely deep as well as more frequent the same sort of effect 
 as is produced by excess of CO 2 , as shown in Chapter II. The 
 
 X F. Walter, Archw /. exper. PatAol. Pharmakol., VII, p. 148, 1877. 
 
172 RESPIRATION 
 
 animals also ultimately became comatose, just as is the case when 
 CO 2 is in great excess ; and finally there were signs of exhaustion 
 of breathing, the breathing ceasing before the heart ceased to beat. 
 
 Another very important result reached in these investigations 
 was that when the experiments were repeated on dogs it was much 
 more difficult to produce the symptoms, and it was found that the 
 amount of ammonia excreted (in combination with acid) in the 
 urine was increased greatly. Under normal conditions the amount 
 of nitrogen excreted as ammonia is small in proportion to the total 
 excretion of nitrogen. Thus in man the amount of ammonia 
 usually excreted in 24 hours is only about 0.7 gram (sufficient, 
 however, to neutralize about 2 grams of H 2 SO 4 ), so that only a 
 small fraction of the total nitrogen is excreted as ammonia. In 
 acid poisoning, however, the fraction becomes a very much 
 larger one in carnivorous animals and in man. Walter found that 
 in dogs the ammonia excretion could be pushed up to several 
 times the normal by giving large doses of acid. 
 
 According to the existing evidence, which originated with 
 Schmiedeberg and his pupils, ammonia is converted into urea in 
 the liver. It appears, therefore, that when acid is administered to 
 carnivorous animals or men, ammonia is not converted into urea, 
 or else nitrogen which normally appears as urea is converted into 
 ammonia and goes to neutralize the acid. If ammonia is admin- 
 istered by mouth as carbonate it is wholly converted into urea, and 
 the excretion of ammonia by the urine may be actually diminished. 
 If, on the other hand, the ammonia is administered in combination 
 with a strong acid as a neutral salt, much of this ammonia appears 
 as salts of ammonia in the urine. Some is, however, converted into 
 urea in the liver, as was recently shown definitely by perfusion 
 experiments. 2 It was found that during health the proportion 
 of ammonia which escapes conversion into urea and consequently 
 appears in the urine depends on the acid-forming or alkali- 
 forming properties of the diet. Thus with a meat diet the pro- 
 portion of ammonia is much higher than with a vegetable diet; 
 and by administering alkalies ammonia may be made to disappear 
 entirely from the urine. 
 
 The varying neutralization of acids by ammonia is therefore 
 one of the means by which the reaction within the body is regu- 
 lated in man and carnivorous animals, while variation in the 
 excretion of acid or alkali in the urine is another. The former 
 means hardly exists in herbivorous animals. But the significance 
 
 'Loffler, Biochem. Zetischr., LXXXV, p. 230, 1918. 
 
RESPIRATION 173 
 
 of the most rapid and effective method of all varying excretion 
 of carbonic acid by the breathing remained hidden till quite 
 recently, although Walter's experiments showed that there is not 
 only a great increase in the breathing, but the amount of carbonic 
 acid present in the arterial blood is reduced in extreme case to 
 about a twelfth of the normal. 
 
 It was discovered by von Jaksch 3 in 1882 that where acetone 
 is present in the urine, as in bad cases of diabetes, verging on 
 coma, or actually comatose, considerable quantities of aceto- 
 acetic acid are also present; and soon afterwards Minkowski 4 
 found that oxybutyric acid, a closely allied substance, is likewise 
 present. The excretion of ammonia had already been shown to be 
 greatly increased, as well as the depth of the breathing and the 
 acidity of the urine, just as in acid poisoning; and indeed it was 
 this that led Minkowski, and Stadelmann before him, to the 
 search for organic acids. Thus all the symptoms point to acid 
 poisoning by the acids mentioned. Shortly after Priestley and I 
 introduced our method of investigating alveolar air, Pembrey, 
 Beddard, and Spriggs investigated the alveolar air in cases of 
 diabetic coma at Guy's Hospital, 5 and found the alveolar CO 2 
 percentage as low as i.i per cent. It went up and down as 
 the patient emerged from or relapsed into coma; and the ad- 
 ministration of sodium bicarbonate warded off the attacks of 
 coma, and at the same time kept the alveolar CO 2 percentage from 
 falling. Investigation of the alveolar CO 2 pressure is now a well- 
 recognized clinical method for estimating the gravity of symptoms 
 in diabetic coma and other states of "acidosis," as well as for 
 judging of the effects of treatment. 
 
 For a long time the degree of alkalinity of the blood was judged 
 from the amount of acid which has to be added to a given volume 
 of it or its serum before an indicator, such as litmus, gives the tint 
 indicative of neutrality. By this method it was found that the 
 blood in acid poisoning or diabetic coma is less alkaline than 
 usual; and all sorts of similar supposed "acidoses" have been dis- 
 covered, although the signs of physiological response to the pres- 
 ence in the body of too much acid might be more or less absent or 
 even contradictory. A few years ago, however, it became evident 
 that the amount of acid required for neutralization is no reliable 
 
 3 Von Jaksch, Bertchte der cLeutschen Chem. Gesellsch., p. 1496, 1882. 
 
 4 Minkowski, Arch. f. exper. Pathol. u. Pharmak., XVIII, pp. 35 and 147, 
 1884. 
 
 6 Beddard, Pembrey, and Spriggs, Journ. of Physiol., XXXI, Proc. Physiol. Soc.. 
 p. xliv, 1904; also XXXVII, p. xxxix, 1908. 
 
174 RESPIRATION 
 
 measure of the blood alkalinity. Even a strong solution of sodium 
 bicarbonate is but feebly alkaline ; but the amount of acid which 
 must be added to it to render it neutral is as great as if the sodium 
 were present as caustic soda, and is thus no measure of the actual 
 alkalinity of the solution. The carbonic acid united with the soda 
 prevents it from being at all strongly alkaline, but at the same time 
 does not completely neutralize it, and all weak acids have the same 
 properties. They may thus be said to be "buffer" substances, since 
 they prevent a strong acid from neutralizing at once a weakly 
 alkaline solution. A great deal of the strong acid has to be added 
 before the weak alkalinity is neutralized. The same applies to weak 
 alkalies, mutatis mutandis. 
 
 Now the blood and tissues are full of buffer substances. In the 
 first place, as already seen in Chapter V, carbonic acid is present 
 in combination. Haemoglobin and various other proteins are also 
 present ; and it has been well known for a long time that proteins 
 act as both acid and alkaline buffers, so that the neutral point in a 
 solution containing proteins is very difficult to ascertain sharply 
 by means of ordinary indicators. The color alters gradually in 
 either direction as the neutral point for any particular indicator 
 is approached. It was shown in Chapter V that in the alkaline 
 blood haemoglobin and other proteins act as weak acids more than 
 sufficient in amount to combine with the bases not already com- 
 bined with strong acids, and that the presence of these proteins 
 along with carbonic acid determines the manner in which the 
 alkali in blood takes up and gives off CO 2 with varying partial 
 pressures of this gas. The amount of acid required to produce 
 neutrality is thus in itself no measure of the degree of alkalinity 
 in blood, but depends on the amount of the various buffer sub- 
 stances, including carbonic acid in combination with alkali ; and 
 they may vary considerably in amount under different conditions. 
 This has been pointed out very clearly by L. J. Henderson. 6 
 
 It may be desirable at this point to remind the reader as to the 
 conception of acidity and alkalinity to which chemical and physi- 
 co-chemical investigation has led during the last thirty years. The 
 phenomena of electrolysis revealed to Faraday the fact that the 
 constituents of any "electrolyte," such as copper sulphate, are 
 torn asunder during electrolysis into definite fragments, of which 
 one kind travels toward the anode, and the other to the cathode. 
 These fragments he called "ions," because it is their movement 
 towards either anode or cathode, and the fact that each of them has 
 
 " L. J. Henderson, Ergebn. der Physiol., VIII, p. 254, 1909. 
 
RESPIRATION 175 
 
 a definite electrical charge, that determines the phenomena of 
 electrolysis and the exact quantitative relationship between-the 
 current passed through a cell containing an electrolyte in solution 
 and the splitting up of the electrolyte into its constituents. Van't 
 Hoff and Arrhenius brought Faraday's conception into relation 
 with osmotic pressure and various other phenomena connected 
 with solutions. 
 
 Osmotic pressure was first measured accurately by the botanist 
 Pfeffer. 7 He used a semi-permeable membrane (i.e., a membrane 
 which allowed the solvent water, but not the dissolved substance, 
 to pass) which had been originally discovered by Moritz Traube 
 in i867, 8 though Traube had not seen how to apply this membrane 
 for measuring osmotic pressures. Some years later van't Hoff 9 
 made the brilliant discovery that in dilute solutions of sugar and 
 other substances, the osmotic pressure is practically the same as 
 the pressure which the solute would have if its molecules were 
 present alone in the gaseous form at the same temperature. There 
 must thus be a fundamental connection between molecular con- 
 centration, osmotic pressure, and gas pressure ; also between mo- 
 lecular concentration and the vapor pressures, boiling points and 
 freezing points of solutions, as had already been empirically shown 
 by the investigations in particular of Raoult. Van't Hoff believed 
 that osmotic pressure, etc., were due in some way to the molecular 
 bombardment of the solute molecules, and therefore vary as their 
 concentration per liter of solution ; and this theory has served for 
 the building up of the theory of solutions as it is still represented 
 in current textbooks of physical chemistry. In reality this theo- 
 retical interpretation was not even justified by Pfeffer's data if 
 concentration per liter is considered, and breaks down entirely for 
 concentrated solutions. The theory is also quite unintelligible 
 mechanically, since the bombardment pressure of the solute mole- 
 cules would be in the wrong direction for explaining the phe- 
 nomena. Hence many persons regarded van't Hoff's theory with 
 the greatest suspicion ; but the fact that it seemed to answer ad- 
 mirably as a means of prediction in the case of dilute solutions, 
 and to cover an enormous mass of facts, has led to its very general 
 acceptance, though other attempts have been made to substitute 
 for it some more intelligible conception. 
 
 In iQiS 10 I showed quite clearly, as I think, that van't Hoff's 
 
 T Pfeffer, Osmottsche Untersuchungen, 1877. 
 
 8 Traube, Archiv f. (Anat. .) Physiol., p. 87, 1867. 
 
 9 Van't Hoff, Zeitschr. f. physik. Chemie, I, p. 481, 1887. 
 "Haldane, Bio-Chemical Journal, XII, p. 464, 1918. 
 
1 76 RESPIRATION 
 
 conception of osmotic pressure was mistaken. It is neither the con- 
 centration per liter of the solute molecules, nor that of the solvent 
 molecules, that determines osmosis, but the diffusion pressure 
 of the solvent. Water passes through a semi-permeable mem- 
 brane into a solution, because the diffusion pressure of pure water 
 is greater than that of the diluted water in the solution. The 
 osmotic pressure is not the excess of diffusion pressure of water 
 outside the solution, but the external mechanical pressure required 
 to equalize the two diffusion pressures, although in sufficiently 
 dilute solutions this mechanical pressure is practically the same 
 as the excess of diffusion pressure of water. 
 
 In a solution, just as in a gas mixture, the molecules are free 
 to move about; and, just as in a gas mixture, the mean free 
 space round each molecule is the same because the mean energy 
 of external movement is the same for each molecule. Hence the 
 free space in which water molecules are free to diffuse is in pro- 
 portion to the total number per liter of molecules present. This 
 space is of course greater per molecule of solvent in a solution 
 than in the pure solvent. Hence the pure solvent diffuses into the 
 solution unless the external pressure on the solution is raised 
 sufficiently to equalize the two diffusion pressures. 
 
 When osmotic pressure, vapor pressures, boiling points, etc., 
 are calculated in terms of this theory instead of van't Hoff's theory, 
 the experimentally ascertained values agree with the theory, 
 whereas this is not the case, as is now well known, with van't 
 Hoff's theory, except in the case of very dilute solutions. Thus 
 for solutions of cane sugar, and allowing for the fact that at 
 temperatures near oC. cane sugar is present in solution as a penta- 
 hydrate, the osmotic pressures at oC. calculated from the con- 
 centrations on the new theory and the pressures actually observed 
 by the Earl of Berkeley and Mr. Hartley at Oxford are as follows : 
 
 OSMOTIC PRESSURE IN ATMOSPHERES 
 
 Grams Cane Sugar 
 
 Observed 
 
 Calculated 
 
 Calculated on 
 
 per 100 cc. 
 
 
 
 van't Hoff's theory 
 
 3-32 
 
 2.23 
 
 2.24 
 
 2.17 
 
 9-59 
 
 6.85 
 
 6.85 
 
 6.29 
 
 18.26 
 
 14.21 
 
 14.17 
 
 11-95 
 
 25.81 
 
 21.87 
 
 21.80 
 
 16.90 
 
 28.13 
 
 24-55 
 
 24.44 
 
 18.41 
 
 54-24 
 
 67.74 
 
 67.66 
 
 35-48 
 
RESPIRATION 177 
 
 The vapor pressures, boiling points, and freezing points of 
 sugar solutions show a similar agreement between observations 
 and the new theory, as pointed out in detail in my paper. 
 
 To physiologists the main advantage of the new theory is that, 
 as will be pointed out in detail in later chapters, it enables us to 
 utilize the kinetic theory of matter in unifying our conceptions 
 of a great number of physiological phenomena. 
 
 The osmotic pressures observed by Pfeffer and others for dilute 
 salt solutions were far greater than corresponded to van't Hoff's 
 theory. This became quite intelligible when Arrhenius pointed 
 out in I88; 11 that the discrepancy could be cleared up on the as- 
 sumption that solutions of electrolytes are ionized to a greater or 
 less extent. Their osmotic pressures are not merely due to the 
 concentration (or, in terms of the new theory just referred to, the 
 diffusion pressure) of complete molecules of the solute, but also 
 to the concentrations of the ions present, as indicated by the vary- 
 ing electrical conductivities of different strengths of the solutions. 
 This explanation of Arrhenius was received at first with some 
 incredulity, but is now universally accepted, as the evidence in 
 favor of it is overwhelming. A dilute solution of sodium chloride, 
 for instance, is not now regarded as a solution of NaCl molecules, 
 but, practically speaking, of sodium and chlorine ions. Similarly 
 a dilute solution of hydrochloric acid is a solution of hydrogen and 
 chlorine ions. 
 
 lonization may be regarded as a tearing apart of the molecules 
 of the electrolyte in solution on account of the molecular affinity 
 of H 2 O molecules for the atoms of the electrolyte molecules ; and 
 in accordance with this conception the ions are not stray atoms or 
 other fragments of molecules, but molecular compounds with 
 molecules of water. In pure water itself the molecules are also to 
 a certain extent ionized, as indicated by, among other things, the 
 conductivity of pure water. The products of this ionization are 
 hydrogen and hydroxyl (HO) ions, combined with molecules of 
 water. 
 
 The acidity of a solution is due to preponderance of hydrogen 
 ions, and the alkalinity to preponderance of hydroxyl ions; and 
 when the concentrations of hydrogen and hydroxyl ions are equal 
 the solution is neutral. As, however, hydrogen and hydroxyl ions 
 are constantly reacting with one another according to the equation 
 
 H + HO * H 2 0, 
 the product of the concentrations of hydrogen and hydroxyl ions 
 
 11 Arrhenius, Zeitschr. f. -phystk. Chemie, I, p. 631, 1887. 
 
178 RESPIRATION 
 
 remains the same, in accordance with the law of mass action, how- 
 ever acid or alkaline a solution may be. Hence the concentration 
 of hydroxyl ions diminishes in proportion as that of hydrogen 
 ions increases, and vice versa. 
 
 All acids and' bases combine with one another in chemically 
 equivalent proportions, but different acids and alkalies vary very 
 greatly in the extent to which they are ionized. The "strengths" 
 of different acids and alkalies were found by the electrical con- 
 ductivity method to depend upon the extent of their ionization. 
 The "strong" acid HC1 is, for instance, very completely ionized 
 into hydrogen and chlorine ions, and the "strong" base NaHO is 
 similarly ionized into sodium and hydroxyl ions ; while "weak" 
 acids, such as carbonic acid, or weak bases, such as ammonia, are 
 very slightly ionized. 
 
 Water itself is slightly ionized into hydrogen and hydroxyl 
 ions, and can thus act as either a very weak acid towards bases or 
 a weak base towards acids. In the case of strong or highly ionized 
 acids and bases this property of water is practically of no account, 
 as the ionization of water is so very small ; but in the case of weak 
 acids or bases the water competes appreciably with the acid or 
 base. For instance in the case of potassium cyanide, a compound of 
 an extremely weak acid with a very strong base, the following re- 
 action occurs : 
 
 KCN + H 2 * KOH + HCN. 
 
 Thus free KOH and free HCN are both present in a solution of 
 this salt. But the KOH is highly ionized into K and HO ions, 
 while the HCN is hardly ionized at all. Hence HO ions pre- 
 dominate, and the solution is alkaline. Carbonic acid is not such a 
 weak acid as hydrocyanic acid; but the same relations hold, so 
 that both carbonates and bicarbonates form solutions which are 
 distinctly alkaline ; and bicarbonate solutions are still slightly 
 alkaline, even though much free carbonic acid is present, as in the 
 case of blood in the living body. 
 
 The ordinary indicators appear to be extremely weak acids 
 or bases which change color on combination. When the only 
 other acids or bases present are strong ones, the change of color 
 is of course very sharp ; but with other weak acids or bases present, 
 the change is gradual and the complete color change does not 
 occur until the solution is distinctly alkaline or acid. This is be- 
 cause the indicator competes with other weak acids for the base ; 
 and different indicators compete in varying degrees. Thus dif- 
 ferent indicators turn with different degrees of slight variation 
 
RESPIRATION 179 
 
 from the true neutrality point where hydrogen and hydroxyl ions 
 are equal in concentration, as in pure water. 
 
 The relative diffusion pressures, or (to use the incorrect lan- 
 guage of the still generally accepted van't HofTs theory of osmotic 
 pressure, etc.) the relative concentrations of any particular sort 
 of ion, in different solutions, can be measured by the differences 
 of potential communicated to a suitable electrode dipped in the 
 solutions. Thus with a hydrogen electrode hydrogen ion con- 
 centration can be measured directly; and this method was ap- 
 plied, soon after its discovery, to the measurement of the hydrogen 
 ion concentration (and therefore indirectly also of the hydroxyl 
 ion concentration) of blood. The earlier attempts gave the result 
 that the blood was neutral in reaction, and remained neutral even 
 in acidosis. The physiological signs of acidosis were, however, 
 very clear, as already explained. The electrometric method in its 
 earlier form was thus far too rough for physiological work. 
 
 It was mentioned in Chapter I that the experiments of Geppert 
 and Zuntz on the hyperpnoea following- muscular contractions in 
 animals showed a great diminution in CO 2 and a slight excess of 
 oxygen in the arterial blood during the hyperpnoea. They there- 
 fore concluded that neither excess of CO 2 nor want of oxygen can 
 be the cause of the hyperpnoea ; and they sought for the cause in 
 some acid substance present in the blood, since acids were known 
 to stimulate the breathing. The search made for the acid substance 
 did not, however, lead to any definite result ; and the experiments 
 of Priestley and myself on man brought us back to CO 2 as the 
 stimulus to the increased breathing. The improbability of any 
 organic acid being the stimulus to the breathing seemed to us to 
 be in any case very great. No acid other than CO 2 is given off in 
 the expired air, and organic acids, etc., are not appreciably oxi- 
 dized in the blood itself. It did not therefore seem possible to un- 
 derstand how the air hunger of muscular exertion could be re- 
 lieved, as it undoubtedly is, by increased breathing. In any case 
 the diminished proportion of CO 2 in the arterial blood in these 
 experiments was entirely discounted by the fact that this dimin- 
 ished proportion continued to exist for at least an hour after the 
 hyperpnoea had passed off. We thought that in Geppert and 
 Zuntz's experiments owing to defective circulation in the artifi- 
 cially stimulated muscles of the animal some lactic acid had been 
 produced and thrown into the blood, thus greatly reducing its 
 power of combining with CO 2 . Thus, although the pressure of 
 CO 2 was perhaps actually higher in the arterial blood and caused 
 
1 8o RESPIRATION 
 
 hyperpnoea, the amount of CO 2 contained in the blood was much 
 less. We also thought that owing to the diminished CO 2 carrying 
 power of the blood there might be an increased rise of CO 2 pres- 
 sure in the tissues. This explanation was, however, somewhat 
 strained and unsatisfactory, as was pointed out in Chapter II. We 
 had correctly divined the main cause of the greatly diminished 
 proportion of CO 2 in the arterial blood in those experiments, but 
 not the whole cause. 
 
 In a series of experiments by Boycott and myself on the effects 
 of low atmospheric pressure in a steel chamber on the alveolar 
 CO 2 pressure 12 we found that on returning from low pressure 
 the alveolar CO 2 pressure, which had been lowered by the hyperp- 
 noea caused by the low atmospheric pressure, did not return at 
 once to normal, but remained low for some time. Ogier Ward, 
 who was working in conjunction with us, found the same thing 
 and in much more marked and persistent degree, on returning to 
 ordinary pressure after a stay on Monte Rosa. 13 Galleotti, 14 and 
 also Aggazotti, 15 had already found that the titration alkalinity of 
 the blood is diminished by exposure to low pressure in a steel 
 chamber or at high altitudes. It was also known from older experi- 
 ments made in Hoppe-Seyler's laboratory by Araki 16 that in con- 
 ditions of acute want of oxygen (CO poisoning, etc.) large quanti- 
 ties of lactic acid are produced in the body. Putting together all 
 these facts, and the results of Walter's experiments on acid poison- 
 ing, we drew the conclusion that what the respiratory center re- 
 sponds to is the combined effect of carbonic acid and other acids on 
 the reaction of the blood. It seemed no longer possible to maintain 
 the hypothesis that CO 2 acts specifically in exciting the respira- 
 tory center. The long duration of the lowering of alveolar CO 2 
 pressure after exposure to want of oxygen seemed intelligible on 
 the theory that excess of lactic acid had been produced owing to 
 the anoxaemia, and that the sodium or potassium lactate thus 
 formed had been excreted by the kidneys, thus robbing the body 
 of alkali and leaving the blood correspondingly less alkaline a 
 deficiency which it required some time to make up. 
 
 This conclusion was further strengthened by the observation of 
 Douglas and myself, that after an excessive muscular exertion 
 
 11 Boycott and Haldane, Journ. of Physiol., XXXVII, p. 355, 1908. 
 "Ogier Ward, Journ. of PhyswL, XXXVII, p. 378, 1908. 
 "Galleotti, Arch. Ital. de BioL, XLI, p. 80, 1904. 
 "Aggazotti, Ibid., XLIV, 1905. 
 
 "Araki, Zeitschr. f. physiol. Chemie, XV, p. 335, 1908; also XVI, p. 425; 
 XVII, p. 311 ; XVIII, p. 422. 
 
RESPIRATION jgi 
 
 the alveolar CO 2 pressure remains low for about an hour. 17 We 
 attributed this to the effect on the respiratory center of lactic acid 
 
 10 15 20 25 30 35 40 45 50 55 60 65 70_ 75 
 
 Figure 54. 
 
 Blood, + Serum, O Corpuscles of same sample. El Blood, 
 X Serum of another sample. / Blood of another sample. 
 Another sample of blood. Same sample with acetic acid added. 
 
 O 8 parts Na 2 HPO 4 and 2 parts KH 2 PO 4 . Q Equal 
 IS IS 
 
 parts Na 2 HP0 4 and KH 2 PO 4 . KC1 solution. 
 15 10 
 
 given off into the blood by muscles in which the work had been 
 far in excess of the possible oxygen supply. The correctness of 
 
 "Douglas and Haldane, Journ. of PhysioL, XXXVIII, p. 43 1, 1909. 
 
1 82 RESPIRATION 
 
 this inference was shortly afterwards established by Ryffel, 18 who 
 had meanwhile worked out a new and very convenient method of 
 determining small amounts of lactic acid in blood and urine. 
 
 The methods of determining hydrogen ion concentration in the 
 blood were at that time still too crude to permit of testing these in- 
 ferences by direct determinations, but shortly afterwards the elec- 
 trometric method was greatly improved by Sorensen and particu- 
 larly by Hasselbalch of Copenhagen. In 1912 Hasselbalch and 
 Lundsgaard 19 published curves showing the variations of hydro- 
 gen ion concentration with variations in CO 2 pressure at body 
 temperature in ox blood, and Lundsgaard 20 repeated the experi- 
 ments with human blood. Figure 54 shows graphically their re- 
 sults for blood and other liquids. For convenience' sake the results 
 for hydrogen ion concentration are plotted, not directly in terms of 
 gram molecules per liter, but in terms of the negative power of TO 
 representing this value. This mode of notation, introduced by 
 Sorensen, is represented by the symbol PH, and since the negative 
 power increases with diminution of hydrogen ion, or increase of 
 hydroxyl ion concentration, the curve rises with diminution of 
 hydrogen ion concentration. 
 
 At body temperature the point of neutrality corresponds to a 
 PH about 6.78, as indicated by the thick line in the figure. It will 
 be seen from the curves that even with a far higher pressure of 
 CO 2 than exists in the living body the neutral point is not 
 reached. This is partly due to the fact that the proportional ioniza- 
 tion of carbonic acid becomes less and less with increasing con- 
 centration, just as is the case with other acids, including even 
 strong ones. The lower curve (for neutral potassium chloride solu- 
 tion) shows this clearly. Thus sulphuric acid when pure is quite 
 devoid of acid properties and does not attack metals, because it is 
 practically not ionized at all. This can be understood on the 
 theory, already alluded to, that ionization in aqueous solutions is 
 brought about through a reversible reaction with the water mole- 
 cules. 
 
 The influence of a buffer substance (disodium phosphate) in 
 hindering changes of hydrogen ion concentration is shown very 
 strikingly in the two curves for phosphate solutions. In blood, as 
 already pointed out, various buffer substances, including haemo- 
 globin with other proteins, and the phosphate in the corpuscles, 
 are present. The curve for acidified blood shows that even when 
 
 18 Ryffel, Journ. of Physiol., XXXIX, Proc. Physwl. Soc., p. xxix, 1910. 
 
 19 Hasselbalch and Lundsgaard, Bwchem. Zeitschr., XXXVIII, p. 77, 1912. 
 
 20 Lundsgaard, Btochem. Zeitschr., XLI, p. 247, 1912. 
 
RESPIRATION 183 
 
 blood is rendered distinctly acid these buffer substances still act 
 very efficiently. The haemoglobin acts as an alkali, whereas it 
 always acts as an acid in blood within the living body. 
 
 In order to test whether it is really to difference in PH that the 
 respiratory center normally reacts, Hasselbalch made the experi- 
 ment of altering the resting alveolar CO 2 pressure by changing 
 the diet. A meat diet, consisting largely of proteins containing 
 sulphur and phosphorus which break down into free sulphuric and 
 phosphoric acid, is evidently an acid-forming diet as compared 
 with a vegetable diet, which contains less protein and a relative 
 abundance of salts of organic acids which break up in the body so 
 as to yield carbonates. Hasselbalch found that with the acid meat 
 diet the resting alveolar CO 2 pressure was 4.4 mm. lower, and 
 then proceeded to compare the PH of the blood in the two condi- 
 tions. The results were as follows : 21 
 
 Alv. 
 
 CO2 Pressure 
 
 PH of blood 
 
 PH of blood at 
 
 
 mm. Hg. 
 
 at 40 mm. CO2 
 
 existing alveolar 
 
 
 
 pressure 
 
 CO2 pressure 
 
 Meat Diet 
 
 38.9 
 
 7-33 
 
 7-34 
 
 Vegetable Diet 
 
 43-3 
 
 7.42 
 
 7.36 
 
 It will be seen that at 40 mm. CO 2 pressure the blood sample 
 taken with the meat diet was distinctly more acid than with the 
 vegetable* diet, but that at the existing alveolar CO 2 pressure the 
 two values for PH were identical, at least within the limit of ac- 
 curacy of the method of measurement. Hence the respiratory 
 center had regulated the alveolar CO 2 pressure in such a manner 
 as to keep the PH of the blood almost constant. 
 
 There is other evidence pointing in the same direction. Barcroft 
 found that on the Peak of Teneriffe the dissociation curve of 
 human blood appeared to be normal, provided that the curve 
 was investigated, not at the normal sea level alveolar CO 2 pressure 
 of about 40 mm., but at the existing resting alveolar CO 2 pres- 
 sure. 22 We got a similar result at a greater height on Pike's 
 Peak, 23 as did also Barcroft and his co-workers on Monte Rosa. 24 
 
 21 Hasselbalch, Btochem. Zeitschr., XLVI, p. 416, 1912. 
 
 22 Barcroft, Journ. of Physiol., XLII, p. 44, 1911. 
 
 23 Douglas, Haldane, Henderson, and Schneider, Phil. Trans. Roy. Soc., (B) 
 203, p. 201, 1913- 
 
 24 See Chapter XVII, of Barcroft, The Respiratory Function of the Blood, 1913. 
 
1 84 RESPIRATION 
 
 As already pointed out this curve is shifted to the right or left with 
 varying alkalinity, and the shifting is a moderately delicate index 
 of the variation (Chapter III). Peters, 25 working with Barcroft, 
 has shown that the shifting with variations in CO 2 pressure de- 
 pends on the shifting of PH. Hence the constancy of the dissocia- 
 tion curve appeared to be a direct index of the constancy in PH of 
 the blood. The lowering of alveolar CO 2 pressure at high altitudes 
 seemed therefore to be just sufficient to keep the PH of the blood 
 steady in so far as direct methods enable us to measure the degree 
 of steadiness. As will be seen below, however, there is physio- 
 logical evidence that the blood is actually more alkaline at high 
 altitudes. More recently Hasselbalch and Lindhard have made 
 direct electrometric measurement of PH in a steel chamber after 
 exposure of sufficient duration to the low pressure, and their 
 measurements give practically the same result. 26 The resting al- 
 veolar CO 2 pressure on Pike's Peak was about 27 mm., or 13 mm. 
 below that at sea level. Raising the alveolar CO 2 pressure on Pike's 
 Peak to 40 mm. would have caused the extremest panting. 
 
 As soon as the results of Hasselbalch and Lundsgaard were 
 published, it was possible to estimate quantitatively the delicacy 
 with which the respiratory center responds to variations in the 
 reaction of the blood : for the delicacy of the reaction of the center 
 to variations of CO 2 pressure was known from our previous ex- 
 periments, while the curves of Hasselbalch and Lundsgaard made 
 it possible to convert variations of CO 2 pressure into variations of 
 PH in the blood. Some confusion arose, however, owing to the 
 fact that Lindhard, 27 and Hasselbalch and Lindhard, 28 had mean- 
 while published experiments which seemed to indicate that the 
 respiratory center in man is commonly far less sensitive to CO 2 
 than Priestley and I had found. The matter was therefore rein- 
 vestigated by Campbell, Douglas, Hobson, and myself. 29 We 
 found that the Danish observers had been deceived, owing to a 
 faulty modification of the method of sampling the alveolar air. 
 The fresh experiments gave practically the same results as Priest- 
 ley and I had obtained, so we could make the calculation accord- 
 ingly. 
 
 A rise of 0.2 per cent or 1.5 mm. in the CO 2 pressure of the 
 
 "Barcroft, The Respiratory Function of the Blood, p. 316, 1913. 
 
 28 Hasselbalch and Lindhard, Biochem. Zeitschr., 68, p. 293, 1915. 
 
 87 Lindhard, Journ. of Physiol., XLII, p. 337, 1911. 
 
 88 Hasselbalch and Lindhard, Skand. Arch. f. Physiol., XXVIII, 1911. 
 
 29 Campbell, Douglas, Haldane, and Hobson, Journ. of Physiol., XLVI, p. 301, 
 1913- 
 
RESPIRATION 185 
 
 alveolar air and arterial blood causes an increase of about 100 per 
 cent in the resting alveolar ventilation, and from Figure 54 it 
 will be seen that this corresponds to a difference of .012 in PH. 
 This difference, large as its physiological effect is, cannot be de- 
 tected with certainty by the electrometric method, or by indicators, 
 and is quite undetectable by the shifting of the dissociation curve 
 of oxyhaemoglobin. Nevertheless a twentieth of this difference 
 would produce an easily measurable effect on the breathing or 
 alveolar CO 2 pressure. The astounding delicacy of the regulation 
 of blood reaction is thus evident. No existing physical or chemical 
 method of discriminating differences in reaction approaches in 
 delicacy the physiological reaction. Unfortunately, however, the 
 quantitative significance of our calculation has not yet been ap- 
 preciated. The blood within the living body is still treated as if 
 its reaction were not only variable, during rest, as it is, but capable 
 of showing the variations by the existing very rough chemical and 
 physical reactions. One might as well try to cut delicate histo- 
 logical sections with a blunt carving knife, as try to demonstrate 
 ordinary very minute changes in blood reaction by the existing 
 physical and chemical methods. 
 
 It was discovered by Christiansen, Douglas, and myself, as 
 previously set forth, that the reduction of oxyhaemoglobin, as 
 this occurs in the course of the circulation, has an effect re- 
 sembling that of the addition of alkali to the blood. Thus the 
 CO 2 pressure of the blood in the systemic capillaries is pre- 
 vented from rising nearly as high as it would otherwise do. The 
 hydrogen ion concentration of the blood is also prevented from 
 rising in correspondence with the actual greatly restricted increase 
 in CO 2 pressure. Accordingly the actual increase of hydrogen ion 
 concentration in mixed venous as compared with arterial blood 
 must be very small. In this way the extraordinarily delicate regu- 
 lation of the reaction of arterial blood becomes much more intelli- 
 gible, as venous blood must be very little less alkaline than arterial 
 blood. In determining the hydrogen ion concentration of blood by 
 the ordinary electrometrical method it is necessary to reduce the 
 blood first, as the presence of oxygen interferes with the action of 
 the hydrogen electrode. 30 Thus the determination is made on re- 
 duced, or by Barcroft's method on partially reduced, blood, but 
 with a CO 2 pressure corresponding to that of arterial blood. It is 
 
 80 Peters, Journ. of Physiol., 48, Proc. Phys. Soc., p. vii, 1914. It is probable 
 that owing to incomplete reduction the values obtained by Hasselbalch have been 
 slightly too low. 
 
1 86 
 
 RESPIRATION 
 
 evident, therefore, that the value obtained for the hydrogen ion 
 concentration is lower than that which exists in either arterial or 
 venous blood in the living body. To investigate the amount of this 
 difference Parsons 31 adopted the method of determining the hy- 
 drogen ion concentration, not in whole blood, but in its serum, of 
 which the hydrogen ion concentration is not altered when free 
 oxygen is removed. Using this method, he found that with normal 
 blood the PH at a constant pressure of CO 2 at anywhere near the 
 alveolar CO 2 pressure is greater by .038 in the oxygenated than 
 the reduced blood. Figure 55 shows his results. From them and 
 
 8.6. 
 6-5 
 
 8-4 
 83 
 8.2 
 81 
 
 e-o 
 
 79 
 78 
 77 
 76 
 75 
 74 
 73 
 
 72 
 7/ 
 
 \ 
 
 \ 
 
 MM. HG 
 
 o 
 
 10 
 
 20 
 
 30 
 
 40 
 
 50 
 
 60 
 
 70 
 
 80 
 
 Figure 55. 
 
 Curve R, completely reduced blood. Curve O, fully oxygenated 
 blood. X, direct measurements on reduced blood without removal of 
 corpuscles. H, Hasselbalch's curve. 
 
 from Figure 26 (Chapter V) it is possible to calculate what 
 the difference for normal blood between the PH of arterial and 
 mixed venous blood is, assuming that the venous blood has lost 
 a certain proportion of its oxygen and simply gained a corre- 
 sponding proportion of CO 2 . If the venous blood had lost all its 
 
 "Parsons, Journ. of Physwl., LI, p. 440, 1917. 
 
RESPIRATION 
 
 oxygen the difference would be .07, as shown in Figure 56 from 
 Parsons's paper. Assuming, however, that the mixed venous blood 
 loses normally a fourth of its combined oxygen (see Chapter X), 
 the difference is only .0175 a difference which can hardly be 
 detected except by physiological methods, and which corresponds 
 to a rise of only 0.3 per cent in the alveolar CO 2 percentage. 
 
 It might be supposed that in order to obtain the true PH of 
 arterial blood under abnormal conditions all that is necessary is to 
 add a constant to the value obtained for reduced blood; and that 
 consequently the ordinary methods of determining PH (whether 
 electrometrically or from indications given by the dissociation 
 
 7-6 
 
 7-5 
 
 7-3 
 
 7-2 
 
 1-1 
 
 40 50 60 70 
 
 TOTAL C0 t CONTNT(c.c/ iQoc J OF BLOOD 
 
 Figure 56. 
 
 The slope of the line AC shows the rate at which the PH 
 of blood increases as its content of CO 2 increases in the 
 capillaries. 
 
 curve of oxyhaemoglobin) give reliable indications of any altera- 
 tion in the PH of the arterial blood. There is, however, no evidence 
 at present that this is the case, and there is in fact other evidence 
 pointing in the opposite direction. 
 
 If, in the first place, the proportion of haemoglobin in the blood 
 is altered, there will presumably be an alteration in the difference 
 between the PH of fully oxygenated and of reduced blood. Apart 
 altogether from this, however, there may be another kind of al- 
 teration in this difference. In the paper by Christiansen, Douglas, 
 and myself, it was pointed out that the probable reason why re- 
 duced blood appears to be more alkaline than oxygenated blood 
 is that on reduction the haemoglobin becomes more aggregated 
 and therefore acts less strongly as an acid. In abnormal blood the 
 degree of increased aggregation may be either increased or di- 
 minished. This will alter the difference in PH between oxygenated 
 and reduced blood, and will also, if our theory as to the cause of 
 
1 88 RESPIRATION 
 
 the peculiar shape of the dissociation curve of the oxyhaemoglobin 
 in blood is correct, alter the shape of the dissociation curve. 
 
 In a quite recent paper Lovatt Evans 32 has shown that the PH 
 of blood as determined colorimetrically by an indicator method 
 is as much as 0.2 higher than when determined electrometrically. 
 He has also shown pretty conclusively that the electrometric 
 method has an error owing to the formation of formate from 
 carbonate by catalytic action at the electrode, so that the PH of 
 blood is higher by 0.2 than appears from the electrometric de- 
 terminations. The new colorimetric method of Dale and Evans 33 
 seems to avoid several defects inherent in the electrometric method 
 as applied to blood. 
 
 On the existing evidence, and allowing for mistaken inferences 
 which have been drawn in ignorance of the peculiar properties of 
 haemoglobin (as it exists in the red corpuscles) in regulating the 
 PH of blood, it seems evident that during health the regulation 
 of the reaction of the arterial blood is carried out with a delicacy 
 and constancy of which we can at present only obtain a real con- 
 ception by physiological observations. The foregoing discussions 
 show that there are at least three regulators of the reaction the 
 lungs, the kidneys, and the liver. We can also now form a general 
 conception of how these regulators act under ordinary conditions. 
 
 The part played by the lungs in this regulation is, quite clearly, 
 to deal rapidly with variations in reaction due to varying pro- 
 duction of CO 2 , and particularly to the rapid variation caused by 
 varying muscular exertion. By keeping the alveolar CO 2 pressure 
 approximately normal, the action of the lungs keeps the arterial 
 CO 2 pressure approximately normal; and so long as the dissocia- 
 tion curve for CO 2 in the blood is also kept normal by other means 
 the reaction of the arterial blood is also kept almost exactly 
 normal. If, however, owing to rapid production of lactic acid in 
 muscles, rapid secretion of gastric or pancreatic juice, or other 
 causes, the dissociation curve for CO 2 is temporarily disturbed, the 
 breathing compensates approximately at once for the disturbance 
 in blood reaction. 
 
 The part played by the kidneys seems also clear. They not only 
 respond to the minutest variations in blood alkalinity by secreting 
 more acid or more alkaline urine, but also tend to keep normal the 
 proportion of soda and potash and other crystalloid substances 
 existing in the blood. In this way the dissociation curve of the CO 2 
 
 "Lovatt Evans, Journ. of Physiol., LIV, p. 353, 1921. 
 "Dale and Evans, Journ. of Physiol,, LIV, p. 167, 1920. 
 
RESPIRATION 189 
 
 in blood is kept normal ; and no physiological phenomenon is more 
 striking than the constancy of this curve under normal conditions. 
 If the proportion of available alkali is temporarily diminished 
 by acid poured out into the blood, the kidneys help to restore it 
 to normal again ; and similarly with excess of alkali. The action 
 of the kidneys is slow compared with that of the lungs; but is 
 apparently still more delicate. As L. J. Henderson was the first to 
 point out clearly 34 the PH of urine is no measure of the total acid 
 excreted in it, since urine, like blood, contains buffer substances. 
 Among these phosphoric acid plays the main part in acid urine, 
 and carbonic acid in alkaline urine. To measure the acid excreted 
 titration must be employed, and in titrating alkaline urine the 
 combining CO 2 must be allowed to escape. 33 
 
 The part played by the liver is to neutralize as far as possible 
 the disturbing effect of any excess of acid or of alkali introduced 
 into the body through the intestines, or formed in the tissues. By 
 allowing more, or less, ammonia to enter the circulation the 
 liver regulates the reaction of the blood ; and the neutral ammonia 
 salts are afterwards eliminated by the kidneys as being foreign 
 substances. The importance of the part played by the liver under 
 normal conditions is evident enough in view of the fact that in 
 man the ammonia excreted daily would just about suffice to neu- 
 tralize all the sulphuric acid formed daily. Like that of the kid- 
 neys, the action of the liver is slow and delicate as compared with 
 that of the lungs. 
 
 Possibly the intestines also play an active part in regulating 
 the blood reaction. It is known, at any rate, that alkali may be 
 eliminated from them in the form of insoluble alkaline phosphates. 
 
 We have now to consider how this joint regulation behaves 
 when the action of one of the regulators is interfered with; and 
 the case of interference with the lung regulation will be considered 
 first. This regulation may be disturbed in various ways, but per- 
 haps most is known at present as to its disturbance owing to the 
 fact that under abnormal conditions the stimulus of anoxaemia in- 
 creases the breathing, and thus disturbs the normal relation be- 
 tween the lung ventilation and the degree of stimulus of the 
 respiratory center owing to varying reaction of the arterial blood. 
 The history of the development of knowledge on this point is very 
 instructive. 
 
 84 L. J. Henderson, Amer. Journ. of Physiol., 21, p. 427, 1908. 
 86 Davies, J. B. S. Haldane, and Kennaway, Journ. of Physiol., LIV, p. 32, 
 1920. 
 
190 
 
 RESPIRATION 
 
 It has already been shown in Chapters VI and VII that until 
 the oxygen pressure of the inspired air is lowered by about a third, 
 or that of the alveolar air to about half (i.e., from about 100 mm. 
 to 50 mm.) there is no marked immediate increase in the breath- 
 ing. The effect on the respiratory center of the very distinct degree 
 
 Cos pressure 
 jnm.cf 800 
 
 Altitude 
 
 7SO 70O 
 
 ISO 
 
 140 
 JSO 
 120 
 
 no 
 
 100 
 90 
 60 
 70 
 60 
 50 
 40 
 30 
 20 
 10 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 
 o/.wt/ 
 30000 
 29000 
 28000 
 27.000 
 25&00 
 25,000 
 24000. 
 215,000 
 22.000 
 21.000 
 20,000 
 19.OOO 
 13,000 
 17.000 
 16,000 
 15.000 
 14,000 
 13,000 
 12,000 
 11.000 
 10,000 
 9,000 
 8,000 
 7,000 
 6.000 
 5,000 
 4.000 
 3000 
 2.0OO 
 1,000 
 
 o 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 / ^ 
 
 I 
 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 7* 
 
 
 
 - 
 
 A/veo/ar 
 
 sgjd 
 
 \ n 
 
 
 
 
 
 
 
 / 
 
 
 
 
 - 
 
 JWV$Si//V 
 
 ^ 
 
 k 
 
 
 
 
 
 // 
 
 
 
 
 - 
 
 
 
 
 x. x 
 
 sX 
 
 
 
 
 / 
 
 ' 
 
 
 
 
 
 
 
 
 
 .K 
 
 \0 
 
 
 
 // 
 "/- 
 
 
 
 
 
 - 
 
 
 
 
 
 
 
 5 
 
 ^ 
 
 f 
 
 
 
 
 
 - 
 
 
 
 
 
 
 
 / 
 
 % 
 
 ^ 
 
 
 
 
 
 - 
 
 At 
 
 veola 
 
 r 
 
 
 
 / 
 
 ^ 
 
 
 N 
 
 > x 
 
 
 
 
 
 pr&fre'Z^' 
 i re^a tc t Q 
 
 * *-2^* 
 
 D O 
 
 ^^ 
 
 ^ 
 
 
 
 
 x 
 
 ~x> 
 
 
 
 
 <7//l 
 P 
 
 vosp/H 
 fSSUt 
 
 r/r 
 e 
 
 ^ 
 
 r 
 
 ^ 
 
 -^ 
 
 ^ 
 
 i^Tj 
 
 .. 
 
 
 ^ 
 
 ^ 
 
 
 
 
 
 / 
 
 
 
 
 
 
 
 "" 
 
 -, 
 
 ii; 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 BOO 75O 7OO 65O 6OO SSO 5OO 4SO 4OO 3SO 3OO 2SO 2OO 
 Atmospheric pressure in mm. of mercury. 
 
 Figure 57. 
 Alveolar gas pressures in relation to barometric pressure or altitude. 
 
 of anoxaemia which is undoubtedly produced, in the manner ex- 
 plained in Chapter VII, is almost entirely masked by the con- 
 trary effect due to extra washing out of CO 2 and consequent 
 lowering of the PH in the arterial blood. But if exposure to the 
 lowered oxygen pressure is continued, not merely for perhaps an 
 hour, but for days or weeks, there is a quite marked increase in 
 
RESPIRATION 191 
 
 the breathing, as shown by a fall in the alveolar CO 2 pressure. 
 This fact, already referred to in connection with the historical 
 development of the theory of regulation of the breathing by the 
 blood reaction, was brought out in full clearness by the investiga- 
 tions carried out in connection with the Pike's Peak expedition 
 by Miss FitzGerald on persons fully acclimatized at different 
 altitudes. 36 Figure 57 represents graphically her results on this 
 subject. It will be seen that in such persons the alveolar CO 2 pres- 
 sure falls regularly with increase of altitude. In other words the 
 breathing increases in a regular ratio with diminution in the oxy- 
 gen pressure of the inspired air. 
 
 What is the cause of this increase? Since the experiments, 
 already referred to, of Boycott, Ogier Ward, and myself, it has 
 been pretty generally assumed that in response to the stimulus of 
 anoxaemia a slight acidosis, sufficient to account for the increased 
 breathing, develops in the blood. This explanation received strong 
 confirmation from the discovery by Barcroft in the Teneriffe 
 experiments that the dissociation curve of the oxyhaemoglobin of 
 the blood at high altitudes is sensibly the same in presence of the 
 existing alveolar CO 2 pressure as at sea level in presence of the 
 alveolar CO 2 pressure existing there. The extra acid, or dimin- 
 ished available alkali, present in the blood seemed just to compen- 
 sate for what would otherwise be increased alkalinity due to the 
 lowered CO 2 pressure. The physiological facts, however, do not 
 correspond with the lactic acid theory. Moreover no excess of 
 lactic acid could be discovered by Ryffel in the urine and hardly 
 any in the blood, of persons exposed to low pressures in a respira- 
 tion chamber or steel chamber, 37 or indeed in persons at high 
 altitudes ; 38 and no other abnormal acid could be discovered in the 
 blood. Hence the theory of an acidosis due to formation of ab- 
 normal acids cannot be substantiated. In the report of the Pike's 
 Peak Expedition we adopted the theory that the anoxaemia alters 
 the activity of the kidneys in such a way that they regulate the 
 blood to a lower level of alkalinity. 
 
 Another, and essentially similar, theory was adopted by Has- 
 selbalch and Lindhard as the result of experiments in a steel 
 chamber. 39 They found that the excretion of ammonia is markedly 
 
 "FitzGerald, Phil. Trans. Roy. Soc., 203 (B), p. 351, 1913; and Proc. Roy. 
 Soc., 88 (B), p. 248, 1914. See also, Yandell Henderson, Journ. of Biol. Chem., 
 1920. 
 
 "Ryffel, Journ. of Physiol., XXXIX (Proc. Physiol. Soc.), p. xxix, 1910. 
 
 88 See Barcroft, The Respiratory Function of the Blood, p. 260. 
 
 "Hasselbalch and Lindhard, Biochem. Zeitschr., 68, p. 295, 1915. 
 
1 92 RESPIRATION 
 
 diminished at the lowered pressure, and were thus led to the theory 
 that the acidosis of high altitudes is due to diminished formation 
 of ammonia by the liver as a consequence of anoxaemia. 
 
 The question was again taken up in a series of experiments in 
 steel chambers by Kellas, Kennaway, and myself, in which care- 
 ful measurements were made of the excretion of acid and am- 
 monia. 40 We found that even with a comparatively slight diminu- 
 tion of pressure there was a great diminution in the excretion of 
 acid and ammonia, and the urine passed to the alkaline side of 
 neutrality. The true explanation of the supposed acidosis then 
 revealed itself to us. The kidneys and liver were responding quite 
 normally, but to an alkalosis, this alkalosis being produced by the 
 increase (largely masked) of breathing caused by the anoxaemia. 
 A similar view of the supposed acidosis of high altitudes was 
 reached, on independent grounds which will be discussed below, 
 by Yandell Henderson. 41 
 
 The increased excretion of alkali and diminished formation of 
 ammonia lead gradually towards a compensation of the alkalosis 
 and simultaneous relief of the anoxaemia, this relief being due to 
 the increased oxygen supply to the lung alveoli, and to other 
 causes discussed in Chapters IX and X. But the final result is a 
 compromise. A certain small degree of anoxaemia and consequent 
 alkalosis still remains, as evidenced by a continued low excretion 
 of ammonia and other physiological symptoms and by the fact 
 that on removal of the anoxaemia there is a quite appreciable 
 immediate rise in the alveolar CO 2 pressure, as was shown for 
 instance, when we breathed air enriched with oxygen after we had 
 become acclimatized on Pike's Peak. The extra excretion of alkali 
 comes to an end, however, as the kidneys cannot reduce the blood 
 alkali further without very serious alteration of the normal balance 
 of salts in the blood. 
 
 The supposed acidosis is thus not an acidosis at all, but the in- 
 complete compensation of an alkalosis. The "adaptation" of the 
 blood so as to relieve the alkalosis and anoxaemia is also nothing 
 but an extension of the everyday adaptations by which alkalosis 
 and anoxaemia are continuously being prevented. The reason why 
 the adaptation takes so long at low atmospheric pressures is simply 
 that it takes a long time for the kidneys and liver to get level with 
 the very prolonged and considerable work thrown on them by 
 
 40 Kellas, Kennaway, and Haldane, Journ. of Physwl., LIII, p. 181, 1919. 
 
 41 Yandell Henderson, Science (N. S.), XLIX, p. 431, 1910; see also the series 
 of papers by Henderson and Haggard, Journ. of Biol. Chem., 1919-1921 incl. 
 
RESPIRATION 193 
 
 progressive increase in the breathing. They are, as it were, pursu- 
 ing in a leisurely manner a goal which is constantly receding from 
 them, so that it is a long time before they finally reach it. The 
 quantity of alkali which has to be removed from the blood and 
 tissues is very large, as a simple calculation will show. 
 
 With the compensation of the alkalosis there also comes com- 
 pensation of any secondary anoxaemia caused by the alkalosis as 
 a consequence of the Bohr effect discussed so fully in Chapters IV 
 and VI. Owing to the increased breathing the percentage satu- 
 ration of the arterial blood is (without any allowance for increased 
 oxygen secretion) as high as at first, while the oxygen pressure in 
 the systemic capillaries is higher (i.e., nearer normal) on account 
 of the decreased alkalosis. Cyanosis may be, however, quite as 
 marked as before. By the administration of acid the adaptation 
 to a lowered oxygenation of the arterial blood could doubtless 
 be hastened. 
 
 The study of responses to the anoxaemia and alkalosis of high 
 altitudes is of great medical interest, since, as already explained 
 in the two preceding chapters, anoxaemia is a very common and 
 often extremely dangerous clinical condition. There can be no 
 doubt that the same responses as occur in healthy persons at 
 high altitudes occur also in patients suffering from anoxaemia. It 
 is therefore important not to misunderstand these responses. 
 During the war, for instance, the intensely dangerous anoxaemia 
 of acute gas poisoning and "shock" was sometimes treated by the 
 administration of alkalies, on the theory, based on nothing but the 
 unintelligent use of a new method of blood examination, that the 
 patients were suffering from "acidosis." Physiological knowledge 
 as to the deadly significance of serious anoxaemia, and the (sup- 
 posed) acidosis as an adaptive change tending towards its com- 
 pensation, was ignored. It is also important to understand that the 
 adaptive changes require time, and that so-called palliative treat- 
 ment, by giving this time, may in reality be curative. 
 
 Another cause of interference with the lung regulation of blood 
 reaction is to place an animal or man in an atmosphere in which 
 the percentage or pressure of CO 2 is so high that the regulation 
 breaks down completely and there is in consequence an excessive 
 and lasting fall in the PH of the blood. This condition was studied 
 recently in animals by Yandell Henderson and Haggard. 42 They 
 made the very important and significant discovery that the acido- 
 
 ** Yandell Henderson and Haggard, Journ. of Biol. Chem., XXXIII, p. 333, 
 1918. 
 
I 9 4 RESPIRATION 
 
 sis thus produced gradually brings about a marked increase in the 
 capacity of the blood for combining with CO 2 . In other words the 
 dissociation curve of the CO 2 in blood, if plotted as in Figure 25, 
 would occupy a higher position. This is evidently a change tending 
 to counteract the diminished blood alkalinity produced by the 
 excess of CO 2 . 
 
 The same observers found that on prolonged and forced arti- 
 ficial ventilation of the lungs, so as to produce a condition of al- 
 kalosis, there is a corresponding diminution in the capacity of the 
 blood for combining with CO 2 . This is also a change towards 
 the normal alkalinity. Thus in an alkalosis produced by excessive 
 removal of CO 2 the available alkali in the blood diminished, while 
 in an acidosis produced by excess of CO 2 the available alkali 
 increased. It is clear that in either case the change is of a character 
 tending to neutralize the change in blood reaction. 
 
 What is the significance of this change? It occurs much too 
 quickly to be capable of explanation as due to an adaptive re- 
 sponse by the kidneys and liver. The probability is, therefore, that 
 it is due to exchange of anions between the tissues and blood in 
 the manner discussed in Chapter IV (Addendum), and is indica- 
 tive, therefore, of very severe alkalosis or acidosis of the tissues. 
 This would help to account for the very dangerous symptoms 
 which Henderson and Haggard found to be an accompaniment of 
 any considerable diminution of the available alkali of the blood, 
 when the diminution was produced by excessive artificial respira- 
 tion. Thus a diminution of about 40 per cent in the capacity of 
 the blood for combining with CO 2 was fatal to the animal. A 
 similar diminution due to the acidosis caused by running quickly 
 up a stair is hardly felt at all. In the latter case the diminution in 
 available alkali in the blood indicates a quite trifling acidosis, 
 while in the former a similar change in the blood indicates a 
 severe and fatal alkalosis. 
 
 These and other experiments 43 of these investigators brought 
 out in a striking manner that it is a complete mistake to regard 
 diminution of the available alkali (or so-called "alkaline reserve") 
 of the blood as a definite sign of acidosis in the living body. The 
 "alkaline reserve" of the blood and whole body is only another 
 name for its "titration alkalinity" ; and it has already been shown 
 above that titration alkalinity is no measure, and not even a sure 
 qualitative indication, of the real alkalinity of the blood. In the 
 
 43 Haggard and Henderson, Journ. of Bwl. Chem., XXXIX, p. 163, 1919; 
 and XLIII, pp. 3, 15, and 29, 1920. 
 
RESPIRATION 195 
 
 experiments of Yandell Henderson and Haggard the animals were 
 suffering from severe alkalosis although the "alkaline reserve^ or 
 titration alkalinity of their blood was greatly diminished; and 
 similarly they were suffering from severe acidosis although the 
 "alkaline reserve" of their blood was greatly increased. 
 
 It was these observations that led Yandell Henderson to the 
 same conclusion which we reached namely, that in the anoxae- 
 mia of high altitudes there is a condition of alkalosis, and not of 
 acidosis, in spite of the greatly reduced titration alkalinity or 
 "alkaline reserve" of the blood. 
 
 A ready method of interfering temporarily with the regulation 
 of blood reaction by the lungs is forced breathing. This can be 
 continued for a considerable time if it is employed in moderation. 
 Leathes 44 found that if forced breathing is continued for some 
 time the urine becomes alkaline to litmus, and the titration alka- 
 linity has still more recently been investigated by Davies, J. B. S. 
 Haldane, and Kennaway. 45 The titration alkalinity is, however, 
 not so striking as after a large dose of sodium bicarbonate has been 
 taken. The same observers found that after a large dose of sodium 
 bicarbonate there was not only a rise of as much as I per cent in the 
 alveolar CO 2 pressure for some hours, but the available alkali in 
 the blood (as shown by the dissociation curve for CO 2 ) was 
 markedly increased, while there was also a great increase in the 
 titration alkalinity of the urine. Large quantities of bicarbonate 
 (readily determined by the blood-gas apparatus) were present in 
 the urine, which effervesced briskly on the addition of acid, though 
 the actual alkalinity of the urine was of course only feeble, since 
 the CO 2 acted as a buffer. The titration alkalinity (after removal 
 of liberated CO 2 ) was equivalent to nearly I per cent of HC1. The 
 ammonia had completely disappeared from the urine, and this 
 was also the case after forced breathing, although such a degree 
 of forced breathing as was practicable did not appreciably dimin- 
 ish the available alkali in the blood within one and one-half 
 hours. A stay of several hours in air containing 5 to 6 per cent of 
 CO 2 was also not sufficient to increase appreciably the available 
 alkali of the blood, although the titration acidity of the urine was 
 increased, along with increased excretion of ammonia. Collip has, 
 however, found that, as might be expected from the change in 
 distribution of acid and alkali between plasma and corpuscles 
 
 "Leathes, Brit. Med. Journ., Aug. 9, 1919. 
 
 46 Davies, J. B. S. Haldane, and Kennaway, Journ. of Physiol., LIV, p. 32, 
 1920. 
 
196 RESPIRATION 
 
 when the PH of blood is altered, the alkaline reserve of the plasma 
 was distinctly diminished by forced breathing. 46 
 
 The blood reaction may, of course, be disturbed in other ways 
 than by interference with respiration. One of these ways is by 
 ingestion of acids or by production within the body of great ex- 
 cess of some organic acid. Walter's experiments, interpreted in 
 the light of our present knowledge, showed the effects of acid 
 poisoning in stimulating to the utmost all the means of diminishing 
 acidosis, including excessive breathing, greatly increased forma- 
 tion of ammonia, and secretion, presumably, of an abnormally 
 acid urine. The titration alkalinity or "alkaline reserve" of the 
 blood and doubtless also of the whole body was evidently dimin- 
 ished very greatly. 
 
 Christiansen, Douglas, and Haldane produced a temporary true 
 acidosis by flooding the blood with lactic acid produced by mus- 
 cular anoxaemia during the heavy exertion of running several 
 times upstairs. In this case two results followed. In the first place 
 there was a fall in the resting alveolar CO 2 pressure, which was, 
 in several experiments, about 39 mm. before the exertion, and 
 30.5 mm. about 10 minutes after the exertion. The blood absorbed 
 about 49 volumes of CO 2 per 100 of blood before the exertion in 
 presence of the existing alveolar CO 2 pressure, and only about 
 28 afterwards. After one and one-half hours both the resting 
 alveolar CO 2 pressure and the absorbing power of the blood for 
 CO 2 had returned to normal. 
 
 In these experiments the capacity of the blood for absorbing 
 CO 2 at a CO 2 pressure of 40 mm. had been reduced by about 40 
 per cent, and the resting alveolar CO 2 pressure by about 20 per 
 cent, corresponding to an increase of about 25 per cent in the 
 lung ventilation. There was thus a very distinct acidosis ; but ref- 
 erence to the calculations already made will show that the acidosis 
 could not have been detected by any existing method of directly 
 estimating hydrogen ion concentration. 
 
 The great drop in the capacity of the blood for combining with 
 CO 2 suggests at first that the blood had become correspondingly 
 inefficient as a carrier of CO 2 from the tissues to the lungs, and 
 that this deficiency could only be made up by a greatly increased 
 circulation rate, if it was made up at all. The truth, however, is 
 that the main difference produced was that the dead weight of 
 CO 2 always carried round by the blood was greatly diminished. 
 As a carrier of CO 2 from the tissues to the lungs, the blood was 
 
 48 Amer. Journ. of Physiol., LI, p. 568, 1920. 
 
RESPIRATION 
 
 197 
 
 nearly as efficient as normal blood. This is due to the fact that, as 
 already explained in Chapter V, the haemoglobin and other pro- 
 teins play the essential part in the actual conveyance of CO 2 from 
 the tissues to the lungs, and can still play this part in spite of what, 
 in a physiological sense, is extreme acidosis. 
 
 The experiments were practically a replica in man of the ex- 
 periments of Geppert and Zuntz on muscular activity in dogs 
 (Chapter I). In discussing these experiments Priestley and I were 
 not aware that a very great diminution of the CO 2 content of the 
 blood may be caused by acidosis without any serious diminution 
 in the capacity of the blood for conveying CO 2 from the tissues to 
 the lungs. The discovery made in 1914 by Christiansen, Douglas, 
 and myself has greatly altered the previously existing ideas as to 
 the conveyance of CO 2 from the tissues. 
 
 The comparatively rapid recovery of the blood after the flood- 
 ing of the body with lactic acid was evidently due to the fact that 
 lactic acid was rapidly oxidized before the slight acidosis actually 
 produced had time to cause any considerable extra excretion of 
 acid by the kidneys, or formation of ammonia by the liver. Had 
 the acidosis been produced by a mineral acid it would probably 
 have taken far longer to pass off. 
 
 Disturbance of the blood reaction may be artificially produced 
 by the ingestion of acids or alkalies, or even, to a slight extent, by 
 varying the character of the diet. It requires a very large amount 
 of acid or alkali to produce any considerable disturbance. This is 
 partly due to the abundance of buffer substances in the body, but 
 still more to the effective means (variations in lung ventilation, 
 ammonia formation, and excretion of acid or alkali by the kid- 
 neys) which the body possesses of active defence against dis- 
 turbance of reaction. If the administration of acids or alkalies is 
 used medicinally as a means of assistance in the regulation of the 
 blood reaction, the large doses required must be borne in mind. 
 Small doses cannot but be practically useless. The amelioration of 
 the physiological symptoms of acidosis or alkalosis will form the 
 safest guide to what is required ; but it is evidently very important 
 not to mistake alkalosis for acidosis, or the hyperpnoea of acidosis 
 for the abnormal breathing caused by anoxaemia or an exhausted 
 or "neurasthenic" respiratory center. There are no short cuts to 
 decisions on such a subject. A physician must be a real physician, 
 and must have learned to be one by study of how the living body 
 behaves of what its <iW is, to use the old expression of Hip- 
 pocrates. 
 
1 98 RESPIRATION 
 
 As the kidneys are essentially concerned in the regulation of the 
 reaction within the body, it is evident that failure of the kidneys 
 may cause serious disturbance of reaction. As, moreover, normal 
 human urine is acid, and presumably is so in all animals if food is 
 not taken, the disturbance will naturally be in the direction of 
 producing acidosis. Hyperpnoea and other symptoms suggestive 
 of acidosis are commonly met with as an accompaniment of serious 
 inflammation of the kidneys; and these symptoms are now com- 
 monly attributed to acidosis. One peculiarity of them is that there 
 may be little or no increase in the ratio of ammonia to total nitro- 
 gen excreted. 
 
 Considerable new light is thrown on the causes of acidosis by 
 quite recent experiments of J. B. S. Haldane. 47 The experiments 
 consisted in taking large doses of NH 4 C1 during two or three 
 days, so that an abnormal percentage of ammonia was present in 
 the blood. As a result there were very pronounced respiratory 
 and other symptoms of acidosis, including a marked fall in the 
 available alkali of the blood. Owing to the excess of ammonia in 
 the blood part of the ammonia of the NH 4 C1 had been converted 
 into urea, setting free much HC1 into the blood. The normal 
 response in which the liver sets free ammonia into the blood on 
 the approach of acidosis was of course reversed, and though the 
 urine was very acid the kidneys were unable by themselves to 
 cope effectively with the HC1, so that acidosis resulted. 
 
 A further result was that the supply of phosphate in the body 
 began to run short, so that the kidneys could not excrete so much 
 acid as usual for a corresponding acidosis. When neutralized 
 sodium phosphate was taken the excretion of acid was much 
 increased, and the acidosis passed off correspondingly more 
 rapidly. 
 
 These experiments are of special interest, as they revealed a 
 practicable method of artificially producing marked symptoms 
 of acidosis in man. Previous attempts to do so by drinking large 
 quantities of dilute HC1 or acid sodium phosphate had failed 
 owing to the efficacy of the physiological means of regulation. 
 It seems likely that in the acidosis of Bright's disease the forma- 
 tion of ammonia by the liver is checked by the accumulation of 
 ammonia in the blood owing to the inefficiency of the kidneys. 
 Hence the ratio of ammonia to total nitrogen in the urine is not 
 increased. 
 
 It follows from the facts brought forward in this chapter that 
 
 4T J. B. S. Haldane, Journ. of Physiol., LV, 1921. 
 
RESPIRATION 199 
 
 the regulation of alveolar and arterial CO 2 pressure resolves itself 
 into regulation of the blood reaction, and that the blood reaction 
 itself is a normal which is constantly being regulated within 
 marvelously narrow limits so narrow that the variations, evident 
 though they are made by physiological reactions, cannot be fol- 
 lowed adequately by existing physical and chemical methods. 
 
 At this point it seems desirable to consider and criticize some of 
 the indirect means which have been used for estimating variations 
 in the hydrogen ion concentration of the blood. In recent years the 
 capacity of the blood, or of its serum, for combining with CO 2 has 
 commonly been taken as an index of hydrogen ion concentration, 
 this capacity being also alluded to as a measure of the "alkaline re- 
 serve" of the blood. It is evident that the "alkaline reserve" of the 
 blood is only another name for the "titration alkalinity" when CO 2 
 is allowed to escape. It is also evident from facts described above 
 that the alkaline reserve is increased in conditions of acute acidosis 
 due to excess of CO 2 , and diminished in conditions of acute al- 
 kalosis due to excessive lung ventilation caused by artificial res- 
 piration or anoxaemia. Hence although the alkaline reserve is 
 diminished in acidosis due to the presence of abnormal acids in 
 the blood, a diminution in alkaline reserve cannot be regarded as 
 by itself an index of acidosis. There is, in fact, no necessary con- 
 nection between diminution in alkaline reserve or titration alka- 
 linity and diminution in blood alkalinity. 
 
 Another indirect method which has been used for estimating 
 variations in alkalinity is observation of one or more points in the 
 dissociation curve of the oxyhaemoglobin of the blood in presence 
 of the existing alveolar CO 2 pressure. This method is due to 
 Barcroft and his pupils, and is based on the following facts. ( I ) 
 As was shown in Chapter IV, each point in the dissociation curve 
 of oxy- or CO-haemoglobin in blood is simply displaced to a pro- 
 portional distance to the right or left on varying within wide limits 
 the partial pressure of CO 2 . Thus only one constant in the equa- 
 tion expressing the curve is altered. (2) It was shown by Peters, 48 
 and this had been completely confirmed by Hasselbalch, 49 that the 
 alteration in the constant depends, in cases where only the CO 2 
 pressure is varied, on alterations in the hydrogen ion concentra- 
 tion, and can thus be used as a measure of it. Barcroft and others 
 have therefore used the alteration in the constant as a measure in 
 all cases of variation of hydrogen ion concentration in the blood. 
 
 48 Barcroft, The Respiratory Functions of the Blood, p. 316. 
 
 49 Hasselbalch, Biochem. Zeitschr., 78, p. 132, 1916. 
 
200 RESPIRATION 
 
 In persons at high altitudes, for instance, the constant is appar- 
 ently quite normal in presence of the existing alveolar CO 2 pres- 
 sure; and from this fact it was inferred that the hydrogen ion 
 concentration of the blood is also normal, as already mentioned. 
 On the other hand, in persons who have shortly before undergone 
 some excessive muscular exertion the constant is very markedly 
 altered ; and from this fact a corresponding increase of hydrogen 
 ion concentration in the blood is inferred. The same method has 
 been employed for estimating variations of hydrogen ion con- 
 centration in the blood of patients. 
 
 When, however, the facts are examined more closely it appears 
 that there must be a flaw in the reasoning. In the case of persons 
 who have completed some severe muscular exertion in a few min- 
 utes before, there is no physiological evidence of anything but a 
 most trifling acidosis, such as could not possibly be detected by 
 alterations in the constant of the dissociation curve. The breathing 
 is only increased to such an extent as to reduce the alveolar CO 2 
 pressure by about a fifth. This only corresponds to an acidosis 
 equivalent to what would be produced by a rise of 0.3 mm. in the 
 alveolar CO 2 pressure ; and such a rise would be entirely inappreci- 
 able in its effect on the dissociation curve of oxyhaemoglobin. The 
 rise apparently indicated by the alteration in the constant is enor- 
 mously greater. Hence it appears that there must be some other 
 cause for the alteration than rise in hydrogen ion concentration. 
 This other cause is probably operative in many cases of patho- 
 logical acidosis. 
 
 Another indirect method of measuring hydrogen ion concen- 
 tration has been proposed by Hasselbalch. 50 He showed quite 
 clearly that when the pressure of CO 2 is varied in blood or even 
 serum the hydrogen ion concentration, as separately determined, 
 is proportional to the ratio of combined CO 2 (which, as already 
 explained, is a measure of the bicarbonate present) to free CO 2 
 when allowance is made for the percentage ionization of the free 
 CO 2 and bicarbonate. This corresponds with the fact that the blood 
 behaves as if more alkali were constantly being added to it in pro- 
 portion as its reaction approaches the neutral point. It is very re- 
 markable how closely Hasselbalch's law holds for different kinds 
 of blood and in blood serum, in spite of great differences in the 
 dissociation curves for CO 2 . Figure 58 from Hasselbalch's paper 
 is also very interesting as showing (for fresh ox blood) the dif- 
 ferences in the dissociation curves for serum, blood, and corpuscles. 
 
 60 Hasselbalch, Biochem. Zeitsckr., 78, p. 112, 1916. 
 
RESPIRATION 
 
 201 
 
 These differences are just what might be expected in view of the 
 action of haemoglobin as a weak acid in alkaline solutions. In spite 
 of the great differences in the dissociation curves of blood and 
 serum Hasselbalch's law held good. He therefore applied it as a 
 means of calculating the hydrogen ion concentration of corpuscles 
 and of abnormal blood, and seemed justified in doing so. 
 
 Vol % 
 
 
 
 
 
 . ' 
 
 " 
 
 
 
 
 ^= 
 
 im-* 
 J-J 
 
 ' 
 
 
 <: 
 
 .- ' 
 < 
 
 .- >" 
 
 ^ 
 
 -'*" 
 
 
 
 
 , ( 
 
 U 
 
 if 
 
 / 
 
 ^ 
 
 
 ^" 
 
 M 
 
 ^- 
 
 
 
 
 
 1 
 
 / 
 
 /> 
 
 ^ff^ 
 
 
 
 
 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 ^ 
 
 ** 
 
 
 
 
 
 
 
 
 //I 
 
 JT 
 
 
 
 
 
 
 
 
 
 
 10 tO 30 40 50 60 70 SO SO 
 
 . Serum, 38 
 
 Blood, 1 8 
 Blood, 38 
 Blood corpuscles, 38 
 Serum, 18 
 
 Figure 58. 
 COa dissociation curves (from Hasselbalch loc. cit.) 
 
 Nevertheless this method, like that of Barcroft and Peters, 
 seems to break down with abnormal blood. As an example of 
 abnormal blood he took, from the paper already referred to by 
 Christiansen, Douglas, and myself, experiments in which Douglas 
 had flooded his blood with lactic acid by running quickly a number 
 of times up and down the laboratory stairs at intervals during 
 about a quarter of an hour. As a consequence his blood had lost 
 about 40 per cent of its normal power of combining with CO 2 , and 
 his resting alveolar CO 2 pressure was diminished by about a fifth. 
 The samples were taken about ten minutes after the last ascent of 
 the stairs, and all sensible hyperpnoea had passed off. From the 
 data given, Hasselbalch calculates, in accordance with the law 
 he had discovered for the same blood at varying pressures of CO 2 , 
 that the PH of Douglas's arterial blood had fallen by .12. This 
 would, in accordance with the data given above as to the effects 
 of increase of PH on the breathing, suffice to increase the breath- 
 ing to about ten times its resting value. Indeed Hasselbalch evi- 
 dently believed that there must have been such an increase, since 
 he speaks of the immensely increased breathing being unable to 
 compensate for the decrease in PH. The breathing was, however, 
 perfectly quiet and apparently normal, though the lowering of 
 
202 RESPIRATION 
 
 the alveolar CO 2 pressure showed that it was about a fourth deeper 
 than it otherwise would have been. On the physiological evidence, 
 therefore, the fall in PH was only about .003, instead of .12, or 
 only one-fortieth as much as calculated. From this example it 
 would seem to follow that Hasselbalch's method, when extended 
 to abnormal blood, is as unreliable as that of Barcroft and Peters. 
 Further investigation as to methods of determining hydrogen ion 
 concentration in abnormal blood seems to be much needed. 
 
 Except by observation of physiological reactions, there seems at 
 present to be no method of estimating in a reliable manner the 
 small variations in PH which are of so much physiological impor- 
 tance. Hasselbalch estimates that a difference of .03 can be detected 
 in single determinations by the electrometrical method ; but this is 
 a very large difference, corresponding to an increase of 250 per 
 cent in the breathing. The colorimetric method by means of indi- 
 cators is equally rough. Time and effort will continue to be wasted 
 on futile measurements until the extreme fineness of the physio- 
 logical regulation of PH in the blood and tissues is more fully 
 realized. 
 
 REACTION OF THE BLOOD IN EIGHT DIFFERENT WOMEN 
 
 BEFORE AND 
 
 AFTER 
 
 CHILDBIRTH 
 
 PH. at 
 40 mm. 
 Before 
 
 CO2 pressure 
 After 
 
 Alveolar CO 2 
 pressure 
 Before After 
 
 PH at alveolar 
 COi pressure 
 Before After 
 
 7.40 
 7.40 
 7-45 
 
 7-44 
 7.48 
 7-45 
 
 31.0 
 27.7 
 
 35-6 
 
 42.2 
 
 43-5 
 398 
 
 7-44 7-43 
 7.49 7.46 
 
 7.48 7-45 
 
 7-39 
 
 7-43 
 
 32.5 
 
 43-5 
 
 7.42 7.42 
 
 7-39 
 7-38 
 7.38 
 
 7-35 
 
 7-44 
 7-45 
 7-43 
 7.38 
 
 32.7 
 27-7 
 30.3 
 33-8 
 
 37-7 
 33-5 
 38.3 
 37-3 
 
 7-43 7-45 
 7-45 7-49 
 7.4i 7-44 
 7.38 7.40 
 
 Mean 7.39 
 
 7-44 
 
 3i.3 
 
 39-5 
 
 7-44 7-44 
 
 On account of various sources of error, already alluded to, in 
 the electrometrical or other measurements of PH, we are still with- 
 out much very exact information as to the permanent steadiness 
 during health of the alkalinity of the blood under resting condi- 
 tions. In this connection some very interesting observations have 
 been made by Hasselbalch and Gammeltoft on the PH of the 
 
RESPIRATION 203 
 
 blood during and after pregnancy. 51 It had already been found 
 by Hasselbalch and others that the alveolar CO 2 pressure is much 
 lower than normal during pregnancy. Taking advantage of this 
 fact, they determined the PH of arterial blood before and after 
 childbirth with the results shown in the accompanying table. 
 
 Allowing for the probable errors in determining the PH and 
 alveolar CO 2 pressure, these figures seem to show that the fall in 
 alveolar CO 2 pressure compensates within the limits of accuracy of 
 the electrometric method for the fall in the PH of the blood which 
 would otherwise occur. The mean of the first two columns shows 
 that this fall in PH would have been 0.05, whereas the compen- 
 sating fall in alveolar CO 2 pressure was 8.2 mm. as shown by the 
 mean for the second two columns. Hence a difference of o.oi in 
 PH corresponds to a difference of 1.6 mm. of CO 2 pressure, or 
 0.23 per cent of CO 2 in alveolar air. We have already seen, how- 
 ever, that a change of about this amount in alveolar CO 2 pressure 
 is sufficient to cause either apnoea or doubling of the alveolar 
 ventilation according to its direction. Even under the most favor- 
 able conditions it is hardly possible at present to determine differ- 
 ences in PH within the body to within 0.03 in single observations; 
 but by measuring the variations in lung ventilation as compared 
 with production of CO 2 we have an index of change in PH which 
 is at least 50 times as sensitive as the existing direct electrometric 
 method, exact as this is in comparison with older methods. 
 
 Although the measurements of PH showed no change in the 
 alkalinity of the blood during pregnancy, yet the fall in alveolar 
 CO 2 pressure indicated that there was an increase of 25 per cent in 
 the lung ventilation per unit of CO 2 given off. This, therefore, 
 would correspond to an "acidosis" to the extent of a PH of 0.003 
 an amount far too small for direct measurement. That it was 
 acidosis which caused the increase in the breathing was shown by 
 the fact that the increase was accompanied by an increase of about 
 20 per cent in the proportion of nitrogen excreted as NH 3 to total 
 nitrogen excreted in the urine. The authors conclude that there is 
 an increased acid production in the body during pregnancy (or 
 perhaps an increased drain of alkali from the body of the mother) , 
 but that it is compensated by increased breathing and formation 
 of NH 3 . It is true that relatively to the degree of accuracy at 
 present attainable in determining the PH of blood the compensa- 
 tion is perfect. But if the compensation were really perfect we 
 should be landed in the position of the vitalists of assuming effects 
 
 "Hasselbalch and Gammeltoft, Biochem. Zeitschr., 68, p. 206, 1915. 
 
204 RESPIRATION 
 
 produced without any measureable cause. In reality the acidosis 
 is not completely compensated, and the incompleteness is only 
 hidden by the extreme roughness of the method of measurement 
 in comparison with the fineness of the physiological reaction. 
 
 The table seems to indicate that the normal PH is not quite the 
 same, though very nearly the same, in different individuals. For 
 the present, however, this conclusion is rather doubtful, in view 
 of the fact that the measurements were for imperfectly reduced 
 blood. We have seen already that in spite of the accuracy of regu- 
 lation there are individual differences in the normal alveolar CO 2 
 pressure, the normal composition of haemoglobin, and the normal 
 dissociation curve of blood for CO 2 . As regards every detail of 
 structure and function we may be certain of rinding similar differ- 
 ences when the measurements are made with sufficient accuracy; 
 and this doubtless applies also to even the PH of the blood. 
 
 We have already considered one cause which alters the PH to 
 which the respiratory center regulates. This cause is anoxaemia. 
 At high altitudes the body is in the long run projected to a large 
 extent from the effects of the alkalosis thus produced, because the 
 kidneys and liver still react almost true to the normal PH. There 
 can be no doubt that other causes, such as the action of anaes- 
 thetics or poisons, or of other small changes in the composition of 
 the blood, would have a similar effect in altering the standard 
 to which the PH regulation of the arterial blood is set. This 
 question, and the question how the PH is regulated, not merely in 
 the arterial blood, but in the systemic capillaries, will be deferred 
 to Chapters X and XIV. 
 
 We can now see much more clearly why it is that the resting 
 alveolar CO 2 pressure is not quite steady in spite of the extreme 
 sensitiveness of the respiratory center to the minutest variation 
 in alveolar CO 2 pressure. There are various causes tending to 
 disturb the constancy of the reaction of the blood ; and the respira- 
 tory center, and not merely the kidneys and liver, must do its share 
 in compensating for these disturbances. Hence the alveolar CO 2 
 pressure cannot remain completely steady during rest. One of 
 these causes is the secretion of acid or alkaline digestive juices. 
 On account of the secretion of acid gastric juices the alveolar 
 CO 2 pressure rises distinctly very soon after a meal. The effects of 
 a meal on alveolar CO 2 pressure have been investigated recently 
 by Dodds. 52 He found that there is normally a sharp rise varying 
 in different individuals, but usually amounting to about 4 mm. half 
 
 82 Dodds, Journ. of Physiol., LIV, p. 342, 1921. 
 
RESPIRATION 205 
 
 an hour after the meal. This is rapidly followed by an equally 
 marked fall below normal, culminating about one and one-half 
 hours after the meal, with a subsequent rapid return to normal. 
 
 Bennett and Dodds 53 have found that the rise of alveolar CO 2 
 just after a meal is closely related to the concentration and rate 
 of secretion of the gastric hydrochloric acid as indicated by 
 samples taken from the stomach. In cases where there is little or 
 no secretion of HC1 the rise in alveolar CO 2 is absent, though the 
 fall due to alkaline secretion into the intestine is still present. 
 Another cause of variation in alveolar CO 2 pressure is the charac- 
 ter of the diet. With an alkali-forming vegetable diet the alveolar 
 CO 2 pressure is quite considerably higher than with an acid-form- 
 ing meat diet. This was brought out very clearly in some of the 
 experiments of Hasselbalch alluded to above; and he showed at 
 the same time that the reaction of the urine varied in correspond- 
 ence with the changes in alveolar CO 2 pressure. 
 
 During starvation the body is living on what amounts to an 
 acid-forming diet, and Higgins 54 has shown that during starva- 
 tion the alveolar CO 2 pressure falls. Perhaps the most striking 
 effects are obtained with a carbohydrate-free diet. This leads to 
 the formation within the body of a certain amount of aceto-acetic 
 and oxybutyric acids, as in severe diabetes. Higgins, Peabody, 
 and Fitz 55 showed that there is a striking fall in alveolar CO 2 
 pressure, together with a very large elimination of oxybutyric and 
 aceto-acetic acid by the kidneys, and an accompanying large in- 
 crease in ammonia excretion and excretion of acid. 
 
 All the available evidence points, therefore, to the conclusion 
 that practically speaking the regulation of breathing in man dur- 
 ing rest under normal conditions is regulation of the blood re- 
 action. This very important conclusion is the outcome of the 
 present chapter. 
 
 Addendum. Within the limits of the present book it is un- 
 fortunately impossible to deal in detail with the mass of quite 
 recent literature bearing on the regulation of blood alkalinity. 
 Some of this literature is based on assumptions with which, for 
 the reasons already given, I am unable to agree : while other parts 
 of it are concerned with details as to which it seems difficult for 
 the present to form definite judgments. In general, however, it 
 
 K Bennett and Dodds, Brit. Journ. of Exper. Pathol., II, p. 58, 1921. 
 "Higgins, Publication No. 203, Carnegie Institution of Washington, p. 168, 
 
 - 
 
 Higgins, Peabody, and Fitz, Journ. of Med. Research, XXXIV, p. 263, 1916. 
 
206 RESPIRATION 
 
 does not appear to me that anything which has recently been 
 published, points to any important modification of the conclu- 
 sions embodied in this chapter. In view of the great confusion 
 which evidently exists as to the subject, it may, nevertheless, be 
 useful to indicate more explicitly the reasons for regarding the 
 words "acidosis" and "alkalosis" as denoting deviations towards 
 the acid or alkaline side respectively of the normal reaction or 
 hydrogen-ion concentration within the body. 
 
 Acidosis and alkalosis are now frequently regarded as condi- 
 tions in which, whether or not there is an alteration in actual 
 reaction, the "alkaline reserve" of the blood plasma is diminished 
 or increased. This definition originated in a paper by Van Slyke 
 and Cullen in which they pointed out the ease with which varia- 
 tions in the "alkaline reserve," or total capacity of the blood 
 plasma for combining with CO 2 can be determined experimentally, 
 and the advantages of using oxalated blood plasma in place of 
 whole blood for the purpose. 56 Though they stated clearly that 
 variations in alkaline reserve are no direct measure of the varia- 
 tions in actual reaction of the blood, they, very unfortunately as 
 I think, proceeded to define "acidosis" as simply a condition in 
 which the alkaline reserve of the blood is diminished. It is, how- 
 ever, to variations in reaction, and not in the conveniently meas- 
 ured alkaline reserve of the plasma that the body is reacting 
 in conditions of acidosis or alkalosis; and to define acidosis or 
 alkalosis as anything else than a deviation towards the acid or 
 alkaline side of the normal reaction seems to me quite unjustifiable. 
 
 The confusion has been added to by the general failure to 
 realize the extreme delicacy of physiological regulation of re- 
 action, as compared with the comparative roughness of our present 
 means of directly measuring changes in reaction. Thus in cases 
 where there are all the physiological signs of acidosis, the avail- 
 able means of direct measurement may show no sign of the 
 change; and hence it has been quite wrongly assumed that no 
 change exists. This has contributed towards an acceptance of the 
 definition of acidosis as a condition, not of increased hydrogen- 
 ion concentration. within the body, but of diminished alkaline re- 
 serve. The picturesque expression "alkaline reserve" is evidently 
 an unfortunate one in so far as it suggests a reserve of alkali not 
 in actual use. The alkali weakly combined in the body is in reality 
 always in physiological use, and the most urgent symptoms of 
 acidosis appear long before the alkaline reserve disappears. 
 
 "Van Slyke and Cullen, Journ. of Biol. Chem., XXX, p. 289, 1917. 
 
RESPIRATION 
 
 207 
 
 As was shown above, a difference of .012 in the PH of the blood 
 is sufficient to double the resting breathing, or cause apnoea. This 
 difference in PH corresponds to a difference of only about one 
 part by weight of ionized hydrogen in a million million parts of 
 blood. A continued difference of o. I in PH would in all probability 
 cause danger to life. This is a much lower limit than has commonly 
 been assumed. By forced breathing we can, it is true, produce a 
 greater difference in the PH of arterial blood, and maintain this 
 difference for an hour or more without loss of consciousness. The 
 difference, however, applies only to the arterial blood. As will be 
 shown in Chapter X, slowing of the circulation protects the tissues 
 to a large extent from great rises in PH. It is possible, also, that ac- 
 tive secretion of CO 2 by the lungs, as well as quickening of the cir- 
 culation, protects similarly against fall in the PH of the tissues. 
 Nevertheless, as Yandell Henderson has so clearly shown, when 
 efficient forced respiration is kept up in animals for a sufficient 
 time, not only do coma and progressive failure of circulation ensue, 
 but so much damage is done that it is impossible to recover the ani- 
 mal on restoring the PH of the blood by administering CO 2 , just as 
 it is impossible to recover a patient who has suffered for a sufficient 
 time from acute anoxaemia. That progressive and often irrepar- 
 able damage ensues also during a condition of excessive acidosis 
 is suggested by the phenomena of CO 2 poisoning and clinical 
 acidosis. To what extent the damage during alkalosis is due di- 
 rectly to the rise in PH, or to the accompanying anoxaemia, we 
 cannot at present say; and perhaps the question is at bottom 
 merely academic. When the forced breathing is of oxygen instead 
 of air the effects are much less marked, as mentioned above; but 
 this may be because the circulation can be shut down more effec- 
 tively when oxygen is breathed, and that hence the rise in PH in 
 the tissues is diminished. 
 
CHAPTER IX 
 Gas Secretion in the Lungs. 
 
 IN the lungs the blood is separated from the alveolar air by two 
 layers of living tissue, namely the capillary endothelium and the 
 alveolar epithelium. What part in respiratory exchange is played 
 by these very thin layers of living tissue? Is this part purely me- 
 chanical? In other words, do these layers behave towards the 
 respiratory gases as any very thin non-living moist membrane 
 would behave? Or do the living membranes play an active part in 
 the process? We must now face this interesting, but also contro- 
 versial subject. 
 
 There has been a tendency to assume that because these mem- 
 branes are very thin they cannot play any active part. But it is not 
 so long since even membranes consisting of cubical or columnar 
 epithelial cells were supposed only to play a passive part in the 
 separation of material; and the presumption that a thinner mem- 
 brane of flattened cells cannot play an active part has come down to 
 us from the time, about the middle of last century, when physico- 
 chemical theories became dominant in physiology, and secretion in 
 general was supposed to be a mere mechanical process like filtra- 
 tion or diffusion. Another prevalent presumption is that though 
 liquids or dissolved solids may be actively secreted, gases probably 
 pass through living membranes by simple diffusion. 
 
 So little information about gas secretion is usually to be found 
 in physiological text books that it may be useful, before discussing 
 gas secretion by the lungs, to give some account of gas secretion 
 as it is now well known to exist in the swim bladder of fishes. 
 
 The swim bladder is morphologically a diverticulum of the 
 alimentary canal, like the lungs. In some classes of fishes there is 
 an open duct from the swim bladder into the alimentary canal, 
 but in other classes this duct is closed. Quite evidently, the main 
 function of the swim bladder is to make the specific gravity of the 
 fish about equal to that of the water it displaces when the fish is 
 at a certain depth. With a certain amount of gas in its swim 
 bladder the fish will just float at a certain depth. It is, however, 
 in a position of unstable equilibrium : for any movement upwards 
 will cause expansion of the air, so that the fish will tend to rise 
 with increasing velocity towards the surface ; and any movement 
 
RESPIRATION 209 
 
 downwards from the position of equilibrium will similarly tend 
 to make the animal sink with increasing velocity to the bottom. 
 When fishes are stunned by an explosion under water, about half 
 of them float to the top, and the other half sink to the bottom. 
 One has only to place a goldfish in a large and tall bottle of 
 water provided with a perforated cork through which a thick 
 walled tube containing water passes to another small bottle of 
 water, in order to see how the fish deals with the situation. If the 
 pressure in the large bottle is raised by raising the small bottle 
 the fish will at first begin to sink, but will immediately turn its 
 nose upwards and swim upwards, so as to reestablish its position 
 of unstable equilibrium; and conversely if the large bottle be 
 lowered. It was formerly believed that a fish compresses or relaxes 
 its swim bladder when it wishes to go downwards or upwards. 
 That this is not the case was shown by Moreau 1 in a series of 
 beautiful experiments. A fish is really confined temporarily to 
 about a certain depth by its swim bladder ; for if any cause tends 
 to make it leave this depth the animal's response to the stimulus 
 of expansion or contraction of its swim bladder soon brings it 
 back to its proper depth. 
 
 The goldfish has an open duct to its swim bladder, so if the 
 pressure is greatly diminished, as by connecting the large bottle 
 to a filter pump, the air of the swim bladder comes bubbling out 
 of the animal's mouth. If the pressure is now restored to normal 
 the animal sinks to the bottom, and after a few fruitless efforts 
 to swim upwards lies helpless on its side. If it is left there for some 
 time, however, it gradually becomes more buoyant, and after a 
 certain number of hours it will be swimming about as usual, with 
 its swim bladder full of gas. If a fish has a closed swim bladder, 
 and the gas from this is removed by means of a hypodermic 
 syringe, the fish also sinks at first, but soon refills its swim bladder 
 with gas. How is this gas produced, and what is it? It cannot have 
 been swallowed as air, as the fish has been lying in water at the 
 bottom all the time, or has a closed swim bladder. This brings 
 us to gas secretion. 
 
 About the beginning of last century the eminent French physi- 
 cist Biot was engaged in survey work in the Mediterranean, and 
 was attracted by the observation that fishes caught with a line at 
 great depths come to the surface and lie helpless with their swim 
 bladders distended with gas and sometimes projecting out through 
 the mouth. He determined to analyze the gas, and having intro- 
 
 1 Moreau, Memoires de PAysiologie, Paris, 1877. 
 
210 RESPIRATION 
 
 duced some of it, along with excess of hydrogen, into a glass 
 "eudiometer" he passed a spark. Instead of the mild explosion 
 usual in air analyses, there was a violent explosion which broke 
 his instrument. He then knew that he had made a most significant 
 discovery, as the gas he was analyzing must be nearly pure oxy- 
 gen. He got another eudiometer and made a number of analyses of 
 gas from the swim bladder. The results showed that while the 
 gas taken from the swim bladder of a fish near the surface often 
 contained less oxygen than ordinary air, that taken from fishes 
 caught at great depths contained nearly pure oxygen. 2 Biot had 
 discovered oxygen secretion. 
 
 To illustrate the real significance of his observations we may 
 take an analysis made much more recently by Schloesing and 
 Richard, 3 in connection with which the depth from which the 
 fish was taken is definitely stated, and was 4,500 feet. They found 
 that the gas contained 84.6 per cent of oxygen, together with 3.6 
 per cent of CO 2 and n.8 per cent of nitrogen. The latter gases 
 are, however, quite likely to have mostly got in by diffusion during 
 the delay before the sample was taken. Now the pressure at 4,500 
 feet is 136 atmospheres. Therefore the oxygen pressure in the 
 
 Q i fZ 
 
 swim bladder was at least I36x = 115 atmospheres, while 
 
 i oo 
 
 the oxygen pressure in the sea water was only about 21 per cent 
 of an atmosphere, and, in the blood circulating in the capillaries 
 round the swim bladder, certainly very much less. At a moderate 
 estimate the oxygen pressure on the inside of the wall of the 
 swim bladder was at least 1,000 times greater than in the cap- 
 illaries outside. 
 
 In the monograph already referred to, Moreau described a 
 number of experiments showing the conditions under which oxy- 
 gen secretion into the swim bladder occurs. He found, for instance, 
 that if a fish confined in an open cage was sunk to a considerable 
 depth, so that its specific gravity became greater than that of the 
 water, it gradually secreted oxygen so as to restore the balance ; 
 and similarly if its swim bladder had been emptied by puncturing. 
 The simple experiment on the goldfish which I have just described 
 is of the same nature. Moreau even found that if a weight was 
 attached to one fish in an experimental tank, and a float to another 
 fish, so that the first fish was for the time glued to the bottom, and 
 the second to the surface, both fishes would soon be swimming 
 
 a Biot, Memoir es de la Societe d'Arcueil, 1807. 
 3 Comptes rendus, Vol. 122, p. 615, 1896. 
 
RESPIRATION 21 1 
 
 about again quite unconcerned in the tank, their respective swim 
 bladders having, compensated by secretion or absorption of gas 
 for the disturbance in equilibrium caused by the sinker or float. 
 
 Such facts as these pointed to the conclusion that the gas secre- 
 tion is under the control of the nervous system ; but this was not 
 clearly demonstrated by Moreau. It was not till sixteen years 
 later that Bohr showed that the secretion after emptying the 
 swim bladder by puncture ceases after the branch of the vagus 
 supplying the swim bladder is cut. 4 I well remember the interest 
 with which I saw this experiment when Bohr showed it while he 
 was staying with me in Oxford a few months before he published 
 his paper on the subject. Dreser 5 had meanwhile already shown 
 that the secretion of oxygen, like that of saliva, sweat, etc., is 
 excited by the action of pilocarpine. 
 
 It is clear that a fish may require to get rid of gas from its 
 swim bladder, as well as to secrete gas. If the duct is open, there 
 is of course no difficulty in getting rid of gas ; but it is different 
 
 Figure 59. 
 Diagram of arrangement of "oval." 
 
 if the duct is closed. The oxygen might, conceivably, be secreted 
 backwards; but often there is a large percentage of nitrogen in 
 the gas, and there might be trouble about this. It was discovered 
 by Jager 6 that in fishes with a closed swim bladder there is an 
 oval window-like area on the dorsal side of the swim bladder 
 (Figure 59). Over this area there is nothing but a thin layer of 
 flattened cells between the air of the swim bladder and an under- 
 lying layer containing a close network of capillaries. This thin 
 layer seems to permit free diffusion outwards of the gas in the 
 swim bladder. Assuming this to be the case, the oxygen will 
 freely diffuse into the blood capillaries, where, as already seen, 
 
 4 Bohr, Journ. of Physwl., XV, p. 499, 1893. 
 
 6 Dreser, Arch. f. Exper. Pathologie, XXX, p. 160. 
 
 e Jager, Pfliiger's Archiv, XCIV, p. 65, 1903. 
 
212 
 
 RESPIRATION 
 
 the oxygen pressure is very low. Nitrogen and CO 2 , on the other 
 hand, will diffuse inwards if their partial pressure is less inside the 
 swim bladder than in the blood, and outwards in the converse 
 case. The pressure of nitrogen in the blood is doubtless about 79 
 
 Figure 60. 
 
 Section through secreting gland of swim bladder of Sciaena aquila, showing 
 the epithelial body and underlying layer of capillary network (f) with gas bubbles 
 distending the gas ducts of the epithelial body (Jager). 
 
 Figure 61. 
 
 More highly magnified portion of epithelial body shown in 
 Figure 60. A distended gas duct, with surrounding secreting 
 cells (Jager). 
 
RESPIRATION 
 
 213 
 
 per cent of an atmosphere, as it is in sea water ; so whenever the 
 oxygen percentage is sufficiently reduced by diffusion to make 
 the nitrogen pressure in the swim bladder more than 79 per cent 
 of an atmosphere, the nitrogen will follow the oxygen out through 
 the "oval" ; as will the CO 2 , and from a similar cause. But Jager 
 found also that the "oval" can be opened or closed by the relaxa- 
 
 Figure 62. 
 
 (X 330) Folds of the swim bladder epithelium of Gobius niger. 
 C.R.M., capillaries of the rete mirabile. I.C.C., intracellular capil- 
 lary (Woodland). 
 
 tion or contraction of a ring of unstriped muscle surrounding its 
 periphery. When this ring is contracted the "oval" is covered up 
 by a layer of the ordinary lining membrane of the swim bladder. 
 Thus not only secretion, but also absorption of gas from the swim 
 bladder, is under complete physiological control. 
 
214 
 
 RESPIRATION 
 
 On microscopic section of the wall of the swim bladder we find 
 that at most parts it is lined by flattened epithelial cells similar in 
 outward appearance to those covering the oval. At certain parts, 
 however, this flattened epithelium passes into a layer consisting of 
 cubical or columnar epithelial cells, and forming the so-called 
 "epithelial body" (Figures 60, 61), or else a convoluted layer of 
 columnar epithelium (Figure 62). In the glandular structure 
 ducts containing gas may be seen (Figures 60 and 61) in certain 
 species of fishes, and the gland is evidently an oxygen-secreting 
 gland. The true glandular structure was one of Johannes Miiller's 
 many discoveries about glands. 
 
 R.M. 
 
 G.E. 
 
 Figure 63. 
 
 Diagram of circulation in rete mirabile of eel. R.M. rete mirabile. G.E. gland 
 epithelium. Arterioles and arterial capillaries continuous lines. Venules and venous 
 capillaries interrupted lines (Woodland). 
 
 Beneath the glandular structure is a mass of red blood vessels, 
 forming a structure which attracted the attention of anatomists 
 hundreds of years ago 7 and came to be known as a rete mirabile. 
 The arrangement of the blood vessels in this "red body" was re- 
 cently studied by Woodland, 8 who established the fact that the 
 rete mirabile is an arrangement in which the arterioles passing 
 to the gland break up into capillaries which come into intimate 
 contact with corresponding venous capillaries from the venules 
 coming from it (Figure 63). What is the significance of this? 
 The arrangement reminds us of that in a regenerating furnace, 
 where the heat carried away in the waste gases is utilized to heat 
 
 T Redi, Observations sur les animaux vivans contenus dans les animaux vivans. 
 Florence, 1684. 
 
 8 Woodland, Proc. Zool. Soc. of London, p. 183, 1911. 
 
RESPIRATION 
 
 215 
 
 the incoming air. Nevertheless it seems hardly probable that the 
 arrangement is for heat regeneration. The blood passes to the 
 gland with, presumably, the main physiological object of supply- 
 ing oxygen, and venous blood in returning is already spent as 
 regards its supply of oxygen. Nevertheless I think we can now 
 suggest an explanation. It was discovered by Barcroft and King 9 
 that at low temperatures the influence of CO 2 in expelling oxygen 
 from haemoglobin is much greater, relatively speaking, than at the 
 temperature of warm-blooded animals. The difference is so great 
 as to suggest that the dissociation of oxyhaemoglobin in the tis- 
 sues of cold-blooded animals is practically dependent, not on fall 
 
 S.G. 
 
 END 
 
 Figure 64. 
 
 (X 1000). Transverse section through anterior end of rete mirabile of 
 Gobius niger, showing the peculiar endothelium (END) of the arterial 
 capillaries (A) as compared with the venous capillaries (V) (Woodland). 
 
 of oxygen pressure, but on rise of CO 2 pressure. It seems probable, 
 therefore, that the function of the rete mirabile is to enable venous 
 blood to communicate part of its CO 2 to the arterial blood. The 
 effect of this will be to raise the CO 2 pressure of the blood sup- 
 plied to the gland, and so raise the oxygen pressure. There may 
 be active secretion of CO 2 into the arterial capillaries; and this 
 
 9 Barcroft and King, Journ. of Physiol., XXXIX, p. 374, 1909. 
 
2i6 RESPIRATION 
 
 hypothesis is supported by the existence in the arterial capillaries 
 of a very peculiar thickened endothelium figured clearly by Wood- 
 land (Figure 64). 
 
 Another very interesting case of gas secretion occurs in Arcella 
 discoides, which is a microscopic unicellular organism found in 
 rivers and ponds. It has a more or less transparent shell, shaped 
 something like the top of a mushroom, with an opening where the 
 stalk should come. Through this opening it protrudes delicate 
 pseudopodia, by means of which it can creep about (Figure 65). 
 
 Figure 65. 
 
 Arcella raising itself by developing bubbles. Two bubbles 
 visible through shell, and pseudopodia projecting through 
 lower opening. 
 
 When a living and healthy arcella is examined in a drop of water 
 under the microscope, the presence of one or more gas bubbles 
 inside its protoplasm can at times be observed, particularly if by 
 accident or design the animal has been turned on its back, with 
 the opening of its shell upwards. The bubbles of course make the 
 animal lighter, so that it rises towards the surface of the water, 
 and also comes right-side up, after which they rapidly disappear 
 again. The occurrence of these phenomena was described many 
 years ago by Engelmann. Quite recently Dr. Bles took up the 
 subject again at my suggestion, as it looked as if oxygen want 
 was in some indirect way the real stimulus to the formation of the 
 bubbles, just as it is (as we shall presently see) the stimulus to 
 oxygen secretion in the lungs. He elicited the very interesting 
 fact that a quite slight fall in the normal oxygen pressure of the 
 surrounding water is sufficient to cause the immediate formation 
 of gas bubbles in the arcella, and thus cause it to rise to where 
 presumably there is more oxygen. It seems probable, also, from 
 other observations made by him later, that the bubbles which are 
 apt to develop when the animal is placed on its back are a conse- 
 quence of stimuli produced by internal want of oxygen owing to 
 increased oxygen consumption during its efforts to right itself. 
 
RESPIRATION 217 
 
 Before going further let us try to form some sort of conception 
 as to what is occurring in a gland cell during the secretion of 
 oxygen. On the side of the cell next the lumen of the duct we have 
 a pressure of oxygen which may be 1,000 times as great as on the 
 side next the capillaries ; and yet oxygen may be passing inwards 
 from the capillaries towards the duct. The cell is permeable to 
 oxygen : for oxygen is passing through it. Yet the oxygen cannot 
 be free to dissolve in the ordinary way in the "protoplasm" of 
 the cell : for if this were the case the oxygen would run backwards 
 through the cell like water through a sieve. At a pressure of 1 1 5 
 atmospheres, to go back to our concrete example, 100 volumes of 
 water at ioC would take up 430 volumes of oxygen (measured 
 at o and 760 mm.) ; and if the oxygen were as freely soluble in 
 the cell water as in ordinary water the swim bladder would leak 
 outwards at a quite hopeless rate. If we start by looking upon 
 "living protoplasm" as a mere solution and suspension of colloid 
 and other material, we may as well give up the attempt to get any 
 insight whatever into even the most rudimentary physiological 
 processes. 
 
 When we take a broad general view of the phenomena of life, 
 one of the most fundamental facts that appears is that the com- 
 position of each organism or part of an organism is distinctly 
 specific. The percentage and nature of each of the substances 
 which we can recover on disintegrating the living tissue are spe- 
 cific ; and the more we learn about the nature of these substances 
 the more clearly does this specific character emerge. It is evidently 
 no mere accident that muscle yields so much potassium, so much 
 phosphoric acid, so much water, and so much of various proteins. 
 These substances must be present in some kind of combination in 
 the living "substance" ; and if so the living substance cannot be 
 regarded as a mere solution of free molecules. The molecules are 
 in some sense bound, as they are in a solid ; and in so far as this is 
 the case the living substance must in certain respects behave as a 
 solid, impervious to the free passage of material by diffusion. 
 The layer of thin flattened epithelium lining appears to be gas- 
 tight (to oxygen at least) except where it covers the oval. At this 
 point the layer allows gas to pass freely. 
 
 From this point of view we can understand why the living cells 
 of the oxygen-secreting gland should be gas-tight, or nearly so, 
 against diffusion backwards, but we have not yet considered how 
 the gas passes forward through them during secretion; and if 
 
218 RESPIRATION 
 
 "living material" behaved like an ordinary solid no such explana- 
 tion would be forthcoming. But evidently a living cell does not 
 behave like an ordinary solid : for it is constantly taking up and 
 giving off material, not merely during secretion, but at every 
 moment of its existence. This is evident from a general considera- 
 tion of the phenomena of nutrition, and becomes still more evident 
 if by altering the environment of a cell we disturb the labile 
 balance between living cells and their surrounding liquids. In the 
 secretion of oxygen and many other substances, such as urea, 
 sugar, salts, etc., the substance taken up on one side of the cell is 
 given off in the same form on the other side. In the processes of 
 ordinary nutrition, on the other hand, the taking up and giving 
 off may be on the same side of the cell, and the substance given off 
 may be in a different chemical form from that taken up. We have 
 no reason to believe, however, that there is any fundamental dis- 
 tinction between the taking up and giving off during ordinary 
 nutrition and during secretion. Nearly a century ago Johannes 
 Miiller, at the end of his famous memoir on secreting glands, 10 
 after pointing out that his observations negatived the mechanical 
 theories of secretion then current, suggested that secretion must 
 be regarded as a process akin to growth, the only difference being 
 that whereas in ordinary growth the material deposited tends to 
 remain where it is, in secretion it is always being carried away 
 and replaced. Johannes Miiller's theory was bound up with his 
 vitalistic physiology, and the clue which he was grasping at was 
 swept from the hands of physiologists by the wave of mechanistic 
 speculation which passed over physiology about the middle of 
 last century. But now that we know from nearly a century of 
 painful experimental investigation what to the genius of a great 
 biologist like Miiller was evident enough, that mechanical theories 
 of secretion are impossible, we can take up the clue again. 
 
 When oxygen (or indeed any other substance entering into 
 cell metabolism) is taken up on one side of the cell, we are led by 
 the experimental facts to assume that the oxygen enters into 
 easily dissociable chemical combination. Were this combination 
 not easily dissociable we could not understand why a cell should 
 be so enormously sensitive, as we shall see later that it is, to 
 changes in the concentration of oxygen and other substances in 
 its immediate environment. Now all we know about cell metab- 
 olism points to the conclusion that the balance of stability at any 
 one part of the cell depends on the balance of stability at other 
 
 "Johannes Miiller, De Glandularum Secernentium Structura Penitiori, 1830. 
 
RESPIRATION 219 
 
 parts. The taking up of oxygen, for instance, depends on a host 
 of conditions in the environment, such as the concentrations, or, 
 more correctly, the diffusion pressures, of ions of different sorts, 
 and of various other substances which are, or may be, passing 
 into and out of the cell. A minute trace of pilocarpine, for instance, 
 will set the oxygen-secreting cell violently taking up oxygen on 
 one side, and giving it off on the other; and probably we could 
 paralyze the oxygen secretion at once by reducing the concentra- 
 tion of calcium ions in the cell environment. 
 
 In a secreting cell the rate of secretion, other conditions being 
 favorable, depends on the concentration of the dissolved material 
 to be secreted. This we can see with the utmost clearness in the 
 case of the kidney or intestinal epithelium. The rate of secretion 
 also depends on the concentration of the dissolved material on 
 the excretory side, as we can also see in the case of the kidney. 
 Clear evidence on. this point is summarized by Ambard in his 
 book La physiologie des reins, Paris, 1920. We are thus led 
 to the conclusion that the stability of the oxygen combination 
 on one side of the oxygen-secreting cell depends, other things 
 being equal, on the stability of the oxygen combination at the 
 other side, and that in proportion as the oxygen combination 
 at one surface becomes increased, the oxygen combination at the 
 opposite surface becomes more ready to release oxygen towards 
 the cell environment. It also seems probable that as we proceed 
 from the absorbing to the secreting side of the cell, the tendency 
 to give off oxygen becomes greater and greater. A cell of sub- 
 stantial thickness is therefore required to produce a large differ- 
 ence in oxygen pressure. The combination which dissociates itself 
 on the excretory surface will, if the concentration of oxygen at 
 that surface is not so high as to stop the dissociation, be constantly 
 resatu rating itself in part from the combination lying deeper in 
 the cell. Thus oxygen will travel from the absorbing to the se- 
 creting side of the gland cell, just as urea, or sodium, or phosphoric 
 acid, will travel from the absorbing to the secreting side of other 
 kinds of secreting cells. We can also imagine how, in the course 
 of their passage, chemical transformations may occur in the 
 transported material, so that, for instance, an intestinal cell which 
 takes up fatty acid may deliver fat on the other side, or a cell 
 which takes up sugar may transform it into fat, or amino acids 
 into proteins, or oxygen into CO 2 and water, or may perform any 
 of the numerous other syntheses or disintegrations with which 
 physiologists are familiar. 
 
220 RESPIRATION 
 
 In the arcella, bubbles, probably consisting largely of oxygen, 
 appear and disappear within the cell body, according to the ex- 
 isting physiological conditions. It seems probable that the bubbles, 
 for the development of which a high internal oxygen pressure 
 will be needed, occur in interstices of the living substance, due to 
 the presence of inclosed liquid or solid substances. In these inter- 
 stices the gas pressure can rise up to the point at which it pro- 
 duces disruption and bubble formation. Gas bubbles have not 
 hitherto been observed in the cells of oxygen-secreting glands, 
 although certain microscopic appearances have been taken for 
 such bubbles. 
 
 The well-known transparent larva of Corethra possesses two 
 gas floats : one near the anterior, and the other near the posterior 
 end of the larva. The gas is enclosed in chitinous bladders de- 
 veloped from the tracheal system and partially rigid, with cells 
 on their external walls. If the pressure of the water is increased 
 the larva begins to sink owing to diminution in the capacity of 
 the bladders, but regains its equilibrium in two or three minutes; 
 and conversely if the pressure is diminished. This looks, therefore, 
 like a case of gas secretion. Krogh showed, however, in a beautiful 
 series of experiments 10A that there is no gas secretion, but secretion 
 of liquid out of or into the bladders, so as to compensate for the 
 alteration in their capacity. The larva can equilibrate itself in 
 this way since the bladders are partially rigid. In deep water, for 
 instance, the gas pressure is kept the same as that of the atmos- 
 phere, and hence much less than that of the surrounding water. 
 The gas pressures inside and outside the bladders are thus the 
 same, and simple diffusion of gases is not modified by gas secretion. 
 
 Having to some extent cleared our ideas by the consideration 
 of undoubted cases of gas secretion, we can now proceed to dis- 
 cuss the evidence as to gas secretion by the lungs. As mentioned 
 already, Ludwig had the idea (in which he was right) that prob- 
 ably something occurs in the lungs to facilitate the escape of CO 2 , 
 and possibly the absorption of oxygen ; and this idea appeared in 
 the work of some of his pupils. It was a time when physiological 
 research was very active in Germany; and friendly, or some- 
 times anything but friendly, shots were often exchanged between 
 the leading laboratories. The Leipzig idea was accordingly put 
 to the test by Pfliiger and his pupils at Bonn, and for the purpose 
 Pfliiger devised an instrument which he called the aerotonometer, 
 its object being to measure the partial pressures or tensions of the 
 
 10A Krogh, Skand. Archiv. /. Physiol., XXV. p. 183, 
 
RESPIRATION 221 
 
 gases contained in venous and arterial blood, so that these pres- 
 sures could be compared with one another and with the corre- 
 sponding pressures in the air of the lungs. The aerotonometer 
 consisted of two tubes immersed in a water bath at body tempera- 
 ture, and closed below by a mercury seal. In one tube was placed 
 a mixture containing a smaller percentage of CO 2 and greater 
 percentage of oxygen than corresponded to the partial pressures 
 expected in the blood ; and in the other tube a mixture containing 
 a higher percentage of CO 2 and a lower percentage of oxygen. 
 The blood from the animal was then allowed to trickle down the 
 inside of the tubes, so that it should as far as possible equalize its 
 gas tensions with those in the tubes, either by taking up or giving 
 off CO 2 or oxygen. In a successful experiment the blood gave off 
 CO 2 and absorbed oxygen in one tube, and vice versa in the other, 
 so that the gas pressures of the blood were defined within narrow 
 limits on the analyses of the gases in the two tubes. The sample of 
 lung air was obtained by another ingenious instrument, the "lung 
 catheter," by means of which a bronchus could be blocked off and 
 a sample of the gas in the lungs drawn off as soon as the air thus 
 confined had reached a constant composition. 
 
 The conclusion drawn from the actual experiments by Pfliiger 
 and his pupils was that there was no average difference in gas 
 pressures between the venous blood and the air inclosed beyond 
 the blocked bronchus; and therefore no evidence of any giving 
 off of CO 2 or absorption of oxygen except by simple diffusion. 11 
 
 The question was taken up again by the late Professor Bohr of 
 Copenhagen, one of Ludwig's pupils. 12 Bohr improved the aeroto- 
 nometer, so that a large stream of arterial blood could be run 
 through it and back to the animal, the blood of which had first 
 been rendered incoagulable by injecting peptone or leech extract. 
 He obtained the result that while usually the CO 2 pressure in the 
 arterial blood is not less than in the alveolar air, and the oxygen 
 pressure not greater, yet sometimes this relation is reversed. From 
 these results he concluded that active secretion of oxygen from 
 the lung air into the blood, and of CO 2 from the blood into the 
 lung air, may both occur. Owing to the many possibilities of 
 error the results were not very convincing, however; and Fred- 
 ericq 13 of Liege soon afterwards made a further series of experi- 
 
 11 Pfli'tger's Archiv, IV, p. 465 ; VI, p. 65 ; VII, p. 23, 1871-1873. 
 
 12 Bohr, Skand. Arch, of Physiol., p. 236, 1891. 
 18 Fredericq, Arch, de Biol., XIV, p. 105, 1896. 
 
222 
 
 RESPIRATION 
 
 ments, all of which seemed to tell in favor of Pfliiger's interpreta- 
 tion. 
 
 About fifteen years later the aerotonometer was greatly im- 
 proved by Krogh, who was then Bohr's assistant. He very greatly 
 diminished the volume of air exposed to the blood in the aeroto- 
 nometer, thus rendering it far quicker in its action ; and ultimately 
 he succeeded in working with a single bubble of air, round which a 
 stream of blood could play, the bubble being afterwards analyzed 
 with the help of a graduated capillary tube into which it could be 
 sucked up and measured before and after its CO 2 and oxygen 
 had been removed by suitable reagents. 
 
 Figure 66. 
 
 Krogh's micro-aerotonometer, showing inlet and outlet 
 for blood, lower part of measuring tube, and air bubble. 
 
 Before his death Bohr published some experiments made with 
 Krogh's aerotonometer, and apparently showing distinctly that 
 the pressure of CO 2 in the venous blood could be less than in the 
 expired air, although CO 2 was being given off in the lungs ; and 
 that the arterial CO 2 pressure could also be less than that of the 
 expired air. Krogh himself, however, took the view that there 
 were errors in these experiments, and published, along with M. 
 Krogh, the results of a careful series of experiments on animals 
 under conditions which were much more nearly normal than in 
 
RESPIRATION 223 
 
 any previous experiments. 14 The arterial oxygen pressures were 
 always very distinctly below the oxygen pressures at the same 
 time in the alveolar air; while the arterial CO 2 pressures were 
 sensibly equal to those in the alveolar air. There was never any 
 approach to excess of arterial over alveolar oxygen pressure, or 
 of alveolar over arterial CO 2 pressure, even when these pressures 
 were varied considerably by altering the composition of the in- 
 spired air. Krogh, therefore, rejected Bohr's conclusions that 
 there is active secretion of oxygen or CO 2 in the lungs, and con- 
 cluded in favor of Pfliiger's view that the exchange of gases in 
 the lungs is entirely due to diffusion. The following table shows 
 the results of a typical experiment in which the alveolar oxygen 
 pressure was varied during the experiment, the alveolar air and 
 blood samples being taken nearly simultaneously. 
 
 TIME 
 
 TENSION OF CO 2 IN 
 
 TENSION OF 
 
 OXYGEN IN 
 
 
 Alveoli 
 
 Blood 
 
 Alveoli 
 
 Blood 
 
 1.36-43 
 
 3-6 
 
 3.7 
 
 12. 
 
 IO.O 
 
 2. 10-12 
 
 3.o 
 
 3-5 
 
 18.0 
 
 15.0 
 
 3-03- 3-07 
 
 2-5 
 
 2-5 
 
 12.0 
 
 ii-5 
 
 Before following this long controversy further, I should like 
 to point out a fallacy in the interpretation of the aerotonometer 
 results. The conclusion of Pfliiger that diffusion alone explains 
 the giving off of CO 2 in the lungs was wholly fallacious, as has 
 already been shown in Chapter V. The oxygen reaching the 
 blood in the lungs helps to drive out CO 2 ,* and under certain con- 
 ditions which are very apt to occur during physiological experi- 
 ments on animals, and may easily be produced in man, the venous 
 CO 2 pressure may be lower than that of the alveolar air, although 
 no secretion at all may be occurring. In the lung-catheter experi- 
 ments the oxygen supply to the lungs was blocked off, so that the 
 blood could not take up oxygen. As a consequence the CO 2 pres- 
 sure in the confined air must have been considerably lower than if 
 oxygen had been present. In reality Ludwig was right, and 
 Pfliiger was wrong. This source of fallacy does not in any way 
 invalidate Krogh's conclusion that the arterial CO 2 pressure is 
 not, under normal conditions, lower than the alveolar CO 2 pres- 
 sure. I think this conclusion is correct ; and it agrees, as he points 
 out, with all the indications given by the work of Priestley and 
 
 14 A. and M. Krogh, Skond. Arch. f. Physwl., XXXII, p. 179, I9 io. 
 
224 RESPIRATION 
 
 myself on the regulation of breathing in accordance with the 
 alveolar CO 2 pressure. 
 
 When Bohr's original experiments on the question of secretion 
 by the lungs were published in 1891, I was just beginning the 
 serious study of mine gases and the physiological effects of vitiated 
 air; and his results interested me greatly. A year or two later 
 Lorrain Smith and I made a visit of several weeks to Copenhagen, 
 and carried out some research work in the laboratory under 
 Bohr's direction, thus learning a great deal which we could not 
 have learned in England about existing methods of blood-gas 
 investigation, and, far more important, getting into personal 
 touch with Bohr himself. I should like to take this opportunity of 
 saying how much we, and indirectly other physiologists in Great 
 Britain and America, have owed to Bohr and the Copenhagen 
 School of physiologists. 
 
 The difficulties of the aerotonometer method of determining the 
 oxygen pressure of arterial blood were very evident, and I cast 
 about in my mind for some better method. Soon afterwards I 
 began investigating the action of carbon monoxide in mines, and 
 the results of this investigation, and the colorimetric method of 
 blood examination, which I worked out during the investigation, 
 suggested a new means of attacking the problem which Ludwig 
 had originally suggested. 
 
 The general principle of this method has already been ex- 
 plained in Chapter IV, and depends on the fact that within wide 
 limits the relative proportions in which haemoglobin is shared 
 between oxygen and CO are proportional to the relative partial 
 pressures of the two gases when allowance is made for their rela- 
 tive affinities for the haemoglobin. Hence if the proportions in 
 which oxygen and CO are shared in the haemoglobin of the 
 blood when equilibrium is established are known, as well as the 
 pressure of CO, the pressure of oxygen can be calculated. To 
 measure the oxygen pressure in the arterial blood it is therefore 
 only necessary to allow a man or animal to breathe a constant small 
 percentage of CO until absorption of CO stops, owing to a balance 
 having been struck between oxygen pressure and CO pressure in 
 the blood passing through the lung alveoli. The percentage satu- 
 ration of the haemoglobin with CO is then determined, and the 
 arterial oxygen calculated from a knowledge of the relative affini- 
 ties of the two gases for haemoglobin, as determined outside the 
 body. 
 
 The method seemed simple in principle, but it turned out to 
 
RESPIRATION 225 
 
 be as full of pitfalls in practice as the use of the blood pump, 
 aerotonometer, or spectrophotometer. What misled us most were : 
 (i) the assumption that Hiifner's oxyhaemoglobin dissociation 
 curve, then and for many years later quoted in every textbook, 
 was at least approximately correct; (2) the assumption that all 
 haemoglobin is alike as regards its relative affinities for oxygen 
 and CO; (3) ignorance at first of the powerful action of bright 
 light on the dissociation of CO haemoglobin, and of the influence 
 of temperature; (4) failure at first to realize how long it takes 
 to saturate blood or blood solution outside the body with air con- 
 taining low percentages of CO. There were probably also some 
 errors in the colorimetric titrations, owing chiefly to our not taking 
 precautions which subsequent experience showed to be necessary, 
 against decomposition of blood solutions during long experiments. 
 The first experiments were made by Lorrain Smith and my- 
 self 15 on men, the subject of the experiment going through the 
 lengthy process of breathing air containing a definite small per- 
 centage of CO, until absorption of CO ceased, as shown by the 
 analyses of blood samples. The results led us to the conclusion 
 that the normal resting arterial oxygen pressure was considerably 
 above that of the alveolar air; and corrections, made afterwards 
 for the causes of error just referred to caused this conclusion to 
 stand out still more clearly. Subsequent experience leads me to 
 the conclusion that we had become acclimatized more or less to 
 want of oxygen by frequently breathing CO, so that at the time 
 we were no longer ordinary normal subjects. We were at any 
 rate breathing with complete impunity a percentage of CO which 
 would under ordinary circumstances cause very unpleasant symp- 
 toms. On trying the next year and once or twice subsequently to 
 repeat one of the experiments, we were surprised to find that the 
 former percentages were too high for us, and we suspected that 
 there must have been some error about the percentages breathed 
 in the first series of experiments. On reconsidering the matter I 
 cannot see how there could have been an error about the per- 
 centages breathed. It now seems practically certain that we had 
 become acclimatized, and had consequently developed during 
 the experiments a considerably higher arterial oxygen pressure 
 than normal persons would have had, or than we ourselves would 
 have had, if we had not absorbed so much carbon monoxide as in 
 the experiments, and thus become somewhat short of oxygen. 
 
 "Haldane and Lorrain Smith, Journ. of Phystol., XX, p. 497, 1896. 
 
226 RESPIRATION 
 
 Our next experiments 16 were on various small animals 
 chiefly mice. Small animals are specially convenient, as their 
 blood becomes saturated within a few minutes to its maximum 
 extent for any percentage of CO in the air. These experiments 
 again gave an apparently higher oxygen pressure in the arterial 
 blood than in the alveolar air. When the percentage of CO was in- 
 creased, so that the animals began to show symptoms of consid- 
 erable oxygen want, the difference between arterial and alveolar 
 oxygen pressure became much greater. On the other hand, when 
 the animals were breathing a mixture of oxygen and CO there 
 was still a large apparent excess of arterial over alveolar oxygen 
 pressure. This result was a great disappointment to us, as we had 
 hoped that when oxygen was breathed, active secretion of oxygen 
 inwards would cease. The fact that it apparently did not do so 
 ought to have aroused our suspicions of the correctness of the 
 measurements. The phenomena observed when the oxygen per- 
 centage, or the barometric pressure, was diminished, led us, apart 
 from the measurements, to conclude that secretion of oxygen in- 
 wards became more active; but in our measurements of oxygen 
 pressure we were depending on the substantial correctness of 
 Hiifner's dissociation curve; and when this curve was subse- 
 quently found to be totally incorrect our measurements had also 
 to be abandoned as incorrect. 
 
 During the next few years knowledge as regards the dissocia- 
 tion of haemoglobin had greatly increased, thanks to the work of 
 Bohr, Zuntz and Loewy, Barcroft, and others, as well as our own 
 work, as described in Chapter IV. Douglas and I now took up the 
 old subject again, but with far more complete knowledge of the 
 material we were dealing with. 17 Dr. Krogh had also kindly in- 
 formed me in a letter of some experiments he had made (subse- 
 quently published) 18 showing that in the blood of a rabbit the 
 relative affinities for haemoglobin of oxygen and CO were dif- 
 ferent from those in the ox ; and we found, as already mentioned 
 in Chapter IV, that this is not only so for different classes of ani- 
 mals, but also, and in a most marked degree, for different indi- 
 viduals of the same species. 
 
 We therefore had to modify the method. Each animal was ex- 
 posed for a sufficient time to a definite percentage of CO in a 
 bottle, and then drowned in situ. Some of its blood was then 
 
 19 Haldane and Lorrain Smith, Journ. of Physiol., XXII, p. 231, 1897. 
 " Douglas and Haldane, Journ. of Physiol., XLIV, p. 305, 1912. 
 "Krogh, Skand. Arch. /. Physiol., XXXII, p. 255, 1910. 
 
RESPIRATION 
 
 227 
 
 placed, undiluted and at body temperature, in the saturator, and 
 thoroughly saturated in presence of some of the same mixture of 
 air and CO that the animal had been breathing. The percentage 
 saturations with CO of the haemoglobin in the blood taken 
 straight from the animal, and in that from the saturator, were 
 then determined, and the arterial oxygen pressure calculated in 
 the usual way. The following table shows the results. 
 
 
 
 Duration 
 
 Percentage saturation Arterial oxygen 
 
 
 
 of exp. in 
 
 of haemoglobin 
 
 pressure in per- 
 
 Animal used 
 
 Percentage 
 
 minutes 
 
 
 with CO 
 
 centage of the ex- 
 
 
 of CO 
 
 
 s~^ 
 
 *^^ -* 
 
 ^ existing 
 
 
 
 
 In vivo 
 
 In vitro atmosphere* 
 
 Mouse 
 
 .Ol6 
 
 60 
 
 26.2 
 
 17.2 
 
 12.2 
 
 
 
 .0165 
 
 50 
 
 26.7 
 
 19-5 
 
 13.9 
 
 
 
 .018 
 
 45 
 
 26.0 
 
 I8. 5 
 
 13-5 
 
 
 
 .019 
 
 33 
 
 19.7 
 
 12-5 
 
 I2.I 
 
 
 
 .025 
 
 43 
 
 25.6 
 
 I 7 .6 
 
 13-0 
 
 
 
 .046 
 
 40 
 
 29.1 
 
 22.7 
 
 15.0 
 
 
 
 053 
 
 40 
 
 37-7 
 
 30.2 
 
 16.2 
 
 M 
 
 .IOO 
 
 32 
 
 45-0 
 
 43-0 
 
 19-3 
 
 
 
 .129 
 
 3i 
 
 56.4 
 
 56.3 
 
 20.8 
 
 
 
 .198 
 
 
 
 57-6 
 
 56.5 
 
 2O.O 
 
 9) 
 
 .213 
 
 13 
 
 59-1 
 
 75-5 
 
 44-Tt 
 
 
 
 .244 
 
 12 
 
 67.3 
 
 71.7 
 
 25-7t 
 
 
 
 255 
 
 60 
 
 60.1 
 
 62.8 
 
 23-3 
 
 
 
 .260 
 
 25 
 
 67.0 
 
 64.7 
 
 18.9 
 
 
 
 .262 
 
 20 
 
 66.4 
 
 73-7 
 
 28.2 
 
 H 
 
 275 
 
 25 
 
 66.5 
 
 76.9 
 
 35-9 
 
 Rabbit 
 
 .029 
 
 140 
 
 28.0 
 
 18.7 
 
 12.4 
 
 Same rabbit 
 
 .191 
 
 ISO 
 
 58.2 
 
 56.0 
 
 19.1 
 
 * Calculated without reduction for 
 
 aqueous 
 
 vapor in the 
 
 alveolar air. 
 
 f Mouse died. 
 
 On looking down this table it will be seen that as long as the 
 percentage of CO did not exceed about .03 per cent, or the per- 
 centage saturation of the blood did not go over about 28 per cent, 
 the arterial oxygen pressure was only about that of the alveolar 
 air, assuming that the alveolar air of a mouse has about the same 
 composition as human alveolar air. But as the percentage of CO 
 in the air, or the percentage saturation of the blood, rose, the 
 arterial oxygen pressure rose, first to about that of the inspired 
 air, and then, in most cases, far above it sometimes to double. 
 
228 RESPIRATION 
 
 We then repeated the old experiments with oxygen which had 
 disappointed Lorrain Smith and me so much. The results were 
 as follows: 
 
 EXPERIMENTS WITH MIXTURES OF OXYGEN AND CO ON MICE 
 
 Percentage saturation Oxygen pressure in percentage 
 Duration of haemoglobin with CO of the existing atmosphere* 
 
 Percentage of exp. in ^ " _ ^~ f 
 
 
 * 
 
 of CO minutes In vivo In vitro Arterial blood. 
 
 Inspired air 
 
 0.16 30 31.3 29.6 
 
 77-4 
 
 83.9 
 
 0.61 30 57.0 54.6 
 
 66.6 
 
 73-5 
 
 I.I5 30 71.4 70.8 
 
 83.1 
 
 85.6 
 
 1.47 30 69.0 75.0 
 
 96.3 
 
 71-5 
 
 * Calculated without reduction for aqueous vapor. 
 
 
 
 It will be seen that as long as the saturation of the blood with 
 CO did not exceed about 60 per cent, the arterial oxygen pressure 
 was about 7 per cent below that of the inspired air, just as the 
 alveolar oxygen pressure would be. With over 60 per cent satura- 
 tion, however, the animals began to suffer from oxygen want, 
 and the arterial oxygen pressure went just as high above that of 
 the inspired air as in animals breathing ordinary atmospheric 
 air. The old experiments were wrongly calculated, because the 
 relative affinities of haemoglobin for oxygen and CO are on an 
 average different in mouse blood from what they are in human 
 blood or in the ox blood which we then took as a fixed standard. 
 This led us to calculate the arterial oxygen pressure about 50 per 
 cent too high in both the ''normal" and the oxygen experiments. 
 Moreover the "normal" experiments were not normal, since the 
 percentage saturations of the blood were about 40 per cent, and 
 therefore too high to give normal results such as those of the first 
 five experiments in the previous table. If one recalculates the 
 average results of the old experiments in the light of this new 
 knowledge they give just the same result as the new experiments. 
 
 The general, and absolutely sharp and definite, result of these 
 experiments is that with very low percentages of CO there was 
 no evidence of active secretion of oxygen inwards, but that with 
 higher percentages of CO there was perfectly clear evidence of 
 active secretion. This active secretion began to show itself as soon 
 as the CO percentage was sufficient to cause symptoms of CO 
 poisoning, which symptoms, as shown in Chapter VII, are simply 
 
RESPIRATION 229 
 
 those of oxygen want : moreover the secretion did not appear if 
 oxygen was breathed along with the CO, until a much higher 
 saturation of the blood with CO was reached. Pure oxygen, as 
 already shown in Chapter VII, provides a certain supply of dis- 
 solved oxygen to the blood independently of the oxygen carried by 
 the haemoglobin, and thus prevents, to a large extent, the oxygen 
 want which would otherwise be caused by the CO. 
 
 Now the oxygen want is in the tissues, and not in the lungs. 
 Hence the stimulus to secretion originates in the tissues. This 
 stimulus is almost certainly something carried by the blood from 
 the oxygen-starved tissues to the lungs or central nervous system. 
 One might perhaps suppose that whenever the respiratory center 
 is excited, nervous impulses pass down secretory fibers in the vagus 
 nerve and excite secretion in the lungs. Lorrain Smith and I tested 
 this hypothesis, and found that when the respiratory center was 
 excited by excess of CO 2 there was not the slightest rise in the 
 arterial oxygen pressure. Hence the secretion has no direct con- 
 nection with the ordinary activity of the center in producing 
 respiratory movements; and the stimulus to secretion is not a 
 hydrogen ion stimulus. 
 
 We also made a series of determinations on man. In view of 
 the results of the mouse experiments we were anxious to work 
 with low percentages of CO ; but if we had used the old method 
 which Lorrain Smith and I had employed, it would have taken 
 so long before equilibrium was reached between the CO in the 
 air and that in the blood that our experiment could hardly have 
 been completed during winter daylight. We therefore adopted 
 the course of quickly absorbing as much CO as would saturate 
 the blood to the desired extent, and then breathing in and out of 
 a small air space, in which the oxygen and CO 2 percentage was 
 kept constant. Under these conditions CO must, of course, be 
 given off into the air of the space, and as this air is breathed again 
 and again, equilibrium between the CO in the air and that in the 
 blood must establish itself very quickly. The method finally 
 adopted was as follows (see Figure 67). 
 
 The subject, wearing a nose clip, breathes through the mouth- 
 piece A, inhaling through the inspiratory valve B, and expiring 
 through the valve C. The expired air passes through a rubber 
 pipe of large caliber to the tin vessel D, which is filled with small 
 fragments of solid caustic soda, and is made of such a size (di- 
 ameter 23 cms., depth 12 cms.) that the whole of the carbonic 
 acid in the expired air is effectively removed. Another rubber 
 
230 
 
 RESPIRATION 
 
 pipe leads the outgoing air current from D to the bottle E of 12 
 liters capacity, which is connected by another pipe with the in- 
 spiratory valve B. The entrance and exit pipes of E are so ar- 
 ranged that the incoming air current is directed to the bottom of 
 the bottle, while the subject inhales air from the top. The arrows 
 
 r "= 
 
 Figure 67. 
 Apparatus for determining the arterial oxygen pressure in man. 
 
 indicate the direction of the air current caused by the subject's 
 respiration in the main circuit. Two side pipes lead into the rubber 
 pipe connecting D with E. One of these, G, is of large bore and 
 short, and is connected with a vulcanized rubber gas bag of con- 
 siderable size, such as is utilized on Clover's ether apparatus. 
 This bag serves only to accommodate each expiration, as the rest 
 of the apparatus is indistensible, and at the end of inspiration 
 the bag collapses entirely. The other side pipe F serves for the 
 admission of oxygen. The oxygen supply is so arranged that oxy- 
 gen enters the main air circuit automatically to fill up the defi- 
 ciency caused by the absorption of oxygen by the subject at each 
 breath. It is essential in a closed system of small size that the 
 oxygen supplied shall be pure; the small amount of nitrogen 
 contained in ordinary cylinder oxygen renders its use inadmis- 
 sible. We therefore in all the later experiments used oxygen 
 made by the action of water on "oxylith" in the generator H. The 
 current of oxygen is controlled by the tap at the top of the gen- 
 erator, and passes along a pipe past a blow-off valve to air J, 
 through a small gas meter K and thence through a water valve 
 
RESPIRATION 231 
 
 M to enter the main air circuit at F. The height of the water above 
 the orifice of the pipe in M is about 2 mm. greater than in J, and 
 the oxygen therefore passes out to air through the valve J unless 
 a slight negative pressure is set up in the main air circuit, when 
 it will pass by preference through M. Such a negative pressure 
 obtains in the main air circuit only at the end of an inspiration, 
 and depends upon the fact that the whole volume of air in the cir- 
 cuit is diminished by the amount of oxygen absorbed at the last 
 breath of the subject, as the carbonic acid expired is removed. The 
 meter records, therefore, the actual oxygen consumption by the 
 individual. Interposed between the meter and the valve M is a 
 small rubber bag L, such as is used in a small sized football. This 
 serves as a reservoir for the oxygen, and enables a free and sudden 
 supply to be drawn into the air circuit. Without this it would be 
 necessary to run the oxygen from the generator at an excessive 
 and wasteful rate, and the slight resistance of the meter might 
 be felt. In practice the oxygen supply is so adjusted that it is just 
 escaping continuously to air through J, so as to insure that the bag 
 L is filled to constant pressure; otherwise the readings of the 
 meter will not accurately represent the oxygen consumption. 
 
 A Haldane gas analysis apparatus N is attached directly to the 
 air pipe leading from the bottle E to the inspiratory valve, so that 
 samples of the inspired air may be withdrawn at intervals during 
 the experiment for analysis. The extremity of a vacuous gas 
 sampling tube O is inserted into the pipe between the expiratory 
 valve and the caustic soda tin, not far from the former, for the 
 purpose of obtaining a sample of alveolar air by Haldane and 
 Priestley's method. By means of the tap P, connected with the 
 laboratory water supply, a large volume of air can be displaced 
 from the bottle E through the pipe R, and used for filling satu- 
 rating vessels, etc. Before each experiment the apparatus is tested 
 for air-tightness by disconnecting the oxygen supply pipe at F 
 and substituting a water manometer for it, and then producing a 
 positive or negative pressure by blowing in air or sucking it out 
 through the mouthpiece. The whole apparatus is readily blown 
 out with fresh air by disconnecting the return air pipe from the in- 
 spiratory valve and blowing through the mouthpiece with a pair 
 of bellows. 
 
 We found that the percentage of oxygen in the air in the ap- 
 paratus falls by about 0.8 per cent during the first five minutes of 
 an experiment, doubtless owing to the rise of temperature caused 
 by the breathing, which will hinder the entrance of oxygen. After 
 
232 RESPIRATION 
 
 this the oxygen percentage shows oscillations, which however do 
 not exceed I per cent. Such oscillations are unavoidable, seeing 
 that the oxygen supply must be influenced in this method by the 
 depth of the individual breaths : the percentage could only re- 
 main absolutely constant if the depth of breathing was itself 
 constant. For the same reason the oxygen consumption should not 
 be determined over a shorter period than five minutes. 
 
 One great advantage of this apparatus is that it is very easy 
 to subject oneself to atmospheres containing different percentages 
 of oxygen by means of it. To obtain an atmosphere poor in oxygen 
 all that is necessary is to uncouple the oxygen supply from the 
 valve M and breathe into the apparatus. Air now enters through 
 F instead of oxygen, and breathing is continued until analysis of 
 the inspired air shows that the required degree of oxygen de- 
 ficiency has been produced. If the oxygen supply is now reestab- 
 lished the artificial atmosphere produced will remain constant. 
 To obtain an atmosphere rich in oxygen, the gas may be blown in 
 through the orifice for the alveolar air sampling tube, leaving the 
 mouthpiece free for the escape of the displaced air from the re- 
 turn air pipe. 
 
 The total volume of the air in our apparatus is about 15 liters, 
 and we may therefore presume that the whole of it goes through 
 the alveoli of a resting adult subject in three minutes. We have 
 on a number of occasions breathed into the apparatus for an 
 hour with the greatest comfort, the percentage of oxygen mean- 
 while varying only within the limits mentioned above. 
 
 The time during which the subject breathed into the respiration 
 apparatus in our experiments has varied on different occasions 
 from twenty minutes to one hour. So far as we could ascertain the 
 shorter time was sufficient to establish equilibrium of concentra- 
 tion of the carbon monoxide in the blood and in the air breathed, 
 though we have as a rule adopted a period in excess of this" as a 
 matter of precaution. In our earlier experiments we passed about 
 2 cc. of CO into the air in the respiration apparatus before begin- 
 ning to breathe into it, in order that the percentage of this gas 
 present at the start might approximate to its final value. As this 
 procedure had no influence on the result of the experiment we 
 gave it up, and the respiration apparatus thereafter always con- 
 tained air free from CO at the commencement of the experiment. 
 
 Analyses of the inspired air were made several times during the 
 course of the experiment, as it was naturally important for our 
 purpose that the composition of the inspired air should show none 
 
1 1 
 
 
 ^-Cj *w 
 ^ VA 
 
 -^ *> 
 
 
 .^ ^ 
 
 K ^ 
 
 f >8 
 
 *cX S 
 
 -S ^ 
 
 K "** tv 
 
 ^ -^"S 
 
 
 "3 5 
 
 Q Q 
 
 
 **! 
 
 5 * 5 
 
 5' !* 
 
 
 ion 3 U 
 
 ^ ? s ^ 
 
 
 * " 
 
 'S 5 * ^ 
 
 
 
 
 
 o oil 
 
 ^. o^loi 
 
 
 14.2 
 
 (104.4) 
 
 
 I4.I 
 
 91.6 
 
 
 12.8 
 
 102.7 
 
 
 13.2 
 
 93.4 Normal oxygen 
 
 
 16.2 
 
 1 1 0.8 atmosphere. Rest. 
 
 
 14.45 
 
 (92.0) 
 
 \ 13-9 
 
 98.9 J 
 
 Mean 
 
 99.1 
 
 : 21.7 
 
 100.2 High oxygen 
 
 : 2 3.i 
 
 98.5 atmosphere. Rest. 
 
 6.9 
 
 H5.4 ' 
 
 8.5 
 
 1 2 1 .6 Low oxygen 
 
 8.2 
 
 128.1 atmosphere. Rest. 
 
 
 6.2 
 
 112. 1 
 
 
 16.0 
 
 124.8 Moderate work 
 
 
 
 with one arm. 
 
 
 19.9 
 
 131.7 ] Normal 
 
 
 20.9 
 
 135.0 \ Severe work ' oxygen. 
 
 
 14.9 
 
 128.0 \ with one arm. 
 
 
 24.8 
 
 128.0 J 
 
 
 15.1 
 
 147.6 Moderate work 
 
 
 
 with one arm. 
 
 
 
 Low oxygen. 
 
RESPIRATION 233 
 
 but minimal variations. Shortly before the close of the experi- 
 ment a sample of blood was withdrawn from the subject's ringer 
 into a capsule, and defibrinated with a platinum wire. Five-hun- 
 dredths cc. of this blood was then introduced into the saturating 
 vessel in the manner described in our paper. Immediately after- 
 wards two further small samples of blood were taken from the 
 subject's fingers as a rule one from each hand the blood being 
 received into small test tubes quite full of water, which were im- 
 mediately corked. These samples served for the colorimetric 
 determination of the degree of saturation of the blood with carbon 
 monoxide. A last sample of the inspired air was then taken, and 
 a sample of the alveolar air. Breathing into the apparatus was 
 continued for about two minutes in case the composition of the 
 air in the respiration apparatus had been altered by the deep ex- 
 piration necessary to afford the alveolar air sample : some car- 
 bonic acid, for instance, might have got through the caustic soda 
 tin. The experiment then terminated, and the mouthpiece of the 
 respiration apparatus was at once closed. The saturating vessel 
 containing the blood was as soon as possible filled by displacement 
 with some of the air remaining in the respiration apparatus, which 
 was expelled for this purpose from the bottle E by the arrange- 
 ment indicated at P and R. While the saturating vessel was being 
 rotated in the water bath at 38 the determination of the degree 
 of saturation with carbon monoxide of the samples taken from 
 the ringers was proceeded with. During this time also the analyses 
 of the alveolar air, and air from the respiration apparatus, were 
 completed and when necessary the analysis of a sample from the 
 saturating vessel. After the saturating vessel had been rotated for 
 half an hour or more, it was removed from the water bath and 
 the degree of saturation with carbon monoxide of the blood con- 
 tained in it was determined. All the data for calculating the 
 oxygen pressure of the arterial blood and contrasting it with that 
 of the alveolar or of the inspired air were then at our disposal. 
 
 Our first experiments on man were taken up with determining 
 the arterial oxygen pressure under as normal conditions as pos- 
 sible, and we especially wished to guard against the effects of 
 deficiency of oxygen. We therefore employed a low saturation 
 (23 per cent) of the blood with CO and made sure that the res- 
 piration apparatus contained a normal atmosphere by ventilating 
 it freely with fresh air before the experiment. All the experiments 
 were made with the subject sitting at rest. 
 
234 RESPIRATION 
 
 The results of these experiments are collected in the accom- 
 panying table. 
 
 The figures show quite distinctly that under normal circum- 
 stances when the subject is at rest the arterial oxygen pressure in 
 man corresponds exceedingly closely to the pressure of oxygen in 
 the alveolar air. In fact in no single instance does the value of the 
 arterial oxygen pressure differ from the alveolar by a greater 
 amount than can be accounted for by the experimental error of 
 the method. 
 
 We then tested the effect of raising the alveolar oxygen pressure 
 considerably above the normal value by filling the respiration 
 apparatus with an atmosphere rich in oxygen. The results of the 
 experiments are also given in the table. Here again the figures 
 show that the arterial and alveolar oxygen pressures have prac- 
 tically identical values. In these experiments on man we were 
 content to use only a moderate increase of the alveolar oxy- 
 gen pressure, for the higher the oxygen pressure is raised the 
 less proportional difference is there between the inspired air and 
 the alveolar air. A point will therefore eventually be reached 
 when the determination of the difference of tint between the blood 
 withdrawn from the body and that saturated with the inspired 
 air in vitro will fall almost within the experimental errors of the 
 method. It should be noted that in these experiments the car- 
 bonic acid in the alveolar air had precisely its normal value, 
 namely 5.6 per cent when measured dry, and we have therefore 
 no reason to suppose that the alveolar air samples were other 
 than normal. 
 
 Having obtained thus results which indicated that during rest 
 under normal conditions the transference of oxygen through the 
 pulmonary epithelium occurs without active secretory interven- 
 tion of the alveolar epithelium, we were naturally anxious to test 
 the matter further under conditions in which some amount of 
 deficiency of oxygen might affect the subject. The necessary de- 
 ficiency of oxygen was obtained by exposing the subject to an 
 atmosphere containing a considerably lower percentage of oxy- 
 gen than the normal. The experimental procedure was precisely 
 the same as before, save that we filled the respiration apparatus 
 before the start with an appropriate atmosphere by the method 
 described above. The results of these experiments are collected 
 in the middle part of the table. 
 
 The partial pressure of oxygen in the air breathed corresponded 
 
RESPIRATION 235 
 
 to an altitude of 15,000 feet or over; yet we noted that a 23 per 
 cent saturation of the blood with carbon monoxide was tolerated 
 without inconvenience. One of the subjects was liable to head- 
 ache when his blood was saturated to 25 per cent or more with 
 carbon monoxide, but this was in no wise accentuated in these 
 experiments. That deficiency of oxygen was exerting its custom- 
 ary effect on the respiration is indicated by the low value of the 
 alveolar carbonic acid percentage. Both the subjects noticed dis- 
 tinct hyperpnoea for some time after commencing to breathe into 
 the respiration apparatus, and that this was accentuated on the 
 slightest movement. The face remained of a distinctly bluish 
 color throughout the experiment, but the blueness passed away 
 if the hyperpnoea became exaggerated for a short time by mus- 
 cular movement. On rebreathing normal air at the close of the 
 experiment well-marked Cheyne-Stokes breathing was once or 
 twice observed, indicating that the want of oxygen had induced a 
 real hyperpnoea which had lowered the general carbonic acid 
 pressure in the body considerably. 
 
 In calculating the arterial oxygen pressures from the experi- 
 mental data of these experiments, it was necessary to make allow- 
 ance for the fact that the arterial blood was not fully saturated 
 with oxygen and CO, while the blood from the saturator must have 
 been almost completely saturated, as the oxygen pressure in the 
 air of the saturator was considerably higher, and hardly any CO 2 
 was present. For the correction required under these circum- 
 stances I must refer to our original paper. 
 
 On looking at the results of the four experiments it will be seen 
 that in every case the arterial was above the alveolar oxygen pres- 
 sure. The mean difference seems to be outside the limits of experi- 
 mental error, but only amounts to 8 mm. 
 
 A further series of experiments was made with the subject 
 doing muscular work. Preliminary experiments made with the 
 work done on a tricycle ergometer had shown that when the 
 breathing was greatly increased difficulties arose with the appa- 
 ratus. We therefore decided to make use of work with only one 
 arm. This enabled us to push the work to the point of fatigue, 
 when want of oxygen would be produced in the muscles, with 
 formation of lactic acid. That lactic acid was actually formed is 
 indicated by the low alveolar CO 2 percentages. The work appa- 
 ratus which we employed was of the simplest description. It 
 consisted of a lever which could be moved backwards and for- 
 wards, and transmitted its motion by means of a connecting rod to 
 
236 RESPIRATION 
 
 a small table carrying a weight which slid to and fro upon a 
 smooth plank, to one end of which the lever was pivoted. 
 
 The work apparatus was placed upon the ground adjacent to 
 the chair on which the subject sat, so that he could move the lever 
 and yet breathe comfortably into the respiratory apparatus. By 
 increasing the weight the amount of work done by the subject 
 could be raised. It was not possible to measure the actual work 
 done in mechanical units, but we could do so in physiological 
 units by observing, by means of the small gas meter, the effect 
 on the oxygen consumption of the subject per minute. What we 
 term "moderate work" in the tables below was sufficient to raise 
 the total oxygen consumption to one and a half times its resting 
 value, while "severe work" doubled the resting oxygen consump- 
 tion. Work which doubles the resting oxygen consumption is only 
 equivalent to walking on the flat at two miles per hour, and does 
 not sound particularly severe, but we found it sufficiently tiring 
 when it was performed by one arm only, and kept up for half an 
 hour at a time. 
 
 The lower part of Table III shows the results of the work ex- 
 periments. These results are very striking : for the arterial oxygen 
 pressure was on an average 4.4 per cent, or 32 mm. of mercury, 
 above the alveolar oxygen pressure, and in two experiments was 
 8.5 and 15.6 mm. above the oxygen pressure of the inspired air 
 (allowing for aqueous vapor). 
 
 In the last experiment on the table the effects of muscular work 
 and low oxygen in the inspired air were combined. It will be seen 
 that the arterial was 33.5 mm. above the alveolar oxygen pressure, 
 whereas with a low oxygen in the inspired air and no work the 
 arterial never exceeded the alveolar oxygen pressure by more than 
 13 mm. As already mentioned, it was noticed that when work 
 was done while a low oxygen percentage was being breathed the 
 lips and face lost the bluish color due to the low oxygen, and 
 became of a normal red color. It was also noticed many years ago 
 by Loewy 25 that even a slight muscular exertion produced a 
 marked improvement in the subjective symptoms of want of oxy- 
 gen in a steel chamber at low atmospheric pressure. Our results 
 on the arterial oxygen pressure during muscular exertion furnish 
 an evident clue to these observations. 
 
 "Loewy. Untersuchungen u. d. Respiration und Circulation, Berlin, 1895, 
 p. 1 6. The fact that Geppert and Zuntz (PfLiiger's Archiv, XLII, p. 189, 1888) 
 found a little more oxygen in arterial blood during work than during rest may 
 point also in the same direction. 
 
RESPIRATION 237 
 
 In a former chapter I have referred to some of the results of 
 the expedition to Pike's Peak undertaken in 1911 by Professors 
 Yandell Henderson and Schneider, Dr. Douglas, and myself. 26 
 Part of our object was to determine whether the want of oxygen 
 due to the rarefied air at 14,000 feet did not produce active secre- 
 tion of oxygen inwards. We used the same method as at Oxford, 
 taking every precaution against errors. The results were quite 
 unmistakable. We found that as soon as acclimatization to the air 
 was established the arterial oxygen pressure became considerably 
 higher than that of the alveolar air. The next table shows our 
 results. In ordinary resting experiments on acclimatized persons, 
 the arterial oxygen pressure was on an average about 70 per cent 
 above the alveolar oxygen pressure. When, however, air extra 
 rich in oxygen was breathed, so that the alveolar oxygen pressure 
 rose to about what it is at sea level, the difference between arterial 
 and alveolar oxygen pressure fell to 8 or I o per cent, even during 
 the short period of an experiment. In a subject investigated im- 
 mediately on arrival at the summit by the cogwheel railway the 
 arterial was only about 15 per cent above the alveolar oxygen 
 pressure, whereas three days later, after acclimatization, the ex- 
 cess was 100 per cent. 
 
 The Pike's Peak results threw much new light on oxygen secre- 
 tion by the lungs, and on the former experiments at Oxford. It 
 was evident, that not only is oxygen want a stimulus to active 
 oxygen secretion by the lungs, but that the response to the stimulus 
 improves greatly with practice or "acclimatization," just as is 
 the case with other physiological responses. We can now see why 
 some experiments for instance those which Lorrain Smith and I 
 made jointly on ourselves, indicated oxygen secretion, while other 
 experiments in which the physical and chemical conditions seemed 
 to be the same gave negative results. It was the physiological con- 
 ditions which were different. In the latter experiments we were 
 not acclimatized against anoxaemia. 
 
 It is easy to see the physiological advantage of oxygen secretion 
 as a defense against the anoxaemia of high altitudes and similar 
 conditions, or against carbon monoxide poisoning; but its uses 
 under ordinary conditions, where nothing but pure air at about 
 ordinary atmospheric pressure is breathed, are not so obvious. It 
 is clear that as the arterial haemoglobin is nearly saturated with 
 oxygen, during rest, at any rate, without any active secretion, 
 
 29 Douglas, Haldane, Yandell Henderson, and Schneider, Phil. Trans. Roy. Soc., 
 (B) 299, p. 195, 1913. 
 

 
 d, d 
 
 G 
 
 00 
 
 ^3 4H ^ 
 hb bX) 
 
 O 
 
 cL 
 
 O-, OQ 
 wo g 
 
 
 
 rt 
 
 
 
 SS& 
 
 a a 
 
 
 
 O 
 
 
 O2 tension of arterial 
 
 OO VO vo O\ HH 
 
 TfOO Ol O 
 
 tx rj- 
 
 "t 
 
 tX Tt 
 
 
 blood in mm. H g (at 
 37 moist) 
 
 ON iO co OiOO 
 00 00 O O ON 
 
 01 01 10 M 
 O ON iO co 
 
 tH M 
 
 tx TJ- 
 tx o 
 
 covo 
 00 ON 
 
 Ol HH 
 
 1000 
 
 
 Oz tension of alveolar 
 
 ^" O O CO \f** 
 
 CO IO Ol VO 
 
 ON Th 
 
 Tj- CO 
 
 vo tx 
 
 
 air in mm. H g {at 
 
 CO ONOO ON T}- 
 
 VO CO 
 
 VO 00 
 
 CO Ol 
 
 IO O 
 
 
 
 37 'moist) 
 
 iO TfOO ONVO 
 
 VO vo CO Ol 
 
 M 
 
 10 VO 
 
 ^ 10 
 
 * M- 
 
 GO 
 
 
 
 
 
 
 
 * 
 W 
 
 Oz tension of arterial 
 
 
 
 
 
 
 H 
 
 blood in percentage of 
 
 ON ON M 00 O 
 
 ON tx 10 tx 
 
 00 co 
 
 CO Tf 
 
 ON tx 
 
 to 
 W 
 
 the existing atmosphere 
 without allowing for 
 
 i-i O ovo xJ" 
 Ol 01 Ol 01 Ol 
 
 ^t" Ol co 1-1 
 01 01 M co 
 
 co 10 
 
 M 01 
 
 O CO 
 01 01 
 
 Ol ON 
 
 O 
 
 aqueous vapor 
 
 
 
 
 
 
 M 
 
 
 
 
 
 
 
 
 
 rv & 
 
 10 
 10 tX IO O tX 
 
 
 
 vo oo 
 
 10 IX 
 
 h-5 
 
 &**0 '3 
 
 01 01 txvo o 
 
 M ON tX 
 
 VO 00 
 
 HH 
 
 ON ON 
 
 5 
 
 !|.i - 
 
 01 01 H H 01 
 
 Ol 01 HH H) 
 
 HH HH 
 
 01 01 
 
 M 
 
 EH 
 ~ 
 
 V. b "^ ^ 
 
 10 
 
 IO IO 
 
 
 
 10 
 
 
 * - 
 
 O IO 01 I-H 10 
 
 CO Tj- co O 
 
 O CO 
 
 01 vo 
 
 00 tx 
 
 <$ 
 
 2 ~^ ** 
 
 ^ 8 
 
 M O ONOO tx 
 
 ON ONOO ON 
 
 00 vo 
 
 ONOO 
 
 01 vo 
 
 1 
 
 
 
 
 
 
 
 w 
 
 OH 
 
 -^ v^S 
 
 1-1 Ol ON O CO 
 ONOO O "^ 10 
 
 00 M vo O 
 
 Ol Tf IX Tl- 
 
 txvo 
 VO 00 
 
 HH CO 
 
 VO vo 
 
 CO Ol 
 
 OsVO 
 
 co 
 
 J ^ ! 
 
 vo vo tx tx vO 
 
 vo vo Tf ts. 
 
 VO 10 
 
 tx tx 
 
 tx tx 
 
 W 
 
 "** Q ^ 
 
 
 
 
 
 
 a 
 
 ^ ^ 
 
 
 
 
 
 
 HH 
 
 i>^ "^. 
 
 co vo co co O 
 
 w co O 
 
 txvo 
 
 HH O 
 
 VO vo 
 
 PH 
 
 ^ <5 
 
 ON M 01 00 
 
 Ol OO HH oi 
 
 tx 10 
 
 vo tx 
 
 M OO 
 
 
 
 CO hH h-l Tf IO 
 
 vo rj- OO ON 
 
 covo 
 
 M 
 
 HH ON 
 
 
 
 M M Ol 01 M 
 
 01 
 
 
 HH HH 
 
 
 
 KCS 
 
 Q 
 
 Id 
 
 . 
 
 CQ 
 
 * 
 
 
 i 
 
 ^ ^ ^ ^ 
 CJ 
 
 - 
 
 >"~ 
 
 w 8 
 
 + 
 
 
 
 ON O ^ VO 01 
 
 HH 00 HH CO 
 
 ON ON 
 
 HH 00 
 
 Tf tX 
 
 
 <& 
 
 M 01 01 01 
 
 01 01 HH 
 
 Ol 
 
 CO 
 
 
 
 r\ 
 
 ^ bb 
 
 >~ hi 
 
 >-> bi) 
 
 * ^ 
 
 hi 
 
 
 ^s 
 
 3 ^ 
 
 ^ 3 
 
 *~r* 3 
 
 
 
 
 
 >> <! 
 
 A < 
 
 HV< 
 
 Z< 
 
 < 
 
RESPIRATION 239 
 
 hardly anything could be gained by secretion, since any additional 
 oxygen which could be added to the blood would be trifling in 
 amount unless with an enormous secretory pressure such as has 
 never been found experimentally. We can thus readily under- 
 stand why there is no secretion during rest under normal condi- 
 tions, as our experiments clearly showed to be the case. It was 
 only during work that the experimental results showed secretion ; 
 but as a matter of fact the increase found in the arterial oxygen 
 pressure above the alveolar oxygen pressure would be of very 
 little service in charging the blood further with oxygen ; and this 
 brings us back to the original question. 
 
 In the first place it must be pointed out that the experiments 
 which Douglas and I made on oxygen secretion during muscular 
 work were carefully arranged in such a way as to demonstrate the 
 existence of secretion if any secretion occurred during muscular 
 work. We knew already that the stimulus to oxygen secretion 
 came from anoxaemia of the tissues. We knew, also, that the 
 only probable function of oxygen secretion was, not to raise the 
 arterial oxygen above that of the alveolar air, but to prevent a 
 serious fall, such as otherwise might take place during work suffi- 
 cient to increase very greatly the oxygen requirements of the 
 body. But we had no index of what the fall might be in the absence 
 of secretion. We therefore made the experiments in such a way 
 that tiring work, such as would presumably furnish the stimulant 
 to oxygen secretion, was done with one arm only. The oxygen 
 requirements of the body were in this way only increased to a 
 very moderate extent, so that oxygen secretion would have every 
 chance to raise the arterial oxygen pressure above the alveolar 
 oxygen pressure, just as in CO poisoning or at high altitudes dur- 
 ing rest. It was also much easier to make the experiments accu- 
 rately when the oxygen intake was not greatly increased. 
 
 Since our original experiments, and those on Pike's Peak, were 
 carried out, a good deal of both direct and indirect evidence has 
 accumulated in confirmation of our conclusions, and must now be 
 referred to. In Chapter VII the very clear physiological evidence 
 was summarized, showing that there is, in persons who are not in 
 good physical training, considerable anoxaemia during hard mus- 
 cular exertion. This is not merely local anoxaemia in the muscles 
 with the associated formation of lactic acid described in Chapter 
 VIII : for if the work is not too hard the respiratory symptoms 
 indicating anoxaemia are still present during the work, but there 
 is no trace of a subsequent fall in the resting alveolar CO 2 pres- 
 
240 RESPIRATION 
 
 sure. This fall is the physiological indication of lactic acid, and 
 runs parallel, as already mentioned, with the presence of much 
 lactic acid in the blood and urine. As Douglas and his pupils have 
 found, 27 there is, as a matter of fact, practically no increase in 
 the lactic acid present in the urine during moderate work. Thus the 
 anoxaemia cannot well t> due to anything else but imperfect 
 saturation of the arterial blood with oxygen ; and that this is the 
 actual cause is directly shown by the fact that a very moderate 
 increase in the oxygen percentage of the air breathed relieves the 
 symptoms. 
 
 In ordinary persons not in good physical training a very mod- 
 erate diminution in atmospheric pressure is quite sufficient to 
 cause a noticeable excess of hyperpnoea on any considerable ex- 
 ertion, such as climbing or walking fast. This is very evident on 
 going by train to some place four or five thousand feet above sea 
 level; and the cause is, without a shadow of doubt, imperfect 
 oxygenation of the arterial blood. At ordinary atmospheric pres- 
 sure we are accustomed to a certain degree of hyperpnoea and 
 exhaustion with a given degree of muscular exertion. That this 
 is in part dependent on imperfect saturation of the arterial blood 
 is only revealed by the fact that in air at a higher atmospheric 
 pressure (as in the case of workers in compressed air, and prob- 
 ably in deep mines), or when air enriched with oxygen is 
 breathed, the same work becomes much easier, at any rate to many 
 persons. 
 
 The observations of Dr. Henry Briggs, described in Chapter 
 VII, show that there is a striking difference in this respect be- 
 tween men in good physical training and ordinary persons, as the 
 former class get no benefit from air enriched with oxygen unless 
 the work is excessively hard, while the latter get great benefit, 
 shown, not only by the much greater ease and comfort with which 
 they perform the work, but by the smaller amount of air which 
 they require to breathe. The corresponding difference at high 
 altitudes is perfectly familiar to mountaineers. The man who is in 
 good training is free from the hyperpnoea, mountain sickness, and 
 other effects of high altitudes to a far greater extent than the man 
 who is not in training; and this evident fact has often led 
 mountaineers to the mistaken conclusion that mountain sickness 
 has nothing to do with altitude or anoxaemia, but is simply a sign 
 of imperfect training. 
 
 "Campbell, Douglas, and Hobson, Phil. Trans. Roy. Soc., (B), Vol. 210, p. x, 
 1920. 
 
RESPIRATION 241 
 
 All the facts just mentioned confirm the direct evidence in 
 favor of oxygen secretion in the lungs. Part of the exhaustion of 
 hard physical work is due to imperfect saturation of the arterial 
 blood with oxygen and the consequent effects on the respiratory 
 center and central nervous system as already described in Chap- 
 ters VI and VII. In persons who are in good physical training 
 these effects are in abeyance because as one part of physical train- 
 ing the lung epithelium has become much more capable of re- 
 sponding to the stimulus calling forth increased secretion of 
 oxygen, just as in the case of a man who has become acclimatized 
 to a high altitude, or to breathing air containing a small per- 
 centage of CO. It is the training of the lung epithelium, and not 
 anything else, that makes the specific difference. This is shown 
 at once by the fact that acclimatization to high altitudes or CO 
 poisoning takes place whether a man takes exercise or not. 
 
 In this connection I may mention the result of an experiment 
 which I made for a specific object during the war. It seemed 
 desirable to find out how soon the fall in oxygen percentage in the 
 air of a submarine would begin to have serious effects. I there- 
 fore shut myself in an air-tight respiration chamber which was 
 provided with the same sort of purifier for absorbing the CO 2 
 produced by respiration as was used in British submarines. The 
 oxygen percentage was also allowed to fall at the same slow rate 
 as that at which it had been found to fall in the most crowded 
 submarines then in use. After a few hours a light would no longer 
 burn in the air, and in a few more hours even a lighted pipe 
 handed in through a small air lock would no longer keep alight. 
 After 56 hours the oxygen percentage had fallen below 10. I then 
 terminated the experiment as the purifier was failing, and the 
 immediate object of the experiment, which was to find out whether 
 the air in a submarine would last easily for 48 hours without any 
 addition of oxygen, had been attained. I had no trace of mountain 
 sickness or any other symptom of anoxaemia, and my lips were 
 just as red as usual, though from other experiments described 
 in Chapter VII, I knew that without acclimatization I should have 
 broken down hopelessly in the existing atmosphere. A laboratory 
 attendant who afterwards went into the chamber along with me 
 became blue and uncomfortable, and finally collapsed and had to 
 be pulled out hurriedly. 
 
 In this experiment the fall in oxygen percentage had been so 
 slow that acclimatization had kept pace in me with the fall in 
 oxygen percentage, just as when a man ascends only very gradu- 
 
242 RESPIRATION 
 
 ally to a high altitude. There is, however, much more in this 
 acclimatization than mere increase in the power of oxygen secre- 
 tion, since there is also the gradual adjustment of blood reaction 
 to increased breathing, as explained fully in Chapter VIII. 
 
 In a more recent series of experiments by Kellas, Kennaway, 
 and myself 28 these two effects were separated. One of our objects 
 was to see how far acclimatization to high altitudes could be ob- 
 tained by discontinuous exposures to low barometric pressures. 
 This question is of course of considerable importance to airmen, 
 in whom the exposures are discontinuous. The effects produced 
 before acclimatization on Dr. Kellas, myself, and others, by an 
 exposure to 320 or 330 mm. barometric pressure, are described 
 in Chapter VI. To obtain acclimatization we used the method of 
 exposing ourselves for six to eight hours to atmospheric pressures 
 of 500, 430, and 360 mm. on three successive days. We found, how- 
 ever, that our resting alveolar CO 2 pressure had always returned 
 to normal before the morning after each successive exposure. 
 Thus there was no lasting adjustment of blood reaction to in- 
 creased breathing, as any change in this direction had disappeared 
 by the morning. There was also no lasting increase in our haemo- 
 globin percentages. Any acclimatization obtained must therefore, 
 apparently, be due to increased power of oxygen secretion. 
 
 The result of the experiment was that there was marked ac- 
 climatization, but limited in amount. When unacclimatized I had 
 been totally disabled, and had lost all memory, at a pressure of 
 320 mm., as already described. But on the last day of the ac- 
 climatization we stayed at 315 mm. for a considerable time, during 
 which, though we were distinctly blue, I could quite easily con- 
 tinue to do gas analyses and other operations, and move about 
 as usual, with no loss of memory afterwards of what had occurred. 
 In this experiment my son, Captain J. B. S. Haldane, acted as an 
 unacclimatized control. He came in with us and stayed for some 
 time at 366 mm. ; but after two hours he was so much affected that 
 we had to let him out. His breathing had become increasingly 
 rapid and shallow, and he had gradually sunk into a stupefied 
 condition. After coming out he could remember hardly anything 
 of the last hour in the chamber. 
 
 It is clear from this experiment that airmen, so long as they 
 retain their health, and remain at high altitudes pretty fre- 
 quently, must be capable of acquiring a considerable degree of 
 
 28 Haldane, Kellas, and Kennaway, Journ. of Physiol., LIII, p. 181, 1919- 
 
RESPIRATION 243 
 
 acclimatization. This acclimatization was long ago noted by 
 Glaisher in connection with his occasional high balloon ascents. 
 An equal degree of acclimatization can undoubtedly be main- 
 tained in a simpler manner by good physical training; and at 
 heights of less than about 20,000 feet an airman in good physical 
 training should have little difficulty from anoxaemia. It must be 
 noted, however, that even a small degree of the neurasthenia with 
 shallow breathing described in Chapter III renders an airman 
 totally incapable of going to any considerable height without an 
 oxygen apparatus. 
 
 Our acclimatization experiment indicated that with complete 
 acclimatization, including adjustment of the blood reaction to 
 the increased breathing, and increase in the haemoglobin per- 
 centage, a man could probably, if the mere physical difficulties 
 were not too great, reach the summit of Mount Everest without 
 breathing anything else than ordinary air, though he would 
 quite certainly die at this altitude if he were not acclimatized. 
 
 It was pointed out in Chapters III and VII that, on account of 
 the imperfect distribution of air in the lungs, the average alveolar 
 oxygen pressure is, even during rest under normal healthy condi- 
 tions, no certain guide to the oxygen pressure of the mixed 
 arterial blood. During heavy work this must be so to an increased 
 degree, since, although the expansion of the lungs is much better, 
 the rate at which oxygen is absorbed is enormously greater. 
 Meakins and Davies 29 have recently made exact determinations 
 of the percentage saturation with oxygen of the haemoglobin in 
 the arterial blood of a number of healthy persons, and found it to 
 vary from 94 to 96 per cent in different persons, the variation 
 depending probably on the differences in the oxyhaemoglobin 
 curves which Barcroft discovered (Chapter IV). In my own case 
 the saturation was 94.3 per cent. This is not much lower than 96 
 per cent, the saturation which would be expected if my arterial 
 blood were fully saturated to the oxygen pressure of the mixed 
 alveolar air. If, however, we look at the dissociation curve of 
 the oxyhaemoglobin of human blood, we see that 94.3 per cent 
 saturation corresponds to an oxygen pressure of only 11.2 per 
 cent of an atmosphere, as compared with 13.2 per cent in the 
 alveolar air. Thus the oxygen pressure in the mixed arterial blood 
 is very distinctly less than in the alveolar air ; and this is the sort 
 of result which the aerotonometer gives, as already explained. 
 
 "Meakins and Davies, Journ. of Pathol. and, Bacter., XXIII, p. 453, 1920. 
 
244 RESPIRATION 
 
 On the other hand the carbon monoxide method gives, during 
 rest under normal conditions, exactly the same oxygen pressure 
 in the arterial blood as in the alveolar air. This difference in the 
 results by the two methods used to be rather a puzzle, and was 
 explained by me as probably due, either to a process of rapid but 
 slight oxidation in the blood itself, or to a little blood getting 
 through the lungs without exposure to alveolar air. Our shallow 
 breathing experiments, and the neurasthenia cases, showed 
 clearly enough why the mixed arterial blood is not fully saturated 
 to the alveolar pressure; but why does the carbon monoxide 
 method not show this? A little consideration will show the reason. 
 The carbon monoxide method gives the average arterial oxygen 
 pressure of all the portions of arterial blood leaving the lung 
 alveoli, just as the "alveolar air" gives the average oxygen pres- 
 sure of all the portions of air in the alveoli of the air-sac system. 
 But the oxygen pressure of the mixed arterial blood cannot be 
 deduced, as fully explained in Chapter IV, from the average of 
 the oxygen pressures in the blood leaving the alveoli. It is this 
 average that the carbon monoxide method gives. Hence for the 
 purpose of deducing the oxygen pressure of the mixed arterial 
 blood the carbon monoxide method has exactly the same defects 
 as the method of inferring this value from the oxygen pressure of 
 the alveolar air on the assumption (perfectly valid for resting 
 conditions at ordinary atmospheric pressure when pure air is 
 breathed) that diffusion equilibrium is established between al- 
 veolar air and blood. For the purpose, however, of deciding 
 whether or not active secretion of oxygen is occurring, the carbon 
 monoxide method is perfectly valid. It gives just the information 
 needed; and for this purpose it is far more reliable than the 
 aerotonometer method, which has always given misleading in- 
 formation on the question of diffusion equilibrium for oxygen, 
 and made it appear as if diffusion equilibrium is never attained, 
 even during complete rest. 
 
 To those who pin their faith, as regards the secretion question, 
 to the aerotonometer results, I may perhaps point out that if they 
 were accepted as evidence they would completely wreck the dif- 
 fusion theory. For if diffusion equilibrium is not even obtained 
 under resting conditions under normal barometric pressure it 
 would be quite inconceivable on the diffusion theory that anything 
 approaching to diffusion equilibrium would be obtained during 
 muscular work, and particularly at high altitudes. Yet on Pike's 
 
RESPIRATION 245 
 
 Peak is was possible to do hard muscular work with .the lips re- 
 maining quite red. 
 
 It will easily be seen on consideration that as the barometric 
 pressure, or the oxygen percentage of the inspired air, is pro- 
 gressively reduced, the difference in percentage saturation be- 
 tween the mixed arterial blood and blood completely saturated at 
 the existing alveolar oxygen pressure will increase more and more 
 if diffusion alone determines the saturation of the blood in the 
 lungs, and will tend in the same direction even if active secretion 
 assists diffusion. We can thus easily explain why some of the 
 persons who ascended Pike's Peak were very blue in the face, and 
 why fainting or partial loss of consciousness were common occur- 
 rences. We can also understand why some persons become more 
 or less unwell at first on going to an altitude of only four or five 
 thousand feet, and why in all persons there is a distinct physio- 
 logical reaction to anoxaemia, as shown by lowering of the al- 
 veolar CO 2 pressure and rise in the haemoglobin percentage. This 
 physiological reaction would be difficult to understand if there 
 was uniform saturation of the haemoglobin in all the alveoli. We 
 must conclude that whether or not a person is acclimatized to a 
 low barometric pressure the percentage saturation of the mixed 
 arterial haemoglobin with oxygen is distinctly diminished, though 
 the amount of the diminution is not indicated by the carbon mon- 
 oxide method. 
 
 In the process of oxygenation of the blood in the lungs, the 
 oxygen has to pass from the alveolar air through a thin layer of 
 living tissue into the blood and into the corpuscles. This process 
 must take some time. To the genius of Christian Bohr we owe the 
 principle of a method by which the time may be estimated, in so 
 far as the process is one of diffusion. In connection with the ab- 
 sorption of oxygen by the lungs it is not possible to measure the 
 rate at which, with a given diffusion pressure, oxygen passes 
 inwards, because we do not know the mean diffusion pressure. 
 We can, as will be shown later, measure the oxygen pressure of 
 the venous blood, as well as that of the alveolar air and arterial 
 blood ; but we do not know how quickly the blood becomes satu- 
 rated in its passage along the alveolar capillaries. Hence we can- 
 not estimate the mean difference in oxygen pressure required for 
 the diffusion inwards of a given quantity of oxygen in a given 
 time. In the case of absorption of CO present in the air in a low 
 proportion the conditions are quite different, however: for we 
 can determine the percentage of CO in the alveolar air, and the 
 
246 RESPIRATION 
 
 rate at which the gas is absorbed, while, for short experiments, 
 the difference in CO pressure between the alveolar air and the 
 blood is constant. In this way we can tell how much CO is ab- 
 sorbed per minute with a given pressure difference; and from 
 this, allowing for the greater solubility and slightly lower dif- 
 fusibility of oxygen, we can calculate the rate at which oxygen 
 diffuses in with the same pressure difference. 
 
 Bohr's original calculations (based on rather rough experi- 
 ments made by myself for another purpose) were not very ac- 
 curate ; but the matter was reinvestigated by A. and M. Krogh, 30 
 and still more recently by M. Krogh. 31 A. and M. Krogh found 
 that for adults about 25 cc. of oxygen will diffuse inwards per 
 minute for every I mm. of difference in oxygen pressure during 
 rest, and about 35 cc. during work. The estimate of M. Krogh is 
 considerably higher ; but I do not think that the method which she 
 used was at all reliable, for the following reasons. The method 
 consisted in taking in a deep breath of air containing a small 
 percentage of CO. Part of this breath was then breathed out, and 
 a sample of the alveolar air taken. The rest of the breath was 
 held for a measured interval of time, after which a second sample 
 of alveolar air was taken, and the percentages of CO in the two 
 samples very accurately determined. From the fall in the per- 
 centage of CO between the two samples the rate of absorption of 
 the CO was then calculated. 
 
 If the difference between the percentages of CO in the two 
 samples represented absorption of CO, the method would be a 
 correct one. Actually, however, it is quite impossible, as I have 
 convinced myself by repeated experiments with various gas mix- 
 tures, to secure an even distribution of a gas through the lung air 
 by taking in a single deep breath. The first alveolar sample con- 
 tains an undue proportion of the atrial air containing a higher 
 initial percentage of CO, while the second sample comes ex- 
 clusively from the alveoli of the air-sac system, in which the per- 
 centage of CO was never nearly so high as in the atria. Thus the 
 apparent absorption of CO during the interval of holding the 
 breath is much greater than the actual absorption. The method is 
 thus fallacious; and the same criticism applies to a number of 
 other Copenhagen experiments with regard to alveolar air, the 
 dead space in breathing, etc. 
 
 Taking, however, the earlier estimate of A. and M. Krogh, it 
 
 30 A. and M. Krogh, Skand. Arch. f. Physiol., XXIII, p. 236, 1910. 
 
 31 M. Krogh, Journ. of Physwl., XLIX, p. 271, 1915. 
 
RESPIRATION 247 
 
 can be calculated 32 that during rest at normal atmospheric pres- 
 sure, the arterial blood passing through an average alveolus 
 would easily be saturated by simple diffusion to the oxygen pres- 
 sure of the air in the alveolus. During considerable muscular work, 
 however, this would not be the case ; and the arterial blood would 
 emerge incompletely saturated. That there should be some an- 
 oxaemia during considerable exertion is therefore exactly what 
 might be anticipated on the diffusion theory, even without any 
 allowance for the effects of uneven distribution of air and blood 
 among different alveoli. When allowance is also made for this 
 factor, the presence of anoxaemia during even very moderate 
 exertion at ordinary atmospheric pressure in persons not physi- 
 cally fit is just what might be expected; and at high altitudes the 
 anoxaemia would be so serious as to make any considerable ex- 
 ertion impossible but for active secretion. 
 
 All the facts, therefore, and not merely our direct measurements, 
 go towards showing that oxygen secretion is a most important 
 physiological factor, not merely under exceptional circumstances, 
 but during ordinary life at sea level. It is probably also an im- 
 portant factor under pathological conditions, though on this sub- 
 ject our knowledge is still almost a blank, owing to lack of 
 observations. The only relevant observations are those of Lorrain 
 Smith. 33 His experiments, when due allowance is made for the 
 errors already referred to in our calculations, showed that either 
 a rise of body temperature or a severe infection paralyzed the 
 power of oxygen secretion in response to CO poisoning. When 
 lung inflammation was produced by exposing the animals to a 
 high pressure of oxygen (see Chapter XII) the arterial oxygen 
 pressure fell to values which, when corrected, are much below 
 that of the alveolar air. In this case it is evident that not only 
 active secretion, but also diffusion of oxygen inwards, was inter- 
 fered with. The animals were incapable of muscular exertion and 
 thus showed symptoms similar to those of phosgene poisoning, as 
 described in Chapter VII. 
 
 A significant determination has quite recently been published 
 by Harrop 34 of the percentage saturation of human arterial blood 
 with oxygen, first during rest, and then just after exhausting 
 work. The results were 95.6 per cent during rest, and 85.5 per 
 cent just after the exertion. The deficiency found in the blood just 
 
 32 Douglas and Haldane, Journ. of Physiol., XLIV, p. 337, 1913. 
 88 Lorrain Smith, Journ. of Physiol., XXII, p. 307, 1898. 
 34 Harrop, Journ. of Exper. Med., XXX, p. 246, 1919. 
 
248 RESPIRATION 
 
 after exertion is far greater than could be accounted for by ex- 
 perimental errors. 
 
 As already mentioned, the aerotonometer experiments of Krogh 
 indicated that the arterial CO 2 pressure is the same as that of the 
 alveolar air. The manner in which the respiratory center responds 
 to the slightest increase or diminution in the alveolar CO 2 pres- 
 sure, and the quantitative correspondence between rise in alveolar 
 CO 2 pressure and response of the respiratory center, point most 
 clearly to the conclusion that within pretty wide limits there is no 
 active secretion of CO 2 outwards in the lung, or active retention 
 of CO 2 when the lungs are over-ventilated. In individual experi- 
 ments Bohr obtained results which seemed to point to active 
 secretion of CO 2 outwards. The latest of these were made with 
 Krogh's small aerotonometer; but Krogh has pointed out how 
 easily errors may arise with this instrument; and in view of all 
 the facts I think his criticism of Bohr's experiments is probably 
 correct. 
 
 If we calculate, by Bohr's method, the rate of diffusion of CO 2 
 from the alveolar air into the blood, the result is that for the same 
 difference in partial pressure CO 2 , in consequence of its much 
 greater solubility, must diffuse outwards about 20 times as rapidly 
 as oxygen diffuses inwards. Against this, however, must be set the 
 fact that the initial difference in CO 2 pressure between the venous 
 blood and alveolar air is only about a tenth of the corresponding 
 difference in oxygen pressure. On balance, however, there is prob- 
 ably little hindrance, even during hard work, to the establishment 
 by diffusion of practical equilibrium in CO 2 pressure between the 
 alveolar air and arterial blood. We have already seen that the 
 giving off of CO 2 in the lungs is dependent in great part on the 
 saturation of the haemoglobin with oxygen. Hence the giving off 
 of CO 2 is to a large extent under the control of oxygen absorp- 
 tion, and so of oxygen secretion when this occurs. 
 
 Apart from this there seem to me to be strong reasons for sus- 
 pecting that although active secretion of CO 2 , like active secretion 
 of oxygen, does not occur under ordinary conditions, it does occur 
 when high pressures of CO 2 exist in the arterial blood, and the 
 body is threatened by the excess of CO 2 . As yet there is no direct 
 evidence on this subject ; but the reasons are as follows : ( I ) When 
 a small volume of oxygen is rebreathed as long as possible, or 
 even when the breath is held as long as possible after filling the 
 lungs with oxygen, the percentage of CO 2 in the alveolar air 
 mounts up much higher and more rapidly than can well be ac- 
 
RESPIRATION 249 
 
 counted for from any probable rise in the pressure of CO 2 in the 
 venous blood. Examples of experiments in this direction are given 
 in the paper by Christiansen, Douglas, and myself. (2) It ap- 
 pears that men in good training and with the power of oxygen 
 secretion well developed are capable of standing a much higher 
 percentage of CO 2 in the inspired and alveolar air than other 
 men. In my experience with self-contained mine-rescue apparatus, 
 and similar devices, I have often been struck with the greater 
 sensitiveness to CO 2 of myself and other sedentary workers in 
 comparison with men in good physical training, although nearly 
 pure oxygen was being breathed. These observations suggest very 
 strongly that along with the power of oxygen secretion the power 
 of secretion of CO 2 is developed by muscular exertion. (3) In the 
 experiments of Paul Bert 35 on the blood gases when increasingly 
 high percentages of CO 2 were breathed by animals, it appeared 
 that with increase in the CO 2 percentage the CO 2 in the arterial 
 blood often showed little or no increase. It seems very dif- 
 ficult to explain these results apart from active secretion of CO 2 
 coming into play progressively, and particularly in view of the 
 experiments of Henderson and Haggard on the increased CO 2 - 
 absorbing capacity of the blood when excess of CO 2 is breathed 
 (Chapter VIII). 
 
 In view of the absence, as yet, of direct measurements, it seems 
 unnecessary to discuss this question further; but I may point out 
 that just as the opponents of the oxygen-secretion theory have 
 been mistaken in drawing general conclusions from experiments 
 in which oxygen secretion was either absent or could not be dem- 
 onstrated, it is very probable that they have been equally mistaken 
 over secretion of CO 2 . Bearing in mind Johannes Miiller's argu- 
 ment as to the analogy between secretory activity and ordinary 
 metabolic processes, it seems quite likely that the active transport, 
 not only of oxygen, but also of CO 2 , is a phenomenon which oc- 
 curs in all living cells. 
 
 Not only do oxygen and CO 2 diffuse through the lung epithe- 
 lium into or out of the blood, but also other gases, such as nitrogen, 
 hydrogen, methane, carbon monoxide, etc., so that their partial 
 pressures become exactly equal in the body and the alveolar air. 
 But how is it that oxygen is sometimes actively secreted inwards, 
 and that the oxygen pressure may be greater in the blood without 
 the oxygen leaking back by diffusion into the alveolar air just as 
 
 "Paul Bert, La Pression barometrlque, p. 985. 
 
250 RESPIRATION 
 
 other gases leak in or out? We must, I think, suppose that the 
 structure of the alveolar epithelium is not homogeneous but may 
 be divided into a reticulum of living structure and a plasma filling 
 the interstices, just as is the case with the body as a whole. The 
 diffusion will take place through the plasma, while the living sub- 
 stance behaves as a solid towards diffusion, as in the case of the 
 secreting cells of the swim bladder. Not only oxygen but also 
 other gases will diffuse through the plasma; but during secretion 
 of oxygen the living substance behaves like the protoplasm of the 
 swim bladder, taking up oxygen on one side of the cell, and giv- 
 ing it off at a higher pressure on the other. The oxygen will tend 
 to diffuse backwards if, as in experiments with a high percentage 
 of CO, the oxygen pressure becomes higher in the blood than in 
 the alveolar air; but some, at least, of this oxygen will be caught 
 on its way and returned. 
 
 This general conception throws light in other directions. For 
 let us suppose the direction of the oxygen secretion to be reversed, 
 so that the lung epithelium, instead of absorbing oxygen, hinders 
 its passage. Nitrogen and other inert gases will still be able to pass 
 inwards freely by diffusion. We shall thus have nitrogen going 
 through, without oxygen. Now let us suppose that the epithelium 
 has an excretory function; and let us apply the general concep- 
 tions, above set forth, to the glomerular epithelium of the kidney. 
 We can imagine the living substance of this epithelium holding 
 back, by an active process, all the normal constituents of blood, 
 particularly water, if their normal diffusion pressures are not ex- 
 ceeded, but otherwise letting them through. All the known facts 
 seem to confirm Bowman's original conclusion that the water of 
 the urine is usually almost entirely separated in the glomeruli. It 
 seems also clear that as shown by Ludwig and his pupils, the 
 process of separation is dependent on blood pressure, like a filtra- 
 tion process. If we suppose that the passages through which the 
 liquid is filtered are not permeable by the proteins of the blood, 
 we have an explanation, as pointed out by Starling, of why a cer- 
 tain minimum blood pressure is needed. The liquid separated might 
 be little different from pure water, whereas the blood plasma 
 contains salts in considerable amount. Such a liquid could not be 
 separated by simple filtration, and numerous other facts are 
 against the simple filtration theory. I think that all the facts con- 
 form with the theory that the glomerulus is a filter, but with a 
 living framework, and that the action of this living framework, 
 is to pick out and return to the blood what belongs to its normal 
 
RESPIRATION 251 
 
 composition, the rest being allowed to pass. In this process the 
 glomerular epithelium will of course be doing work; but every 
 living tissue seems to be always doing work, even when it is "rest- 
 ing." During a glomerular diuresis there may be no extra work for 
 the epithelium to do, and it will simply act as a filter, just as the 
 lung epithelium during rest under normal conditions acts like a 
 nonliving membrane. Barcroft and Straub have shown that during 
 certain kinds of diuresis there is no increased consumption of 
 oxygen by the kidney, and therefore presumably no work done by 
 the kidney in the process of separation of the extra urine formed. 36 
 It is probable that under normal conditions a pure filtration 
 diuresis of this type never occurs at all; but the possibility of 
 producing it experimentally throws much light on the mode of 
 action of the glomeruli and also of the lung epithelium. Possibly 
 the substances carried to the lungs during anoxaemia act in the 
 same way as a diuretic drug acts on the kidneys. 
 
 In concluding this long chapter I must make some reference 
 to criticisms which have been made on our experiments. Part of 
 these criticisms are the evident outcome of a natural conservative 
 desire to save some remnant of the old mechanistic theory of 
 glandular secretion. The lungs and the kidney glomeruli were 
 the last remaining strongholds that there seemed much hope of 
 defending, and I can admire the spirit which has animated the 
 defenders. It is different, however, with the criticisms made by 
 my friend Mr. Barcroft in his recent book, 37 as he fully ac- 
 knowledges the difficulties of the diffusion theory and the inherent 
 probability of secretory activity in the lungs. 
 
 He bases these criticisms on the work of his pupil, Mr. Hart- 
 ridge. The latter devised a new and thoroughly sound method of 
 determining the percentage saturation of the blood with CO by 
 delicate measurements of the shifting in position of the absorption 
 bands of oxy- and CO-haemoglobin; and he showed clearly that 
 his method, although it requires elaborate apparatus, is capable 
 of giving accurate results. Armed with this method he proceeded 
 to repeat, as he thought, some of the experiments (not yet pub- 
 lished except in a short abstract) of Douglas and myself on man. 
 Unfortunately he modified the method in essential respects, neither 
 taking precautions that the subject was breathing a constant per- 
 centage of oxygen, nor using whole blood in the saturator, nor 
 experimenting in a way calculated to elicit any evidence of active 
 
 * Barcroft and Straub, Journ. of Phystol., XLI, p. 145, 1911. 
 " Barcroft, The Respiratory Function of the Blood,, p. 204. 
 
252 RESPIRATION 
 
 secretion during work. His experiments did not appear to show 
 any active secretion, and it would have been extraordinary if 
 they had. 
 
 I now come to the main point of Barcroft's criticisms. Hartridge 
 had at first calibrated his instrument by ascertaining its readings 
 with what he believed to be known mixtures of oxyhaemoglobin 
 and CO-haemoglobin. He subsequently found that his calibrations 
 had been quite incorrect; and in order to secure correct calibration 
 he finally had recourse to the very tedious method of pumping 
 out the CO and oxygen from the blood mixture after adding 
 ferricyanide, and determining the CO and oxygen by analysis, 
 using the general method which I followed in originally testing 
 the accuracy of the ferricyanide method for blood gases. In his 
 paper 38 Hartridge says of his first method that "experiments 
 made since to discover the cause of the error have shown that 
 with the method of mixture employed complex interactions take 
 place between the two portions of solvent." Let us expand this 
 somewhat mystic statement. He was working with blood diluted 
 with water to about a twentieth. One portion of this he saturated 
 with CO, and another portion with air. These were then mixed. It 
 was apparently expected that the result would be a mixture con- 
 taining half the haemoglobin saturated with CO and the other 
 half with oxygen. Now if one dilutes blood to a twentieth and 
 saturates with CO, the solution will contain about one volume of 
 CO in combination with haemoglobin to two and one-half in 
 simple solution ; and when this is mixed with an equal proportion 
 of the solution saturated with air the CO in simple solution in the 
 first part will straightway combine with the haemoglobin in the 
 second part, and turn out the oxygen, the result being that prac- 
 tically the whole of the haemoglobin combines with CO. With 
 the method first adopted by Hartridge it was clearly impossible 
 for him to calibrate his instrument. 
 
 Our colorimetric method of determining the saturation of 
 haemoglobin with CO had repeatedly been tested against mix- 
 tures previously prepared, the most scrupulous precautions (de- 
 scribed in three different papers) being, however, taken to avoid 
 errors arising from the solubility of CO. Barcroft, however, infers 
 that because Hartridge's calibration failed with the method of 
 mixtures, ours was presumably also inaccurate : whereas Hart- 
 ridge's final calibrations were made with the blood pump, which 
 
 38 Hartridge, Journ. of Physiol., XLIV, p. 9, 1912. 
 
RESPIRATION 253 
 
 is an "objective method," and therefore the only trustworthy one. 
 Hence, Barcroft argues, Hartridge's experiments, so far as they 
 go, furnish the only reliable evidence about oxygen secretion, as 
 to which they give a negative result. As a matter of fact there is 
 not a shadow of doubt that our method of testing the colorimetric 
 method was at least as exact as the final method used by Hartridge. 
 
 Barcroft's reference to objective methods recalls to my mind 
 what happened when Hartridge came to Oxford to demonstrate 
 his method. It was apparently an "objective method," dependent, 
 like Hiifner's spectrophotometric method, on the exact positions 
 of absorption bands in the spectra of oxyhaemoglobin and CO 
 haemoglobin bands of which the "exact positions" can be quite 
 easily photographed. A solution of blood was prepared for dem- 
 onstration; and Hartridge, the late Professor Gotch, and I went 
 into a dark room and proceeded first to determine the zero point 
 on the scale of the apparatus. First one, and then the others of us 
 determined the zero point. But the results were all different, 
 though each one of us always got the same result. We stood there 
 in the dark, each suspecting the others of want of accuracy, but 
 afraid to say so. Suddenly the truth dawned on us. Even the 
 position of an absorption band is subjective ! 
 
 And then, if our ears could have caught it, we might have 
 heard a gentle but kindly laugh. It came from a Spirit that flits 
 round old university walls and even wanders sometimes into 
 laboratories. It was the Spirit of Humanism that laughed, and it 
 always laughs when men find out with Socrates that what is ob- 
 jective is also subjective. 
 
 Addendum.. Barcroft and his associates 39 have recently made a 
 very carefully planned attempt to see whether any evidence of oxy- 
 gen secretion could be obtained by analyses of the arterial blood. 
 Barcroft himself was the subject of the experiment, and he re- 
 mained for a week in a respiration chamber in which the oxygen 
 percentage was gradually lowered, until on the last day there was 
 only about 1 1 per cent of oxygen in the air, corresponding to an 
 altitude of 18,000 feet, or about 17,000 if allowance is made for 
 the presence in the air of about 0.5 per cent of CO 2 . There was 
 thus apparently every chance of acclimatization occurring. On the 
 other hand very little acclimatization seems to have actually oc- 
 curred, as the subject was very unwell, with slight rise of tempera- 
 
 39 Barcroft, Cooke, Hartridge, T. and W. Parsons, Journ. of PhysioL, LIII, 
 p. 450, 1920. 
 
254 RESPIRATION 
 
 ture, on the last day or two, and was in a fainting condition at the 
 end, just before the samples of arterial blood were taken. 
 
 Samples of arterial blood were taken, firstly during rest, and 
 later during work on a bicycle ergometer of about 380 kilogram- 
 meters per minute, which would increase the respiratory exchange 
 about three or four times. The haemoglobin of the sample during 
 rest was found to be 88. 1 per cent saturated with oxygen. Analyses 
 of the arterial blood were made, both by the ferricyanide method 
 and with the pump, and agreed closely. Samples of alveolar air 
 were also taken, and part of the arterial blood saturated with air 
 of about the same composition. The saturation of the haemoglobin 
 of this blood, when corrected for the slight difference in oxygen 
 pressure between the air in the saturator and the sample of 
 alveolar air, was found to be 91 to 92 per cent, which is distinctly 
 higher than the saturation of the arterial blood. The oxygen pres- 
 sure of the sample of alveolar air was, however, quite unac- 
 countably high. It was 68 mm., instead of about 45 mm. which was 
 the value actually found in a determination made a few hours 
 previously, and was also the value to be expected from the curve 
 shown in Figure 98 of this book. Had the actual alveolar gas pres- 
 sures corresponded with those of the sample, the respiratory quo- 
 tient would have been about 2 ; and such a quotient occurs only 
 during forced breathing, which could not have occurred. It seems, 
 therefore, that there must have been some mistake about the al- 
 veolar sample; but what this was is far from clear. If the actual 
 alveolar oxygen pressure had been about 45 mm., as would cor- 
 respond to the alveolar CO 2 pressure, the oxygen saturation of 
 the blood from the saturator would have been considerably lower 
 than that of the arterial blood. The experiment is thus inconclusive, 
 apart altogether from the question as to whether the subject was 
 acclimatized at all, or to what extent. 
 
 The experiment during work is much more consistent. The 
 arterial haemoglobin was found to be only 83.5 per cent saturated 
 with oxygen. A lower saturation during work of the character 
 chosen corresponds well with all our observations on Pike's Peak 
 and at Oxford. Unacclimatized persons became very blue in the 
 face on Pike's Peak with comparable work ; and even after acclima- 
 tization there were clear indications of some anoxaemia. In me, for 
 instance, the alveolar oxygen pressure rose about 8 mm., and the 
 alveolar CO 2 pressure fell, on walking at 4 miles an hour; and 
 this, as we pointed out, indicated arterial anoxaemia. The haemo- 
 globin of the blood exposed to the alveolar air in the saturator 
 
RESPIRATION 255 
 
 gave a saturation of 89.2 per cent, which is 5.7 per cent higher 
 than the saturation of the arterial blood. This result furnishes no 
 evidence of secretion, but to show that there was actually no secre- 
 tion it would, I think, be necessary to make a control experiment 
 on a person who had spent only a short period in the chamber, and 
 was undoubtedly unacclimatized. 
 
 Barcroft and his associates consider that the results of the 
 experiments were against the secretion theory. In this I cannot 
 agree with them. It seems to me evident that if there was any 
 acclimatization in these experiments it was very imperfect, and 
 not comparable to the acclimatization commonly observed at high 
 altitudes, and closely studied by us on Pike's Peak. Acclimatization 
 occurs much more readily in certain persons than in others, and 
 seems also to be greatly affected by accompanying conditions. An 
 experiment in which marked acclimatization occurred in myself 
 in a respiration chamber was referred to above. On endeavoring 
 to repeat this experiment in the summer of 1920 there was no 
 effective acclimatization, and on account of severe symptoms of 
 anoxaemia, accompanied by blueness of the lips, etc., I had to 
 stop before the oxygen pressure had fallen to quite as low a point 
 as on Pike's Peak, or to nearly as low a point as in the previous 
 experiment where no pathological signs of anoxaemia were pro- 
 duced. It was about a week before I recovered from the effects of 
 this unsuccessful experiment. The weather was hot, and the 
 chamber correspondingly uncomfortable. I was also several years 
 older. In this experiment my arterial blood was analysed by Pro- 
 fessor Meakins, who found the haemoglobin to be considerably 
 below its normal saturation with oxygen. There was evidently 
 little or no acclimatization. 
 
 I should like to correct here one or two misunderstandings which 
 occur in the paper of Barcroft and his associates. Through a mis- 
 reading of the paper by Douglas and myself he concluded that 
 on lowering the oxygen pressure of the inspired air to what cor- 
 responded to about the oxygen pressure on Pike's Peak we found 
 in a short experiment at Oxford that by the carbon monoxide 
 method the arterial oxygen pressure was 70 mm. above the al- 
 veolar oxygen pressure. The actual difference was only trifling 
 (about 8 mm.), as shown in the table reproduced above. It re- 
 quired prolonged acclimatization to produce as great a difference 
 as even 35 mm. There is also a misunderstanding as to our experi- 
 ments on the effects of work. Though we made no observations by 
 the carbon monoxide method on the effects of work such as was 
 
256 RESPIRATION 
 
 employed by Barcroft, all the other observations referred to in 
 the present chapter tend to show that except, perhaps, when physi- 
 cal training or acclimatization is very effective, the arterial oxygen 
 saturation during such work is lower than during rest. 
 
 Clear evidence is brought forward by Barcroft and his associ- 
 ates that no appreciable loss of dissociable oxygen occurs in ar- 
 terial blood which is allowed to stand for a short time. In the 
 Pike's Peak report we concluded that such a loss probably occurs. 
 The chief reason for this conclusion was that the aerotonometer 
 always gives a lower oxygen pressure than that deduced on the 
 diffusion theory from the alveolar oxygen pressure, or indicated 
 by the carbon monoxide method during rest under ordinary baro- 
 metric pressure. As explained above, however, there is now an- 
 other and very clear explanation for this ; and since the investiga- 
 tion by Meakins, Priestley, and myself on the effects of shallow 
 breathing I have altogether ceased to believe in the presence, to 
 any extent which would upset a blood-gas or aerotonometer de- 
 termination, of "reducing substances" in blood. I am in entire 
 agreement with Barcroft's criticism of the old experiments by 
 which Pfliiger believed that he had demonstrated the existence 
 of reducing substances in fresh arterial blood. It may also be men- 
 tioned here that in some unpublished experiments Douglas and I 
 were unable to obtain any evidence by blood-gas analysis of the 
 presence of reducing substances, even in blood which was com- 
 pletely reduced by prolonged stoppage of the circulation in the 
 arm. 
 
CHAPTER X 
 Blood Circulation and Breathing. 
 
 ALTHOUGH it does not fall within the scope of this book to deal 
 in detail with the physiology of the circulation, yet the connection 
 between breathing and circulation is so specially intimate that a 
 chapter must be devoted to this subject. Physiology is most em- 
 phatically not a subject which can be divided off into water-tight 
 compartments. 
 
 We have seen that it is with the composition of the arterial 
 blood that breathing is essentially correlated; but it has also been 
 shown in successive chapters that the amount and composition of 
 the blood returning from the tissues to the lungs play a most es- 
 sential part in determining the composition of the arterial blood, 
 and are thus intimately correlated with breathing. If, moreover, 
 the blood supply to the brain and other tissues is insufficient, or 
 the blood is abnormal in composition, the breathing is affected in 
 various ways. On the other hand circulation is intimately de- 
 pendent on breathing. If the breathing is hindered the circulation 
 is quickly affected; and, as Yandell Henderson was the first to 
 show, excessive breathing brings about failure of the circulation. 
 Thus we cannot at all fully understand how the breathing is regu- 
 lated and what part it is playing unless we understand the dis- 
 tribution of the circulating blood and the means by which its 
 composition in the tissue capillaries is regulated. 
 
 It seems evident that the most urgent and immediate need for 
 an adequate blood supply to any part of the body arises from the 
 necessity for a continuous supply of fresh oxygen. If the supply 
 of oxygen to the arterial blood is cut off in a warm-blooded animal 
 by placing it in nitrogen or hydrogen, loss of consciousness oc- 
 curs as soon as the store of oxygen in the lungs and venous blood 
 is washed out. In man eight or ten breaths suffice for this during 
 rest, and still fewer breaths during exertion. In very small ani- 
 mals, with their rapid breathing and circulation, two or three 
 seconds are sufficient; and a few seconds afterwards the heart is 
 paralyzed also. The important effects of even a slight diminution 
 in the pressure of oxygen in the arterial blood have been made 
 clear in preceding chapters. 
 
258 RESPIRATION 
 
 A second, but somewhat less urgent, need is for a continuous re- 
 moval of carbonic acid or any other acid product formed in the 
 tissues. We can probably express this generally as a need for pre- 
 venting an abnormal proportion of hydrogen ions to hydroxyl 
 ions. The effect on the central nervous system of a sudden flooding 
 with CO 2 , without deficiency of oxygen, is almost as striking, 
 though not so immediately dangerous to life, as the effect of 
 deprivation of oxygen. The results of even a slight variation in 
 arterial CO 2 pressure have often been referred to already. 
 
 Other conditions in the blood besides the diffusion pressures of 
 oxygen and CO 2 or other acid products are just as important to 
 life. For instance there are the diffusion pressure of water (inac- 
 curately identified with osmotic pressure) and the diffusion pres- 
 sures of the ions of various inorganic salts, on the importance of 
 which the investigations of Ringer and many others have thrown 
 much light. But none of these values vary in the same rapid 
 manner as the diffusion pressures of oxygen and CO 2 do ; and of 
 ordinary nutrient substances present in blood, the tissues them- 
 selves appear to possess a store which can be drawn on if the 
 supply from the blood fails for a time. The results of perfusion 
 experiments continued with the same blood indicate that if only 
 the blood is properly aerated it continues for a very long time to 
 support life in the tissues. 
 
 It would seem, therefore, that the regulation of circulation 
 through the tissues must in the main be determined in correlation 
 with the need for supplying oxygen and removing CO 2 . There 
 are evidently, however, cases where some other factor determines 
 the circulation rate. For instance, the skin circulation is de- 
 termined to a large extent in relation to the regulation of body 
 temperature; and the circulation through an actively secreting 
 gland is probably determined to a considerable extent in corre- 
 lation with local excess or deficiency of water or dissolved solids. 
 
 We can form a general idea as to what changes in gaseous 
 composition determine the circulation rate through the tissues if 
 we compare the arterial blood with the mixed venous blood re- 
 turning to the lungs. As regards this point, analyses showing the 
 difference in composition have already been quoted in Chapter V, 
 and indicate that, in the animals experimented on, the blood in 
 its passage through the tissues had lost about a third of its avail- 
 able oxygen, and gained the amount of CO 2 which would cor- 
 respond to the loss of oxygen when allowance is made for the 
 existing respiratory quotient of the animal. If we applied these 
 
RESPIRATION 259 
 
 results to man, and interpreted them in the light of the thin line 
 in the dissociation curves of oxyhaemoglobin shown in Figure 28 
 (assuming that the haemoglobin of arterial blood is 95 per cent 
 saturated) and the thick line in the corresponding curve for CO 2 
 (Figure 26) it would appear that the average pressure of oxygen 
 in the venous blood is about 5.2 per cent of an atmosphere, or 
 40 mm. of mercury, and the average pressure of CO 2 about 47 mm. 
 The experiments were, however, made on animals, while the dis- 
 sociation curves (the only accurately determined ones) are for 
 human blood. Moreover the animals, owing to operative disturb- 
 ances, anaesthetics, etc., were more or less under abnormal con- 
 ditions. Hence the inferences -just drawn are mere approxima- 
 tions. The very great variability in the CO 2 content of the samples 
 of arterial blood from animals of the same species, as compared 
 with the constancy of CO 2 content in the case of man under normal 
 resting conditions, is in itself very significant. The history of the 
 investigations detailed in the preceding chapters is sufficient to 
 warn us of the necessity for reaching more than rough approxima- 
 tions in physiological investigation, and for expecting that physio- 
 logical regulation of the circulation may turn out to be something 
 just as delicate and definite as regulation of respiration. It is to 
 measurements in man, rather than in animals, that we must look 
 for information of sufficient physiological accuracy, just as it has 
 been through measurements in man that our definite information 
 as to the regulation of breathing has been obtained. 
 
 The difficulty as regards human experiments has till quite 
 recently been that of suitable methods. We can easily measure 
 the blood pressure, pulse rate, etc., in man; but the information 
 thus obtained is extremely limited in value and almost impossible 
 to interpret satisfactorily in the absence of information as to the 
 rate of blood flow. Direct measurements of the rate of blood flow 
 in animals have been carried out by means of the Ludwig 
 "Strohmuhr" and the improved forms of it which have been 
 applied to measuring the blood flow through the aorta; but the 
 operative disturbance is far too serious to allow of sufficiently 
 definite results being obtained. Valuable information of a rough 
 kind was obtained by Zuntz and Hagemann 1 in experiments in 
 which the gases of the venous and arterial blood were determined 
 in horses, along with the total respiratory exchange, during rest 
 and work. These experiments seemed to show clearly that the 
 
 1 Zuntz and Hagemann, LancLwirtsch. Jahrb., 27, Supplem. Bd. Ill, 1898. 
 
260 
 
 RESPIRATION 
 
 general circulation rate is considerably increased during muscular 
 work, so that, in spite of the enormous increase in consumption of 
 oxygen and production of CO 2 in the body, there is still a good 
 deal of oxygen in the venous blood. 
 
 Other very interesting experiments were made on man by 
 Loewy and von Schrotter. 2 They succeeded in introducing a 
 modified Pfliiger lung catheter (Figure 68) into a branch bron- 
 chus or one of the two main bronchi in man. The supply of fresh 
 
 Figure 68. 
 
 Lung-catheter as used by Loewy and von Schrotter. The lung-catheter con- 
 sists of a central inner tube open at the lower end, and an outer tube ending 
 below in a distensible bulb which can be blown up by the rubber bag when the 
 end of the catheter is placed in position in a bronchus. By means of the syringe 
 and glass sampling tube a sample of gas from beyond the bulb can be collected 
 over mercury free of air. 
 
 air to the corresponding part of a lung, or whole lung, was thus 
 completely cut off and remained so for long periods. The breath- 
 ing, however, went on quite quietly and naturally, just as before, 
 even though all the air usually distributed to the two lungs was 
 going to only one lung. It is very significant that so little dis- 
 turbance in breathing, etc., was produced; but the fact is quite 
 easily intelligible now in the light of the preceding chapters. The 
 
 2 Loewy and von Schrotter, Die Blutcirculation beim Menschen, 1905. 
 
RESPIRATION 261 
 
 lung which remained connected with fresh air was receiving 
 much more fresh air than usual, so that the proportion of CO 2 
 in the arterial blood from this lung would be reduced practically 
 in proportion to its increased ventilation. This blood would mix 
 with the venous blood from the other lung, and in this way form 
 a mixture in which the proportion of CO 2 was about normal. The 
 arterial blood from the ventilated lung would, in virtue of the 
 higher pressure of oxygen and lower pressure of CO 2 , contain 
 slightly more oxygen than usual, while the blood from the un- 
 ventilated lung would contain considerably less. The result would 
 be a mixture containing an abnormally low proportion of oxy- 
 gen, but not sufficiently low to cause any marked immediate dis- 
 turbance. Even with a whole lung blocked off, the haemoglobin 
 of the mixed arterial blood would be at least 85 per cent saturated 
 with oxygen instead of 95 per cent, so that the effect on the breath- 
 ing would be no greater than the probable effect, hardly notice- 
 able at the time, of breathing air containing 14 per cent of oxy- 
 gen, or ordinary air at a height of about 1 1,000 feet. 
 
 Analyses of the air in the blocked lung showed that after a 
 comparatively short interval of time the percentages of oxygen 
 and CO 2 became steady, and were, in different individuals, about 
 5.3 per cent of oxygen and 6.0 per cent of CO 2 , corresponding 
 respectively to 37.5 mm. and 42 mm. These values are evidently 
 the pressures of oxygen and CO 2 in the venous blood. The low 
 value of the venous CO 2 pressure was quite unintelligible at the 
 time, since the average arterial CO 2 pressure is about 40 mm. as 
 shown above. The experiments of Christiansen, Douglas, and 
 myself (Chapter V) showed, however, that the true venous CO 2 
 pressure is in reality only a little higher than the arterial CO 2 
 pressure; and if we allow for the fact that the breathing was 
 presumably slightly increased by the stimulus of want of oxygen 
 the result is just what might be expected. The venous oxygen 
 pressure would be somewhat lower than usual, since the arterial 
 blood was incompletely saturated with oxygen. Hence both the 
 oxygen pressure and the CO 2 pressure would be below normal. 
 The results of these experiments were nevertheless of the highest 
 interest. 
 
 It is evident that if by any means we can measure the rate of 
 blood flow through the lungs, and at the same time measure the 
 intake of oxygen and discharge of CO 2 from the blood, we can 
 calculate how much oxygen a given volume of the blood gains, 
 and how much CO 2 it loses, in the lungs ; and in this way we can 
 
262 RESPIRATION 
 
 indirectly calculate how far the gain and loss vary under different 
 conditions. A rough method devised by Yandell Henderson for 
 measuring the relative rates of the blood flow was used in the 
 Pike's Peak expedition, and served to indicate that the rate of 
 blood flow remained practically normal in spite of the great alti- 
 tude. Another method, the principle of which was tried, though 
 without success, by Henderson on Pike's Peak, was about the same 
 time independently worked out and extensively used by Krogh 
 and Lindhard at Copenhagen. 3 This method gives absolute and not 
 merely relative results. The principle of the method is that the 
 lungs are filled by a very deep breath with a mixture containing a 
 considerable percentage of nitrous oxide, a gas which is very solu- 
 ble in blood. A sample of alveolar air is taken after an interval of 
 five seconds to allow the lung tissue to become saturated with the 
 nitrous oxide, and after a further interval during which the breath 
 is held, another alveolar sample. By determining the nitrous 
 oxide in the two samples, and also the total volume of gas in the 
 lungs, we find out how much nitrous oxide has been absorbed. 
 Knowing the solubility of nitrous oxide in blood, and assuming 
 also that the blood leaving the lungs is fully saturated with nitrous 
 oxide to the existing partial pressure of the gas, we can calculate 
 from the loss of nitrous oxide how much blood has passed through 
 the lungs in the given time interval. The experiment must be 
 carried out so rapidly that the venous blood continues to be free 
 of nitrous oxide. 
 
 There are various sources of probable error in this method, but 
 in the hands of Krogh and Lindhard it gave fairly consistent 
 results. They found that during rest the amount of blood circula- 
 ting through the lungs of an adult man varies from about 2.8 
 to 5 liters per minute, and that the arterial blood loses about 30 
 to 60 per cent of its available oxygen on an average, and during 
 considerable work about 50 to 70 per cent. The following table 
 gives calculated volumes of blood passing through the lungs, and 
 calculated percentage losses in the available oxygen of the blood 
 as it passes round the tissues. 
 
 It will be seen that, allowing for the fact that the haemoglobin 
 of arterial blood is only 95 per cent saturated with oxygen, the 
 haemoglobin of the venous blood was apparently only 38 per cent 
 and 53 per cent saturated in the two resting experiments. The 
 flow of blood through the lungs during work appeared to be as 
 
 1 Krogh and Lindhard, Skand. Arch. /. Physiol., XXVII, p. 100, 1912. 
 
RESPIRATION 263 
 
 much as six times as great as during rest. As the pulse rate only 
 went up to about double the normal, the volume of blood expelled 
 from the heart at each systole must, if these results were reliable, 
 have been trebled. This would be just as striking an increase as 
 occurs in the depth of breathing during muscular work. The values 
 for utilization of the available oxygen of the arterial blood are 
 
 Subject Work in kg.m. 
 
 Calculated, blood- 
 
 Percentage utilization 
 
 per minute 
 
 flow Liters per 
 
 of available O of 
 
 
 minute 
 
 arterial blood, 
 
 J. L. 
 
 2.8 
 
 60 
 
 458 
 
 9.8 
 
 73 
 
 i minute ajter 
 
 
 
 work 
 
 4-45 
 
 44 
 
 A. K. o 
 
 2.92 
 
 46 
 
 446 
 
 16.0 
 
 47 
 
 552 
 
 17.6 
 
 Si 
 
 not very far from those obtained in the horse by Zuntz and 
 Hagemann, but do not agree at all well with those of Loewy and 
 von Schrotter in man. In the case of six experiments on different 
 individuals where approximate data were available the latter 
 observers calculated a utilization of rather less than 20 per cent 
 during rest. 
 
 During or since the war several other observers have used the 
 method of Krogh and Lindhard, and obtained more or less similar 
 results. These observers include Boothby, 3A as well as Newburgh 
 and Means 3B in America. Lindhard 30 has also published a 
 number of additional results, which give, on the whole, a distinctly 
 higher rate of circulation, and lower percentage utilization of 
 oxygen, during rest. 
 
 The subject had meanwhile been approached by a quite different 
 method by Yandell Henderson. 4 He used dogs for his experi- 
 ments, and placed a recording plethysmograph round the heart 
 after removing the pericardium. By this method he found that 
 the volume of blood discharged per heartbeat was approximately 
 the same, whether the heart was beating faster or slower. Thus 
 within wide limits the volume of blood discharged per minute 
 
 3A Boothby, Amer. Journ. of Physiol., XXXVIII, p. 383, 1915. 
 
 8B Newburgh and Means, Journ. of Pharm. and, Exp. Therap., VII, p. 4, 1915. 
 
 30 Lindhard, PfHiger's Archiv. 
 
 4 Yandell Henderson, Amer. Journ. of Physiol., XVI, p. 325, 1906. 
 
264 RESPIRATION 
 
 appeared to depend almost entirely on the pulse rate. He concluded 
 that under normal conditions the heart is, practically speaking, 
 always adequately filled during diastole, although under abnormal 
 conditions the filling may become inadequate for instance when 
 the carbon dioxide of the blood is greatly reduced by excessive 
 artificial respiration. If we apply Henderson's conclusions to man 
 it is evident that they cannot be reconciled with those of Krogh 
 and Lindhard. On Henderson's theory the increased absorption of 
 oxygen and discharge of CO 2 from the blood passing through the 
 lungs during muscular exertion must be due to a very large ex- 
 tent to greater utilization of the oxygen in the blood passing 
 round the body, and a corresponding increase in its charge of CO 2 . 
 The rate of circulation can only be increased in proportion to in- 
 creased pulse rate, the discharge of blood per systole remaining 
 about the same. 
 
 There is no question that the systolic discharge may, at least 
 under abnormal conditions, vary enormously. This was very 
 clearly shown by the experiments of Starling and Patterson, 5 with 
 a "heart-lung preparation" i.e., a preparation in which the 
 only circulation was through the lungs and heart, the lungs being 
 ventilated so as to insure full oxygenation of the blood. By vary- 
 ing the venous blood pressure, the systolic discharge could be 
 varied tenfold, without any variation in the pulse rate. It does not 
 follow, however, that there are corresponding variations in systolic 
 discharge in normal men and animals with the organic regulation 
 of circulation not thrown out as in the case of a heart-lung prepa- 
 ration. 
 
 In the nitrous oxide method there are various sources of pos- 
 sible very serious error which can hardly be discussed in detail 
 here. In order to get a more direct and accurate insight into the 
 venous gas pressures and their relation to blood flow, a new 
 method was introduced by Christiansen, Douglas, and myself. 6 
 In the first application of this method we simply determined the 
 CO 2 pressure of the venous blood after oxygenation but without 
 its losing any CO 2 . As we had already discovered (see Chapter 
 V), this pressure is higher by an easily calculable amount than 
 that for the unoxygenated venous blood. Mixtures containing 
 about the required percentage of CO 2 were prepared by adding 
 CO 2 to air. A deep breath of one of these mixtures was taken in 
 
 5 Starling and Patterson, Journ. of Physwl., XLVIII, p. 357, 1914. 
 9 Christiansen, Douglas, and Haldane, Journ. of Phystol., XLVIII, p. 244, 
 1914. 
 
RESPIRATION 265 
 
 after previously expiring deeply. After two seconds part of the 
 air in the lungs (about i l / 2 liters) was expired, so as to obtain a 
 sample of alveolar air. The rest of the breath was held for five 
 seconds and a second sample of alveolar air was then taken. If 
 these two samples gave practically the same percentage of CO 2 , 
 the CO 2 in the alveolar air was evidently in pressure equilibrium 
 with the CO 2 of the oxygenated venous blood. If too much CO 2 
 were present in the alveolar air the second sample would contain 
 less CO 2 than the first, and if too little, more. We were thus using 
 the whole of both lungs as an aerotonometer. For any particular 
 person it was easy to find the mixture which gave equilibrium. 
 With the help of Figure 26 (Chapter V) we could then calculate 
 the CO 2 content of the venous blood and the true value of the 
 venous CO 2 pressure. We could also calculate how much CO 2 the 
 blood had' taken up in passing round the body if we knew the 
 normal alveolar CO 2 pressure. The following table shows the 
 results obtained during complete rest in a sitting position with the 
 four subjects investigated. 
 
 Subject 
 
 Arterial COz pressure 
 
 Venous COz pressure 
 
 Difference 
 
 
 in mm. Hg. 
 
 in mm. H g. 
 
 
 J. C. 
 
 34-9 
 
 41.8 
 
 6.9 
 
 J. S. H. 
 
 40.6 
 
 45-6 
 
 5-0 
 
 C. G. D. 
 
 39-7 
 
 44-4 
 
 4-7 
 
 J. G. P. 
 
 40.4 
 
 45-1 
 
 4-7 
 
 Mean 
 
 38.9 
 
 44-2 
 
 5-3 
 
 Reference to Figure 26 shows that on an average the venous 
 blood had only taken up about 24 per cent of the CO 2 which it 
 would have taken up if all its available oxygen had been used up. 
 Hence the blood had only lost about 24 per cent of its oxygen in 
 passing round the circulation; and in the three male subjects the 
 proportion lost was only about 21 to 22 per cent. This indicates a 
 much faster circulation rate during rest than the nitrous oxide 
 method had shown. 
 
 At the outbreak of war, Dr. Douglas and I were engaged 
 in carrying these experiments further; but as he volunteered at 
 once for active service they were interrupted; and owing to the 
 disorganization following the war they are not yet completed, 
 though I was able to carry them on up to a certain point with 
 help from Dr. Mavrogardato, and to communicate a number of 
 
266 RESPIRATION 
 
 results to the Physiological Society in 1915. We had been engaged 
 in measuring directly both the true venous CO 2 pressure and 
 oxygen pressure just after forced breathing, so as to discover the 
 effects of lowered CO 2 pressure on the circulation. We found that 
 the apparent venous oxygen pressures were incredibly high 
 70 mm. or even more. On further investigation it became evident 
 that after a single deep expiration, followed by a single deep in- 
 spiration of the gas mixture, the air in the alveoli was not properly 
 mixed. At the end of the forced breathing there would be nearly 
 20 per cent of oxygen in the alveolar air. With one deep inspira- 
 tion of the mixture, the air in the air-sac system of alveoli was 
 mingled with air from the inspired mixture, but an even mixture 
 in all parts of the alveolar system was not obtained, so that the 
 air-sac alveoli contained considerably more oxygen than the rest 
 of the alveoli. As a consequence the second alveolar air sample, 
 taken more exclusively from the air-sac alveoli, contained more 
 oxygen than the first, in spite of the fact that it had remained 
 longer in the lungs. It was evidently necessary, therefore, to take 
 two or, in the case of forced breathing, three successive deep 
 breaths of the mixture before holding the breath and taking 
 the samples. When this was done the results were quite consistent, 
 and showed that the venous CO 2 pressures as determined directly 
 during rest confirmed the calculated values previously obtained; 
 while the venous oxygen pressures, when interpreted in the light 
 of the thin-line curve of Figure 28, corresponded very closely 
 with the percentage oxygen loss of the blood as calculated indi- 
 rectly from the venous CO 2 pressure. Moreover, not only the 
 venous CO 2 pressure, but also the venous oxygen pressure, was 
 considerably lower at the end of forced breathing. 
 
 The following are examples of two typical experiments carried 
 out on myself at the end of ten minutes' rest on a chair. 
 
 No. i, 26/2/15. Bar. 762 mm. 
 
 Mixture used contained 6.21 per cent of CO 2 and 5.73 per cent of 
 oxygen. 
 
 First alveolar sample 2" after last deep inspiration, 6.43 per cent 
 of CO 2 and 6.18 per cent of oxygen. 
 
 Second alveolar sample 5" after first sample, 6.47 per cent of CO 2 
 and 6.22 per cent of oxygen. 
 
 Therefore venous CO 2 pressure = 6.47 per cent = 46.16 mm. and 
 oxygen pressure 6.22 per cent = 44.5 mm. 
 
 Normal alveolar CO 2 percentage (mean of inspiratory and expira- 
 tory samples) 5.64 per cent = 40.3 mm. 
 
RESPIRATION 267 
 
 Metabolism (by Douglas Bag method) = 330 cc. of CO 2 and 379 
 cc. of oxygen (at o and 760 mm.) per minute. 
 
 As the venous CO 2 pressure was 6.O mm. above the arterial, 
 the blood (calculating from Figure 26) had gained 4.2 per cent 
 by volume of CO 2 . Hence the circulation rate calculated from CO 2 
 
 was - = 7.9 liters per minute. As the venous oxygen pressure 
 
 was 44.5 mm. and this corresponds, calculating from Figure 28, 
 to 73 per cent saturation of the haemoglobin, the blood had lost 
 about 22 per cent of its combined oxygen. Adding the correspond- 
 ing small amount of dissolved oxygen this corresponds to a loss 
 of about 4.3 volumes per cent of oxygen. Hence the circulation 
 
 rate, calculating from the oxygen, was - = 8.8 liters per 
 
 43 
 minute. 
 
 No. 2. 27/2/1$. Bar. 752 mm. 
 
 Mixture used contained 6.26 per cent of CO 2 and 5.26 per cent of 
 
 oxygen. 
 First alveolar sample 2" after last deep inspiration, CO 2 = 6.26 
 
 per cent and O 2 = 6.25 per cent. 
 Second alveolar sample 5" after first sample, CO 2 = 6.30, O 2 = 
 
 6.09 per cent. 
 Therefore venous CO 2 pressure = 6.30 per cent = 44.4 mm. ; and 
 
 oxygen pressure 6.09 per cent = 42.9 mm. 
 
 Normal alveolar CO 2 pressure (mean) = 5.55 per cent = 39.1 mm. 
 Metabolism 332 cc. of CO 2 and 374 cc. of oxygen absorbed (at o 
 
 and 760 mm.) per minute. 
 
 As the venous CO 2 pressure was 5.3 mm. above the arterial, the 
 blood (calculating from Figure 26) had gained 3.7 volumes per 
 cent of CO 2 . Hence the circulation rate calculated from CO 2 was 
 
 - = 9.0 liters per minute. As the venous oxygen pressure was 
 o / 
 
 42.9 mm., and this corresponds (Figure 28) to 70 per cent satura- 
 tion of the haemoglobin, the blood had lost about 25 per cent of 
 combined oxygen or about 4.9 volumes per cent of oxygen. Hence 
 
 the circulation rate, calculating from the oxygen, was = g.o 
 
 47 
 liters per minute. 
 
 If we take these two experiments together, the circulation rate 
 determined from the CO 2 was 8.45 liters per minute, and from 
 
268 RESPIRATION 
 
 the oxygen 8.40 liters, the general mean being 8.4 liters. As my 
 pulse rate was 80 to 85 per minute this means that just about 100 
 cc. of blood were delivered at each heartbeat; and as my blood 
 volume is about 4.8 liters (see p. 280 of the Pike's Peak Expedi- 
 tion's Report) a volume of blood equal to that in the whole body 
 was passing round every 35 seconds. 
 
 This is a much higher rate than has usually been calculated in 
 recent years, but not higher than what the data of Loewy and 
 von Schrotter indicate. There are so many sources of probable 
 error in the nitrous oxide method, 7 that I do not think that much 
 stress can be laid on the lower estimates which this method has 
 given during the resting condition. Nevertheless it is already 
 evident from our experiments that considerable individual dif- 
 ferences exist in the resting circulation rate in man; and it is 
 probable that under abnormal conditions both the circulation 
 rate and the delivery per beat vary considerably even in persons 
 of the same weight. 
 
 At different times we have found very little difference in the 
 resting venous gas pressures of the same individual. These gas 
 pressures seem to be not much less steady during rest under 
 normal conditions than the arterial gas pressures. It is very dif- 
 ferent, however, during exertion. The smallest muscular exertion 
 raises the venous CO 2 pressure, and the rise is far more than 
 corresponds to the comparatively slight rise in arterial CO 2 pres- 
 sure as measured in the ordinary way in the alveolar air. Hence 
 it is now perfectly certain that the general circulation rate does 
 not increase in anything like direct proportion to increased me- 
 tabolism. Even with moderate exertion (about a third the maxi- 
 mum possible) on a Martin's ergometer, the difference between 
 arterial and venous CO 2 pressure became about two and one-half 
 times as great as usual, so that the venous blood could not be more 
 than about 45 per cent saturated with oxygen. So far as we can 
 calculate there is sometimes more increase in circulation than can 
 be accounted for by increased pulse rate ; but the increase is seldom 
 
 7 For instance, it seems very probable that while the breath is held in perform- 
 ing an experiment the blood flow to the heart, and consequently through the lungs, 
 is temporarily diminished. Krogh and Lindhard, misled, as we believe, by the 
 imperfect mixture of oxygen in the alveolar air in their experiments, estimated that 
 there is a greatly increased absorption of oxygen, and a corresponding abnormal 
 increase in circulation, while the breath is held ; and their results are corrected 
 accordingly. The correction, which is a large one, does not seem to us to be war- 
 ranted, and without it their results come much closer to ours. This is especially 
 true for Lindhard's later results. 
 
RESPIRATION 269 
 
 great. Roughly speaking, therefore, our results confirm those 
 obtained by Henderson on the dog. 
 
 Henderson and Prince have determined in a number of persons 
 the oxygen consumption per beat of the heart, or what they call 
 for brevity "the oxygen pulse." 8 This value is obtained by simply 
 dividing the oxygen consumption per minute by the pulse rate. 
 Figure 69 shows graphically a fairly typical example of their 
 
 g 
 
 25 25oo 
 
 20 2000 
 
 15 '500 
 
 JO 'Ooo 
 
 5 500 
 
 Pulse 60 70 &o 90 100 no 120 130 140 150 160 
 
 Figure 69. 
 
 Subject Y. H., Weight 75 kilos. Haemoglobin 107. In this diagram the 
 broken line expresses the oxygen consumption per minute, the dotted line the 
 CO 2 elimination, and the solid line the oxygen pulse. During the short periods 
 of vigorous exertion and rapid heart rates, the CO2 elimination was increased 
 to a greater extent than the oxygen consumption, the respiratory quotient even 
 rising above unity in some cases, and indicating an excessive blowing off 
 of C0 2 . 
 
 results. It will be seen that with low oxygen consumption per 
 minute the oxygen consumption per beat is low, but increases 
 rapidly up to a maximum as the oxygen consumption per minute 
 increases owing to muscular exertion. When, however, this maxi- 
 mum is reached, further increase of the oxygen consumption per 
 minute causes no increase in the oxygen consumption per beat. 
 Interpreting these data in the light of our own experiments on 
 man, and Henderson's former experiments on the heart of the dog, 
 the increased oxygen consumption per beat is not due to any 
 marked extent to increased output of blood per beat, but to in- 
 creased utilization of the charge of oxygen in the arterial blood. 
 
 8 Yandell Henderson and Prince, Amer. Journ. of Physiol., XXXV, p. 106, 1914. 
 
2;o RESPIRATION 
 
 When this increased utilization reaches its physiological limit, 
 further increase in the oxygen consumption per minute can only 
 be obtained by increase in the rate of heartbeat. 
 
 The mixed venous blood returning to the heart comes from 
 various parts of the body; but during muscular exertion a very 
 greatly increased proportion must come from the muscles. Now 
 there is evidence from a series of experiments by Leonard Hill 
 and Nabarro that the venous blood returning from the muscles 
 contains even during rest far less oxygen and more CO 2 than at 
 any rate the venous blood returning from the brain. 9 Without 
 obstructing the vessels they collected venous blood returning from 
 muscles through the deep femoral vein, and from the brain 
 through the torcular Herophili in the dog. The following table 
 shows the average of about eight determinations in each case. 
 
 OXYGEN, VOLUMES Percentage 
 
 PER CENT 
 
 loss of 
 
 [Muscle 
 Rest \ 
 
 Artery 
 I8.IO 
 
 Vein 
 5-12 
 
 Difference 
 12.98 
 
 oxygen 
 72 
 
 [Brain 
 
 16.81 
 
 13.39 
 
 3.42 
 
 20 
 
 Tonic [Muscle 
 fit \ 
 
 17-05 
 
 3.30 
 
 13.75 
 
 8l 
 
 [Brain 
 
 15-17 
 
 10.22 
 
 4-95 
 
 32 
 
 Clonic ("Muscle 
 fit \ 
 
 18.66 
 
 6.03 
 
 12.63 
 
 6 9 
 
 [Brain 
 
 15.77 
 
 11.46 
 
 4.3i 
 
 27 
 
 It will be seen ( I ) that during rest the blood lost three and one- 
 half times as much of its charge of oxygen in the muscles as in 
 the brain; (2) that during the intense activity of a tonic or clonic 
 fit (produced by absinthe) the percentage loss of oxygen by the 
 blood was only slightly increased in either the brain or the muscles. 
 The animals were anaesthetized with morphia or chloroform, so 
 it is possible that the circulation was less active than in normal 
 animals; but the difference between the brain circulation and 
 that through muscles is none the less striking. 
 
 In the light of these experiments we can see what is presumably 
 happening as regards the mixed venous blood during muscular 
 
 9 Leonard Hill and Nabarro, Journ. of Physiol., XVIII, p. 218, 1895. 
 
RESPIRATION 271 
 
 activity. The chief reason why the oxygen diminishes and CO 2 
 increases so strikingly is that the mixed venous blood contains 
 a much larger proportion of blood from muscles, and that this 
 blood is very poor in oxygen whether the muscles are working or 
 not. During rest the mixed venous blood will contain but little 
 blood from the muscles, and a large proportion from the brain 
 and probably other parts which furnish venous blood relatively 
 rich in oxygen. As indicated by the size of its arteries, the brain 
 has a very rich blood supply, going mainly to the gray matter. 
 Its normal oxygen pressure is evidently very high; and this 
 renders intelligible the fact that it is so sensitive to deficient satu- 
 ration of the arterial blood with oxygen. The rapid circulation 
 explains the promptness of its reaction to changes in quality of 
 the arterial blood. 
 
 The fact that during muscular exertion the mixed venous blood 
 contains much less oxygen and more CO 2 explains why, if the 
 breath is voluntarily held during exertion, the alveolar CO 2 
 percentage shoots up much higher than if it is held for a far 
 longer time during rest. It also explains what would otherwise 
 be a very puzzling fact with regard to congenital heart affections 
 ("morbus coeruleus"). In cases of morbus coeruleus the face 
 becomes intensely blue on muscular exertion. Quite evidently the 
 arterial blood is very imperfectly oxygenated ; and Douglas and I 
 found that the blueness continues even if the patient breathes pure 
 oxygen during the exertion. The blueness is due to part of the 
 venous blood short-circuiting through a congenital direct com- 
 munication between the right and left sides of the heart, so that 
 the mixed arterial blood always contains a certain proportion of 
 unae' rated venous blood. During rest this venous blood contains 
 so much oxygen that the cyanosis is only slight ; but during exer- 
 tion, with much less oxygen in the venous blood, the cyanosis is of 
 course far more marked, and the breathing of oxygen avails very 
 little towards redressing the balance. 
 
 It is evident from the facts just referred to that the increase in 
 blood flow through the lungs during exertion is very much less 
 than the increase in air breathed. At first sight, therefore, it might 
 seem that the regulation of circulation differs fundamentally from 
 the regulation of breathing. A little consideration, however, shows 
 that there are no real grounds for this conclusion. If we take as our 
 measure, not the blood flow through the heart, but the blood flow 
 through individual parts of the body, the facts so far discussed do 
 not point to any other conclusion than that the blood flow, just 
 
272 RESPIRATION 
 
 like the breathing, is delicately regulated in accordance with the 
 local requirements for the supply of oxygen and removal of CO 2 . 
 
 The idea that the local circulation is regulated in accordance 
 with the local CO 2 pressure was brought forward in a very definite 
 form by Yandell Henderson in a series of papers on "Acapnia and 
 Shock." 10 He showed, firstly, that the local circulation and func- 
 tional activity in the exposed intestines depends upon the main- 
 tenance in them of a sufficient pressure of CO 2 , and secondly, that 
 on the removal of an excessive quantity of CO 2 from the body by 
 excessive artificial or natural respiration the circulation fails, 
 whereas excessive ventilation with air to which sufficient CO 2 has 
 been added produces no such effect. These are evidently facts of 
 fundamental importance as regards the regulation of the circula- 
 tion, and as showing the intimate connections between respiration 
 and circulation. On these and other observations he also based the 
 theory that the immediate cause of shock may be excessive res- 
 piratory activity. 
 
 The blood-gas changes caused by excessive artificial respira- 
 tion were first investigated by Ewald in connection with apnoea. 11 
 He not only found that there is a slight excess of oxygen and very 
 large deficiency of CO 2 in the arterial blood, but also (though of 
 this he did not realize the significance) that there is great de- 
 ficiency of both CO 2 and oxygen in the mixed venous blood. The 
 changes in the arterial blood have already been discussed in earlier 
 chapters, and it was pointed out in Chapter VII that owing to the 
 deficiency of CO 2 a state of anoxaemia must, other things being 
 equal, be produced by forced breathing. Ewald's analyses show, 
 however, that there is something more to cause anoxaemia than 
 mere deficiency of CO 2 . The latter would not by itself account 
 for the deficiency of oxygen combined with haemoglobin in the 
 venous blood. In long experiments Ewald found this oxygen down 
 to about a third of the normal, and the CO 2 down to half the 
 normal. Taking into account both the direct effect of deficiency 
 of CO 2 in diminishing the free oxygen present in the venous blood, 
 and the effect in the same direction of the diminished proportion 
 of oxyhaemoglobin present, the artificial respiration must have 
 brought about a condition of very intense anoxaemia in the tis- 
 sues. But the diminution in the proportion of oxyhaemoglobin 
 
 10 Yandell Henderson, Amer. Journ. of Physiol., XXI, p. 126, 1908; XXIII, 
 p. 345, 1909; XXIV, p. 66, 1909; XXV, p. 310, 1910; XXV, p. 385, 1910; 
 XXVI, p. 260, 1910; XXVII, p. 152, 1910; XLVI, p. 533, 1918. 
 
 "Ewald, Pfluger's Archiv., VII, p. 575, 1873. 
 
RESPIRATION 273 
 
 cannot have been due to any other cause than diminution in the 
 circulation rate; and this diminution is shown far more directly 
 by Yandell Henderson's experiments and numerous blood-gas 
 analyses by the ferricyanide method. The diminution in circula- 
 tion goes so far that the venous return to the heart becomes quite 
 inadequate to fill the ventricles. Hence arterial as well as venous 
 pressure finally falls, and the heart itself is inadequately supplied 
 with free oxygen or CO 2 , and gradually fails along with fail- 
 ure in the brain and other parts of the body. 
 
 Slowing of the circulation through the hands during forced 
 breathing was clearly demonstrated by his calorimetric method by 
 G. N. Stewart. 11A 
 
 By means of the new method for determining venous gas pres- 
 sures in man we found that though there is a considerable fall, 
 after forced breathing for about three minutes, in the CO 2 con- 
 tent of the mixed venous blood, there is, relatively speaking, an 
 even greater fall in the oxygen content. The experiments were 
 difficult because of the mental state of the subject. I had to be 
 watched very closely to see that I carried out the proper manipula- 
 tions, and many experiments failed because of gross errors, such 
 as taking in a deep breath of ordinary air from the room. The gas 
 mixture used had to contain less than 4 per cent of oxygen and 
 less than 5 per cent of CO 2 . The fall in oxygen pressure was con- 
 siderably more than could be accounted for as due to the fall in 
 CO 2 pressure on account of the Bohr effect. Hence the circulation 
 rate was diminished. The mental condition was apparently due to 
 marked anoxaemia of the nervous centers ; and it may be remarked 
 that owing to the rapid normal circulation through the brain the 
 effects of the forced breathing must be felt there sooner than else- 
 where. 
 
 We also investigated the effect on the circulation of a moderate 
 excess of CO 2 , sufficient to increase the breathing to about five 
 times the normal. This was easily accomplished in a respiration 
 chamber in which the CO 2 percentage had been raised to a little 
 over 5 per cent. Under this condition there was a slight rise in 
 both my arterial and venous CO 2 pressure; but the difference 
 between them was not diminished. Thus there had been no ap- 
 preciable increase in the circulation rate. It was quite clear that 
 the circulation does not increase with increased arterial CO 2 pres- 
 sure in a manner corresponding to the increase of breathing. The 
 
 UA G. N. Stewart, Amer. Journ. of Physiol., XXVIII, p. 190, 1911. 
 
274 RESPIRATION 
 
 breathing had increased five times or more, but the circulation 
 had apparently not increased at all. The pulse, etc., were also 
 hardly affected. With a great excess of CO 2 , however, the ve- 
 nous return to the right heart is evidently much increased. This 
 was first definitely observed by Yandell Henderson, who also 
 makes the, to me, interesting remark that he first noted the signs 
 of increased circulation rate on myself, while I was nearly over- 
 come by accumulation of CO 2 in a mine- rescue apparatus, without 
 any deficiency of oxygen. 12 Similarly, great deficiency of CO 2 , as 
 in forced breathing or excessive artificial respiration, will dim- 
 inish the circulation rate; and it seemed probable that great in- 
 crease in the oxygen pressure in the tissues (though this is difficult 
 to produce except under the high atmospheric pressures referred 
 to in Chapter XII) would have a similar effect. 
 
 That this effect is actually produced in man is indicated by the 
 results of quite recent experiments by Dautrebande and myself. 13 
 We found that when pure oxygen was breathed, particularly under 
 a barometric pressure increased to two atmospheres, the breathing 
 increases, as shown by a fall in alveolar CO 2 pressure, and there 
 is a simultaneous slowing of the pulse. This indicated a slowing 
 of circulation through the brain, such as would compensate for 
 the high oxygen pressure of the arterial blood. The slowing would 
 of course raise the pressure of CO 2 in the brain, and thus increase 
 the breathing. It would also explain the fact that though oxygen 
 at two atmospheres pressure has a rapid poisonous action on the 
 lungs and other living tissues directly exposed to it (see Chapter 
 XII), there are no evident cerebral symptoms until oxygen at 
 much higher pressures is breathed. 
 
 The responses involved in the chemical control of the venous 
 return to the right heart were found by Henderson and Harvey to 
 be peripheral, but independent of the vasomotor nerves and nerve 
 endings. In the "spinal" cat they found that slow injections of 
 adrenalin, and other prolonged vasomotor stimulations, cause a 
 maintained elevation of arterial pressure, but only an evanescent 
 rise of venous pressure. Ventilating the lungs with air rich in CO 2 
 (with ample oxygen) has, on the contrary, in the absence of the 
 medullary vasomotor center, no appreciable direct effect upon 
 arterial pressure, but induces a gradual, sustained and large eleva- 
 tion of venous pressure. They note also that during this action 
 
 "Yandell Henderson and Harvey, Amer. Journ. of PAysiol., XLVI, p. 533, 
 1918. 
 
 18 Dautrebande and Haldane, Journ. of Physiol., LV, p. 296, 1921. 
 
RESPIRATION 275 
 
 the veins are always relaxed, as well as distended ; and they con- 
 sider that the easier escape of the blood from the tissues, due to 
 relaxation especially of venules, is the cause of the larger venous 
 return and consequent rise of venous pressure. Recently Hender- 
 son, Haggard, and Coburn 14 have shown that inhalation of air 
 containing 6 or 8 per cent of CO 2 has a powerful restorative effect 
 upon the circulation, and particularly upon the venous pressure, 
 in patients after prolonged anaesthesia and major surgical opera- 
 tions. 
 
 With great deficiency of oxygen there is also at first a very 
 marked increase in the circulation rate. This is shown by the 
 greatly increased pulse rate, deep blue flushing of the skin, etc., 
 and great rise of venous blood pressure when air very deficient in 
 oxygen is breathed. In rapid poisoning by CO there is the same 
 flushing of the skin and distention of large veins, though the color 
 is now red and not blue. The increased pressure in the great veins 
 causes the distention of the right side of the heart and rapid pro- 
 duction of oedema of the lungs so characteristic of acute asphyxia, 
 although but for the fact that the heart muscle is lamed by the 
 anoxaemia there would probably be no over-distention. As Star- 
 ling and Knowlton found, oedema of the lungs and over-disten- 
 tion of the right side of the heart are very quickly produced by a 
 quite moderate increase of the ordinary very low venous pressure 
 at the entry to the heart. 15 With moderate oxygen deficiency, pro- 
 duced rapidly, there are, just at first, distinct signs of increased 
 circulation as well as of increased respiration; but very soon the 
 increased washing out of CO 2 from the blood moderates both the 
 breathing and circulation, and after a short time the circulation, 
 as well as the breathing, quiets down, so that unless the anoxaemia 
 is considerable the increased pulse rate and other signs of in- 
 creased circulation may have practically disappeared. 
 
 The circulation during and just after forced breathing in man 
 was meanwhile investigated by a quite different method by Hen- 
 derson, Prince, and Haggard. 16 They measured the venous pres- 
 sure by observing the height of the column of blood in a vein of 
 the arm when the subject was placed in a head down position on a 
 sloping board (Figure 70) , thus obtaining a measure of the venous 
 
 14 Henderson, Haggard, and Coburn, Journ. Amer. Med. Assn., LXXIV, p. 783, 
 1920. 
 
 15 Starling and Knowlton, Journ. of Physwl., XLIV, p. 206, 1914. 
 
 18 Yandell Henderson, Prince, and Haggard, Journ. of Pharmac. and, Exper. 
 Therapeutics, XI, p. 203, 1918. 
 
2 7 6 
 
 RESPIRATION 
 
 blood pressure at the entry to the heart. The effect of forced 
 breathing was to cause a great diminution in venous blood pres- 
 sure. Thus the supply of blood to the heart must have become 
 inadequate to fill the right ventricle. Owing, however, to the 
 diminished outflow of blood from the arterial system there was 
 no fall in arterial blood pressure. It seems to be only when the 
 anoxaemia of forced breathing becomes so intense as to affect the 
 heart muscle seriously that the arterial blood pressure falls. 
 
 Figure 70. 
 
 Measurement of venous blood pressure by placing subject in a head-down 
 position. 
 
 Putting all these facts together, it appears that in general the 
 circulation is so regulated as to keep the pressures of both oxygen 
 and CO 2 approximately steady in the venous blood from any 
 particular organ. The regulation is evidently of a double kind, 
 involving both oxygen and CO 2 . If the oxygen pressure goes 
 down and the CO 2 pressure also goes down, as in a pure anox- 
 aemia, there is comparatively little effect on the circulation rate, 
 because increase due to the lowered oxygen pressure is at once 
 counteracted by the effect of diminution due to the lowered CO 2 
 pressure. Similarly, in an atmosphere containing simple excess of 
 CO 2 increased circulation due to the excess of CO 2 pressure tends 
 to be counteracted by decrease due to increased oxygen pressure. 
 During muscular work, on the other hand, there is both a rise of 
 CO 2 pressure and fall of oxygen pressure, and consequently a 
 
RESPIRATION 277 
 
 great increase in blood flow through the muscles, with a corre- 
 sponding increase in venous blood pressure, as Henderson and 
 his colleagues found with the apparatus shown in Figure 7O. 17 
 
 The correspondence between blood flow and amount of work 
 done by a muscle seems to appear clearly in data obtained by 
 Markwalder and Starling for the coronary circulation with vary- 
 ing work of the heart in a heart-lung preparation. 17A The amount 
 of blood pumped by the heart, the aortic blood pressure, and the 
 flow through the coronary vessels, were measured simultaneously. 
 The data show that if the work done is estimated by the amount 
 of blood pumped multiplied by the aortic pressure, the coronary 
 blood flow varied within wide limits in proportion to the work 
 done. The variations in coronary blood flow might, of course, be 
 attributed to the variations in aortic blood pressure, but this inter- 
 pretation does not seem to explain more than a small part of the 
 facts. 
 
 At first sight the regulation of the circulation appears to be 
 different from that of respiration, since in the case of the latter 
 the influence of CO 2 predominates. This, however, is simply be- 
 cause when ordinary air is breathed the oxygen pressure in the 
 tissues is not increased when the breathing increases. In reality, 
 there is no fundamental difference. Whenever anoxaemia is pres- 
 ent the respiratory regulation, as already shown in Chapter VII, 
 works just like the local circulatory regulation. The breathing is 
 not then free to increase in such a way as to compensate approxi- 
 mately for increasing anoxaemia, because increased breathing 
 lowers the CO 2 pressure and this tends to diminish the breathing. 
 Similarly the breathing cannot increase freely with increased 
 CO 2 pressure, because the increased breathing would diminish 
 the anoxaemia. Under deep anaesthesia, when the arterial blood 
 becomes dark, CO 2 has very little effect on the breathing. 
 
 There can be little doubt that in the case of circulation, just as 
 in that of respiration, increase in CO 2 pressure stands simply for 
 increase in hydrogen ion concentration. Hence alkalosis due to 
 deficiency of CO 2 in the systemic capillaries, or acidosis due to 
 excess, will tend to be relieved by the slow acclimatization changes 
 described in Chapter VIII. 
 
 When once the fundamental fact is grasped that the general 
 flow of blood throughout the body is correlated with the gas pres- 
 
 17 Yandell Henderson and Haggard, Journ. of Pharmac. and, Exper. Therap. t 
 XI, p. 197, 1918. 
 
 17A Markwalder and Starling, Journ. of Physiol., XLVII, p. 279, 1913. 
 
278 RESPIRATION 
 
 sures in the capillaries, the whole physiology of the circulation 
 appears in a new light. It is not the heart nor the bulbar nervous 
 centers which govern the circulation rate, but the tissues as a 
 whole; and they govern it with an accuracy and delicacy com- 
 parable to the accuracy and delicacy with which they govern 
 breathing. The heart and vaso-motor system are only the executive 
 agents which carry out the bidding of the tissues, just as the lungs 
 and nervous system do in the case of breathing. 
 
 It appears that the immediate function of the heart is not to 
 regulate the circulation rate, but simply to pass on at a greatly 
 increased pressure the blood supplied to it. The problem of the 
 regulation of the circulation under normal conditions seems in the 
 main to resolve itself into that of the regulation by the tissues of 
 the amount of blood supplied to the heart; and this regulation 
 depends, as we have just seen, to an overwhelming extent on a 
 linked control by the oxygen pressure and hydrogen ion concen- 
 tration in the systemic capillaries. 
 
 Just as in the case of regulation of breathing, so also in the 
 case of regulation of the circulation, the dominant facts have been, 
 and still are, obscured by masses of detail which, in their un- 
 connected form, simply confuse the mind and lead to wholly 
 mistaken judgments. It is difficult to pick a way through all these 
 details, but the salient points concerning the immediate control of 
 the heart's action must now be referred to. 
 
 We owe mainly to Gaskell the demonstration that the muscular 
 fibers of the heart may continue to contract rhythmically apart 
 from nervous control and even when they are separated from one 
 another, just as the rhythmic activity of the respiratory center 
 continues apart from peripheral nervous control. When, however, 
 different parts of the heart are separated from one another, the 
 frequency of the contractions in the different parts is different, 
 the ventricular contracting less frequently than the auricular 
 parts. In lower vertebrates the order of frequency in contractions 
 is sinus venosus, auricle, ventricle, and bulbus arteriosus. More- 
 over the individual fibers in each separated part contract normally 
 in unison with one another so long as they are not separated. In a 
 normal intact heart, however, not only do the individual fibers in 
 sinus venosus, auricles, ventricles, and bulbus arteriosus contract 
 in unison, but so also do all the parts of the heart. 
 
 The explanation of this contraction in unison has been furnished 
 by the physiological and clinical investigations of the last few 
 years. As was shown by Lewis with the help of the string gal- 
 
RESPIRATION 279 
 
 vanometer, each normal contraction starts in what is known as 
 the Keith-Flack node, an island of primitive sinus venosus tissue 
 in the right auricle. Thence it is conducted by primitive muscular 
 tissue to the auricles, and by a bundle of similar muscular tissue, 
 the bundle of Kent or His, to the ventricles. This primitive tissue 
 is distributed (as the fibers of Purkinje) over the ventricles, and 
 has a conduction rate far faster than the rest of the muscular tis- 
 sue of the heart. Thus all parts of the ventricles contract almost 
 simultaneously, and shortly after the almost simultaneous con- 
 traction of all parts of the auricles; while the pace of the whole 
 heart is set by the contractions starting in the Keith-Flack node. 
 Impairment or total failure in the conduction from auricle to 
 ventricle, or from fiber to fiber in auricle or ventricle, explains 
 many of the peculiarities met with in heart affections. 
 
 So long as the contractions of the ventricles are complete, the 
 volume of blood discharged at each beat must depend on the ex- 
 tent to which the right ventricle fills in diastole. This, in turn, 
 depends on the rate at which blood is let through from the arteries 
 to the veins. The difference between arterial and venous pressure 
 is so great that accessory factors such as the pumping movements 
 of respiration can hardly have more than a very minute average 
 influence on the circulation, though they have a marked tempo- 
 rary influence. It is therefore the rate at which the systemic 
 blood is allowed to pass through the tissues into the venous system 
 that determines the amount of blood pumped by the heart; and, 
 as already pointed out, the rate at which blood is allowed to pass 
 through the tissues is determined by their metabolic requirements, 
 and particularly by the amount of blood required to keep the 
 diffusion pressures in them of oxygen and carbonic acid approxi- 
 mately steady. 
 
 It is evident that in the carrying out of this regulation, both by 
 the heart and the blood vessels, the nervous system plays a very 
 important part, just as in the case of regulation of breathing; but 
 the main fact must never be lost sight of that the primary factor 
 in determining the rate of circulation is neither the heart nor the 
 nervous centers specially connected with the circulation, but the 
 metabolic activities of the tissues. At bottom the regulation of the 
 circulation is a chemical regulation, just as in the case of the 
 breathing. 
 
 The frequency and strength of the heartbeats are moderated 
 through the central nervous system, first by the well-known in- 
 hibitory impulses passing to the heart through the vagus nerve, 
 
2 8o RESPIRATION 
 
 and secondly by the equally well-known accelerator impulses 
 passing to the heart through sympathetic branches. Increased 
 liberation of inhibitory impulses has been found to be a direct re- 
 sult of rise of arterial blood pressure (so that the inhibition tends to 
 prevent an excessive rise of arterial pressure and consequent fa- 
 tigue of the heart or over-distention of arteries), but is certainly 
 also a result of rise in oxygen pressure and diminution in CO 2 
 pressure in the blood passing through the brain. An increase of 
 arterial blood pressure will, therefore, owing to the increased 
 rate of circulation, slow the heart. When the arterial blood pres- 
 sure is normal there is a considerable amount of vagus inhibition, 
 so that on section of the vagi the heartbeats quicken. It appears 
 also that this tonic nervous inhibition of the heart is itself reflexly 
 inhibited, either directly or indirectly, by increase of pressure on 
 the great veins opening into the heart. This was recently shown by 
 Bainbridge, 18 who found that, even if the accelerator nerves are 
 cut, increase in venous pressure causes marked quickening of the 
 heartbeats provided that the vagi are still intact. He showed that 
 any considerable increase in venous pressure causes quickening 
 of the heartbeat, and that the quickening depends upon the in- 
 tegrity of the vagus nerves. Part, at any rate, of this effect is due 
 to inhibition of the tonic inhibitory action of efferent vagus fibers. 
 Another part is probably due to reflex excitation of accelerator 
 nerves, but on this point the evidence was not so clear. The action 
 of the heart is not subject to direct voluntary control, but the ef- 
 fects of emotional stimuli on the rate of heartbeat are well known 
 and very evident. 
 
 There is no necessary connection between rate of heartbeat and 
 circulation rate. This has been shown by various experiments, but 
 most strikingly by the experiments of Starling and his pupils on 
 the bodies of animals in which an artificial circulation through 
 the heart and lungs alone had been established, the physiological 
 connections with central nervous system and rest of the body being 
 cut off. In such a "heart-lung preparation" the rate of heartbeat 
 remains steady for long periods if the temperature is kept steady 
 and artificial respiration is maintained; but the flow of blood can 
 be varied within very wide limits by simply varying the rate at 
 which blood is supplied to the right side of the heart. Thus Pat- 
 terson and Starling found that with a pulse rate which was steady 
 at 144 the circulation rate in a heart-lung preparation from the 
 
 18 Bainbridge, Journ. of Physiol., L, p. 65, 1915. 
 
RESPIRATION 281 
 
 dog could be varied from 215 to 2,000 cc. per minute by simply 
 regulating the supply of blood to the right side of the heart. 19 
 
 The heart is thus a pump which is capable of adjusting its out- 
 put without any variation in rate of stroke ; and we might imagine 
 a heart working quite efficiently on this principle, without any 
 -regulation by the nervous system. The circulation would adjust 
 itself automatically in accordance with the rate at which blood 
 was allowed to pass through the systemic capillaries; and the 
 resistance in the arterioles and capillaries would automatically 
 maintain a sufficient arterial blood pressure. 
 
 It is possible that in certain cases of heart disease, where the 
 physiological connection between auricles and ventricles through 
 the bundle of Kent and His is broken, the circulation is main- 
 tained in this way, since in these cases the pulse rate does not 
 change during the very limited amount of muscular exertion 
 which is possible. In normal persons or animals, however, the 
 pulse rate increases very markedly during muscular exertion ; and 
 in persons in whom, owing to some nervous or cardiac abnormality 
 this increase does not occur, the capacity for exertion is very small. 
 We must infer, therefore, that under normal conditions the ca- 
 pacity of the heart for increasing the circulation rate without 
 increase of the rate of heartbeat is very limited far more so than 
 might be inferred from study of a heart-lung preparation. In 
 other words the output of the heart during systole is usually 
 pretty constant under normal conditions, as Henderson was the 
 first to point out. 
 
 We must now consider in more detail how the distribution of 
 blood is regulated. It has been known since the discovery by 
 Claude Bernard of vasomotor nerves that the distribution of 
 blood in the body is regulated through the nervous system. Vaso- 
 constrictor nerves are known to be widely distributed in all parts 
 except the central nervous system, and vasodilator nerves have 
 also been discovered at certain points. There is also a main vaso- 
 motor center in the medulla from which vasoconstrictor impulses 
 radiate, and subsidiary vasomotor centers in the spinal cord. 
 Another and much more direct means of regulating the distribu- 
 tion of blood has recently been discovered by Krogh. 20 He has 
 found by microscopical examination of living capillaries, and by 
 injection of Indian ink, that under resting conditions the great 
 majority of capillaries in muscular and other tissues are firmly 
 
 19 Patterson and Starling, Journ. of PhysioL, XLVIII, p. 357, 1914. 
 
 20 Krogh, Journ. of Physiol., LII, p. 457, 1919. 
 
282 RESPIRATION 
 
 contracted and impermeable to blood, so that neither blood cor- 
 puscles nor even the finest particles of Indian ink can pass 
 through them. Nor is the full arterial blood pressure capable of 
 forcing them open. Whenever the tissue is stimulated to activity, 
 however, these capillaries open wide, so that blood can pass 
 through them freely. He found, for instance, that in muscle of the 
 guinea pig about twenty times as many capillaries were open 
 during activity of the muscle as during rest. The active contrac- 
 tility of capillaries had been directly observed by Roy and Gra- 
 ham Brown in 1880, but the real significance of this observation 
 had not been realized. 
 
 Krogh's observations have thrown a flood of new light on the 
 exchange of gases and other material between the blood and the 
 living tissues : for the opening out of new capillary paths when- 
 ever a greater exchange of material is taking place must facili- 
 tate enormously the exchange, and thus furnish a means of keep- 
 ing the gas pressures in the tissues approximately normal in spite 
 of great variations in metabolism. During muscular work, for 
 instance, the immense increase of capillary paths will greatly 
 facilitate the exchange of oxygen and carbonic acid between the 
 blood and the muscle fibers. There must be a great tendency to 
 fall in the oxygen pressure of the blood passing through the 
 muscle capillaries during muscular work. Unless this fall were 
 approximately compensated for by the opening out of new capil- 
 laries, it is difficult to see how a sufficient oxygen supply could be 
 maintained, as in all probability the oxygen consumption in a 
 muscle during very hard work is twenty or thirty times as great 
 as during rest. We can also now understand much better how it 
 comes about, for instance, that when the skin circulation is cut 
 down to the utmost by vasoconstriction in the prevention of un- 
 necessary loss of heat from the body, the skin, though more or 
 less blue from greatly diminished blood flow, may be still full of 
 blood, as shown by the full blue color. 
 
 Probably it is the stimulus of the presence in excess of certain 
 metabolic products, particularly carbonic acid, and the deficiency 
 of others, particularly oxygen, that determines the relaxation of 
 the capillary walls. There can also be little doubt that the same 
 stimuli, acting reflexly, determine the activity of local vasomotor 
 nerves. Temperature stimuli, or irritation stimuli, appear to act 
 in a similar manner. Stimuli may also act centrally, however, as 
 in the general regulation of body temperature by variations in the 
 skin circulation, or in emotional vasomotor changes. 
 
RESPIRATION 283 
 
 How very powerfully a local stimulus may act on local blood 
 circulation is strikingly shown by a recent experiment of Meakins 
 and Davies. 20A They found that when the arm was immersed in 
 cold water the returning venous blood was completely deprived 
 of oxygen. On the other hand, when the arm was kept in hot water 
 the haemoglobin of the venous blood was 94 per cent saturated 
 with oxygen, as compared with 96 per cent for the arterial blood. 
 The oxygen consumption was doubtless much greater in the warm 
 than in the cold skin, so the difference in circulation rate must have 
 been enormous. 
 
 If the regulation of blood distribution in the body were simply 
 a matter of opening the proper sluice gates according to local re- 
 quirements, the matter would be much more simple than it is. 
 Actually, however, the contraction and dilatation of various ar- 
 teries, veins, and capillary tracts must tend to have the effect of 
 varying the total capacity of the blood vessels, with the result that 
 the venous blood pressure at the heart inlet varies, and either too 
 little, or too much, blood is supplied to the heart. As a conse- 
 quence, the arterial blood pressure would either tend to fall too 
 much to secure an adequate supply of blood to the brain and other 
 parts, or else to rise too high. 
 
 There appears to be an elaborate nervous defense against such 
 disturbances. Excessive rise of arterial blood pressure is guarded 
 against, not only by the reflex vagus inhibition already referred 
 to, but also by reflex vasomotor inhibition through the "depres- 
 sor" branch from the cardiac vagus. Excitation of the depressor 
 fibers causes inhibition of the vasomotor center in the medulla and 
 consequent dilatation of arteries and probably veins in the splanch- 
 nic and other areas. Depressor action is brought about (whether 
 directly or indirectly) by excessive arterial blood pressure, so 
 that the pressure is relieved. Deficiency in arterial and venous 
 pressure is guarded against by an opposite "pressor" action re- 
 sulting in excitation of the vasomotor center and consequent rise in 
 blood pressure. A normal stimulus to pressor action of the center 
 is quite evidently deficiency of oxygen combined with excess of 
 carbonic or other acids in the blood supplying the brain. Thus the 
 arterial and venous blood pressures rise very markedly in response 
 to deficiency of oxygen combined with excess of carbonic acid, 
 whether produced by deficient aeration of the blood or circulatory 
 failure. A very important effect of this rise of blood pressure is 
 
 SOA Meakins and Davies, Journ. of Path, and, Bact., XXIII, p. 460, 1920. 
 
284 RESPIRATION 
 
 to concentrate the available blood flow towards the brain. In mus- 
 cular exertion there is also a rise of blood pressure, due partly to 
 the effect on the vasomotor center of excess of CO 2 and deficiency 
 of oxygen in the arterial blood, but perhaps partly also to a gen- 
 eral pressor action complementary to a local depressor action on 
 the arteries and veins concerned in supplying the muscles with 
 blood. 
 
 We may compare the action of the bulbar centers controlling 
 blood pressure and heart rate with that of the respiratory center 
 in its linked responses to direct chemical and peripheral nervous 
 stimuli; but data are not yet available for carrying the com- 
 parison into detail. 
 
 From this general survey of the experimental evidence relating 
 to the regulation of the circulation, it will be seen that the deciding 
 factor in determining the rate of circulation ancl local distribution 
 of blood flow is local or general deficiency or excess in the diffusion 
 pressures of the substances which enter into tissue metabolism, 
 and particularly deficiency or excess in the diffusion pressures of 
 oxygen and carbonic acid. Temperature is also a factor, but per- 
 haps not a different one, since the diffusion pressure of a sub- 
 stance varies as its absolute temperature. 
 
 The regulation of the circulation may be abnormal in various 
 ways, and the present chapter would be incomplete without some 
 reference to this subject. The abnormality may arise from disease 
 or congenital defect of the heart or from operative interference, 
 but is very commonly due to disorder of the nervous regulation, 
 whether or not any organic defect is also present. Another form 
 of abnormal circulation is due to a deficient volume of blood, or to 
 abnormality in its composition. In all these cases the abnormal 
 circulation is reflected in abnormal breathing. Owing to the ab- 
 sence of adequate clinical or experimental investigations it is 
 difficult as yet to deal with this subject in a satisfactory manner, 
 and I can only attempt to discuss it tentatively in the light of what 
 is already known. 
 
 The effect may first be considered of a valvular defect which 
 either causes narrowing of valvular openings (stenosis) or makes 
 a valve incompetent so that there is regurgitation. The effect of 
 this is that, other things being equal, more work is thrown on one 
 or another part of the heart. If this extra work is not serious it 
 may be completely met, and partly by a true hypertrophy of the 
 muscular substance on which the increased work is thrown; but 
 if the extra work is serious the action of the heart as a pump will 
 
RESPIRATION 285 
 
 be limited, so that the increased circulation required during mus- 
 cular exertion cannot be produced. The arterial blood pressure 
 will therefore fall during muscular work of more than a certain 
 amount. In consequence of this the coronary circulation may also 
 be impaired, with possibly dangerous consequences under the ex- 
 isting circumstances ; and there will be faintness along with hy- 
 perpnoea, owing to slowed circulation and hence diminished oxy- 
 gen pressure and increased CO 2 pressure in the capillaries of the 
 brain. During rest, however, or such muscular exertion as is pos- 
 sible without abnormal symptoms, the circulation will be carried 
 on in a normal manner. 
 
 The alveolar CO 2 pressure in a number of cases of valvular 
 heart disease was investigated by Miss FitzGerald, and found to be 
 normal except in cases confined to bed with serious symptoms. 21 
 The absence of any fall in the alveolar CO 2 pressure constituted 
 good evidence of the absence of any impairment of the circula- 
 tion during rest. In cases with serious symptoms even during rest 
 there was a marked fall in the alveolar CO 2 pressure. This is also 
 the case in congenital heart affections, when the alveolar CO 2 
 pressure may be as low as 20 mm. 22 
 
 We can see what is happening in these cases. Owing to the im- 
 paired or short-circuited circulation the oxygen pressure in the 
 tissues falls and the CO 2 pressure tends to rise. This, however, 
 increases the breathing, and so prevents the rise of CO 2 pressure 
 by abnormally diminishing the CO 2 pressure of the arterial blood 
 leaving the lungs. The fall in oxygen pressure cannot, however, 
 be prevented in this way, as the increased breathing will not 
 materially increase the oxygen in the arterial blood. Some anox- 
 aemia will therefore be present, and will probably show itself by 
 the color of the skin and lips, as well as by more frequent, and 
 possibly shallower, breathing, and other symptoms of anoxaemia. 
 The alkalosis produced by the increased breathing due to anox- 
 aemia will gradually be compensated for by increased excretion 
 of alkali and diminished formation of ammonia, just as at a high 
 altitude (see Chapter VII) ; and this will tend to diminish the 
 real anoxaemia though without diminishing the cyanosis. Unless 
 the breathing became shallow no material relief could be looked 
 for owing to active secretion of oxygen inwards by the lung ep- 
 ithelium, as this would only slightly increase the oxygen in the 
 
 21 FitzGerald, Journ. of Pathol. and Bact., XIV, p. 328. 
 
 * French, Pembrey, and Ryffel, Journ. of Physiol., XXIX, Proc. P&ysiol. Soc., 
 p. ix, 1909. 
 
286 RESPIRATION 
 
 arterial blood; but some relief may come from compensatory in- 
 crease in the percentage of haemoglobin in the blood. In a bad 
 heart case the heart has usually broken down owing to either some 
 more or less acute infection or to too much muscular exertion; 
 and usually the main question is whether, and to what extent, the 
 heart will recover with rest and the passing off of the infection. 
 
 In many heart affections the defect is in the nervous regulation 
 of the heart, either without or with a valvular defect. The ac- 
 celerator, inhibitory, depressor, or pressor reflexes may be acting 
 excessively. Cases with evident defects of nervous control have 
 been very common during the war, under such names as "soldier's 
 heart," "disordered action of the heart," "neurasthenia," etc. In 
 the commonest form of this defect there is very abnormal increase 
 in pulse rate on slight exertion or emotional and other stimuli ; 
 and accompanying the increase there is pain and hyperalgesia in 
 the areas where pain is usually felt in heart affections. The exag- 
 gerated cardiac reflexes seem to be similar to the exaggerated 
 Hering-Breuer respiratory reflex in the same cases, and to be due 
 to the same causes (see Chapter III). Reflexes and nervous or 
 emotional responses of all kinds are exaggerated in these cases of 
 neurasthenia; and the exaggeration of cardiac reflexes is fre- 
 quently only one symptom of a condition of general neurasthenia. 
 The pain is probably only an expression of fatigue produced by 
 the over-frequent heartbeats. 
 
 A similar condition is very commonly present as an accompani- 
 ment of valvular defect; and the associated shallow breathing 
 may cause very serious secondary anoxaemia in the manner al- 
 ready described in Chapter VII. This seems to be the explanation 
 of the orthopnoea and Cheyne-Stokes breathing so often seen in 
 bad heart cases, and also explains the marked effects of oxygen 
 inhalation in relieving the symptoms. Continuous inhalation of 
 air enriched with oxygen is likely to prove a very valuable remedy 
 in promoting recovery where failure of the respiratory center is 
 complicating defects of circulation. 
 
 A very interesting investigation demonstrating a relation be- 
 tween vascular disturbances in the lungs and the Hering-Breuer 
 reflex has recently been published by J. S. Dunn, 22A who was 
 working at the time in conjunction with Barcroft. He produced 
 multiple embolism of pulmonary arterioles by intra-venous injec- 
 tion of starch granules. When only a moderate degree of embolism 
 was produced (so as not to cause immediate death) he observed 
 
 MA Dunn, Quart. Journ. of Med., XIII, p. 129, 1920. 
 
RESPIRATION 287 
 
 an extraordinary increase in frequency and diminution in depth 
 (to half or even a fourth) of respiration. At the same time the 
 rate of circulation (measured by a very perfect blood-gas method 
 described in the same journal by Barcroft, Boycott, Dunn and 
 Peters) was not diminished, nor was the venous blood pressure 
 raised, or the arterial pressure disturbed : nor was there appreci- 
 able deficiency of oxygen or excess of CO 2 in the arterial blood. 
 But when the vagi were cut the respirations slowed down and 
 became normally deep at once. It appears, therefore, that the 
 Hering-Breuer reflex (Chapter III) was enormously exaggerated 
 as a result of the disturbed pulmonary circulation. Just at first the 
 breathing was stopped, which suggests that the respiratory move- 
 ments were jammed completely by the exaggerated reflex. These 
 experiments throw a quite new light on the intense and exhausting 
 dyspnoea caused by pulmonary embolism, and also in cardiac cases 
 where there is rapid breathing without other cause. How the vagus 
 nerve endings are excited is not yet clear. The discovery of a drug 
 capable of controlling their action would evidently be an important 
 advance in therapeutics. 
 
 In defective circulation owing to loss of blood the primary 
 cause of breakdown appears to be that, in spite of contraction of 
 arterioles and venules owing to pressor reaction of the vasomotor 
 center, there is not sufficient blood to fill the large veins and ade- 
 quately supply the right side of the heart. As a consequence the 
 arterial blood pressure falls and the circulation slows down, with 
 consequent anoxaemia acting most seriously on the brain, and 
 affecting the breathing in the manner already explained in con- 
 nection with valvular affections where compensation is imperfect. 
 The natural remedy for this condition would appear at first sight 
 to be a pressor excitation of the vasomotor center, just as the 
 natural remedy for arterial anoxaemia due, say, to low atmos- 
 pheric pressure, appears at first sight to be increased breathing 
 and increased circulation rate. But just as the increased breathing 
 and circulation rate in arterial anoxaemia is to a large extent pre- 
 vented by the counter-balancing effect of the alkalosis thereby 
 produced, so also is the full pressor response to anoxaemia due to 
 fall in blood pressure. The breathing is already stimulated by the 
 diminished blood circulation in the brain, so that the arterial 
 blood is so alkaline as to quiet down the vasomotor center, in spite 
 of the anoxaemia. Benefit may be expected from the administra- 
 tion of CO 2 or even of acids ; but the main need is for increase in 
 the volume of the blood. This increase comes naturally, provided 
 
288 RESPIRATION 
 
 that fluid is supplied ; and the great thirst which results from loss 
 of blood is an expression of the need for fluid. But time is required 
 for this natural process of recuperation, and meanwhile the patient 
 may die. 
 
 Fluid may be supplied quickly by the intravenous injection of 
 Ringer's Solution, but this plan is rather ineffective, since the 
 injected liquid leaks out from the vessels quickly. Bayliss there- 
 fore introduced his now well-known gum-saline solution for use 
 in cases of loss of blood and similar conditions. 23 The gum does 
 not leak out at all readily from the vessels, and in virtue of the 
 osmotic pressure which it produces it keeps the salt solution from 
 leaking out. The gum thus plays the same part in this respect as 
 the proteins of the blood plasma, but is free from the occasional 
 toxic properties of the proteins in blood transfused from another 
 person, although it seems to be sometimes not free from disadvan- 
 tages. It might seem at first sight as if the injection of gum saline 
 must, other things being equal, be very inferior in its effects to 
 transfusion of blood, since there is no haemoglobin in the salt solu- 
 tion. But unless the loss of blood has been enormous there is no 
 great need for haemoglobin. Increased rate of circulation will 
 make up for diminished power of the blood to carry oxygen and 
 CO 2 , as explained more fully on page 293. 
 
 The conditions known as "wound-shock," "surgical shock," 
 "anaesthetics shock," and shock from burns, have given rise to 
 much discussion and investigation. When "shock" is fully de- 
 veloped, the arterial blood pressure is very low, the pulse feeble, 
 the lips and skin leaden colored, and the breathing shallow and 
 often rapid, or sometimes periodic. It appears at present as if this 
 general condition can be brought about in several different ways; 
 and Yandell Henderson's investigations have thrown a clear light 
 on certain of the causes of shock. It will be convenient to consider 
 these first. 
 
 He showed in the first place that a condition of shock can be 
 brought about in animals by continued excessive ventilation of the 
 lungs. This of course greatly reduces the CO 2 in the arterial 
 blood, thus producing a state of alkalosis. The response to this is 
 slowing of the circulation, and consequent great anoxaemia, as 
 already explained. The slowing of the circulation tends, of course, 
 to diminish the alkalosis in the tissues, but only at the expense of 
 producing most formidable anoxaemia. The alkalosis is also com- 
 
 23 Bayliss, Intravenous Injection in Wound Shock, 1918. 
 
RESPIRATION 289 
 
 bated by the body in other ways, one being the prompt stoppage 
 of ammonia formation and the excretion of alkaline urine, as 
 already explained ; and, whether in consequence of this or of other 
 causes, the so-called "alkaline reserve" of the blood decreases 
 greatly, as Henderson and Haggard showed (Chapter VIII). 
 Nevertheless the anoxaemia and alkalosis cannot be overcome. 
 The circulation rate steadily diminishes; the heart, in consequence, 
 probably, of anoxaemia, begins to fail, apart altogether from its 
 inadequate supply of venous blood ; and finally there is complete 
 failure of the heart. If, however, the forced breathing is stopped 
 before cardiac failure has occurred, death may occur from pro- 
 longed apnoea and consequent acute asphyxia, as mentioned in 
 Chapter II. When the condition of shock has developed suffi- 
 ciently, the animal cannot be saved by adding CO 2 to the air 
 breathed ; but in the earlier stages this procedure is quite effective. 
 The hopeless condition to which the animal is reduced by the 
 forced artificial respiration is probably analogous to the condition 
 produced in various ways by prolonged anoxaemia, as in very 
 severe CO poisoning, or in a patient who has been allowed to 
 suffer for long from severe arterial anoxaemia. It is probably the 
 anoxaemia rather than the alkalosis that produces the serious 
 effect, since, as already mentioned, forced breathing of oxygen is 
 more easily tolerated than forced breathing of air. 
 
 A condition of shock produced by forced artificial respiration 
 is, of course, not a natural occurrence; but Henderson showed 
 that excessive respiration can be produced by natural means in 
 two ways : firstly, by powerful afferent stimuli, as by electrical 
 stimulation of the sciatic nerve, even in the presence of anaesthesia 
 sufficient to abolish consciousness ; and secondly, by the action of 
 ether in doses not sufficient to anaesthetize an animal completely. 
 The afferent stimuli, or the ether, increase the breathing to such 
 an extent as to diminish greatly the CO 2 in the arterial blood, 
 thus producing great alkalosis or acapnia, with concomitant anox- 
 aemia. By these means, therefore, a condition of shock may easily 
 be produced in a patient; and it seems probable that in this way 
 the condition generally known as shock is frequently produced as 
 a matter of fact. 
 
 Clinical evidence seems, nevertheless, to indicate that in many 
 ordinary cases of wound shock there has been no excessive 
 breathing. On the other hand there are many facts indicating 
 that the symptoms are due to absorption from injured tissues of 
 
290 RESPIRATION 
 
 harmful disintegration products, 24 and Dale and Laidlaw have 
 shown that similar symptoms are caused by the action of histamine 
 produced by tissue disintegration. 25 In "histamine shock" the 
 venous return to the heart is inadequate, just as in acapnial shock, 
 and blood appears to stagnate in dilated capillaries so that the 
 rest of the vascular system is imperfectly filled with blood. Dale 
 and Laidlaw regard the dilatation of capillaries as a primary ac- 
 tion of the poison. The respiratory center seems, also, to be affected 
 very quickly, so that artificial respiration is needed to keep the 
 animal alive. How far the failure of the respiratory center is 
 consequent on failure of the circulation, or vice versa, it seems 
 difficult at present to say ; but the shallow breathing and leaden 
 cyanosis in shock are indicative of advancing failure of the re- 
 spiratory center, and appear to be clear indications for early 
 and continuous oxygen administration, if the condition cannot be 
 dealt with by removing its cause or in other ways. To remedy the 
 imperfect filling of the vessels and consequent failure of the circu- 
 lation, there is an equally clear indication for the intravenous 
 injection of gum-saline solution. Whether the administration of 
 air containing CO 2 would be of service, as in shock due to simple 
 alkalosis, is not yet known. If the respiratory center is injured by 
 a poison from the injured tissues it may be unable to respond 
 properly to the CO 2 . 
 
 Dale found that the danger from histamine shock may be enor- 
 mously increased by the administration of an anaesthetic. Many of 
 Henderson's observations seem to point in the same direction as 
 regards acapnic shock. These investigations throw much light on 
 the fatal accidents of anaesthesia. 
 
 In connection with circulation and breathing it is important to 
 consider the manner in which the volume and haemoglobin per- 
 centage of the blood adjust themselves under varying conditions. 
 They are fairly constant within about five per cent under ordinary 
 conditions for any individual, and the volume of blood in a mam- 
 mal bears a pretty constant ratio to the body weight. This propor- 
 tion does not depend upon size or ratio of body weight to surface, 
 since it is about the same in large as in small mammals. Thus 
 in the rat or mouse the proportion is about the same as in man. 
 
 In a small warm-blooded animal such as a mouse the metabolism 
 per gram of body weight is enormously greater than in a large 
 
 84 Report No. VIII of Surgical Shock Committee (Special Report No. 26 of 
 Medical Research Committee), 1919. 
 
 "Dale and Laidlaw, Journ. of Physiol., LII, p. 355, 1919. 
 
RESPIRATION 
 
 291 
 
 animal such as a man, and roughly speaking is proportional to 
 the ratio of external surface to body weight. As was shown by 
 Dr. Florence Buchanan, 26 the pulse rate and respiration rate vary 
 in about the same proportion. Thus in a canary the pulse rate, as 
 recorded photographically by means of the capillary electrometer, 
 was about 1,000 per minute, the rate, as compared with that in 
 man, being greater in proportion to the more rapid metabolism. 
 The circulation rate in a small animal is thus enormously greater 
 than in a large animal, and indeed must be so ; but the proportions 
 
 12.0 
 
 10-0 
 
 S-o 
 
 6.0 
 
 4.0 
 
 2.0 
 
 40O 
 
 600 1200 I60O 200O 2400 2BOO 
 
 Weight of Rabbits in Grammes 
 
 szoo 
 
 Figure 71. 
 
 Blood volumes of rabbits in cc. of blood per 100 grams of body weight. 
 The curve shows what the blood volumes would be if they varied in the pro- 
 portion of body surface to body weight. The dots and crosses show average 
 results of actual determinations by the modified Welcker method. Dots repre- 
 sent results of Boycott : crosses of Dreyer and Ray. The numbers indicate 
 number of determinations for each group of observations. 
 
 between the different parts of animals, including the blood, do 
 not depend on differences in size of the animals. From a very 
 limited number of experiments on animals, Professor Dreyer of 
 Oxford 27 drew the extremely improbable conclusion that in ani- 
 
 29 Buchanan, Science Progress, July, 1910. 
 
 "Dreyer and Ray, P kilos. Trans. Royal Society, B, CCI, p. 138, 1910; also 
 Dreyer, Ray, and Walker, Skand. Arch. f. Physiol., 28, p. 299, 1913. 
 
292 RESPIRATION 
 
 mals of the same species the blood volume is a function of the 
 ratio of body surface to mass, and even inferred that the carbon 
 monoxide method of determining blood volume (appendix) 
 must be incorrect because it showed no such relation in experi- 
 ments published by Douglas. 28 The matter was afterwards re- 
 investigated in rats by Chisolm, 29 and by Boycott. 30 Figure 71 
 shows the results of Boycott and of Dreyer (all obtained by the 
 modified Welcker method) in rabbits of different sizes. It will 
 be seen that there is no difference between them, and that, al- 
 though young rabbits have usually a somewhat higher proportion 
 of blood than older ones, the increased proportion does not vary 
 with the proportion of body weight to surface. The circulation 
 rate must, other things being equal, be faster in a smaller animal 
 with its higher proportional metabolism, but an increased pro- 
 portional dead weight of blood would be no advantage, but a 
 disadvantage. 
 
 When the volume of blood is reduced by considerable bleeding, 
 there is at first a fall in arterial, and doubtless also in venous, 
 blood pressure ; but soon the blood pressure is restored. The first 
 effect of the bleeding is probably to evoke partial compensation 
 by a pressor excitation of the vasomotor center. This is probably 
 due to diminished circulation rate and consequent fall in oxygen 
 pressure and increase of CO 2 pressure in the medulla. Very soon, 
 however, the blood volume is more or less restored by taking up 
 of liquid from the tissues and intestines. The blood is thus diluted ; 
 but the diluted blood fills up the blood vessels and completely re- 
 stores the blood pressure. After a delay of many days or perhaps 
 several weeks, the hydraemic blood is restored to normal by re- 
 production of the missing corpuscles. 
 
 Similarly when blood is transfused from another animal of 
 the same species there is at first a rise of both venous and arterial 
 blood pressure. Soon, however, the volume of blood is reduced 
 by disappearance of most of the extra plasma. The remaining 
 blood then contains an excess of red corpuscles, and these are only 
 got rid of in the course of some days or weeks. 
 
 The changes which occur were followed by Boycott and Doug- 
 las with the help of the carbon monoxide method of determining 
 the blood volume in living animals. 31 They found that on repeated 
 
 28 Douglas, Journ. of Physiol., XXXIII, p. 493, 1906. 
 
 29 Chisolm, Quart. Journ. of Exper. Physiol., IV, p. 208, 1911. 
 
 30 Boycott, Journ. of Pathol. and Bacter., XVI, p. 485, 1912. 
 
 81 Boycott and Douglas, Journ. of Pathol. and Bacter., XIII, p. 270, 1909. 
 
RESPIRATION 
 
 293 
 
 bleeding the reproduction of the red corpuscles becomes more and 
 more rapid, so that finally the animal can reproduce the lostxor- 
 puscles very rapidly. Similarly on repeated transfusion the animal 
 can get rid of the transfused corpuscles more and more rapidly. 
 It thus becomes adapted to either bleeding or transfusion. 
 
 In an animal in which as a result of bleeding or similar causes 
 the proportion of haemoglobin in the blood is abnormally low the 
 oxygen pressure must fall more rapidly than usual if the rate of 
 circulation is unaltered, as the blood passes through the tissues. 
 In accordance with what has been already said, this will naturally 
 tend to be more or less compensated for by an increased rate of 
 circulation. But this can occur freely without the opposing effect 
 due to the production of alkalosis, since owing to the diminished 
 percentage of haemoglobin the pressure of CO 2 would also be 
 
 OXFOGD 
 
 1 
 P 
 
 SUMMIT OF PIKES PEAK 
 
 COLORADO 
 SPRINGS 
 
 NEW HAVEN 
 
 
 OXFORD 
 
 HALDANE 
 
 I3U 
 120 
 
 
 
 
 
 
 ^ 
 
 AT"*" 
 
 o~_. 
 
 
 
 * . 
 
 <^ 
 
 
 
 
 
 
 
 
 
 
 
 
 no 
 
 A * 
 
 
 
 ^ 
 
 f** 
 
 ..^ 
 
 ^> 
 
 ^ 
 
 V 
 
 \, 
 
 " 
 
 N>- 
 
 -= 
 
 
 
 oX 
 
 
 ^, 
 
 ^ 
 
 
 
 
 JWWw'-' 
 
 0^ 
 
 ^\ 
 
 A 
 
 / 
 
 
 
 
 v^ 
 
 ~~~ 
 
 "' 
 
 v_ 
 
 
 -s^ 
 
 i* 
 
 fl 
 
 
 
 
 ** 
 
 < 
 
 
 90 
 
 * \r^\ 
 
 A/ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Figure 72. 
 
 Ordinates represent percentages of the average haemoglobin percentages 
 obtained before ascending the Peak (Oxford and Colorado Springs) on the 
 particular subject. Continuous thick line = total oxygen capacity or total 
 amount of haemoglobin. Continuous thin line = percentage of haemoglobin. 
 Interrupted line = blood volume. The values in Oxford before the start of 
 the expedition are plotted without relation to time. 
 
 too high unless the circulation rate were increased. An increased 
 circulation rate is thus the natural response to a diminished haemo- 
 globin percentage. 
 
 We know from observations on persons living at high altitude 
 that one result of the shortage of oxygen caused by the diminished 
 barometric pressure is that the percentage of haemoglobin and of 
 red corpuscles in the blood rises (see Chapter XIII). In different 
 individuals the rise varies considerably. Thus in persons who had 
 been living for some weeks on the summit of Pike's Peak we found 
 that the haemoglobin percentage varied from 113 to 153 per cent 
 of the normal. The rapidity with which the change occurs varies 
 also greatly in different individuals. Figure 72 shows the rate 
 
294 
 
 RESPIRATION 
 
 at which the change occurred and disappeared in one of the 
 members of the Pike's Peak expedition, and Figure 73 shows the 
 far faster rate of increase in haemoglobin in Mr. Richards, a 
 mining engineer who kindly made for me a careful series of 
 
 observations on himself on going to a mine in Bolivia at a height 
 of 15,000 feet. Figure 72 also shows the changes in blood volume 
 and total haemoglobin in the body (total oxygen capacity). It 
 will be seen that after the first few days the blood volume in- 
 
RESPIRATION 295 
 
 creases, so that the total haemoglobin in the body increases more 
 than the percentage of haemoglobin. Thus the corpuscles dcrnot 
 simply increase at the expense of the space occupied by plasma, 
 but the total space occupied by the blood is increased. It seems 
 probable, however, that when a rapid increase in the percentage 
 of haemoglobin occurs, as shown in Figure 73, the increase is 
 mainly brought about at first by disappearance of plasma owing 
 to a pressor reaction of the vasomotor center, with consequent in- 
 creased filling of the capillaries and resulting loss of liquid from 
 the blood. In acute anoxaemia produced by asphyxial conditions 
 there appears to be a rapid loss of fluid from the blood, and this 
 is probably due to a pressor reaction. Schneider and his colleagues 
 have recently observed that in a considerable proportion of airmen 
 exposed for a quite short time to low pressures of oxygen there is 
 a small but quite appreciable rise in the haemoglobin percentage. 32 
 
 There appears to be no doubt that the cause of the increased 
 total amount of haemoglobin and red corpuscles in the body at 
 high altitudes is increased activity of the bone marrow in forming 
 red corpuscles. On this point direct evidence was obtained by 
 Zuntz and his colleagues. 33 They found that in dogs the blood- 
 forming red marrow was markedly increased at a high altitude. 
 The stimulus to this increase was undoubtedly fall in the oxygen 
 pressure of the blood, and it is doubtless in the same way that in- 
 creased formation of red corpuscles is brought about by loss of 
 blood, especially if repeated. From the experiments of Boycott and 
 Douglas on repeated blood transfusions, we can also infer with 
 great probability that with increased oxygen pressure in the tis- 
 sue capillaries, owing to an increased proportion of haemoglobin, 
 there is a corresponding increase in the blood-destroying tissues. 
 The proportion of haemoglobin in the blood appears, therefore, 
 to be dependent on the oxygen pressure in tissue capillaries. This 
 inference is confirmed by the fact that, as Nasmith and Graham 
 showed, 34 the haemoglobin percentage rises markedly in animals 
 which are kept exposed to a small percentage of CO. 
 
 In cases of chronic heart disease, and more particularly in cases 
 of congenital heart defects accompanied by cyanosis, there is often 
 a great increase in the total haemoglobin and also in the blood 
 volume. Thus in a congenital case of "Morbus coeruleus," brought 
 
 82 Gregg, Lutz, and Schneider, Amer. Journ. of Physiol., L, p. 216, 1919. 
 
 33 Zuntz, Loewy, Muller, and Caspar!, Hohenklima und Bergwanderungen, 
 Berlin, 1906. 
 
 34 Nasmith and Graham, Journ. of Physwl., XXXV, p. 32, 1906. 
 
296 RESPIRATION 
 
 to us by Dr. Parkes Weber, Douglas and I found that the haemo- 
 globin percentage was increased 80 per cent; the blood volume 
 IOO per cent; and the total haemoglobin 260 per cent; 35 and we 
 found similar increases in another case. Lorrain Smith had al- 
 ready found a considerable increase in a non-congenital heart 
 case with chronic cyanosis. 36 
 
 In some cases (so-called idiopathic polycythaemia) where there 
 is neither exposure to a lowered oxygen pressure nor any heart 
 or lung affection, the haemoglobin percentage and number of 
 red corpuscles per unit volume is greatly increased. On determin- 
 ing the blood volume in two of these cases I found it greatly in- 
 creased. Boycott and Douglas examined three other cases with 
 a similar result. 37 In the most marked of these cases the haemo- 
 globin percentage was 1 76 per cent of the normal, and the blood 
 volume nearly three times the normal, so that the amount of 
 haemoglobin in the body was about five times the normal. Idio- 
 pathic polycythaemia is accompanied by a bluish tint of the skin, 
 and this suggests that from some cause there is slowing of the 
 circulation and consequent anoxaemia of the tissues, to which the 
 increased haemoglobin percentage is a natural response. 
 
 It is clear that increase in the haemoglobin percentage will tend 
 to diminish the tissue anoxaemia at high altitudes or in cases of 
 heart affections; for the blood can pass more slowly (or at a more 
 normal rate at high altitudes) through the capillaries before a 
 given fall in the oxygen pressure occurs. This compensation is 
 never complete, however; for if it were there would be no stimu- 
 lus to the increased concentration of haemoglobin. An undue rise 
 of CO 2 pressure in the tissues is also prevented by the increased 
 haemoglobin percentage. 
 
 When the red corpuscles and haemoglobin are increased 60 or 
 80 per cent the viscosity of the blood is very greatly increased, 
 and a good deal of stress has been laid on this increased viscosity 
 as a hindrance to circulation. Nevertheless persons with their hae- 
 moglobin percentage increased 50 per cent at high altitudes are 
 capable of the severest muscular exertion ; and there is no indica- 
 tion in them of any circulatory impairment. When we consider 
 the manner in which the circulation is normally regulated, as 
 
 "The details of this case are given by Parkes Weber and Dorner, Lancet, 
 Jan. 21, 1911. 
 
 38 Lorrain Smith and McKisack, Trans. Path. Soc. of London, LIII, p. 136, 
 1902. 
 
 87 Boycott and Douglas, Guy's Hospital Reports, LXII, p. 157. 
 
RESPIRATION 297 
 
 explained above, it seems evident that anything but a very ex- 
 treme increase in viscosity will at once be compensated for by 
 more free opening of arterioles and capillaries. The resistance 
 to flow of blood in the living body is regulated physiologically, 
 and cannot for a moment be compared to the mechanical resist- 
 ance in a system of lifeless tubes. 
 
 The rapid variations in blood volume from diminution or in- 
 crease in the vasoconstrictor (pressor) influence of the vaso- 
 motor center is perhaps shown most strikingly by the effects on 
 the blood of section of the spinal cord below the vasomotor center 
 in the medulla. Cohnstein and Zuntz found that very quickly 
 after section and consequent fall of blood pressure the proportion 
 of red corpuscles fell to about half, while the proportion rose 
 rapidly again on stimulation of the cord just below the section, 
 with consequent rise of blood pressure. 38 The blood appears to 
 take up. or lose plasma rapidly when the capacity of the blood 
 vessels is diminished or increased. 
 
 It was discovered by Lorrain Smith with the help of the carbon 
 monoxide method that in chlorosis and in secondary "anaemias" 
 the blood volume is increased without any diminution, or with 
 only a very slight one, in the total haemoglobin in the blood. The 
 anaemia is thus in reality a hydraemia or dilution of the haemo- 
 globin. 39 Boycott and I found the same condition in the ''anaemia" 
 of ankylostomiasis. 40 Miss FitzGerald found later that in chlorosis 
 the alveolar CO 2 pressure is not diminished but normal, so that 
 in this form of anaemia there appears to be no anoxaemia during 
 rest. 41 These facts suggest that the apparent anaemia is due to 
 some cause leading to abnormal dilation and consequent increased 
 capacity of the blood vessels, with the natural sequence of hydrae- 
 mia, but so that the oxygen pressure in the tissues is not dimin- 
 ished. Possibly, therefore, the anaemia is produced through the 
 vasomotor nervous system, or through substances, or the deficiency 
 of substances, which act primarily on the blood vessels. The facts 
 that salts of iron have a striking curative action in chlorosis, and 
 that iron is a constituent of haemoglobin, have led to the idea that 
 the anaemia is caused by the absence of sufficient iron for a normal 
 formation of haemoglobin; but in the cure of chlorosis by iron 
 Lorrain Smith could find no appreciable increase in the total 
 
 38 Cohnstein and Zuntz, P finger's Archiv., 88, p. 310, 1888. 
 
 39 Lorrain Smith, Trans. Pathol. Soc. of London, LI, p. 311, 1900. 
 
 40 Boycott and Haldane, Journ. of Hygiene, III, p. 112, 1903. 
 
 41 FitzGerald, Journ. of Pathol. and Bacterial., XIV, p. 328, 1910. 
 
298 RESPIRATION 
 
 amount of haemoglobin in the body. The characteristic dyspnoea 
 and faintness on exertion in chlorosis, etc., are probably due to 
 the impossibility of sufficiently increasing during exertion the 
 already greatly increased circulation. 
 
 In pernicious anaemia and the anaemia of haemorrhage, Lor- 
 rain Smith found a very marked diminution of the total haemo- 
 globin present ; but often enough the blood volume was increased 
 above normal. 
 
 Although the intimate connection between breathing and cir- 
 culation is already very evident, many points in the connection 
 are still uncertain or obscure. There is an abundant field for 
 clinical and physiological investigation in elucidating this sub- 
 ject, though it must always be remembered that not only are 
 breathing and circulation closely dependent on one another, but 
 they are dependent also on other physiological activities. 
 
 Addendum. The experiments by Douglas and myself on the 
 regulation of the circulation in man have now been completed, and 
 are in course of publication. A very complete series, in which 
 Douglas was himself the subject, shows that during complete rest 
 the mixed venous blood had only utilized about 19 per cent of its 
 available oxygen, and gained a corresponding charge of CO 2 - 
 During hard work, with the oxygen consumption increased about 
 nine times, about 65 per cent of the arterial oxygen was utilized. 
 The pulse rate was increased about 2.6 times, and as the utiliza- 
 tion of the arterial oxygen was increased 3.4 times, the output of 
 blood per heartbeat was practically the same during hard work 
 as at complete rest, and the blood flow had simply increased in 
 proportion to the increase of pulse rate. 
 
 Various other subjects, including myself, had a similar high 
 rate of blood flow (about 8 liters per minute) during rest, but one 
 or two had a markedly lower rate of flow, with the percentage 
 utilization of oxygen as high, in one case, as 33 per cent. In this 
 case the output per beat during rest, and the circulation rate (about 
 4.7 liters per minute) were a good deal lower than in the other 
 subjects, but the output per beat increased to about double during 
 hard work. There are thus considerable individual differences 
 (quite apart from differences in weight) as regards the rate of 
 general blood flow and the particular manner in which the circu- 
 lation adapts itself to varying amounts of work. 
 
 As some doubt has arisen lately as to whether oxygenation of 
 blood within the living body has the same influence on the CO 2 
 
RESPIRATION 299 
 
 carrying power of blood as after the blood has been removed and 
 defibrinated, we made careful observations on this point. The 
 experiments showed clearly that oxygenation produces the same 
 effect in the living body as outside it. 
 
 A full account of the method, and of the results reached by it, 
 will be found in our paper. 
 
CHAPTER XI 
 Air of Abnormal Composition. 
 
 IN the present chapter I propose to describe the mode of occur- 
 rence and physiological effects of the more commonly occurring 
 gaseous constituents of air. The number of noxious gases, vapors, 
 and particulate impurities, which may, under particular circum- 
 stances, be present in air, is of course very large, and only the 
 commoner additions to air can be dealt with here. 
 
 Outside Air. Pure country air, freed from moisture, contains 
 20.93 per cent by volume of oxygen, .03 per cent of carbon diox- 
 ide, and 79.04 per cent of a residue usually designated as "nitro- 
 gen," although of this 79.04 per cent about .94 per cent consists 
 of argon. Very minute traces are also present of hydrogen and 
 various rare gases. Ordinary atmospheric air contains, however, 
 aqueous vapor in varying proportions ; and about I per cent is on 
 an average present in a climate such as that of Great Britain. The 
 composition of dry country air is the same to the second decimal 
 point all over the world. In summer weather the percentage of 
 CO 2 near the ground may be as low as .025 during the day, and 
 as high as .035 during the night, owing to the influence of vegeta- 
 tion, etc. ; and doubtless the oxygen percentage rises or falls 
 correspondingly, though this has not yet been shown directly. 
 
 In towns the composition of the outside air varies surprisingly 
 little from that in the country. The percentage of CO 2 seldom 
 rises above .05, nor does that of oxygen fall below 20.9, even in a 
 large town, like London ; and in summer weather there is hardly 
 any difference between the oxygen and CO 2 percentages of town 
 and country air. In a London park on a summer day the per- 
 centage of CO 2 may fall quite as low as in the country. Consider- 
 ing the great area of a town like London, and the enormous 
 quantity of coal and gas burnt, this fact is very striking, and 
 shows clearly that apart from horizontally-flowing wind there 
 are very active up-and-down movements of the air, and these 
 keep the air of a town pure. It is only in foggy weather that these 
 up-and-down movements cease more or less; and then the im- 
 purities in the air of a large and smoky town may become very 
 
RESPIRATION 301 
 
 appreciable. Russell found, for example, that in London the per- 
 centage of CO 2 might rise to 0.14 during a dense fog. 
 
 Along with CO 2 there are present in the air of towns a number 
 of other impurities. From fires a good deal of unburnt CO passes 
 off. In the air of the underground railways when steam loco- 
 motives were still used, I found that about I volume of CO was 
 present for every 12 volumes of CO 2 . If we assume the same pro- 
 portion for the air of a town, there would be about .01 per cent 
 of CO present in the air of a bad London fog. This would be 
 sufficient in time to saturate the haemoglobin with CO to the ex- 
 tent of about 17 per cent, and might thus produce appreciable 
 effects on persons already in bad health, though healthy persons 
 would not notice any effect. 
 
 Much more appreciable, however, are the effects of the par- 
 ticulate impurities. Ordinary coal contains a good deal of sul- 
 phur; and the sulphur, in the process of combustion, is mainly 
 oxidized to sulphuric acid, which condenses along with water 
 in the form of minute droplets and thus helps to form fog. Of the 
 unpleasant irritant effects of this sulphuric acid one can form a 
 good idea in passing through a railway tunnel, particularly if the 
 train is moving slowly up an incline and the coal burnt contains 
 much sulphur. Those familiar with sulphuric acid fumes in chemi- 
 cal laboratories or factories will at once recognize them in the 
 tunnel air. When badly purified lighting gas is burnt in a room, 
 the same irritant effect is also noticeable to a less degree. In a bad 
 fog in a large town the choking effects of sulphuric acid con- 
 tribute largely to the unpleasant effect of the fog and the manner 
 in which the fogginess of the air persists even when the air is 
 warmed in the interior of a house. There is no escape from this 
 effect unless the air is scrubbed or filtered. The sulphuric acid is 
 also destructive to metal and other materials. 
 
 Besides sulphuric acid the smoky air contains particles of black 
 carbonaceous matter which greatly help absorb the light, and 
 also contains substances which have an unpleasant odor and more 
 or less irritant effect on the air passages. As will be shown below, 
 there is no reason to believe that the continued inhalation of these 
 particles has any deleterious effect on the lungs, and in ordinary 
 town air they are not present in sufficient concentration to be of 
 any direct consequence in other ways to health. Their greatest 
 importance arises from the inconvenience and expense caused by 
 their obstruction of light and the manner in which they dirty 
 clothes, walls, ceilings, and everything else in a house. By the 
 
302 RESPIRATION 
 
 substitution of well-purified gas for coal in fires, or by smokeless 
 combustion of coal, the trouble might be avoided, and indeed has 
 been much diminished within recent years. 
 
 Lower organisms, and particularly plants, are on the whole 
 far more sensitive to impurities in air and other changes in en- 
 vironment than higher animals, and particularly man. The real 
 reason for this is that between the living tissue elements and the 
 outside environment higher organisms possess an internal en- 
 vironment which is not only highly developed, but is maintained 
 with an efficiency which increases with the scale in development. 
 Plants are extremely sensitive to the particulate and other impuri- 
 ties in air and the obstruction of light by smoke and opaque fogs. 
 But few trees and plants can flourish in the air of a town or in- 
 dustrial area. The traces of acid and other impurities present in 
 the air can act more or less directly on their tissue elements, which 
 have very little between them and the external environment. 
 
 Air of Occupied Rooms. In rooms of all kinds where men are 
 present the composition of the air becomes altered, owing to res- 
 piration and evaporation and to any gas or oil lamps which may 
 be burning. Both respiration and lamps consume oxygen and pro- 
 duce CO 2 and moisture. The combustion in the lamps is perfect, 
 so that no CO passes into the air; and unless the gas is badly 
 purified from sulphur the products of combustion have very little 
 unpleasant effect apart from what may be due to heat. It was 
 formerly supposed that some volatile toxic substance is given off 
 in the breath; but the experimental evidence in support of this 
 belief was found to be fallacious, and all attempts to demonstrate 
 the existence of such a substance have failed. Some of the most 
 striking evidence on the subject is afforded by experience in sub- 
 marines, in which a limited volume of air is quite commonly re- 
 breathed until after a few hours a light will not burn and 3 per cent 
 or more of CO 2 may be present. Provided the air remains cool, as 
 it does in a temperate climate owing to the cooling influence of the 
 water, the only effects observed are those due to CO 2 . 
 
 Even in the most crowded and ill-ventilated rooms the pro- 
 portion of CO 2 seldom rises above 0.5 per cent, with, of course, a 
 corresponding drop in the oxygen percentage. From the account 
 already given of the physiology of breathing it is evident that a 
 difference of this order in the composition of the air is in itself of 
 no appreciable importance. The breathing simply becomes very 
 slightly deeper and the composition of the alveolar air and 
 
RESPIRATION 303 
 
 arterial blood remains practically unaffected as regards either 
 CO 2 or oxygen. 
 
 Although apart from CO 2 no appreciable amount of any 
 poisonous substance is given off to the air by the body, various 
 substances which affect the olfactory nerves are given off in minute 
 amounts from persons or furniture in a room. As a rule these sub- 
 stances are only perceived on entering a room, and are not noticed 
 after a short time by those who remain in it. In sensitive persons, 
 however, they may produce an unpleasant reflex effect; and for 
 this reason apart from any other a good ventilation is desirable. 
 When, however, there is no musty furniture, and the bodies and 
 clothing of those present are fairly clean, there is little or no in- 
 convenience from this cause. 
 
 A far more important factor in connection with the physio- 
 logical effects of the air in rooms is temperature, and along with 
 it moisture. The maintenance of a constant internal body tempera- 
 ture depends on constant physiological adjustment between ac- 
 tual heat loss from the body and variations in environmental con- 
 ditions which tend to make the heat loss greater or less than the 
 heat production. The variations in environmental conditions con- 
 sist in variations in temperature, moisture content, and movement 
 of the air, and also variations in the radiant heat gained or lost 
 by the body, apart from the actual temperature of the air. The 
 actual heat loss is regulated physiologically, apart from conscious 
 regulation by variation of clothing, etc., partly by varying the 
 rate of blood circulation through the skin, and partly by varying 
 the amount of water evaporated by the skin. The latter means of 
 regulation does not come into play unless the air is warm, or heat 
 production in the body is greatly increased by muscular exertion. 
 
 When the air of a room is so cold, or the movement of the air 
 is so great, that the skin, or parts of it, become uncomfortably 
 cold, we are always clearly aware of the cause of discomfort. But 
 when the air is so warm as to lead to the skin being uncomfortably 
 warm we are apt to attribute the discomfort to some other cause 
 than the heat. The matter is also complicated by the fact that in 
 different persons the air temperature at which discomfort is 
 felt varies considerably. Thus persons who have been undergoing 
 "open-air" treatment and are accustomed to rooms with open 
 windows feel much discomfort in rooms with closed windows 
 where other persons are just comfortable. Similarly Americans 
 accustomed to the warm air associated with central heating find 
 British houses with fires very uncomfortably cold in winter, while 
 
304 RESPIRATION 
 
 British visitors to America find the warm air of American houses 
 very trying. 
 
 The discomforts of warm or cold air are not usually associated 
 with rise or fall of internal body temperature. When suffering 
 great discomfort from sitting in a very cold room, I have found 
 the rectal temperature slightly raised rather than lowered, and 
 on going to an uncomfortably warm room there was a slight fall 
 in rectal temperature. Persons going unaccustomed into very 
 warm air may become faint or suffer from nausea or headache 
 without any appreciable rise of body temperature. There appears 
 to be a fall of arterial pressure owing to failure on the part of the 
 vasomotor center to compensate for the increased flow of blood 
 through the skin in a warm atmosphere, and this probably ac- 
 counts for the more striking symptoms. In any case persons soon 
 become more or less acclimatized within limits to the effects of 
 warm air. One can observe this in miners who become accustomed 
 to warm places in mines, or in people who become accustomed to 
 Turkish baths. 
 
 It is somewhat noteworthy that men accustomed to hard outdoor 
 work seem to be much less sensitive to heat or cold indoors than 
 other persons. This is probably due to the fact that though they 
 are not accustomed to external heat they are accustomed to what 
 in this reference comes to much the same thing, namely, greatly 
 varied internal heat production, which involves the same capacity 
 for vasomotor adaptation as exposure to external heat or cold. 
 Those who are most affected by external heat or cold indoors are 
 persons who are not only unaccustomed to external heat, but are 
 also unaccustomed to hard muscular exertion. 
 
 Part of the discomfort of warm air in rooms is due to its drying 
 effect on the skin and particularly the upper air passage. Winter 
 air warmed to a temperature of about 70 F. is very dry; and if 
 the skin and upper air passages are kept warm by the air they lose 
 far more moisture than usual and become uncomfortable. With 
 cold air the inside of the nose is kept cool, and during expiration 
 moisture condenses in it, so that it is kept moist in spite of the fact 
 that the cold air contains very little moisture. With warm dry air, 
 on the other hand, there is much evaporation during inspiration 
 and little or no condensation during expiration, so that the nose 
 is apt to become very dry ; and this appears to lead to swelling of 
 the mucous membrane. 
 
 The combination of physiological disturbances produced by 
 warm air in a room is apt to be attributed to chemical impurities 
 
RESPIRATION 305 
 
 in the air. Owing to this fact, and general ignorance as to the 
 physiology, as distinguished from the chemistry, of respiration, 
 too much stress was formerly laid on the chemical purity of the 
 air in rooms. The chemical purity is nevertheless a very important 
 index of the chances of infection through the air from person to 
 person in a room. The more air is passing through the room the 
 less the chances of infection become ; and for this reason as high 
 as possible a standard of chemical purity is desirable where a 
 number of persons, some of whom may be carriers of infection, 
 are present. A reasonable standard to aim at under these circum- 
 stances is that the excess of CO 2 in the air of the room should not 
 be over .02 per cent unless lights are burning, or that about 5 
 cubic feet of air per person and per minute should be supplied. 
 This standard can easily be maintained in ordinary houses with 
 natural ventilation ; and even in the case of crowded buildings a 
 similar standard can be attained by the right application of 
 modern engineering methods. 
 
 When air becomes very warm the regulation of body tempera- 
 ture becomes dependent on increased evaporation from the skin 
 and not merely on variation in the blood flow through it. If mus- 
 cular work is being done this point is soon reached if the air is 
 fairly still. The amount of moisture in the air then becomes very 
 important, as the rate of evaporation from the skin depends on the 
 amount of moisture already present in the air. In still air, or in 
 air moving at any given rate, a temperature is finally reached at 
 which in spite of profuse sweating the skin cannot evaporate 
 water quickly enough to prevent the body temperature from rising. 
 As I showed experimentally in 1905, this temperature is reached 
 when the wet-bulb temperature reaches a certain point. 1 Thus in 
 still air and with hardly any clothing, the body temperature be- 
 gins to rise when the wet-bulb temperature exceeds 88 F (3iC). 
 It does not matter what the actual air temperature is, or the 
 actual percentage of moisture in the air, provided that the wet- 
 bulb temperature reaches 88. Thus it was indifferent whether the 
 air temperature was 88 with the air saturated, or 133 with the 
 air very dry, provided that the wet-bulb temperature was 88. 
 When the wet-bulb temperature was far above 88 the rate of 
 rise of body temperature was proportional to the rise of wet-bulb 
 temperature. 2 
 
 1 Haldane, Journ. of Hygiene, V, p. 494, 1905. 
 
 a Haldane, Trans. Inst. of Mining Engineers, XLVIII, p. 553, 1914. 
 
3 o6 RESPIRATION 
 
 When even moderate muscular work was being done the criti- 
 cal wet-bulb temperature was, even with almost no clothing, at 
 least 10 below 88 in still air. With the ordinary clothing of 
 temperate climates the critical wet-bulb temperature is much 
 lower than without clothing, especially during muscular work. On 
 the other hand, with the air in motion, the critical wet-bulb tem- 
 perature is higher. The beneficial effects of fans, punkahs, etc., 
 during heat is well known. With the wet-bulb temperature above 
 the body temperature, however, the rise of body temperature is the 
 more rapid the more the air is in motion. 
 
 In the climate of Great Britain the wet-bulb shade temperature 
 very seldom rises above 70, even on very warm summer after- 
 noons ; but during heat waves in America a wet-bulb temperature 
 of 75 is not infrequently reached, and cases of hyperpyrexia 
 from the heat then become common. Wet-bulb temperatures of 
 over 80 are of course common in tropical countries, and are met 
 by proper adaptation of clothing and mode of life; but the 
 amount of muscular exertion which is possible with a wet-bulb 
 temperature over 80, except in a good breeze, is limited. In 
 ordinary rooms in a temperate climate, and when ordinary cloth- 
 ing is worn, a wet-bulb temperature of even 65 becomes oppres- 
 sive and likely to cause fainting and headaches in persons not 
 accustomed to heat or heavy muscular exertion. 
 
 In order to obtain a simultaneous measure of the cooling action 
 on the body of air temperature, movement of air, and maximum 
 evaporation from the skin, Dr. Leonard Hill has devised an in- 
 strument known as the katathermometer. This consists of an 
 alcohol thermometer with a very large bulb, which, when an ob- 
 servation has to be made, is heated to about iooF. The flask is 
 jacketed with an absorbent jacket which can be moistened with 
 water. By the rate at which the water cools, a comparative esti- 
 mate can be obtained of the maximum possible combined cooling 
 action on the human body of movement of air, temperature, and 
 evaporation. The actual cooling effect of the air depends, of 
 course, on the physiological responses of the body, but cannot ex- 
 ceed the maximum shown by the wet katathermometer. 
 
 The physiology of temperature regulation lies outside the scope 
 of this book; but temperature effects are so liable to be confused 
 with effects due to chemical impurities in air that it seemed 
 necessary to refer briefly to the physiological disturbances due to 
 warm air. 
 
 The air of occupied rooms is liable to be contaminated by 
 
RESPIRATION 307 
 
 escapes of lighting gas ; and under certain circumstances fatal or 
 very serious accidents from this cause may occur and lighting 
 gas may be used very easily for purposes of suicide or even 
 murder. The great majority of accidental deaths from poisoning 
 by lighting gas have been in bedrooms, owing to the gas being in 
 some way left turned on after being extinguished. In 1899 a 
 Departmental Committee of which I was a member reported on 
 the influence of the use of water gas in connection with poisoning 
 by lighting gas, and I investigated the conditions under which 
 poisoning may occur in bedrooms. 3 
 
 It might be supposed that the sense of smell would always give 
 warning of an escape of lighting gas in a room. On going into a 
 room in which gas is escaping one notices the smell at once, and 
 long before sufficient gas is present to cause any symptoms of 
 poisoning; but a person inside the room when the escape begins 
 may quite probably never notice it. The reason for this is that 
 the sense of smell for any particular substance becomes fatigued 
 very rapidly, and if the proportion of the odoriferous substance in 
 the air is only very gradually increased the smell is never noticed. 
 In this way an escape of gas in a bedroom is often unnoticed. 
 
 When a continuous escape of gas occurs in a room, the per- 
 centage of gas in the air goes on increasing until the rate of es- 
 cape through walls, roof, etc., balances the rate of inflow of gas. 
 In any ordinary room the walls, roof, and floor are permeable to 
 air, and, if any cause such as pressure of wind or difference of 
 temperature between inside and outside tends to produce air 
 currents in and out of the room, the flow of air is surprisingly 
 free. If, for instance, the door and windows are closed and all 
 visible chinks pasted up, it will be noticed that when a fire is lit 
 the chimney draws just as well as before. Large volumes of air 
 are passing up the chimney, and this air comes in through the 
 walls, roofs, etc. Brick and stonework, for instance, are fairly 
 permeable to air, as can easily be shown by suitable means. Small 
 rooms in a dwelling house do not require artificial ventilation, 
 provided the passages, etc., are well ventilated, since the ratio of 
 surface to cubic capacity is high, so that ventilation through the 
 surfaces of the room counts for more in relation to the cubic space 
 per person in the room. 
 
 It will thus be readily seen that what happens in a room when 
 gas escapes continuously will depend on various circumstances, 
 such as the difference in temperature between inside and outside, 
 
 3 Report of the Water-gas Committee, Part. Paper, 1899. Appendix i. 
 
3 o8 RESPIRATION 
 
 the presence of a fire or of central heating by warm air, the 
 amount of wind, etc. But even if there is little or no cause of ex- 
 change of air before the gas escape begins, the escape itself will 
 furnish a cause, since the gas is much lighter than air, so that air 
 to which gas has been added will tend to pass out by the roof. 
 Hence even under conditions least favorable to ventilation, the 
 gas can never accumulate to more than a very limited concentra- 
 tion in the air of a room. 
 
 Another complication in connection with gas escapes is that the 
 gas may or may not mix evenly with the air of a room. Gas escap- 
 ing from a burner passes straight upwards to the roof and there 
 spreads. I found that unless the temperature of the windows and 
 walls was below the air temperature of the room the gas never 
 came down again to any very great extent. With a very rapid 
 escape of gas, as when a burner was completely removed or a 
 pipe cut, this was very marked. It was impossible to obtain a 
 poisonous atmosphere at the ordinary breathing level, but there 
 was a heavy concentration of gas near the roof. The danger of 
 poisoning was to persons in the floor above, and not to those in the 
 room where the escape was occurring. Near the floor level, how- 
 ever, a curious phenomenon was observed. The gas actually 
 present in the air was found to be nearly pure hydrogen. This 
 showed that it was only by diffusion, and not by convection cur- 
 rents, that gas had penetrated downwards. Hydrogen, being much 
 more diffusable than any of the other constituents of lighting gas, 
 had diffused downwards much more rapidly ; and in general it was 
 found that the hydrogen in lighting gas separates off by diffusion 
 very readily, leaving a mixture containing more of CO and the 
 other heavier constituents of the gas. At night, when the windows 
 were cold, and the tendency to convection currents down them 
 was consequently strong, mixture of the gas by convection was 
 much more apt to occur, especially if the escape was at a moderate 
 rate. There was consequently more danger at night to persons 
 sleeping in the room. 
 
 When the percentage of gas was determined at intervals in the 
 air of a room with gas continuously escaping from a burner and 
 mixing by convection currents down the windows, I found that, 
 if the conditions of wind, etc., remained constant, the percentage 
 became constant after a certain time which depended on the size 
 of the room among other conditions, and might vary from about 
 one to three hours according to the size of the room, rate of gas 
 escape, amount of wind, etc. The maximum percentage obtained 
 
RESPIRATION 309 
 
 was 2.7 per cent at the breathing level. With larger escapes of 
 gas this percentage could hardly be increased, as most of the gas 
 remained at the roof. The air at all parts of the rooms tested was 
 examined with a miner's safety lamp to see if the air ever became 
 explosive; but with such escapes as could be produced when 
 burners were not taken off, I never succeeded in obtaining an 
 explosive atmosphere even at the roof. It requires about 8 per 
 cent of lighting gas to render air explosive. 
 
 These experiments had a very definite practical significance in 
 connection with the composition of lighting gas used for domestic 
 purposes : for it is evident that whether or not a dangerous result 
 will ensue from an escape of gas in a room will depend on how 
 poisonous the gas is, and not simply on the time during which the 
 escape continues. The poisonous action of lighting gas largely 
 diluted with air depends exclusively on the CO contained in it. 
 In every case of persons found dead in air containing lighting gas 
 the post mortem appearances are those of CO poisoning, and the 
 percentage saturation of blood as determined by the method de- 
 scribed in the appendix has turned out to be round 80, just as in 
 the case, referred to below, of miners poisoned by CO. Thus, 
 broadly speaking, the danger of poisoning from escape of lighting 
 gas depends on whether the air will be poisonous from CO when 
 less than 2 or 2.5 per cent of gas is present. 
 
 Lighting gas as originally introduced is made by the distillation 
 of bituminous coal, and usually contains about 7 or 8 per cent of 
 CO. With 2 per cent of this lighting gas in the air there would 
 only be about 0.14 per cent of CO ; and this, though a formidable 
 percentage, would not, so far as known, produce fatal effects in a 
 healthy person, as the haemoglobin would, in all probability, not 
 become much more than about half -saturated. To judge from all 
 our present knowledge, and from the results of experiments on 
 animals, about 0.3 per cent would usually be needed to produce 
 death within a few hours. 
 
 Excellent lighting gas can also be made by blowing steam 
 through incandescent coke or coal. The product is what is called 
 "blue" water gas consisting roughly of equal parts of hydrogen 
 and CO. This gives a very hot, though small, flame, and although 
 the flame by itself is "blue" and practically nonluminous, an ex- 
 cellent light is given when a properly adjusted mantle is used. On 
 the other hand the calorific value of a given volume of this gas is 
 very low as compared with ordinary coal gas ; and as the value of 
 gas depends mainly on the heating power of a given volume of it, 
 
310 RESPIRATION 
 
 as well as, to a certain extent, on the luminosity of its flame when 
 no mantle is used, water gas is usually "carbureted" by the ad- 
 dition of cheap oil in a chamber where the oil is "cracked" by 
 means of heat. The product is known as carbureted water gas, 
 and is very largely used as a substitute for ordinary coal gas. It 
 has a luminous flame and more or less satisfactory calorific value, 
 but contains about 30 per cent of CO. 
 
 It is evident that with gas containing 30 per cent of CO, poison- 
 ing will occur very readily with an escape of gas during the night 
 in a house. On inquiring into the deaths from gas poisoning in 
 American towns supplied with carbureted water gas, the com- 
 mittee referred to above found that about 100 to 200 times as 
 many deaths occurred from gas poisoning with a given distribu- 
 tion of gas as in English towns supplied with coal gas only. The 
 gas was also used very extensively for purposes of suicide, and 
 sometimes also as a means of murder. Apart from actual danger 
 from poisoning, there was also the constant anxiety as to danger 
 from gas poisoning. An American mother, for instance, told me 
 that she regularly got up every night to make sure that gas was 
 not escaping where her children were sleeping. The result of the 
 committee's inquiries was to show that if gas is to be used for 
 domestic purposes the percentage of CO in it should be reasonably 
 low ; and in consequence of this finding the use of undiluted car- 
 bureted water gas was discontinued in Great Britain, where, 
 indeed, it had only been introduced in one or two places, though 
 with unfortunate results which led to the inquiry. It should, how- 
 ever, be mentioned that with the general introduction of mantles 
 the danger of poisoning from accidental escapes from burners is 
 considerably diminished, as less gas escapes, and if there is a 
 pilot flame the risk is further greatly diminished. 
 
 Gas poisoning in houses may not only occur from escapes 
 within the house, but also from escapes from street gas mains; 
 and many serious accidents from this cause have occurred, par- 
 ticularly with carbureted water gas. The danger is much increased 
 from the fact that in passing through earth the odoriferous con- 
 stituents (benzene, etc.) of the gas are apt to be more or less 
 absorbed, so that the gas entering the basements of houses is more 
 or less odorless. Probably, also, it may have lost a good deal of 
 its hydrogen by diffusion, and this will make it more poisonous. 
 A large number of persons in several houses and many different 
 rooms may be poisoned by one serious breakage of a main. 
 Pettenkofer recorded an interesting case where, in the times before 
 
RESPIRATION 311 
 
 clinical thermometers, illness through gas poisoning from a broken 
 main was mistaken for a peculiar and rapidly infectious form uf 
 typhus. No smell of gas was noticed at first, and the percentage 
 of CO must have been so low, and perhaps inconstant, that it took 
 some hours before any distinct symptoms of illness were produced. 
 At last the smell became noticeable, probably because the earth 
 through which the gas was escaping had become saturated with 
 the odoriferous constituents, and so ceased to absorb them com- 
 pletely. 
 
 Air of Mines. The air of mines is liable to be contaminated by 
 various gases known to British miners as black damp, fire damp, 
 afterdamp, white damp, and smoke. Of these, black damp is the 
 commonest and most universally present; fire damp is hardly 
 found except in connection with coal or oil; afterdamp occurs 
 only after explosions ; white damp in connection with spontaneous 
 heating of coal ; and smoke in connection with fires or blasting. 
 
 Black damp is distinguished by miners through its character- 
 istic properties of extinguishing lamps without exploding and 
 not causing danger to life provided a lamp will still burn. As 
 ordinary black damp is heavier than air, it was formerly identi- 
 fied with CO 2 . Its true composition was first ascertained in 1895 
 by Sir William Atkinson and myself. 4 It is the residual gas of an 
 oxidation process, and thus consists of nitrogen with anything up 
 to about 21 per cent of carbon dioxide. It is now evident that 
 black damp may be formed by several different oxidation pro- 
 cesses, among which oxidation of timber, of coal, and of iron 
 pyrites (FeS 2 ) are the most important. 
 
 When timber oxidizes in the process of decay, it gives off 
 nearly as much CO 2 as it consumes oxygen. Hence the black damp 
 formed consists of about 80 parts of nitrogen and 20 of CO 2 . 
 Freshly broken coal also oxidizes slowly for some time at ordinary 
 temperatures, but to a very limited extent. The oxidation process 
 is a simple chemical one and not dependent on microorganisms; 
 and extremely little CO 2 is formed. In the oxidation of pyrites, 
 which is also a simple chemical process, no CO 2 is directly formed ; 
 the sulphur is oxidized to sulphuric acid, which partly combines 
 with the iron to form ferrous and ferric sulphates, but may react 
 with calcium carbonate to form calcium sulphate, CO 2 being of 
 course liberated. 
 
 Black damp of one sort or another is found in practically all 
 mines, though in coal mines where there is much fire damp its 
 
 Haldane and Atkinson, Trans. Instit. of Mining Engineers, 1895. 
 
3 i2 RESPIRATION 
 
 presence can often be detected only by analysis, on account of the 
 predominance of fire damp. Occasionally there is so little CO 2 
 present in black damp that it is lighter than air; or it may be 
 lighter than air owing to admixed fire damp. I found that the 
 black damp formed simply in the oxidation of coal at ordinary 
 temperatures contains small percentages of CO, 5 but black damp 
 as ordinarily found in considerable concentrations in mines is 
 practically free from CO. 
 
 The action of black damp on lamps and candles is of much 
 practical importance, particularly as a miner trusts to his lamp 
 to warn him of the presence of black damp or fire damp. A flame 
 is extremely sensitive to any variation in the oxygen percentage 
 in air. If the oxygen percentage is increased the flame becomes 
 brighter and hotter, and substances which are not inflammable 
 in ordinary air may then become readily inflammable. If the 
 oxygen percentage is diminished the flame becomes dimmer and 
 less hot, unless the diminution is due to the addition of an inflam- 
 mable gas to the air. When the oxygen percentage is dimin- 
 ished by the addition of nitrogen or black damp to the air, the 
 light given by a candle or lamp diminishes by about 3.5 per cent 
 for a fall of o.i per cent in the oxygen percentage. 6 With a fall 
 of about 3 to 3.5 per cent in the oxygen an oil or candle flame is 
 extinguished. Aqueous vapor is even more effective than nitrogen 
 in causing extinction of flame. It should be noted that it is to the 
 percentage, and not the partial pressure, of oxygen that the flame 
 is so sensitive, whereas it is the partial pressure that is of physio- 
 logical importance. A fall in the oxygen percentage of 3 per cent 
 is of very little importance to a man, though it extinguishes a 
 flame. On the other hand a flame still burns well when the atmos- 
 pheric pressure is diminished to a third, while a man is soon as- 
 phyxiated. Gas flames may be much less readily extinguished by 
 fall in oxygen percentage than oil or candle flames. Thus a hydro- 
 gen flame may not be extinguished till the oxygen percentage 
 falls to half or even less, the extinction point depending to a con- 
 siderable extent on the velocity with which the gas is issuing from 
 the burner. An acetylene lamp will burn till the oxygen percentage 
 falls to about 12. 
 
 The physiological action of black damp added to air depends 
 within wide limits on the percentage of CO 2 in the black damp, 
 
 * Haldane and Meachem, Trans, Inst. of Mining Engineers, 1899. 
 8 Haldane and Llewellyn, Trans. Inst. of Mining Engineers, XLIV, p. 267, 
 1902. 
 
RESPIRATION 313 
 
 and can be deduced from the data already given as to the physio- 
 logical actions of CO 2 and oxygen. It should be noted that the 
 CO 2 diminishes greatly the risk that would otherwise exist from 
 diminution of the oxygen percentage. This risk is greatly di- 
 minished, owing to the fact that the CO 2 firstly increases the oxy- 
 gen percentage in the alveolar air by stimulating the breathing, 
 and secondly raises the hydrogen ion concentration of the blood, 
 thus increasing the circulation rate and assisting the dissociation 
 of oxyhaemoglobin in the tissue capillaries. There is therefore 
 little or no danger from lack of oxygen till the oxygen percentage 
 in the air falls to 6 or 7 per cent; but if the oxygen falls much 
 lower death occurs from want of oxygen. The very evident effect 
 of the CO 2 on the breathing gives good warning of the danger, 
 so that apart from the ample warning given by a lamp a man is 
 not likely to go into a dangerous percentage of black damp unless 
 he does so suddenly, as in descending a shaft or steep incline. 
 
 In former times miners often worked in air containing so much 
 black damp as to put a great strain on their breathing while they 
 were at work. Air containing, say, 3 per cent of CO 2 doubles the 
 breathing during rest; but this effect is scarcely noticeable sub- 
 jectively. During work, however, the breathing is also about 
 double what it would otherwise be, and the lungs are thus strained 
 to the utmost. Probably a great deal of the emphysema from 
 which old miners used to suffer was due to this cause. 7 
 
 The ordinary fire damp of coal mines is, practically speaking, 
 pure methane (CH 4 ). In a very "fiery" seam as much as 1,500 
 cubic feet of methane may be given off per ton of coal extracted. 
 The methane is adsorbed in the coal, 8 and may come off under a 
 pressure of 30 atmospheres or more. Of other higher hydro- 
 carbons a small amount is also adsorbed in the coal, but held more 
 firmly, so that only in the last fractions of gas coming off from 
 coal can their presence be clearly demonstrated by analysis. 
 No carbon monoxide comes off with the methane, but appreciable 
 quantities of CO 2 and nitrogen are often given off. It occasionally 
 happens, however, that enormous quantities of CO 2 are adsorbed 
 in coal and may come off in very dangerous outbursts. This is un- 
 known in British and American coal fields, but has been met with 
 in France. Sudden outbursts of adsorbed gas, whether methane 
 or CO 2 , can only occur, however, where coal has been locally dis- 
 integrated, as is apt to be the case near a fault. Ordinary solid coal 
 
 * Haldane, Trans. Inst. of Mining Engineers, LI, p. 469, 1916. 
 8 Graham, Trans. Inst. of Mining Engineers, LII, p. 338, 1916. 
 
314 RESPIRATION 
 
 is so impermeable to gas that it only adsorbs or gives off gas 
 very slowly. In the inflammable gas associated with oil fields 
 higher hydrocarbons are present in considerable amount, so that 
 the gas may burn with a luminous flame and has toxic properties. 
 Methane may of course also be produced by the action of bacteria 
 on old timber or other organic matter in the absence of oxygen ; 
 and accidents from the explosion of gas from this source have 
 occasionally occurred in British ironstone mines. 
 
 When about 6 per cent of methane is present in air, the mixture 
 becomes inflammable with an ordinary light, and explodes vio- 
 lently with a somewhat higher percentage. Curiously enough, 
 however, an excess of methane prevents explosion, although plenty 
 of oxygen is still present; and with more than about 12 per cent 
 
 10 
 
 I) 
 
 < 
 
 
 
 /% 
 
 Figure 74. 
 
 Diagram showing outlines of caps visible on an oil flame with different 
 percentage of methane. 
 
 of methane the mixture ceases to be inflammable. This fact limits 
 considerably the direct dangers from explosions of fire damp. 
 
 The presence of nonexplosive proportions of fire damp in air 
 can easily be detected by the appearance of a "cap" on the flame 
 of a lamp. The cap is a pale, nonluminous flame which appears on 
 the top of the ordinary flame. In order to see it properly the ordi- 
 nary flame must be either effectively shaded or lowered till little 
 else than a blue flame is present, as otherwise the light from the 
 ordinary flame produces a dazzling effect which renders the cap 
 invisible, though it can be photographed without difficulty. The 
 length of the cap depends on the temperature and size of the 
 flame, and with the very hot hydrogen flame the test becomes far 
 more delicate, so that as little as 0.2 per cent of methane can be 
 detected easily. Figure 74 shows the outlines of the cap visible 
 
RESPIRATION 315 
 
 with different percentages of methane when an ordinary oil flame 
 is lowered to the extent required in testing. 
 
 To obviate the danger arising from ignition of fire damp mix- 
 tures by lamps, some sort of safety lamp is now always used in 
 fiery mines. A safety lamp may be either an oil lamp constructed 
 on the general principle introduced by Davy, or an electric lamp ; 
 but the latter has of course the disadvantage that it does not indi- 
 cate the presence of fire damp and black damp. 
 
 As regards its physiological properties, fire damp behaves as 
 an indifferent gas like nitrogen or hydrogen. A mixture of 79 per 
 cent of methane and 21 of oxygen has the same physiological 
 properties as air, except that the voice is altered ; and the physio- 
 logical action of methane is simply due to the reduction which it 
 causes in the oxygen percentage. Its action can thus be deduced 
 from the data in Chapters VI and VII. In actual practice the 
 danger from asphyxiation by fire damp is considerably greater 
 than from black damp, since a man going with an electric lamp or 
 no lamp into air progressively vitiated by fire damp has little 
 physiological warning of impending danger. He is in a similar 
 position to an airman at a very high altitude, and if he suddenly 
 falls from want of oxygen he is very likely to die from failure of 
 the respiratory center. 
 
 Afterdamp. Afterdamp is the gas produced as the result of an 
 explosion, and has been known for long to be specially dangerous. 
 In 1895 I made an inquiry into the causes of death in colliery 
 explosions, 9 and found that nearly all (about 95 per cent) of the 
 men who died underground were killed by CO, although a con- 
 siderable number had received such serious skin burns that they 
 could hardly have survived in any case. Death was never due to 
 deficiency in the oxygen percentage of the air, nor to excess of 
 CO 2 , nor, apart from exceptional cases, to more than 2 per cent 
 of carbon monoxide. It was clear that the men had died in air 
 containing plenty of oxygen, and not much carbon monoxide. 
 That carbon monoxide was the actual cause of death was clear 
 from the fact that the venous blood was usually about 80 per cent 
 saturated with CO ; and that death was slow, and therefore due to 
 a low percentage of CO, follows from the fact that about the same 
 saturation was found all over the body. With more than about 
 2 per cent of CO the venous blood has not time to become evenly 
 saturated and the saturation is usually a good deal lower. 
 
 9 Haldane, Report on the Causes of Death in Colliery Explosions and Fires. 
 Parl. Paper C, 8112, 1896. 
 
316 RESPIRATION 
 
 Colliery explosions were formerly attributed simply to ex- 
 plosions of fire damp. About 40 years ago it was first clearly 
 pointed out by Mr. Galloway that this explanation is unsatis- 
 factory, and that the spread of an explosion must be due to coal 
 dust. Further evidence of the predominant part played by coal 
 dust in all great colliery explosions was soon brought forward; 
 and it became clear that many explosions occur in the complete 
 absence of fire damp, the coal dust being originally stirred up and 
 lighted by the blowing out of flame in blasting, and the explosion 
 carried on indefinitely by further stirring up and ignition. In other 
 cases the starting point is some, perhaps quite small, explosion of 
 fire damp, caused by a defective lamp, a spontaneous fire in the 
 coal, or perhaps even by a spark from falling stone. The ease with 
 which coal dust explosions may be produced by blasting when 
 even a very little coal dust is lying on a road, and the astounding 
 violence which they may develop after the flame has traveled 
 about a hundred yards, were strikingly shown in experiments 
 made with pure coal dust at Altofts Colliery under Sir William 
 Garforth's direction. 10 On account of their danger in a populous 
 neighborhood these experiments were transferred to Eskmeals on 
 the Cumberland coast; and finally showed that when an equal 
 weight of shale dust or other similar material was present along 
 with the stone dust the mixture could not be ignited by blasting 
 or gas explosions. 11 
 
 Sir William Garforth's plan of stone-dusting all the roads in 
 collieries with shale dust, so that at no point is there more than 
 half as much coal dust as shale dust, has now been adopted very 
 generally in Great Britain ; and the only serious recent explosions 
 have been in mines where this precaution was not adopted. Stone- 
 dusting is far more efficacious and cheaper than watering the 
 dust; and indeed efficient watering is impossible in many cases, 
 owing to the effect of water on the roof and sides of a colliery 
 road. 
 
 In the Altofts experiments, samples of afterdamp were analyzed 
 by Dr. Wheeler. The following is a typical example. 
 
 Carbon dioxide 11.9 
 
 Carbon monoxide 8.6 
 
 Hydrogen 2.9 
 
 Methane 3.1 
 
 Nitrogen 73.5 
 
 10 Record of British Coal-dust Experiments, 1910. 
 
 11 Reports of the Explosions in Mines Committee, Parl. Papers, 1912-1914. 
 
RESPIRATION 317 
 
 It will thus be seen that pure afterdamp, free from air, may 
 contain as much as 8.6 per cent of CO. Fresh afterdamp also con 
 tains an appreciable percentage of H 2 S (not shown in the analy- 
 sis). This is a very poisonous gas, and o.i per cent will knock a 
 man over unconscious in a very short time. The most immediate 
 effect of fresh afterdamp may be due to H 2 S ; but on this point 
 there is no definite knowledge as yet. 
 
 Considering the deadly composition of pure afterdamp it is at 
 first sight somewhat surprising that in actual colliery explosions 
 the men are not killed at once by the afterdamp, and that the CO 
 is so dilute in the atmosphere that kills them. It must, however, 
 be borne in mind that along the roads of collieries the coal dust 
 is never pure, and often contains so much shale dust that an ex- 
 plosion is not possible. The combustion is probably, therefore, far 
 from complete, so that much air is left, apart from what is drawn 
 in as soon as the air cools. Possibly, also, the percentage of CO in 
 the pure afterdamp is lower. 
 
 Afterdamp is, of course, extremely dangerous to rescuers, and 
 many lives of rescuers have been lost owing to poisoning by CO. 
 They have gone too far into the poisonous air before becoming 
 aware of any danger, and the first symptom noticed is usually 
 faintness and failure of the legs, so that return is impossible. 
 Moreover the mental condition of men beginning to be affected 
 by CO is usually such, as already explained in Chapter VI, that 
 they will not turn back, and are reckless of danger. A lamp is of 
 course useless for indicating the danger. 
 
 In order to give miners a practical means of detecting danger- 
 ous percentages of CO, I introduced the plan of making use of 
 a small warm-blooded animal such as a mouse or small bird. 12 
 Owing to their very rapid general metabolism and respiration and 
 circulation small animals absorb CO far more rapidly than men. 
 Hence they show the effects of CO far more quickly, and can thus 
 be used as indicators of danger, although in the long run they are 
 possibly rather less sensitive to CO than men are. Thus a danger- 
 ous percentage which would require nearly an hour to affect a man 
 at rest will affect the bird or mouse within about five minutes. 
 This test has now come into very general use, and was, for in- 
 stance, largely used during the war by the tunneling companies. 
 It is easier to see the signs of CO poisoning in a bird in a small 
 cage, as it becomes unsteady on its perch, and finally drops, while 
 a mouse only becomes more and more sluggish; but the mouse is 
 
 "Haldane, Journ. of Physiol., XVIII, p. 448, 1895. 
 
318 RESPIRATION 
 
 easier to handle, and less apt to die suddenly and thus leave the 
 miner without any test. The animals recover very quickly as soon 
 as purer air is reached and this greatly increases their value as a 
 test. 
 
 After an explosion it is very necessary to have some test for 
 CO. The ventilation system is thrown out of action owing to doors 
 and air crossings being blown in. On the other hand it is very im- 
 portant to get in as soon as possible in case men are still alive, and 
 in order to deal with any smoldering fires left by the explosion. 
 
 When air in a mine is for any reason not safe to breathe, self- 
 contained breathing apparatus are now frequently employed. It 
 is beyond the scope of this book to describe these apparatus in 
 detail ; 13 but it may be mentioned that the usual principle employed 
 is that the wearer breathes through a mouthpiece into and out of 
 a bag, the nose being closed by a noseclip. Into the bag there is 
 directed a stream of oxygen from a steel cylinder carried behind ; 
 and by means of a reducing valve and properly adjusted opening 
 beyond it the stream is kept steady at not less than 2 liters 
 per minute. This is as much as a man uses during pretty hard 
 exertion. If he uses less, the excess is allowed to blow off. 
 If he uses more, the oxygen percentage in the bag may fall 
 rather low, or the bag may become flat before the end of a full 
 inspiration. In the former case he will begin to pant more than 
 usual, but will not fall over so long as the 2 liters are coming in. 
 If less than about 2 liters are coming in he will be liable to fall 
 over, owing to a rapid fall in the oxygen percentage. If the bag 
 begins to go flat he will notice this, and either turn on more oxy- 
 gen through a by-pass, or exert himself less. The carbon dioxide 
 in the expired air is absorbed by a purifier containing caustic 
 alkali. 
 
 In another form of apparatus the delivery of oxygen is gov- 
 erned by the state of fullness of the bag; but in applying this 
 principle there is the difficulty that the oxygen may not be quite 
 pure, and the contained nitrogen may thus accumulate in the bag, 
 or a little nitrogen may leak in from the air at the mouthpiece. 
 
 In still another form use is made of liquid air, of which a large 
 amount can be carried, so that most of the expired CO 2 can be 
 allowed to pass out and only a small purifier is needed. 
 
 13 A thorough discussion of the apparatus in use in America and the principles 
 and practice applicable to it is given in U . S. Bureau of Mines Technical Paper, 
 No. 82, 1917, by Yandell Henderson and J. W. Paul. Numerous investigations, 
 including two full reports by myself, have appeared in Great Britain. 
 
RESPIRATION 319 
 
 Whichever form of apparatus is used it is very necessary that 
 it should be extremely reliable in its action, and that the users 
 should be thoroughly instructed and trained in its proper use and 
 upkeep. A number of lives have been lost or endangered through 
 defective supervision and mode of use, or defective design, of 
 apparatus; and as a consequence of these defects men wearing 
 the apparatus in quite breathable air have often had to be rescued 
 by men without apparatus. With proper and scientific supervision 
 these accidents do not occur, as has been shown again and again 
 during extensive operations in irrespirable air. 
 
 By white damp miners understand a poisonous form of gas 
 coming off from coal which has spontaneously heated. The term 
 seems to have arisen from the fact that steam commonly comes 
 off from the warm coal with this poisonous gas and causes a white 
 mist. By experiments on animals and analyses I have frequently 
 found that the poisonous constituent of the gas was CO. 
 
 Freshly broken coal is, as already mentioned, liable to a slow 
 oxidation process. This of course produces heat, and if sufficient 
 coal is present, so that the heat is not lost as quickly as it is pro- 
 duced, the coal will heat, and the heated coal will oxidize faster 
 and faster until at last it is red hot or bursts into flame if sufficient 
 oxygen is present. It is for this reason that coal may be a danger- 
 ous cargo on long voyages, and that coal cannot be stacked safely 
 in very high heaps. In many seams there is great trouble and no 
 little danger from spontaneous heating of broken coal under- 
 ground ; and the residual gas coming off from heated coal is often 
 called white damp. The higher the temperature of coal which is 
 slowly oxidizing, the greater the proportion of CO in the residual 
 gas. The effects of white damp are thus much the same as those 
 of afterdamp ; and the same precautions are required. 
 
 Smoke in mines may come either from fires or from blasting. 
 The smoke from a fire is usually, of course, visible and irritates 
 the air passages and eyes owing to the irritant properties of the 
 suspended particles. If, however, smoke has slowly traveled some 
 distance in a mine, the particles have subsided and the smoke has 
 become a more or less odorless and transparent gas. Many very 
 serious accidents, involving sometimes the loss of 100 lives, have 
 occurred through the poisonous action of smoke from fires in 
 mines. In these cases the deaths have always, so far as hitherto 
 ascertained, been due to CO poisoning. A large amount of un- 
 burnt CO is given off from smoky or smoldering fires, so that the 
 gases from a fire are almost as dangerous as the afterdamp of an 
 
320 RESPIRATION 
 
 explosion. Practically speaking, afterdamp and smoke from fires 
 produce nearly the same effects, and require the same precautions. 
 A fire in the main intake of a mine is a most dangerous occurrence, 
 since the poisonous gas is apt to be carried all over the mine, and 
 to kill all the men in it. To afford a means of dealing with this 
 danger, the ventilating fans provided at British coal mines are 
 now so constructed that the air current can be at once reversed, so 
 as to drive back the smoke. 
 
 Smoke from blasting may contain various poisonous gases, 
 along with CO 2 , according to the nature of the explosive. Some 
 explosives, such as guncotton, give much CO, and some very 
 little ; but all seem, in practice, to give some. Hence there is always 
 risk of CO poisoning where explosives are used in mines, unless 
 the proper precautions are taken. Black gunpowder, as used for 
 blasting, produces both CO and H 2 S ; and in the cases of gassing 
 it is often difficult to decide whether CO or H 2 S has been mainly 
 responsible for the effects. With explosives containing nitro- 
 compounds another and very serious danger is met with. When 
 these explosives detonate properly the nitrogen is given off as 
 nitrogen gas; but when they burn instead of detonating, the 
 nitrogen comes off as nitric oxide, along with CO instead of CO 2 . 
 In practice, owing to defective detonators or other causes, some 
 of the explosive is apt to burn instead of detonating. The nitric 
 oxide then passes into the air and combines with oxygen to form 
 yellow nitrous fumes. These have a somewhat irritant effect at the 
 time, but this is not sufficient to give proper warning of their 
 dangerous properties. The immediate effects are very slight. If, 
 however, enough of the mixture has been inhaled, the result is 
 that after a few hours symptoms of very severe lung irritation 
 appear, and finally oedema of the lungs and great danger to life. 
 I have found that exposure to the fumes from as little as .05 per 
 cent of nitric oxide in air may be fatal to an animal. This subject 
 will be referred to more fully below in connection with poisonous 
 gas used in war. 
 
 Poisoning with CO in mines is so apt to occur, that a few words 
 may not be out of place as to the treatment of CO poisoning. The 
 symptoms and their cause have already been dealt with. The first 
 thing, is, of course, to get the patient out of the poisonous air. In 
 doing so, however, it is important to keep him well covered and 
 avoid in any possible way exposing him to cold. For some reason 
 which is at present not clear, a man suffering from CO poisoning 
 gets much worse on exposure to cooler and moving air, as in the 
 
RESPIRATION 321 
 
 main intake of a mine. If the breathing has stopped artificial res- 
 piration should be applied promptly ; and this can best be done by 
 Schafer's well-known method. If oxygen is available it should be 
 given at once. It immediately increases greatly the amount of dis- 
 solved oxygen in the blood, and also expels far more rapidly the 
 CO from the blood, as will be evident considering the properties 
 of CO haemoglobin. The oxygen will do most good at first, and 
 may be continued with advantage for at least twenty minutes. Suit- 
 able apparatus for giving oxygen can now be obtained easily. Hen- 
 derson and Haggard have recently shown, however, that owing to 
 the great washing out of CO 2 which occurs during the hyperpnoea 
 produced in acute CO poisoning, or perhaps owing to temporary 
 exhaustion of the respiratory center, the breathing is apt to re- 
 main for some time inadequate. 13A They found by experiments on 
 animals that under this condition the removal of CO from the 
 blood is greatly accelerated by adding CO 2 to the air or oxygen 
 inhaled. The desirability of having some safe and practicable 
 means of adding CO 2 to oxygen used in reviving men poisoned by 
 CO seems evident from these experiments. 
 
 A man who has been badly gassed by CO, and has been un- 
 conscious for some time, is sure to have very formidable symptoms, 
 lasting long after all traces of CO have disappeared from the 
 blood. He may never recover consciousness at all; but when he 
 does his nervous system generally is likely to remain very seriously 
 affected for days, weeks, or months, so that he requires to be care- 
 fully watched, nursed, and treated. Mental powers and memory 
 may be much impaired, and the nervous system seems to be in- 
 jured in many different directions. Thus the regulation of body 
 temperature is apt to be imperfect, and symptoms resembling those 
 of peripheral neuritis are common. A condition of neurasthenia, 
 similar to that so often seen during the war, appears to result fre- 
 quently, with the usual affections of the respiratory and cardiac 
 nervous system. In some cases there seems to be acute dilatation of 
 the heart ; and probably almost every organ in the body has suf- 
 fered from the effects of want of oxygen. 
 
 As mines grow deeper and warmer, the importance of the wet- 
 bulb temperature in connection with mine ventilation becomes 
 more and more prominent. The reasons for this will be evident 
 from what has already been said on this subject; especially when 
 the fact that a miner has to do hard physical work is also taken into 
 
 13A Yandell Henderson and Haggard, Journ. of Pharmac. and Exper. Therap., 
 XVI, p. u, 1920. 
 
322 RESPIRATION 
 
 consideration. To this subject I have given very close attention 
 in recent years, and a full general discussion of it will be found in 
 the recent Report of the Committee on Control of Temperature 
 in Mines. 14 
 
 Owing to the nature of their work and the dry conditions in 
 deep and well-ventilated mines, miners are very much exposed 
 to dust inhalation; and the prevalence of "miners' phthisis" 
 among certain classes of miners led me to the investigation of the 
 effects of dust inhalation. Both men and animals are in general 
 more or less exposed to dust inhalation. The problem presented 
 by dust inhalation in mining and other dusty occupations is thus 
 only a part of a general physiological problem as to how the dust 
 inhaled along with air is dealt with by the body. It is evident that 
 if the insoluble dust which is constantly being inhaled by civilized 
 men, particularly in towns and in dusty occupations, accumulated 
 in the lung alveoli, the effects would in time be disastrous. There 
 is, however, no evidence that such effects are ordinarily produced. 
 The lungs of a town dweller, for instance, are more or less black- 
 ened by smoke particles, but remain perfectly healthy; and the 
 same applies to the lungs of coal miners and of persons engaged 
 in many other very dusty occupations. In other cases, however, 
 such as certain kinds of metalliferous mining, steel grinding, 
 pottery work, etc., the effects of continuous inhalation of the dust 
 are disastrous. Why have certain kinds of insoluble dust no cumu- 
 lative bad effect on the lungs? Why, on the other hand, have other 
 kinds such disastrous cumulative effects? When the first question 
 is answered the second becomes relatively easy. 
 
 It is in the production of phthisis (pulmonary tuberculosis) 
 that the continued inhalation of a dangerous variety of dust shows 
 its effects most clearly. The following table, which I compiled 
 from the statistics of the Registrar General for England and 
 Wales, shows the marked contrast between different occupations 
 as regards the effects of dust inhalation in producing phthisis. 
 Two dusty occupations are included coal mining and tin mining. 
 Of the two, coal mining is much the dustier occupation. It will be 
 seen, however, that among coal miners there is not only very little 
 phthisis, but even less than among farm workers, and much less 
 than the average for all other occupations. Among tin miners, on 
 the other hand, there is a great excess of phthisis; and detailed 
 
 14 First Report of the Committee on Control of Underground Temperature, 
 Trans. Inst. of Mining Engineers, 1920. 
 
RESPIRATION 323 
 
 investigation has shown clearly that it is to dust inhalation that 
 this excess is solely due. 15 
 
 A very large proportion of the dust in inspired air is caught on 
 the sides of the nasal and bronchial inspiratory passages, from 
 which it is continuously removed by the action of the ciliated 
 epithelium. It is only the very finest particles that penetrate to the 
 lung alveoli. Nevertheless large amounts of dust do, as a matter 
 
 DEATH RATES FROM PHTHISIS PER 1000 
 
 LIVING AT EACH 
 
 AGE PERIOD FOR ENGLAND AND WALES, 1900-1902 
 
 Age period 
 
 15-25 
 
 25-35 
 
 35-45 
 
 45-55 
 
 55-65 
 
 All occupied and retired males 
 
 i.i 
 
 2.1 
 
 2.9 
 
 3-2 
 
 2.6 
 
 coal miners 
 
 0.7 
 
 1.0 
 
 I.I 
 
 i-5 
 
 2.0 
 
 farm workers 
 
 0.6 
 
 I.I5 
 
 1-3 
 
 1.4 
 
 2.6 
 
 tin miners 
 
 0.4 
 
 7.0 
 
 II.7 
 
 16.1 
 
 16.2 
 
 of fact, reach the alveoli. Arnold showed that even what, in human 
 experience, is relatively harmless dust, will produce, if inhaled 
 in very large amount, foci of scattered broncho-pneumonia in the 
 lungs, and that quartz dust is specially apt to produce inflam- 
 matory changes followed by development of connective tissue. 18 
 In connection with the use of shale dust for preventing colliery 
 explosions Beattie showed that neither coal dust nor shale dust 
 produce any harm in animals if the dust is inhaled in the moder- 
 ate quantities comparable to what a miner inhales. On the other 
 hand, the dust from grindstones produces signs of fibrosis. 17 The 
 subject was followed further in my laboratory by Mavrogardato 
 in an investigation undertaken for the Medical Research Com- 
 mittee. 18 This work showed that the very fine particles which 
 reach the alveoli are rapidly taken up by special cells of the al- 
 veolar walls. When coal dust or shale dust was inhaled, these cells 
 soon detached themselves and wandered away with their load of 
 dust particles. Some pass directly into the open ends of the bron- 
 chial tubes, and are thence swept upwards by the cilia. Others 
 pass into lymphatic vessels and are carried to the nodules of lym- 
 
 15 Haldane, Martin, and Thomas, Report on the Health of Cornish Miners, 
 Parl. Paper Cd, 2091, 1904. 
 
 18 Arnold, Untersuchungen uber Staubinhalation und Staubmetastase, 1885. 
 
 17 Beattie, First Report of Explosions in Mines Committee, Parl. Paper, Cd, 
 6307. 1912. 
 
 18 Mavrogardato, Journ. of Hygiene, XVII, p. 439, 1918. 
 
324 RESPIRATION 
 
 phatic tissue surrounding bronchi and then pass right through 
 the walls of the bronchi and are swept out. Others reach the 
 lymphatic glands at the roots of the lungs, and finally seem to pass 
 from there into the blood. In this way the dust is removed from 
 the lungs, and if too much dust is not inhaled the process of re- 
 moval will keep pace with the introduction of dust. The well- 
 known "black spit" of a collier, which continues for long periods 
 when he is not working underground, is apparently a healthy 
 sign showing that dust particles are being removed from the 
 lungs. It seems quite probable, also, that the efficiency of the 
 physiological process for dealing with dust improves with use, 
 like other physiological processes. Moreover the dust-collecting 
 cells appear to be identical with cells which collect and deal 
 with bacteria in the lungs. Possibly, therefore, the somewhat re- 
 markable immunity of colliers from phthisis is connected with 
 their capacity for dealing with inhaled dust particles. 19 
 
 At the end of a few months the lungs of a guinea pig which have 
 been heavily charged with coal dust or shale dust by experimental 
 inhalations are again free from dust. On the other hand this was 
 not the case when the dust inhaled was quartz. Most of the quartz 
 remained in situ, though mainly within the dust-collecting cells. 
 Part had, however, been carried onward to lymphatic glands. The 
 quartz did not seem to excite the cells to wander in the same way 
 as the coal dust or shale dust did ; and it appeared as if this dif- 
 ference in the properties of different kinds of dust explained why 
 some dusts are much more apt than others to produce cumulative 
 ill effects in the lungs. Presumably the quartz particles are so 
 inert physiologically that they do not excite the dust-collecting 
 cells to wander away. Other kinds of dust particles may be equally 
 insoluble, but may also be charged with adsorbed material which 
 makes them physiologically active. Coal, for instance, though very 
 insoluble in water, adsorbs substances of all kinds, and the im- 
 portance of its power of adsorbing gases has already been pointed 
 out. 
 
 Shale dust was found by Dr. Mellor to contain about 35 per 
 cent of quartz. Nevertheless the quartz in shale dust does no harm 
 to the lungs and is eliminated readily. There are many other 
 kinds of stone which contain still more quartz, but also produce a 
 harmless dust. In fact nearly all the dust ordinarily met with is 
 of the harmless variety, and Mavrogardato's investigation indi- 
 
 19 Haldane, Trans. Inst. of Mining Engineers, LV, p. 264, 1918. 
 
RESPIRATION 325 
 
 cated that quartz dust becomes relatively harmless when it is 
 mixed with other dust of the harmless variety. The lung cells 
 appear to clear out the quartz when they are clearing out the other 
 dust. 
 
 It is evident that much further investigation is needed in order 
 to elucidate completely the physiology of dust excretion from the 
 lungs. It is equally evident, however, that this process is under 
 physiological control, just as much as other physiological activi- 
 ties are. 
 
 Air of Wells. The case of the air of wells and other unventilated 
 underground spaces differs from that of mines owing to the fact 
 that no artificial ventilation is provided for. It might be supposed 
 that the air in a well, with only rock or brickwork round it, pure 
 water at the bottom, and the top more or less open, would never 
 be more than slightly contaminated. Experience shows, however, 
 that this is not the case, and that the air in even a shallow well, 
 only a few feet deep, is sometimes dangerously contaminated. 
 In 1896 I investigated this subject, visiting various wells where 
 men had been asphyxiated, in order to see what had happened. 20 
 I found plenty of foul air, and that its composition was similar 
 to that of black damp, and not simply CO 2 , as was then believed. 
 The composition of the gas varied from about 80 per cent nitro- 
 gen and 20 per cent CO 2 to almost pure nitrogen ; and it was quite 
 evident that this black damp or choke damp was simply the 
 residual gas from oxidation processes occurring in the strata 
 round the well. 
 
 Another point which emerged quite clearly was that the state 
 of the air in any well liable to foul air depended entirely on 
 changes in barometric pressure. With a rising barometer the air 
 was quite clear, and with a falling barometer it was foul. Thus 
 any fall in barometric pressure might make a well very dangerous, 
 though an hour before the air was quite pure. Moreover with a 
 falling barometer the well might be brimful and rapidly over- 
 flowing with dangerous gas. The danger to which well sinkers 
 are exposed is thus evident At one well an engine house which 
 covered the top of the well had been built, and sometimes it was 
 unsafe to enter this building owing to the gas, unless doors and 
 windows were wide open. The engine man was much comforted 
 when I lent him an aneroid barometer and thus convinced him that 
 the outbursts of gas were due to natural and not supernatural 
 
 20 Haldane, Trans. Inst. of Mining Engineers, 1896. 
 
326 RESPIRATION 
 
 causes. By always carrying a lighted candle or lamp with him, a 
 well sinker can guard most effectually against the danger from 
 black damp ; but it is quite unsafe to trust to previous tests. 
 
 It is thus evident that a well acts as a chimney communicating 
 with a large air space in the substance of the surrounding rock, 
 or in crevices within it. Air may either be going down this chim- 
 ney or returning; and if the rock contains any oxidizable material 
 such, for instance, as iron pyrites, the returning air or gas has 
 lost more or less of its oxygen, and possibly also gained some CO 2 . 
 If, however, less than about 4 per cent of CO 2 were present in 
 the black damp it would be lighter than air, and thus likely to 
 escape unnoticed. 
 
 An interesting case which came under my notice later may be 
 mentioned in this connection. While a tunnel was being driven with 
 compressed air under the Thames it was found that in a large 
 cold storage on the river bank lamps or candles were extinguished. 
 The air was analyzed for CO 2 , but no noticeable excess was found. 
 On analysis I found the air very poor in oxygen. On further 
 investigation it turned out that air very poor in oxygen, but with 
 practically no excess of CO 2 , was coming up the shaft of a well 
 belonging to the building. 21 The flow did not depend on baro- 
 metric pressure, and nothing of the sort had occurred before the 
 construction of the tunnel began. It was evident, therefore, that 
 the flow was due to compressed air escaping deep down through 
 the London clay from the advancing end of the tunnel, and 
 forcing a way to the well, but at the same time losing oxygen 
 owing to the presence in the clay of oxidizable material such as 
 iron pyrites. The pure black damp contained 99.6 per cent of 
 nitrogen and 0.4 per cent of CO 2 . 
 
 Air of Railway Tunnels. Although the great difficulties form- 
 erly experienced in the ventilation of long railway tunnels have 
 been overcome by the substitution of electric traction for steam 
 locomotives, it may be worth while to record here some of these 
 difficulties. Probably the worst cases were those of single-line 
 tunnels on a stiff gradient in the Apennines. When the wind was 
 blowing in the same direction as a train was traveling on an up- 
 gradient the smoke from the engine or engines tended to travel 
 with the train. Thus the air rapidly became poisonous from the 
 presence of CO, and the oxygen percentage fell so low that some- 
 times lights were extinguished and steam began to fail, owing 
 
 21 Blount, Journ. of Hygiene, VI, p. 175, 1906. 
 
RESPIRATION 327 
 
 to the engine fires burning badly. The passengers could partly 
 protect themselves by closing the windows; but the engine drivers 
 were liable to become unconscious, and at least one very serious 
 accident occurred, owing to a train running on with the men on the 
 engine unconscious. 
 
 In the London Underground Railway there was also much 
 trouble, owing to the great traffic, although there were numerous 
 openings to the street along all parts of the system, and a colliery 
 fan had also been installed at one point. The difficulties were 
 referred to a Board of Trade Committee of which I was a member, 
 and I made numerous analyses of the air. 22 It was never so bad 
 as appeared to have been sometimes the case in the Apennine tun- 
 nels, and the trouble from sulphuric acid and smoke was largely 
 mitigated by the use of Welsh steam coal containing very little sul- 
 phur. The air was often, however, very unpleasant, and many 
 persons were unable to use the railway. At busy times the per- 
 centage of CO 2 might rise as high as 0.8, and of CO to .06 ; but of 
 course passengers and railwaymen were not long enough exposed 
 to this air to suffer from the effects of CO, and repairing work on 
 the line was not carried out except at night. At the end of the 
 inquiry it was agreed to introduce electric traction, and since this 
 was done there has been no further difficulty. The tunnels are close 
 to the surface, and the trains push abundance of air out and in 
 through openings to the outside air. 
 
 In the (London) tubes, which lie much deeper, the ventilating 
 action of the trains proved insufficient by itself to prevent the air 
 from becoming rather unpleasant; and systematic ventilation by 
 fans was therefore adopted. In various other railway tunnels 
 simple shafts are provided; and in the Severn Tunnel there is a 
 nearly central shaft provided with a powerful fan. By these 
 means the air is kept fairly pure. 
 
 Air of Sewers. The air of sewers is perhaps mainly of interest 
 in connection with the time-honored belief that "sewer gas" 
 spreads infection. Some of my earliest scientific work was con- 
 cerned with the air commonly present in sewers, and was started 
 by the late Professor Carnelley and myself 23 at the request of a 
 House of Commons' Committee appointed in consequence of alarm 
 as to the sewers of the House of Commons. 
 
 The air of a sewer has, of course, an unpleasant smell, which, 
 however, is hardly noticed except at the manhole by which access 
 
 M Report of the Committee on Tunnel Ventilation, Parl. Paper, 1897, Appendix i. 
 a3 Carnelley and Haldane, Proc. Roy. Soc., 4-?, p. 501, 1887. 
 
328 RESPIRATION 
 
 is gained to the sewer. The air is saturated with moisture, and 
 may be somewhat warm if much warm water flows into the sewer. 
 Chemically speaking, however, the air is very little contaminated. 
 Even in the sewers of Bristol, where ventilating shafts were re- 
 duced to a minimum, I found only about 0.2 per cent of CO 2 . On 
 determining the number of bacteria in the air we found that 
 fewer were present in the sewer air than outside, but of much the 
 same kinds. In sewers which were well ventilated there were far 
 more than in badly ventilated sewers; and it was evident that 
 nearly all the bacteria came from the outside through the venti- 
 lators. Where there was much splashing, however, a few were 
 thrown into the air. These results, which have been confirmed by 
 other investigators, are just what might be expected. Particulate 
 matter is not given off from moist surfaces apart from mechani- 
 cally acting causes, and any bacteria or other particles driven 
 into suspension in the air of a sewer will tend to fall back again. 
 It is conceivable that infection might be carried by sewer air ; but 
 innumerable other paths of infection are much more probable. 
 
 Although ordinary sewer air is chemically very pure, and not 
 even a trace of H 2 S can be found, accidents to sewermen from 
 foul air are not very uncommon ; and there is no doubt that most 
 of these accidents are due to H 2 S. I investigated one case of this 
 kind where five men had lost their lives at a manhole the last 
 four in attempts at rescue. 24 All the symptoms described, includ- 
 ing irritation of the eyes, were those of H 2 S poisoning ; and though 
 the air was not poisonous when I descended, a little H 2 S was 
 present. When some of the sewage was put into a large bottle and 
 shaken up, H 2 S was found to be present, and a mouse lowered 
 into the bottle showed severe symptoms of H 2 S poisoning. These 
 symptoms were absent when lead acetate was added before shak- 
 ing, or when caustic soda was added. 
 
 It is only when sewage stagnates or deposits solid matter that 
 H 2 S is formed. Any cause that stirs this sewage, or liberates H 2 S 
 from it, may make the air dangerous. About 0.2 per cent will kill 
 an animal within a minute or two ; and o. I per cent will rapidly 
 disable it. H 2 S is thus a good deal more poisonous than CO, and 
 far quicker in its action. 
 
 Another source of danger is lighting gas from leaky street 
 mains. Lighting gas is frequently met with in sewers, and I have 
 several times smelt it in sewers. In one recent case which I in- 
 vestigated two men were killed by CO poisoning from lighting 
 
 *' Lancet, Jan. 25, 1896. 
 
RESPIRATION 329 
 
 gas. There seems to be no evidence of accidents in sewers from any 
 other gas than H 2 S or CO ; but many strange smells are en- 
 countered, and we were once much alarmed by chlorine coming 
 from a bleaching factory. 
 
 Air of Ships. In the compartments of a ship air is specially 
 liable to become foul owing to the air-tight conditions which 
 often exist. In a double bottom compartment, for instance, the 
 whole of the oxygen may disappear, owing to rusting or to ab- 
 sorption of oxygen by drying paint. In an ordinary compartment 
 battened down the same thing may also occur owing to slow ab- 
 sorption of oxygen by articles of cargo, such as grain, wool, etc. 
 Accidents from this cause are not infrequent if men descend with- 
 out first testing the air with a lamp or giving time for ventilation 
 to occur. In coal bunkers fire damp may accumulate in the absence 
 of proper ventilation, or else the oxygen may fall very low. Coal 
 trimmers are occasionally also affected by what appears to be CO 
 poisoning due to small quantities of CO formed at ordinary 
 temperatures in the slow oxidation of coal, as described above. 
 
 The ventilation of passenger and crew spaces on ships was very 
 defective, particularly in rough weather, until fan ventilation was 
 generally introduced. It was forgotten that the rooms in a ship 
 do not ventilate themselves naturally through walls and roof, as 
 a house ashore does. Owing to the close quarters, it is often diffi- 
 cult to ventilate the spaces in a ship properly without causing 
 intolerable draughts. In the mess decks of warships this is specially 
 difficult, as there are hammocks everywhere at night. The matter 
 was investigated recently by an Admiralty Committee of which 
 I was a member and a system introduced by which equal amounts 
 of air can be made to issue from a large number of louvres on the 
 sides of ventilating ducts. In this way the men are supplied with 
 an average of 50 cubic feet of air each per minute, without any 
 unpleasant draught impinging on any one. The temperature, and 
 particularly the wet-bulb temperature in warm weather, can also 
 be controlled very efficiently by this plan. With men perspiring 
 more or less from heat, and giving off perhaps fifty times as great 
 a volume of aqueous vapor as of CO 2 , very ample artificial venti- 
 lation is needed when no other means of ventilation is available. 
 
 Gas Warfare. It would be out of place to attempt to discuss the 
 nature and mode of action of the various substances used in gas 
 warfare ; but a certain number of facts of physiological interest 
 in connection with respiration may be fitly referred to here. 
 
 The first serious gas attacks were made, as is well known, with 
 
330 RESPIRATION 
 
 chlorine, discharged into the air in a good breeze as "drift gas" 
 from cylinders of liquefied gas. The liquefied gas quickly evapo- 
 rated, thus cooling a large body of air which rolled along the 
 ground, producing at the same time a mist if the air was nearly 
 saturated, and passing downwards into every trench. From ac- 
 counts given by officers and men at the time, I estimated that 
 along the lines attacked there was usually about .01 per cent of 
 chlorine in the air; but of course the percentage would vary. At 
 about this and higher concentrations, chlorine has an immediate 
 and severe irritant effect on the air passages, and a less severe 
 action on the eyes. Bronchitis follows if the exposure lasts for 
 more than a very short time, and some time later symptoms of 
 oedema of the lungs appear, owing to the action of the gas on the 
 alveolar walls. The symptoms are then similar to those which 
 follow exposure to nitrous fumes. The men suffering from this 
 condition were deeply cyanosed, with superficial veins about the 
 neck prominent, greatly increased depth and rate of breathing, 
 and a frequent, but usually fairly strong pulse. Intelligence was 
 clouded, but the distress seemed very great. 
 
 On testing a drop of blood by diluting it to a yellow color, 
 saturating with coal gas and comparing the pink tint thus pro- 
 duced with the tint of normal blood similarly diluted, it was evi- 
 dent that there was no decomposition of the haemoglobin. The 
 cyanosis was therefore due to imperfect saturation of the blood 
 with oxygen. That the imperfect saturation was due, not to slow- 
 ing of the circulation, but to imperfect saturation in the lungs, 
 was shown at once by the effect of giving oxygen. This abolished 
 the cyanosis, cleared up the clouded intelligence, but had no great 
 effect on the breathing. On post mortem examination of fatal 
 cases it was found that the lungs were voluminous and greatly 
 congested. Large quantities of albuminous liquid could be 
 squeezed out through the cut bronchi, and there was much em- 
 physema. 
 
 The interpretation of the more dangerous symptoms seems 
 fairly clear. The cyanosis was due to the fact that the blood in 
 passing through the lungs was imperfectly oxygenated, owing 
 mainly to swelling and exudation, which hindered the diffusion 
 of oxygen inwards to the blood. On raising the alveolar oxygen 
 pressure when oxygen was given, the diffusion became much 
 faster and the blood was properly oxygenated. The hyperpnoea 
 remained, however, and was probably attributable to the fact 
 that though much air was entering the lungs, a great deal of it 
 
RESPIRATION 331 
 
 only went into the emphysematous spaces where there was little 
 or no circulation, leaving the rest of the lung imperfectly venti- 
 lated, with an abnormal excess of CO 2 in the alveoli which were 
 permeable to blood, and consequently an abnormal excess of 
 breathing. 
 
 Considering the depth of the cyanosis it was somewhat re- 
 markable that consciousness was not more impaired; but the ex- 
 cess of CO 2 which accompanied the cyanosis would of course 
 facilitate the dissociation of oxyhaemoglobin in the tissue capil- 
 laries, and thus diminish the real anoxaemia. The distention of 
 superficial veins was an indication of the veno-pressor effect of 
 excess of CO 2 combined with failure on the part of the heart to 
 respond normally to the large amount of blood returning to it 
 from the tissues. This failure was evidently due to the anoxaemic 
 condition of the blood supplied to the heart. The failure was pre- 
 sumably most marked in the left ventricle, which has far the most 
 work to do, and the consequence would be a rise of blood pressure, 
 not only in the veins, but also in the right side of the heart and 
 the whole pulmonary circulation. The rise in pulmonary blood 
 pressure would of course tend to aggravate greatly the oedema of 
 the lungs, and would thus in itself be a very serious source of 
 danger. The ease with which oedema of the lungs follows on in- 
 creased venous blood pressure, even when there is no injury to the 
 lungs, has been shown experimentally by Knowlton and Star- 
 ling. 26 
 
 The cause of the greatly increased flow of blood was simply 
 the fact that the arterial blood was in a venous condition, with 
 both a lowered oxygen pressure and raised CO 2 pressure. The 
 perfectly normal effect of this, as pointed out in Chapter X, is to 
 cause dilation of capillaries and increased blood flow through the 
 tissues. Owing, however, to the pressor reaction of the vasomotor 
 center, the arterioles and probably also the venules in most parts 
 of the body except the central nervous system were contracted, 
 and in this way the blood pressure was maintained, so that the 
 pulse was of good strength. 
 
 It was first observed by Macaulay and Irvine of Johannesburg 
 that in the treatment of cases of oedema of the lungs from 
 poisoning by nitrous fumes in mines, great benefit is often ob- 
 tained by free bleeding to the extent of about half a liter. From 
 the foregoing account it is clear that bleeding will reduce the 
 venous and pulmonary blood pressure, and thus also- reduce the 
 
 M Knowlton and Starling, Journ. of Physiol., XLIV, p. 206. 
 
332 RESPIRATION 
 
 tendency to oedema of the lungs. The indication for bleeding is 
 evidently the distention of superficial veins. Bleeding was fre- 
 quently employed in the treatment of the chlorine cases, and with 
 great success. It is evident, however, that if there is no venous 
 distention, bleeding could not be expected to do anything but 
 harm. 
 
 A more radical treatment is the continuous administration of 
 air enriched with oxygen. Unfortunately the problem of con- 
 tinuous administration of oxygen had never been attacked before 
 the war, and no suitable apparatus was available for the early 
 chlorine cases. But in the later stages of the war many cases of 
 lung oedema were successfully treated continuously with oxygen 
 by means of a nasal tube or the apparatus described in Chapter 
 VII. 
 
 The next lung irritant gas used was phosgene (COC1 2 ). This 
 produces dangerous effects in considerably lower concentration 
 than chlorine, and its action is distinguished by the fact that it 
 has relatively less effect on the air passages and eyes and in the 
 end more on the alveolar walls. Thus a man exposed to a danger- 
 ous concentration of phosgene may notice but little irritant effect 
 at the time, or this effect may pass off rapidly, while the dangerous 
 effects on the alveoli only show themselves some hours later. 
 Phosgene was at first used as drift gas; but when drift gas was 
 abandoned as more or less ineffective against the protective 
 measures adopted, and also unmanageable owing to uncertain- 
 ties of wind, etc., phosgene was largely used in shells or bombs. 
 Various other substances with similar toxic properties were also 
 employed. 
 
 A change in the type of the symptoms accompanying lung 
 oedema was now noticed. The deep plum-colored cyanosis and 
 venous distention were usually absent, and bleeding was useless. 
 The cyanosis was still very marked, but was of a pale or "gray" 
 type. The breathing was also shallower, and the pulse feeble and 
 rapid. Many slighter cases were also observed in which no defi- 
 nite lung symptoms were observed, but only general malaise with 
 cyanosis and fainting on any muscular exertion. 
 
 In all these cases it seems evident that the rate of diffusion of 
 oxygen through the alveolar walls was diminished, but without 
 any marked interference with diffusion of CO 2 outwards, so that 
 owing to the hyperpnoea from want of oxygen there would be a 
 deficiency of CO 2 in the arterial blood. This is very intelligible 
 in view of the fact that on account of its greater solubility CO a 
 
RESPIRATION 333 
 
 diffuses outwards from the blood much more readily than oxygen 
 diffuses inwards (see Chapter VIII). The deficiency of CO 2 in 
 the arterial blood would prevent or minimize the true hyperpnoea, 
 and lessen the increase of circulation through the tissue capillaries 
 and the pressor excitation of the vasomotor center. But it would 
 increase the true tissue anoxaemia with a given degree of cyanosis. 
 Anoxaemia in the coronary circulation would also lead to the 
 enfeebled action of the heart, as shown by the very weak and 
 feeble pulse. The symptoms generally were those of a pure anox- 
 aemia with urgent danger of failure of the respiratory center in 
 the manner already referred to in Chapter VI. 
 
 In these cases bleeding was of course useless. On the other 
 hand injection into the blood of saline solution or, still better, 
 gum-saline, seemed likely to be of some use in view of the failing 
 blood pressure. By far the most effective treatment, however, was 
 the continuous administration of air enriched with oxygen, par- 
 ticularly if this was begun early and before there was time for the 
 dangerous effects which continued severe anoxaemia causes. By 
 this means the oxygen pressure in the alveolar air was sufficiently 
 raised to permit of a nearly normal aeration of the arterial blood ; 
 and the administration could be continued till the lung inflamma- 
 tion subsided. 
 
 The chronic after effects on the respiratory center of irritant 
 gases have already been referred to in former chapters. 
 
CHAPTER XII 
 Effects of High Atmospheric Pressures. 
 
 THE foundations of our scientific knowledge of the physiological 
 effects of high and low atmospheric pressures were laid broad 
 and firm b/ the investigations of Paul Bert, collected together in 
 his book, already so often referred to, "La Pression Barome- 
 trique," published in 1878. It will be convenient to consider first 
 the effects of high atmospheric pressures. 
 
 Very high atmospheric pressures are met with in deep diving 
 and in engineering work under water or in water-logged strata. 
 
 Apart from laboratory experiments on animals, the highest 
 atmospheric pressures (up to ten atmospheres) have been met 
 with in deep diving. To understand the conditions under which 
 a diver is placed, it is necessary to understand the design of the 
 ordinary diving dress, which was introduced early last century by 
 Siebe, the founder of the well-known London firm of manu- 
 facturers of diving apparatus. The dress consists of a copper 
 helmet which screws on to a metal corselet, the latter being 
 clamped water-tight to a stout waterproof dress covering the 
 whole body except the hands, which project through elastic cuffs 
 (Figures 75 and 76). Air is supplied to the diver through a non- 
 return valve at the back of the helmet from a stout flexible pipe 
 strengthened with steel wire and connected with an air pump at 
 the surface. The air escapes through an adjustable spring valve 
 at the side of the helmet (Figure 77). The arrangement is thus 
 such that the pressure of air in the helmet is at least equal to, and 
 can, by varying the resistance of the valve, be made greater than, 
 the water pressure at the outlet valve. For every 34 feet of fresh 
 water (or 33 feet or 10 meters of sea water) the pressure in- 
 creases by one atmosphere, or nearly 15 pounds per square inch. 
 At a depth of 33 feet of sea water the diver is therefore breathing 
 air at an excess pressure of one atmosphere, or a total pressure of 
 two atmospheres. It is absolutely necessary that he should breathe 
 compressed air, otherwise his breathing would be stopped in- 
 stantly by the pressure of the water upon the abdomen ; and at a 
 greater depth blood would probably pour from his nose and mouth 
 on account of the squeezing to which all parts of his body, except 
 his head in the helmet, would be subjected. 
 
Figure 75. 
 
 Diving dress, front view, with air pipe and life line, 
 which are connected with the helmet behind. 
 
Figure 76. 
 
 Diving dress, back view, showing attachment of air pipe 
 and life line with telephonic connection ; new pattern, with 
 legs laced up to prevent diver from being capsized and 
 accidentally blown up to surface, or hung in a helpless 
 position. 
 
RESPIRATION 
 
 335 
 
 In order to enable the diver to sink and stand firmly on the 
 bottom, the dress is weighted with 4O-pound leaden weights, 
 back and forward, as shown, with 16 pounds of lead on each 
 boot about 112 pounds of lead in all. Besides the air pipe, the 
 diver is connected with the surface by a so-called life line, which 
 usually contains a telephone wire. He goes down by a rope at- 
 
 Figure 77. 
 Helmet and section of outlet valve. 
 
 tached to a heavy weight which has been lowered to the bottom 
 previously, and on reaching the bottom he takes with him a line 
 attached to this weight, so that he can always find the rope again. 
 As a diver enters the water, the superfluous air in his dress is 
 driven out through the outlet valve by the pressure of the water 
 round his legs and body. The water seems to grip him all round. 
 If the valve is freely open he feels his breathing somewhat 
 
336 RESPIRATION 
 
 labored by the time he gets first under water. The reason of this is 
 that the pressure in his lungs is that of the water at the valve 
 outlet, whereas the pressure on his chest and abdomen is greater 
 by something like a foot of water. He is thus inspiring against 
 pressure, and if he has to breathe deeply, as during exertion, the 
 breathing is apt to become fatigued in the manner described in 
 Chapter III. With another foot of adverse pressure the fatigue is 
 very rapid. One of the first things which a diver has to learn is 
 to avoid the adverse pressure by regulating the spring on the 
 outlet valve, so that the breathing is always easy. The spring 
 regulates at the same time the amount of air in the dress, and 
 therefore the buoyancy of the diver. A practiced diver can thus 
 slip easily, and without exertion, up or down the rope. A pres- 
 sure gauge attached to the air pipe where it leaves the pump 
 indicates the depth of the diver at any moment. 
 
 The breathing is of course easiest when the dress is full of air 
 down to the level of the diaphragm, but when this is so the diver 
 is in danger of being "blown up" ; for if he is crawling on the 
 ground, it may easily happen that the air gets into the legs of 
 his dress. His head goes down so that the excess of air can- 
 not escape readily. He is then blown helplessly to the surface, 
 while his arms are fixed in an outstretched position (see Figure 
 78). His air pipe may be caught by a rope or other obstruction, 
 so that he is hung up in a helpless position with his legs upwards, 
 the excess of air being unable to escape at the valve since it is 
 downwards. In very deep diving there is considerable risk of 
 being blown up ; and to avoid this risk the arrangement for lacing 
 up the legs, shown in Figure 76, was introduced (see also Fig- 
 ure 79). 
 
 In the Denayrouze apparatus, extensively used on the Conti- 
 nent, the air is pumped into a steel reservoir on the diver's back. 
 By means of a reducing valve his air is supplied from the reser- 
 voir according to his requirements. The arrangement is a beauti- 
 ful piece of mechanism, but an encumbrance which gives rise to 
 various inconveniences and dangers, one being that the depth of 
 the diver cannot be read off at the surface, and another that he 
 cannot regulate the pressure in his helmet. 
 
 For engineering work in preparing foundations, etc., on the 
 sea bottom, a diving bell is sometimes employed. This is a heavy 
 metal box, open below, and supplied with compressed air by a 
 pipe (Figure 80). It is lowered to the bottom with the workmen 
 sitting in it, and they can work dry on the bottom. The diving 
 
Figure 78. 
 
 Diver in ordinary dress blown up. His head is down and his arms 
 outstreched. 
 
 Figure 79. 
 
 Diver in laced-up dress purposely blown up. His head 
 is up and his arms free. 
 
Figure 80. 
 
 Diving bell in use at National Harbour Works, Dover. Each bell 
 measures 17x10 feet by 6 YZ feet high, and weighs about 3 5 tons. 
 
 Figure 81. 
 
 Diagram showing use of caisson in making the foundations 
 of a bridge. (After Foley.) 
 
RESPIRATION 337 
 
 bell in its crude original form was invented by Sturmius in the 
 sixteenth century, and further developed by Halley two centuries 
 later. 
 
 The caisson introduced about 1840 by the French engineer 
 Triger, for sinking colliery shafts through water-logged strata 
 near the surface, is a further development of the diving bell. It 
 is now largely used for carrying the foundations of the piers of 
 bridges, etc., through soft ground on the bottom of a river or the 
 sea. The caisson (see Figure 81) is the bottom section of the 
 steel pier, and resembles a diving bell except for the fact that it 
 communicates with surface through a tube occupying the center 
 of the future pier and kept full of compressed air. This tube 
 serves for access and for removal of excavated material. The men 
 excavate the soft bottom so as to allow the caisson to sink down 
 to a secure foundation, and the sections of the pier are added 
 from above and filled with concrete as the caisson sinks. Access 
 to the central tube is through an air lock on surface. The men 
 enter the air lock, close the door, and then let the air pressure 
 rise till they can open the door into the central tube ; and in coming 
 out the reverse process is used. 
 
 In tunneling operations in soft strata under water, the ad- 
 vancing tunnel is kept full of compressed air, so as to hinder the 
 penetration of water into the advancing end, as the steel rings 
 forming the permanent walls of the tunnel are successively put 
 in. The men thus work in an atmosphere of compressed air, to 
 which access is gained through one or more air locks. The tubes 
 and large tunnels under the Thames or deep in the water-logged 
 London clay, and under the Hudson and East Rivers at New 
 York, have been, or are being, constructed by this means. In the 
 sinking of colliery shafts through water-logged strata the freez- 
 ing or cementation processes are now generally used, as, except 
 in strata fairly near the surface, the water pressures are too high 
 for the compressed-air process. 
 
 Various physiological disturbances are associated with ex- 
 posure to compressed air, and these must now be considered one by 
 one. As the pressure rises when a man goes below water, in a 
 diver's suit, or as compressed air enters an air lock through which 
 he is passing to a caisson or tunnel, the first trouble usually noticed 
 is a sense of pressure and pain in the ears. This is due to un- 
 balanced pressure on the membrana tympani, owing to the fact 
 that the Eustachian duct does not open freely so as to equalize 
 the air pressure in the middle ear with the atmospheric pressure 
 
338 RESPIRATION 
 
 outside. The passage is specially liable to be blocked if any 
 catarrh of the air passages is present ; and if the warning pain is 
 disregarded the membrane may burst, though this is not a very 
 serious accident. In men accustomed to compressed air the Eus- 
 tachian tubes open easily, so that no inconvenience is felt, and a 
 diver goes quite easily within two minutes to a pressure of seven 
 atmospheres or more, while one who is not accustomed to com- 
 pressed air may have a long struggle with his Eustachian tubes 
 before he can reach an extra pressure of half an atmosphere. It 
 also happens occasionally that there is trouble with the frontal 
 sinuses. The same difficulties with the middle ear may, of course, 
 be met with by airmen during rapid descents, or even, to a minor 
 extent, in descending a deep mine shaft. 
 
 A man who has reached a pressure of six or seven atmospheres, 
 and is breathing pure air, is perfectly comfortable if he has es- 
 caped ear trouble. His voice is, however, altered by the com- 
 pressed air, and this is so marked that it is often difficult to make 
 out through the telephone what he is saying. At first sight it 
 might seem that an increased mechanical pressure of several 
 atmospheres would in itself be expected to have an appreciable 
 effect on a man or animal. It was commonly supposed, for ex- 
 ample, that the increased pressure on the skin must at first tend to 
 drive blood into the internal organs, producing congestion of the 
 brain, etc., with a converse effect on diminishing the atmospheric 
 pressure. The pressure is, however, transmitted instantly through- 
 out all the liquid and solid tissues of the body, so that this idea 
 was totally fallacious, and indeed ridiculous. As will be seen 
 below, many divers have lost their lives owing to well-meant in- 
 junctions to descend and ascend slowly. As regards other possible 
 effects of a few atmospheres of mechanical pressure, it should be 
 remembered that the intrinsic pressure of water is calculated to 
 be over 10,000 atmospheres. As the tissues are largely composed 
 of water, the addition to this of a few atmospheres of mechanical 
 pressure in the liquid or semi-liquid parts of the body cannot be 
 of much account. 
 
 As Paul Bert showed experimentally, the serious inconveni- 
 ences and dangers to which workers in compressed air are ex- 
 posed are due (apart from easily avoidable effects on the ears) 
 not to the mechanical pressure, but to the increased partial pres- 
 sures of the gases in the air breathed. If the air breathed is pure, 
 the only gases which come into consideration in this connection 
 are nitrogen and oxygen; but if the air is rendered impure by 
 
RESPIRATION 339 
 
 respiration, as is commonly the case in diving, carbon dioxide 
 also comes into consideration. The case of this gas may be con- 
 sidered first, though Paul Bert did not himself allude to it in 
 connection with work in compressed air, as he was not practically 
 familiar with diving. 
 
 Owing to the difficulties frequently experienced by divers in 
 attempts to work at depths over about 12 fathoms a Committee, 
 including myself as the physiological member, was appointed by 
 the British Admiralty to investigate the whole subject of the 
 difficulties and dangers associated with deep diving. 1 It appeared 
 that men who attempted to make any serious exertion when at 
 depths of over about 12 fathoms often became unconscious or 
 greatly exhausted. The symptoms pointed to excess of CO 2 , and, 
 on taking samples from the divers' helmets at about this depth, 
 we frequently found 2 or 3 per cent of CO 2 . This occurred in spite 
 of an apparently abundant supply of air from the pumps, which 
 were working at a much faster rate than was sufficient to keep the 
 diver comfortable at a lesser depth. As explained in Chapter II, 
 the physiological effects of 3 per cent of CO 2 at 1 1 fathoms, or a 
 total pressure of three atmospheres, is equal to that of 3 x 3 = 9 
 per cent at normal atmospheric pressure; so no wonder the divers 
 became unconscious. The pumps were often found to be leaking 
 badly through the piston rings, as many of them were old, and 
 no tests were then employed to detect this leakage. Apart from 
 this cause, however, the air supply was often insufficient. 
 
 It is evident that in order to keep down the pressure of CO 2 
 in the air of the helmet to a proper limit, the amount of air as 
 measured at surface by the strokes of the pump must be increased 
 in proportion to the increase in the total atmospheric pressure in 
 the helmet. The diver at 3 atmospheres pressure, requires, there- 
 fore, three times as much air, and so on in proportion to the 
 pressure. When this was attended to, and the piston rings kept 
 tight, no discomfort whatsoever was experienced at a depth of 
 even 35 fathoms. With a full air supply, hard exertion is actually 
 easier to a diver at some depth than near surface, on account of 
 the higher oxygen pressure, as explained in Chapter IX. 
 
 By far the most serious danger to divers and other workers in 
 compressed air is of a quite different character. From the earliest 
 days of diving and work in compressed air it had been observed 
 that soon after returning to atmospheric pressure the men fre- 
 
 1 Re-port of the Admiralty Committee on Deep Water Diving, Parl. Paper, C. N., 
 1549, 1907. 
 
340 RESPIRATION 
 
 quently became ill, and sometimes died or became paralyzed. 
 The risk of these attacks increased with the pressure and the 
 duration of exposure to it, but they never occurred except on 
 return to atmospheric pressure. Divers are exposed to the highest 
 pressures, and in divers the attacks were of the most dangerous 
 character. In the worst cases the diver began to feel faint a few 
 minutes after return to surface ; soon he became unconscious 
 and his pulse disappeared; and in a few minutes he was dead. 
 In other cases his legs became paralyzed, and cases of "diver's 
 paralysis" used to be not uncommon in British hospitals. In the 
 slighter cases, very common among workers in caissons and tun- 
 nels under construction, there is severe pain, known to the work- 
 men as "bends," in one or other of the limbs, or in the body. 
 Another of the common slight symptoms is itching of the skin. 
 Various other nervous symptoms are also met with, the whole 
 complex being designated as "caisson disease" a somewhat mis- 
 leading name. 
 
 Paul Bert investigated on animals the nature of compressed 
 air illness or "caisson disease," and found that it is due to libera- 
 tion in the blood and tissues of bubbles of gas consisting almost 
 entirely of nitrogen. In the rapidly fatal cases the heart becomes 
 filled with a mass of bubbles which stop the whole circulation. 
 In the cases of paralysis bubbles have obstructed the circulation 
 and so caused necrosis of parts of the spinal cord ; and it is evi- 
 dent that the bubbles may produce the most varied symptoms 
 according to the positions in which they are formed. 
 
 The cause of the bubble formation was evident. At the high 
 pressure the blood in the lungs is exposed to greatly increased 
 partial pressures of nitrogen and oxygen, although, as shown in 
 Chapter II, there is no increased pressure of CO 2 . As, in ac- 
 cordance with Henry's law, liquids take up in simple solution a 
 mass of any gas proportional to its partial pressure, the blood in 
 the lungs takes up in the compressed air an extra amount of nitro- 
 gen and oxygen proportional to the increased pressure. The extra 
 oxygen disappears at once when the blood reaches the tissues, but 
 the extra nitrogen does not disappear, and gradually saturates 
 the whole of the tissues till they are charged with nitrogen at the 
 partial pressure existing in the air breathed. When the external 
 atmosphere is reduced to normal, the internal partial pressure of 
 nitrogen is of course far above the atmospheric pressure. The 
 blood and tissues are therefore supersaturated with nitrogen and 
 bubbles begin to form. These bubbles consist primarily of nitro- 
 
Figure 82. 
 
 Portion of goat's mesentery showing bubbles in blood vessels caused by 
 rapid decompression in 1^2 minutes from 100 Ibs. pressure, after ij^ hours 
 exposure at this pressure. 
 
RESPIRATION 341 
 
 gen, but of course take up a little oxygen and CO 2 from the sur- 
 rounding blood and tissue liquids. If they are formed in the blood 
 they tend to block the circulation on account of the great resist- 
 ance which they cause. Figure 82 is from a photograph of blood 
 vessels in the mesentery of a goat killed by rapid decompression, 
 and shows abundant bubbles in the veins. 
 
 The bubbles are formed, not merely in the blood, but also in 
 the tissues outside it. We found that fat in particular is apt to be 
 very full of bubbles and thus become spongy. It had been found 
 by Vernon in connection with another investigation that gases 
 are much more soluble in oils than in water. In connection with 
 our investigations he determined the solubility of nitrogen in 
 body fats at blood temperature, and found that it is about six 
 times as great as in water. 2 The tendency of fatty substances to 
 act as a special reservoir of dissolved nitrogen is thus intelligible ; 
 and Boycott and Damant 3 afterwards showed that fat animals, 
 other conditions being the same, are considerably more liable to 
 symptoms of caisson disease than spare animals. Not only ordi- 
 nary fat, but the myelin sheaths of nerve fibers, will form reser- 
 voirs of dissolved nitrogen; and for this reason bubbles will tend 
 to be liberated in the white matter of the brain and spinal cord, 
 and inside the sheaths of large nerves. The "bends" and certain 
 other associated symptoms from which workers in compressed 
 air so frequently suffer are probably due to liberation of bubbles 
 from the gas dissolved in the myelin sheaths. It is difficult to un- 
 derstand otherwise the severe pain of "bends." Figure 83 shows 
 the positions of a large number of bubbles found in the white 
 matter at different parts of the spinal cord. 
 
 The increased amount of nitrogen dissolved in the blood at 
 high atmospheric pressures was demonstrated by Paul Bert by 
 blood-gas analyses ; and Hill and Greenwood 4 not only confirmed 
 this, but showed that there is the same excess in the urine. Hill 
 and Macleod also observed directly the sudden appearance of 
 gas bubbles in the capillaries of the frog's web when the animal 
 was decompressed from a high atmospheric pressure. 5 
 
 As a preventive of the occurrence of caisson disease Paul Bert 
 recommended slow and gradual decompression; but his experi- 
 ments in this direction were not very successful, as he had not 
 
 2 Vernon, Proc. Roy. Soc., LXXIX, B, p. 366, 1907. 
 
 3 Boycott and Damant, Journ. of Hygiene, VIII, p. 445, 1908. 
 *Hill and Greenwood, Proc. Roy. Soc., LXXIX, B, p. 21, 1907. 
 'Hill and Macleod, Journ. of Hygiene, III, p. 436, 1903. 
 
342 
 
 RESPIRATION 
 
 2nd cervical. 
 
 3rd dorsal. 
 
 Figure 83. 
 
 Shows the distribution of extravascular bubbles in five regions of the spinal 
 cord of goat 3 (series IV). The animal died of oxygen poisoning during de- 
 compression after 3 hours' exposure at 81 Ibs. in an atmosphere containing 
 36 per cent oxygen. The bubbles are practically confined to the white matter 
 and are there especially concentrated in the boundary zone where the circula- 
 tion is least good. Each diagram is a composite drawing showing all the bubbles 
 in 0.4 mm. length of cord. (After Boycott, Damant, and Haldane.) 
 
RESPIRATION 343 
 
 completely realized the conditions. Slow and uniform decompres- 
 sion was, and still is, also enjoined by various government regu- 
 lations, etc., in different countries, but with only very moderate 
 success; and deaths or paralyses from caisson disease remained 
 common if the extra pressure used was above about 1.5 atmos- 
 pheres. Workers in compressed air had soon discovered that the 
 pain of "bends" can be relieved at once by returning into the com- 
 pressed air; and this became quite intelligible from Paul Bert's 
 experiment. He made some experiments on the curative effects 
 of recompression, but here again he was not very successful, as 
 he applied the remedy only in extreme cases. Medical recompres- 
 sion chambers for the treatment of compressed air illness were 
 first introduced by Sir Ernest Moir in connection with the con- 
 struction of the first East River tunnel at New York, and the 
 Blackwall Tunnel under the Thames, about 1890. They proved 
 strikingly successful when applied to the cases which occurred 
 with the comparatively slow decompression in the air lock. Pa- 
 ralyses and "bends" were relieved at once, even when they had 
 occurred a considerable time after leaving the tunnel. The pro- 
 vision of medical recompression chambers has now become a 
 necessary adjunct of all considerable engineering undertakings 
 at pressures of over about 1.5 atmospheres, and in extensive deep 
 diving operations. Figures 84 and 85 show one of the recompres- 
 sion chambers used in the British Navy. The trouble, however, 
 about the use of recompression chambers is that it is often very 
 difficult to get the patient out without the symptoms recurring. 
 The decompression may require many hours, or even days in bad 
 cases. 
 
 Paul Bert also tried another method of treatment that of 
 administering pure oxygen to his animals. This must hasten the 
 diffusion outwards of nitrogen, while the oxygen itself is ab- 
 sorbed by the tissues. At first sight it might seem as if this plan 
 ought to be very successful, either in treatment or in the pre- 
 vention of bubble formation during decompression. The results, 
 however, were disappointing and from causes which will be made 
 evident below. There seems, however, to be some scope for oxy- 
 gen administration where there is great difficulty in getting a 
 patient out of a medical air lock, and where there is no fear of 
 oxygen poisoning a condition which will be discussed presently. 
 
 When the Admiralty Committee had dealt with the troubles 
 traced to CO 2 , it was faced by the dangers of caisson disease, 
 which of course became much more important after it had been 
 
344 RESPIRATION 
 
 rendered possible for divers to work at great depths without in- 
 convenience. The existing precautions against "caisson disease" 
 were evidently quite insufficient. The divers were officially en- 
 joined to descend and come up at a slow and even rate of about 
 5 feet per minute, but many serious or fatal cases were occurring 
 in spite of this. The problem was to find a safe and reasonably 
 short method. Very slow methods are impractible on account of 
 changes of tides and weather. The whole physiological side of 
 compressed-air illness had therefore to be reconsidered. 
 
 The formation of bubbles depends, evidently, on the existence 
 of a state of supersaturation of the body fluids with nitrogen. 
 Nevertheless there was abundant evidence that when the excess 
 of atmospheric pressure does not exceed about i% atmospheres 
 there is complete immunity from symptoms due to bubbles, how- 
 ever long the exposure to the compressed air may have been, and 
 however rapid the decompression. Thus bubbles of nitrogen are 
 not liberated within the body unless the supersaturation corre- 
 sponds to more than a decompression from a total pressure of 
 2/4 atmospheres. Now the volume of nitrogen which would 
 tend to be liberated is the same when the total pressure is halved, 
 whether that pressure be high or low. Hence it seemed to me 
 probable that it would be just as safe to diminish the pressure 
 rapidly from 4 atmospheres to 2, or 6 atmospheres to 3, as from 
 2 atmospheres to i. If this were the case, a system of stage 
 decompression would be possible, and would enable the diver to 
 get rid of the excess of nitrogen through his lungs far more 
 rapidly than if he came up at an even rate. The duration of ex- 
 posure to a high pressure could also be shortened very consid- 
 erably, without shortening the period available for work on the 
 bottom. 
 
 The whole matter was put to the test in a long series of experi- 
 ments carried out on goats by Professor Boycott, Commander 
 Damant, and myself 6 at the Lister Institute, London, in a large 
 steel chamber which was given for the purpose by the late Dr. 
 Ludwig Mond (see Figures 86 and 87). We found that after 
 very long exposure of a number of the animals at a total pressure 
 of 6 atmospheres sudden decompression to 2.6 atmospheres pro- 
 duced not the slightest ill effect. This decompression is in the 
 proportion of 2.3 to I, and the drop of pressure was 3.4 atmos- 
 pheres. In a corresponding series where the drop of pressure was 
 
 a Boycott, Damant, and Haldane, Journ. of Hygiene, VIII, p. 242, 1908. The 
 Report of the Admiralty Committee contains a short abstract of the work. 
 
Figure 84. 
 
 Outside of naval recompression chamber, showing man- 
 hole for access, and air lock for food. 
 
 Figure 85. 
 Inside of recompression chamber, showing bed for patient. 
 
O +3 
 IH 
 
 ** H ^3 
 
 rl 
 
 > c 
 
 u 
 
 
 
 T3 c 
 
 w .1 | 
 
 2i Jr s 
 
 M .... K e 
 
 || 
 
 cJ bfl O 
 
 ^ S 
 H ^2 
 
 o 
 t3 
 
 Us 
 
RESPIRATION 345 
 
 the same, but from 4.4 to I atmosphere, or in the proportion of 
 4.4 to i, only 20 per cent of the animals escaped symptoms, while 
 20 per cent died, 30 per cent had severe symptoms, and 30 per 
 cent had "bends," quite easily recognized in the animals by their 
 behavior and the manner in which they held the affected limb 
 (Figure 88). It seemed evident, therefore, that it is quite safe to 
 halve the absolute pressure rapidly. Before venturing on such 
 extensive rapid decompressions of divers under water we re- 
 peated the goat experiments on men in the steel chamber, Com- 
 mander Damant and Lieutenant Catto being the subjects. There 
 were no ill effects in a number of experiments, nor in subsequent 
 trials by them under water at sea; and rapid decompression to 
 half the absolute pressure is now the routine practice of divers, 
 and is not known to have ever resulted in harm. 
 
 We were still, however, only at the beginning of the inquiry. 
 It was evident that the whole danger lay in the last stages of the 
 decompression. "On ne paie qu'en sortant," as was remarked by 
 Pol and Watelle, who were the first to give a medical account of 
 the symptoms of caisson disease. 7 The problem was to get divers 
 completely clear of the compressed air without paying. This 
 problem had resolved itself into that of avoiding the critical 
 supersaturation with nitrogen in any part of the body at or before 
 the last stage of decompression. 
 
 Let us consider the process of saturation and desaturation more 
 closely. The blood passing through the lungs of a man breathing 
 compressed air will, in accordance with what has been explained 
 in Chapter IX as to the permeability of the lung epithelium to 
 gas, become instantly saturated to the full extent with nitrogen 
 at the existing partial pressure in the air. When this blood 
 reaches the systemic capillaries, most of the excess of nitrogen 
 will diffuse out and the blood will return for a fresh charge, this 
 process being repeated until at length the tissues are fully charged 
 with nitrogen at the same partial pressure as in the air. But the 
 blood supply to different parts of the body varies greatly, as we 
 have seen. The capacity of different parts of the body for dis- 
 solving nitrogen varies also. Thus the white matter of the central 
 nervous system has but a small blood supply and at the same time a 
 high capacity for storing nitrogen ; and the same remark applies to 
 fat. The gray matter, on the other hand, has an enormous blood 
 supply and no extra capacity for storing nitrogen. Other tissues, 
 such as muscles, may or may not have a great blood supply, ac- 
 
 T Pol et Watelle, Ann. d,' hygiene -pubUqiie, (2), p. 241, 1854. 
 
346 RESPIRATION 
 
 cording to the amount of work a man is doing. We can easily 
 see, therefore, that the time taken for different parts of the body 
 to become saturated with nitrogen will vary greatly. 
 
 Taking into consideration the amount of fatty material in the 
 body, we estimated that the whole body of a man weighing 70 
 kilos will take up about I liter of nitrogen for each atmosphere 
 of excess pressure about 70 per cent more nitrogen than an 
 equal weight of blood would take up. Now the weight of blood in a 
 man is about 6.5 per cent of the body weight; hence the amount 
 of nitrogen held in solution in the body, when it is completely 
 
 saturated with nitrogen, will be about or 26 times as great 
 
 as the amount held in the blood alone. If, therefore, the composi- 
 tion of the body were the same at all parts, and the blood dis- 
 tributed itself evenly to all parts, the body would have received at 
 one complete round of the blood after sudden exposure to a high 
 pressure of air one twenty-sixth of the excess of nitrogen cor- 
 responding to complete saturation. The second round would 
 add one twenty-sixth of the remaining deficit in circulation, i.e., 
 1/26 x 25/26 of the total excess. The third round would add 
 1/26 x (25/26 x 25/26), and so on. On following out this calcu- 
 lation, it will be seen that the body would be half saturated in 
 less than 20 rounds of the circulation, or about ten minutes, and 
 that saturation would be practically complete in an hour. The 
 progress of the saturation would follow the logarithmic curve 
 shown in Figure 89. Actually the rate of saturation will vary 
 widely in different parts of the body ; but for any particular part 
 the rate of saturation will follow a curve of this form, assuming 
 that the circulation rate is constant. 
 
 There is abundant evidence, both from human experience and 
 from experiments on animals, that liability to compressed-air 
 illness increases with duration of exposure. We found that in 
 goats the liability increased up to about 3 hours' exposure, but 
 did not increase further even with far longer exposure. In man, 
 on the other hand, limitation of exposure to 3 hours has been 
 found to diminish the liability distinctly, and we calculated from 
 the goat experiments, taking into account the greater rate of 
 circulation in the goat on account of its much smaller weight (see 
 Chapter X), that in man the liability would increase up to about 
 5 hours' exposure. We had therefore to allow for parts of the 
 body which would only become half saturated in about 1^4 hours, 
 but for nothing slower than this. 
 
Figure 88. 
 "Bends" of foreleg in a goat. 
 
RESPIRATION 
 
 347 
 
 The longer any part of the body takes to saturate, the longer 
 will it also take to desaturate to the point at which it is safe to 
 reduce the pressure to normal. But if we know the pressure and 
 duration of exposure, we can now calculate a safe rate of further 
 decompression after the initial reduction of total pressure to half 
 
 IUU 
 
 
 
 
 . 
 
 
 
 
 
 
 
 ^ 
 
 ^^ 
 
 
 
 
 
 / 
 
 ' 
 
 
 
 
 60 
 50 
 
 
 / 
 
 
 
 
 
 
 / 
 
 
 
 
 
 / 
 
 
 
 
 
 
 30 
 
 10 
 
 n 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 
 
 
 / 
 
 
 
 
 
 
 Multiples of the time required to produce half -saturation. 
 
 Figure 89. 
 
 Curve showing the progress of saturation of any part of the body 
 with nitrogen after any given rise of pressure. The percentage 
 saturation can be read off on the curve, provided the duration of 
 exposure to the pressure, and the time required to produce half satu- 
 ration of the part in question, are both known. Thus a part which 
 half saturates in one hour would, as shown on the curve, be 30 per 
 cent saturated in half an hour, or 94 per cent saturated in 4 hours. 
 
 has been carried out : for we can calculate the rate at which nitro- 
 gen is being carried away from parts which saturate and de- 
 saturate quickly, or from those which do so slowly. We can thus 
 regulate the rate of decompression so that no part of the body 
 is at any time supersaturated to such an extent as to cause risk of 
 bubble formation. In this way tables were calculated for regu- 
 lating the rate of decompression of divers and other workers in 
 compressed air. For the sake of convenience the decompression 
 
348 
 
 RESPIRATION 
 
 rate was calculated in stages, each of which represents a reduc- 
 tion in depth of 10 feet, so that a diver is stopped by signal at every 
 ten feet of ascent. 
 
 Figure 90 represents what is happening during a dive to 28 
 fathoms, with the stay on the bottom limited to 14 minutes, and 
 
 Figure 90. 
 
 Diving to 168 feet by new method: Diver 14 minutes on the bottom and 46 
 minutes under water. The curves from above downward represent, respectively, 
 the variations in saturation of parts of the body which half saturate in 5, 10, 20, 
 40, and 75 minutes; the thick line representing the air pressure. 
 
 the new method carried out of rapid descent and ascent by stages. 
 It will be seen that when the diver reaches surface, the maximum 
 condition of supersatu ration with nitrogen in any part of the body 
 corresponds to only 17/4 pounds per square inch (or 1.17 atmos- 
 pheres) of air pressure. This leaves a margin of safety. Figure 91 
 shows what happened by the old method, with the same time on 
 
RESPIRATION 
 
 349 
 
 the bottom. It will be seen ( I ) that the dive took twice as long a 
 time, and (2) that when he reached surface the maximum super- 
 saturation was 36 Ibs. (2.4 atmospheres), so that he would run a 
 
 15 
 
 10 IS 20 25 30 JS 40 45 SO SS 6O 6S 
 
 Time in minutes. 
 
 75 8O 
 
 Figure 91. 
 
 Diving to 168 feet by old method: Diver 14 minutes on the bottom and 84 
 minutes under water. The curves from above downward represent, respectively, 
 the variations in saturation of parts of the body which half saturate in 5, 10, 20, 
 40, and 75 minutes; the thick line representing the air pressure. 
 
 most dangerous risk. It is evident from the figure that the slow 
 descent and most of the slow ascent were simply adding to the 
 
 s 
 
 f^r- 
 
 _____ 
 
 
 
 
 
 
 
 
 
 xji 
 198 
 165 
 132 
 99 
 
 ^Sk 
 
 **: : . 
 
 ~~~-_, 
 
 *._,^ 
 
 
 
 
 
 
 
 rcssure in atmosp 
 
 Co * 0> C 
 
 ^ 
 
 \x 
 
 ""v; 
 
 "V-. 
 
 ***** ---^^ 
 
 .^ 
 
 
 
 
 
 
 X 
 
 N^X 
 
 ^^.^ 
 
 IT****-. 
 
 
 ~~-~.^ 
 
 
 
 
 "L- 
 
 
 ^C 
 
 Xj.% 
 
 ^^^ 
 
 * 
 
 ^^ 
 
 "^- 
 
 ~___ 
 
 
 
 ^ 
 
 
 
 ^ 
 
 ^.^ 
 
 
 
 ** % _ | 
 
 _^ 
 
 "*-**,, 
 
 33 
 
 
 Absolute p 
 ^o * ' 
 
 
 
 
 "1 1 
 
 T ...._., 
 
 3 
 
 
 
 ^* 
 
 *-* fc- 
 
 
 
 
 
 
 
 
 
 
 
 ? / : 
 
 ' (. 
 
 4 i 
 
 t? 7 5 1Q" V 
 
 Time iu hours. 
 
 Figure 92. 
 
 Theoretical ascents of a diver after a prolonged stay at 213 feet of sea 
 water. Stage decompression in 309 minutes compared with uniform decom- 
 pressions in 309 minutes and in 10 hours. Continuous lines = stage decompres- 
 sion : interrupted lines = uniform decompression. Thick lines = air pressure : 
 thin lines = saturation with atmospheric nitrogen in parts of the body which 
 half saturate in 75 minutes. 
 
350 RESPIRATION 
 
 danger. These figures show also in a clear way, the advantages of 
 cutting down the duration of stay on the bottom. It appears from 
 Figure 90 that with the short stay on the bottom the more slowly 
 saturating parts of the body have not time to reach a dangerous 
 degree of saturation, though they might do so if similar dives 
 were repeated after short intervals on one day. 
 
 With a long exposure to a high air pressure the time required 
 for safe decompression, even by the stage method, becomes much 
 too long for ordinary diving work. Figure 92 shows, for instance, 
 that it would take nearly five hours by the stage method, and ten 
 hours with uniform decompression, for completely safe decom- 
 pression after a stay of some hours under a pressure of 35^2 
 fathoms of water, or an excess pressure of 6J^ atmospheres. 
 In the ordinary diving table, therefore, the stay on the bottom is 
 so limited that the diver can be decompressed safely in half an 
 hour. Nevertheless, it may happen that it is justifiable to stay 
 longer, or that a diver's air pipe is fouled by something on a wreck 
 and even that he cannot be liberated till the tide slackens or turns. 
 To meet such cases a supplementary table was drawn up. These 
 two tables are reproduced below. 
 
 Since the introduction into the British Navy twelve years ago 
 of the method of decompression embodied in the tables, with the 
 corresponding regulations as to air supply and testing of the 
 pumps, deep diving has been conducted with comfort and safety 
 to the divers, so that compressed-air illness has now practically 
 disappeared except in isolated cases where from one cause or 
 another the regulations have not been carried out. When a medi- 
 cal compressed-air chamber is available, it is justifiable to cut 
 down the time for the last wearisome stages of the decompression, 
 and so extend the time on the bottom. This has been cautiously 
 tried under Commander Damant's supervision, but the result was 
 that the divers began to suffer from "bends." These could easily 
 be relieved in the chamber, but much loss of time and incon- 
 venience resulted, and the "bends" were apt to recur. It seemed 
 better to keep the chamber as a precaution against emergencies or 
 unforeseen accidents. I calculated the tables with great care on 
 the theoretical lines borne out by the experiments and in the 
 light of all the available evidence from human experience ; and it 
 appears that the times cannot be cut down without risk of trouble, 
 unless the divers are placed in the chamber as a matter of routine 
 after each dive. 
 
 If a diver develops serious symptoms of compressed-air illness, 
 
to co co Ix co vo 
 
 
 00 O 
 
 
 
 01 
 
 
 M 03 co w co w 
 
 CO -" CO HI co 
 
 -. ro 
 
 CO 
 
 CO 
 
 
 1 
 
 
 
 
 
 
 tx o 10 tx to io 
 
 d 10 03 10 O 
 
 o o 
 
 
 
 h-1 
 
 o 
 
 M 
 
 c 
 
 CO to O IO O 1 O 
 
 O 10 o o tx 
 
 10 tX 
 
 K 
 
 tx 
 
 1 
 
 03 IO IO 01 IO CO 
 
 LO CO 10 CO tO 
 
 co to 
 
 to 
 
 to 
 
 rt 
 
 1 1 I 1 ! 
 
 01 03 03 CO 
 
 03 CO 
 
 CO 
 
 CO 
 
 a 
 
 Mil 1 
 
 
 
 
 
 .2 
 
 1 1 1 1 1 
 
 1 1 " 
 
 1 " 
 
 03 
 
 03 
 
 0) 
 
 5 
 
 III! 1 
 
 1 1 1 
 
 
 
 03 
 
 S 
 
 Mil 1 
 
 1 1 1 
 
 1 ' 
 
 
 
 c 
 
 CO co CO co co co 
 
 CO CO CO CO CO 
 
 CO CO 
 
 CO 
 
 CO 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 2 
 
 
 
 
 
 
 
 
 en en en 
 
 en en 
 
 
 C/5 
 
 CO 
 
 
 j C C G 
 
 C C 
 
 en 
 
 c 
 
 a 
 
 w 
 
 saa aa s 
 
 lO'S * 2 tO '3 Ol 
 M H B i-i C -. 
 
 o 10 o o o o 
 
 -> 03 CO +> co -t-> 
 
 a a ' a 
 
 S 2 S 28 
 
 <O Q O o ^ 
 01 -i-> 03 4-> M 
 
 1 a 
 g* s 
 
 'a 
 
 CO 
 
 'a 
 
 03 
 
 u 
 
 d) 
 
 (L) 
 
 rv 1 1 G, rv 
 
 o 1 o 1 
 
 o * H 
 
 . 
 
 ,. 
 
 
 ^r 1 10 10 ,r 10 nT 1 
 
 P HH 03 P M p 
 
 ^S^iH^ 
 
 5^ 
 
 D 
 
 * 
 
 1 
 
 CO ON T 
 
 <i 
 \ 
 t o to 
 
 "o 
 
 VO 
 
 ^ 
 
 1 
 
 IO IO VC 
 
 1 1 
 
 CO XM C 
 
 5 IX tx 
 7\ NO! O 
 
 ? 
 
 IO 
 
 00 
 
 i 
 
 I 
 
 3 
 
 0) 
 
 Tf KN I. 
 
 M\ tX 
 
 fs,^ 
 
 t-H\ 
 
 00 
 
 U 
 
 co 
 
 ^J- 
 
 
 
 
 
 S 
 
 IO 
 
 VO 
 
 
 03 
 
 
 S 
 
 
 
 
 
 
 Q 
 
 O 03 " 
 
 1" VO CO 
 
 
 
 03 
 
 ^f 
 
 i 
 
 01 03 C 
 
 ^ 03 O3 
 
 CO 
 
 CO 
 
 CO 
 
 
 eb i C 
 
 
 1 
 
 oo 
 
 1 
 
 
 
 1 
 
 03 
 
 
 IH 03 C 
 
 *3 03 03 
 
 03 
 
 CO 
 
 CO 
 
 
 03 
 
 ^- vo co 
 
 O 
 
 01 
 
 ^f. 
 
 
 03 co 
 
 ^- 10 VO 
 
 00 
 
 ON 
 
 
 
 
 CO O C 
 
 I I 
 
 ^1 rj- vo 
 
 1 
 
 co 
 
 *\ 
 o 
 
 s 
 
 03 
 
 
 O 03 c 
 
 o ^ o 
 
 VO 
 
 CO 
 
 ON 
 
 
 
 
 
 
 
 
RESPIRATION 351 
 
 and no compressed-air chamber is available, the best plan is to 
 screw on his helmet and drop him down under water till his 
 symptoms disappear. An unconscious man (who had developed 
 bad symptoms as a result of disregarding orders to stop at the 
 proper stages) soon answered the telephone when he was dropped 
 down in this way. The trouble, however, is to get the man up 
 again safely. A very cautious ascent is needed. When once bubbles 
 of any considerable size have formed it takes a considerable time 
 to get them redissolved. 
 
 The reason why a bubble in the blood or elsewhere in the body 
 tends to disappear, is that the partial pressure of nitrogen in the 
 bubble is greater than in the blood. The blood is saturated in the 
 lungs with nitrogen at a pressure of about 75 per cent of the 
 existing atmospheric pressure. In the venous blood, and there- 
 fore in the tissues, the pressure of oxygen, as shown in Chapter X, 
 is only about 6 per cent and of CO 2 about 6.5 per cent of an 
 atmosphere. There is also a pressure of about 6 per cent of aqueous 
 vapor. As the bubble is at atmospheric pressure and the total gas 
 pressure in the surrounding tissues is only about 75 -}- 18.5 = 
 93.5 per cent of an atmosphere, its nitrogen pressure is above that 
 of the tissues by 6.5 per cent. It must therefore gradually go into 
 solution, and at high atmospheric pressures it will do so all the 
 sooner since the pressures of oxygen and CO 2 do not increase pro- 
 portionally to the atmospheric pressure. If the bubbles are only 
 very small they will probably dissolve very rapidly on recompres- 
 sion; but if they are large, and particularly if they have been 
 formed at places where there is but little circulation, they will take 
 a long time to disappear. Great patience may therefore be needed 
 in treatment by recompression. 
 
 In the experiments made at sea under the direction of the 
 Admiralty Committee, the greatest depth at which trials were 
 made was 35 fathoms. At this depth Commander Damant and 
 Lieutenant Catto were perfectly comfortable, and in all the 
 numerous experimental dives which they made up to this depth 
 with stage decompression, no symptoms whatever of compressed- 
 air illness were observed. This depth was, however, greatly ex- 
 ceeded in the course of operations for the recovery of a United 
 States submarine at Honolulu in 1915. A diving crew had been 
 trained in the new methods at New York, and proceeded to Hono- 
 lulu to assist in getting hawsers in position round the submarine, 
 which was lying at a depth of 50 fathoms (corresponding to an 
 excess pressure of over 9 atmospheres or 135 pounds per square 
 
352 RESPIRATION 
 
 inch). The operations were successful, and these remarkable 
 dives are described in a paper by Assistant Surgeon French, 
 U. S. N., who was one of the medical officers in immediate 
 charge. 8 
 
 Eleven dives were made to depths of from 306 to 270 feet, the 
 time on the bottom being usually about 20 minutes. The stage 
 decompression, which was shortened as a recompression chamber 
 was always ready, occupied about no minutes. When everything 
 went according to plan, as turned out in eight of the dives, there 
 were no symptoms except in one case. One of the divers, however, 
 got foul at a depth of 250 feet and was delayed there about three 
 hours before he could be liberated. When he was freed he came 
 up beyond the proper stopping places, disregarding the telephoned 
 orders. Possibly he was partly stupefied by the prolonged action 
 of the high pressure of oxygen. At forty feet from surface he 
 collapsed. This was about 40 minutes after starting the ascent. 
 He was then pulled up to surface, where he was still able to 
 say a few words before becoming unconscious. His dress was 
 quickly ripped off and he was hurried into the recompression 
 chamber along with the two doctors and the other diver who had 
 rescued him. By this time he was black in the face, his breathing 
 had ceased, and no pulse could be felt at the wrist. Artificial res- 
 piration was at once applied, and at the same time the pressure 
 was run up to 75 Ibs. in 3^/2 minutes, which ruptured both the 
 eardrums of one of the doctors. As 75 Ibs. pressure was reached 
 the patient suddenly recovered and sat up, feeling all right again. 
 He was then gradually decompressed to 20 Ibs. in about iJ/2 
 hours, but at this point severe pain developed, so that the pressure 
 had to be raised again. For the next five hours many attempts at 
 decompression below 20 pounds had to be given up. At last he 
 was very gradually decompressed in about 3 hours in spite of the 
 pain. Soon after being taken from the chamber he was in a very 
 precarious condition, with the pulse no longer palpable. In spite 
 of haematuria, almost complete suppression of urine, extreme 
 pain, and other threatening symptoms, he recovered gradually; 
 and when it was possible to examine his lungs he was found to 
 have double broncho-pneumonia, the result, presumably, of the 
 very high oxygen pressure, as will be explained below. In a few 
 weeks he had completely recovered. 
 
 This case shows clearly the efficacy of recompression even under 
 
 "French, U. S. Naval Medical Bulletin, p. 74, January, 1916. 
 
RESPIRATION 353 
 
 conditions of apparently the most desperate character. It would 
 have taken over four hours to bring him up at all safely by stage 
 decompression, and his blood was certainly full of bubbles before 
 he was got into the chamber. 
 
 The difficulty of safe decompression in the chamber is one that 
 has often been met with before in bad cases. It may be necessary 
 to keep a patient in the chamber for 24 hours or more. 
 
 In work in tunnels or caissons the pressures encountered are 
 not nearly so high as in diving work; but the durations of ex- 
 posure are usually a good deal longer. Hitherto the time given to 
 decompression in the air lock has hardly ever been sufficient to 
 prevent symptoms, though in recent years it has often been suffi- 
 cient to prevent almost entirely the very dangerous symptoms 
 produced by rapid decompression, which leaves most of the body 
 in a condition of supersaturation with nitrogen. On this account 
 most of the symptoms in tunnel workers, etc., consist of the 
 "bends," itching of the skin, etc., due to bubbles in the tissues 
 which saturate and desaturate very slowly. In divers, on the 
 contrary, the symptoms met with before stage decompression was 
 introduced were mostly of a far more serious character, and due 
 to wholesale formation of bubbles in the blood and in tissues which 
 saturate and desaturate fairly quickly. Death or more or less 
 permanent paralysis were therefore common. With shortened 
 stage decompression it is usually the less serious symptoms which 
 appear among divers, and if the stage decompression is shortened 
 these symptoms must be expected. It is unfortunate that stage 
 decompression cannot be introduced in some countries on account 
 of antiquated state regulations enjoining decompression at a 
 constant rate, or even decompression starting very slowly and 
 increasing in rate as atmospheric pressure is approached. 
 
 During decompression, or immediately after it, it is very de- 
 sirable that as much muscular work as possible should be carried 
 out, so as to increase the circulation, and therefore the rate of 
 desaturation, over all parts of the body, and particularly those 
 parts which, owing to muscular exertion during exposure to the 
 high pressure, may have become saturated to a greater extent 
 than would otherwise be the case. For this reason the naval divers 
 were enjoined to keep their arms and legs moving as much as 
 possible during the stoppages at each stage. Bornstein has more 
 recently brought forward evidence collected at the Elbe tunnel 
 works that muscular exertion just after decompression diminishes 
 greatly the liability to "bends." 
 
354 RESPIRATION 
 
 It is probable that the bubbles first formed in supersaturated 
 blood and tissues are extremely small and comparatively harm- 
 less. One can observe the formation of these minute bubbles in 
 water which has stood in a pipe under pressure in contact with 
 air. When the tap is opened the water comes out milky with 
 minute bubbles, but no large bubbles are present. The smallness 
 of the bubbles leaves time to deal with cases of sudden decompres- 
 sion. Thus a diver who is blown up accidentally from a great 
 depth comes to no harm if he is sent down again at once or very 
 quickly got under high pressure in a recompression chamber. The 
 small bubbles already formed seem to go into resolution at once. 
 With any delay, however, the bubbles become larger and more 
 difficult to redissolve. In the diver referred to above bubbles had 
 evidently formed long before he reached surface and was recom- 
 pressed. 
 
 In the case of workers in tunnels and caissons it is practically 
 very difficult, and undesirable in various ways, to keep the men 
 very long in an air lock during decompression. Another plan 
 seems much better, and has been partially carried out in recent 
 years in tunnels under construction at New York. 9 The very high 
 pressures needed to keep the advancing face secure are only 
 employed in a section close to the face, this section being separated 
 from the rest of the tunnel by a steel air dam. If the total air 
 pressure in the advanced section is not more than I J4 times that in 
 the rest of the tunnel, the men can come through the air lock with- 
 out any delay. Let us suppose that the excess pressure is 35 Ibs. 
 at the face and 7.5 Ibs. in the rest of the tunnel. The total atmos- 
 pheric pressure is thus 50 Ibs. at the face and 22.5 Ibs. in the rest 
 of the tunnel. It is evident, therefore, that the men who have been 
 working at the face can come straight through either air lock, 
 even after very long shifts, provided that they are kept for a 
 sufficient time (fully an hour) in the low-pressure part of the 
 tunnel before coming through the second lock. If there were ar- 
 rangements for washing, changing, and meals in the low-pressure 
 section, this hour could be profitably employed. A six-hour shift 
 could be worked at the face, with an interval for a meal in the 
 low-pressure section, and there would be no blocking of the air 
 locks. The men could also go home at once, without the risk of 
 symptoms developing later. A plan of this kind, modified to suit 
 the varying conditions at different undertakings, seems to afford 
 the best means of solving the difficulties with air locks ; but exist- 
 
 * Japp, Trans. Intern. Congress on Hygiene, Section IV, Washington, 1912. 
 
RESPIRATION 355 
 
 ing state regulations might need modification to enable the im- 
 provement to be introduced. In any case there is now no justifica- 
 tion for imperiling men's lives by methods of decompression which 
 are known to give imperfect protection. 
 
 At present the tendency of the supervising medical officers is 
 to shorten the periods of work at the face under high pressure; 
 and of course the period of decompression may then be shortened 
 also. While this may cover the physiological aspects of the prob- 
 lem, it is evidently very uneconomical as compared with the 
 method above suggested. 
 
 Not only may increased partial pressures of nitrogen and CO 2 
 cause trouble, but also increased pressure of oxygen. The poison- 
 ous action of oxygen at high partial pressure was discovered by 
 Paul Bert; and his numerous and very thorough experiments on 
 the subject are described in his famous book. There is a popular 
 belief, based on the supposed similarity between life and com- 
 bustion, that the breathing of oxygen at a high partial pressure 
 must quicken the processes of life, and Paul Bert's experiments 
 on the effects of a high partial pressure seem to have been begun 
 with the view of testing this belief. He found that when the partial 
 pressure of oxygen exceeds three or four atmospheres, very re- 
 markable tonic convulsions are produced in warm-blooded ani- 
 mals, and they soon die. More remarkable still, perhaps, their 
 body temperature falls in the compressed oxygen, and the con- 
 sumption of oxygen and production of CO 2 are markedly di- 
 minished. The oxygen acts as a poison. 
 
 He then extended his observations to other forms of life besides 
 warm-blooded animals, and proved conclusively that for life in 
 every form, including the very lowest, oxygen at high pressure 
 is a poison. Plants, infusoria, and bacteria are killed just as 
 certainly as the higher animals. His experiments left no doubt that 
 it is the partial pressure of oxygen, and not mere mechanical 
 pressure, that matters. When air was used instead of pure oxygen, 
 the pressure required to produce fatal effects was nearly five times 
 as great as when pure oxygen was used, but the pressure of oxy- 
 gen was the same. He also found that oxygen pressures of less 
 than one atmosphere would kill or retard the growth of various 
 small organisms of different classes in the animal kingdom, and of 
 plants ; and he came to the conclusion that any increase over the 
 normal oxygen pressure of ordinary air is more or less detrimental 
 to living organisms directly exposed to it. He had discovered a 
 biological fact of the most far-reaching significance. 
 
356 RESPIRATION 
 
 It is usually not till the oxygen pressure in the air reaches more 
 than three atmospheres that warm-blooded animals show marked 
 immediate symptoms of oxygen poisoning. This we can under- 
 stand. The extra oxygen taken up in the arterial blood is nearly 
 all in simple physical solution, as Paul Bert showed by blood-gas 
 analyses of the arterial blood. At three atmospheres of oxygen the 
 blood will only take up about seven volumes of oxygen in solution. 
 On the other hand, the blood commonly loses about as much oxy- 
 gen in its passage through the capillaries. It is also indicated 
 by the results of experiments described in Chapter X, that the 
 effect of the increased oxygen is to slow the circulation, so 
 that more oxygen than usual is lost. Hence the oxygen pres- 
 sure will probably be very little above normal in the tissues 
 or venous blood until the oxygen pressure in the arterial blood 
 is over three atmospheres. As was shown in Chapter VII, ani- 
 mals in which the haemoglobin has been thrown out of action 
 by CO or nitrite poisoning are still a little short of oxygen when 
 they are breathing oxygen at two atmospheres pressures. We can 
 therefore easily understand why so high an oxygen pressure as 
 three or four atmospheres is needed before the nervous system 
 and other tissues are markedly affected by the oxygen. 
 
 In his experiments on warm-blooded animals Paul Bert had, 
 however, overlooked one thing which his other experiments might 
 have led him to look for. Although the tissues generally in a 
 higher animal are protected from the high pressure of oxygen, 
 since they have round them that wonderfully constant internal 
 environment which protects them from so many variations in the 
 external environment, yet the cells lining the air passages and 
 lungs are exposed to the high oxygen. It was discovered by Lor- 
 rain Smith in i899 10 that oxygen at a pressure quite insufficient 
 to affect the nervous system appreciably will, if time is given, 
 produce fatal inflammation of the lungs. The higher the pressure 
 of the oxygen, the sooner this appears. The lungs are filled with 
 exudation, so that they sink in the fixing fluid, a general oedema 
 similar to that in phosgene poisoning being produced. Probably 
 the animals only survive as long as they do in the compressed 
 oxygen because they get sufficient oxygen in spite of the oedema. 
 As Lorrain Smith showed, the oedema protects them against the 
 effects of very high oxygen pressure on the nervous system. At an 
 oxygen pressure of 180 per cent of an atmosphere (that to which 
 
 10 Lorrain Smith, Journ. of Physwl., XXIV, p. 19, 1899. 
 
RESPIRATION 357 
 
 the American diver referred to above was exposed for three 
 hours) one of the animals died from lung inflammation in 7 hours. 
 
 The higher the oxygen pressure the more rapidly was the fatal 
 inflammation produced. The lowest oxygen pressure at which fatal 
 pneumonia was observed was 73 per cent of an atmosphere, after 
 4 days' exposure. At 40 per cent no ill effects were observed. It is 
 evident from these observations that when oxygen is used con- 
 tinuously for therapeutic purposes the percentage ought not to be 
 increased more than is really necessary. A lung that is already 
 inflamed may be extra sensitive to an unusually high oxygen 
 pressure. At an oxygen pressure corresponding to 57 fathoms of 
 water we found that out of seven goats one died in three hours 
 from pneumonia, while the others were also affected, but re- 
 covered on decompression. At an oxygen pressure corresponding 
 to 40 fathoms we could not detect in ourselves any subjective 
 symptoms during short exposures ; but quite probably such symp- 
 toms might appear after longer exposure, and the behavior, 
 described above, of the experienced American diver seems sug- 
 gestive of this. 
 
 Although oxygen at high pressure acts generally as a poison, 
 yet as shown in Chapter IX, the living swim bladder may contain 
 oxygen at a pressure of 100 atmospheres without harm to the 
 cells lining its walls. These cells are apparently "acclimatized" to 
 the oxygen, just as the cells lining the stomach wall are acclima- 
 tized to hydrochloric acid. It is not improbable that the lungs are 
 capable of acquiring some degree of acclimatization or immunity 
 to the effects of a high pressure of oxygen ; but on this point there 
 are as yet no observations. 
 
CHAPTER XIII 
 Effects of Low Atmospheric Pressures. 
 
 VERY low atmospheric pressures are met with on mountains or 
 high plateaus and in ascents by balloons or aeroplanes to great 
 altitudes. Mountain sickness, one of the characteristic effects of 
 low atmospheric pressures, was known long before atmospheric 
 pressure and the composition of the atmosphere were understood. 
 It was commonly attributed to poisonous emanations. A good 
 account of earlier records of it is given by Paul Bert. His experi- 
 ments on animals and men showed clearly that the physiological 
 effects produced by low atmospheric pressure are simply the re- 
 sult of the diminished partial pressure of oxygen. The nature of 
 these effects and the manner in which they are produced have been 
 described generally in Chapters VI and VII in connection with 
 the symptoms and causes of anoxaemia. It remains, however, to 
 discuss the subject in detail. 
 
 Although Paul Bert's very important conclusion that the physi- 
 ological actions of oxygen and other gases depend on their partial 
 pressures has often been referred to in preceding chapters, no 
 very definite account has been given of his experiments. It will 
 be convenient to summarize them here, and at the same time refer 
 to certain points on which later investigation has thrown new 
 light. 
 
 By studying the conditions producing death in animals (chiefly 
 sparrows) confined in a closed vessel at varying atmospheric pres- 
 sures and with varying compositions of the initial air breathed, 
 Paul Bert proved that if the pressure of oxygen was not sufficiently 
 high to produce oxygen poisoning, death was due either to in- 
 creased pressure of CO 2 or to diminished pressure of oxygen. At 
 ordinary barometric pressure, and with ordinary air inclosed in 
 the vessel, death occurred when the oxygen percentage fell to 
 about 3.5. At half the ordinary pressure 7.0 was the fatal oxygen 
 percentage, so that the partial pressure of oxygen was the same; 
 and so on down to pressures of a third or even a fourth of an 
 atmosphere. If the vessel was filled with air highly enriched with 
 oxygen and the pressure was reduced to a fourth, or even a tenth, 
 the result was the same as regards the fatal partial pressure of 
 
Figure 93. 
 
 Paul Bert's apparatus for showing the effects of varying 
 low pressures of oxygen and CO2. The tap B is connected with 
 an air pump, and D with a bag of oxygen or nitrogen, while C 
 connects with a mercury manometer. 
 
 Figure 94. 
 
 Paul Bert's twin steel chamber for studying in man the 
 effects of very low atmospheric pressures with respiration of 
 oxygen. 
 
RESPIRATION 359 
 
 oxygen. On the other hand if the vessel was filled with the en- 
 riched air and left at ordinary barometric pressure, death oc- 
 curred when the percentage of CO 2 reached about 26, although 
 the oxygen pressure was far above the danger point ; and similarly 
 if the vessel was filled with compressed air at a pressure not suffi- 
 cient to cause oxygen poisoning. The cause of death depended 
 simply on whether the partial pressure of 3.5 per cent of an 
 atmosphere of oxygen or 26 per cent of an atmosphere of CO 2 
 was reached first. The mere mechanical pressure had no influence. 
 When, however, the partial pressure of oxygen was raised to the 
 dangerous limits referred to in Chapter XII, death was due to 
 oxygen poisoning, or hastened by it ; and the results suggest that 
 increase of the circulation rate, owing to the presence of CO 2 , with 
 consequent increase of the partial pressure of oxygen in the tissues, 
 increased the poisonous action of the oxygen, though Paul Bert 
 was unaware of the action of CO 2 on the circulation. 
 
 Figure 93 shows an apparatus used by Paul Bert for showing 
 that it is the diminished pressure of oxygen, and not simply the 
 diminished barometric pressure, that affects an animal. The fol- 
 lowing are the notes of an experiment on a sparrow. 
 
 "At 3.20 pressure reduced to 250 mm. in a few minutes. On 
 further reduction to 210 mm. the animal turned round and round, 
 fell down, and was at the point of death. I restored the normal 
 pressure by letting in air enriched with oxygen ; the animal re- 
 covered immediately and appeared lively and well. The air in the 
 bell jar now contained 35 per cent of oxygen. At 3.30 pressure 
 reduced to 180 mm. when the animal again became very ill. Pres- 
 sure again restored to normal by letting in oxygen, when the ani- 
 mal recovered at once. The air now contained 77.2 per cent of 
 oxygen. On again reducing the pressure the animal did not fall 
 over till 100 mm. pressure was reached. Immediate recovery on 
 restoring the pressure by letting in oxygen. The air now contained 
 87.2 per cent of oxygen. On reducing the pressure to 100 mm. at 
 3.50 the animal did not seem at all in danger; but at 80 mm. it fell 
 over in a dying condition. It recovered at once on letting in oxy- 
 gen. The air now contained 91.8 per cent of oxygen, and at 4.05 
 the pressure was reduced to 75 mm., when the animal again be- 
 came very ill, so that there was only just time to open the taps and 
 let it recover." This experiment shows very clearly that in air 
 greatly enriched with oxygen the barometric pressure could be 
 reduced to about a third of what was possible in ordinary air. 
 
 It was evident that oxygen could be used to avert the very 
 
RESPIRATION 
 
 dangerous effects of the rarefied air in balloon ascents ; and Paul 
 Bert proceeded to test this on himself in a steel chamber which 
 he had procured. The arrangement is shown in Figure 94. In this 
 chamber he not only studied in himself and others the subjective 
 and other effects of low barometric pressure when ordinary air 
 was breathed, but also showed that by breathing oxygen all these 
 effects could be prevented in man, down to very low pressures. 
 Figure 95 is a diagram showing the variations of pressure in one 
 
 76 _ 
 70 _ 
 
 60 _ 
 50 _ 
 40 _ 
 
 10-20 30 
 
 Figure 95. 
 
 Tracing showing Paul Bert's pulse rate during a decompression experiment 
 in his steel chamber. Upper line = barometric pressure in centimeters. Lower 
 line = pulse rate. At o the breathing of oxygen was begun and continued 
 till the end of the experiment. 
 
 of his experiments, and the striking effect on his pulse when he 
 began the continuous breathing of oxygen. The oxygen abolished 
 at once the various symptoms, of which an account was given in 
 Chapter VI. 
 
 I have frequently verified in steel chambers, and also when air 
 very poor in oxygen was being breathed, Paul Bert's statements 
 as to the effects of oxygen. He noted the sudden increase in ap- 
 parent brightness of light and loudness of sounds, the return of 
 powers of memory and of intellectual powers, etc. As illustrating 
 how even one who is perfectly familiar with the effects on vision 
 of rapid relief of anoxaemia may be deceived by the subjective 
 
RESPIRATION 361 
 
 effect, I may mention a recent personal experience. Dr. Priestley 
 and I had gone to a barometric pressure of about 360 mm. in a 
 steel chamber to test a piece of apparatus; and, being anxious to 
 test our Eustachian tubes, we opened the inlet tap full, so as to 
 raise the pressure to nearly normal within about a minute, as in a 
 nose dive of about 18,000 feet. Our ears were all right, but I was 
 alarmed to see the filament of the electric lamp suddenly become 
 intensely bright, as if it were about to fuse ; and on hastily pushing 
 the door open at the end of the decompression I inquired what had 
 gone wrong with the voltage. The appearance was of course only 
 subjective. I had forgotten the increase of oxygen pressure, and 
 had only been thinking of the mechanical effect on the eardrums. 
 
 Nothing in subsequent investigation has shaken Paul Bert's 
 conclusions as to the effects of gases being dependent on their 
 partial pressures, though the scientific world has taken a long 
 time to assimilate his reasoning, so that much of what has been 
 subsequently written on the subject of high and low atmospheric 
 pressures has been simply out of date. On a number of points, 
 however, later investigations have thrown new light. To take one 
 quite minor point first, the action of CO in air does not depend 
 upon its partial pressure, since the higher the pressure of an 
 atmosphere containing CO is raised the more innocuous does the 
 CO become, from the causes already discussed in Chapters IV 
 and VII. But at a constant partial pressure of oxygen the physio- 
 logical action of CO depends upon its partial pressure. There may 
 be other apparent exceptions to Paul Bert's rule, but we may be 
 confident that they will also turn out to be only apparent. 
 
 In his experiments Paul Bert took into direct account only the 
 pressure of oxygen and other gases in the inspired air. But we 
 have already seen that what directly matters is the gas pressures 
 in the alveolar air. When the barometric pressure is lowered the 
 alveolar oxygen pressure falls at a greater proportional rate than 
 the oxygen pressure of the inspired air. This is because, even 
 though the breathing is increased, which would in itself tend to 
 keep up the alveolar oxygen pressure, and may nearly prevent 
 the alveolar CO 2 percentage from rising, the percentage of 
 aqueous vapor is constantly rising. At a barometric pressure of 
 47 mm. no air at all would enter the lungs, since the pressure of 
 aqueous vapor would be 47 mm., and the liquids of the body would 
 from this cause alone be just about their boiling point; as a matter 
 of fact they would boil at a higher pressure, as they contain much 
 free CO 2 . At a pressure of 100 mm. in an atmosphere of pure 
 
362 RESPIRATION 
 
 oxygen, the alveolar air in situ would contain 47 per cent of H 2 O ; 
 probably about 20 per cent of CO 2 ; and 33 per cent of oxygen, 
 with a partial pressure of about 4.3 per cent of an atmosphere or 
 33 mm. of mercury. This pressure of oxygen is only one twenty- 
 third of that in dry oxygen at atmospheric pressure, though the 
 oxygen pressure in the inspired oxygen is only reduced to a little 
 over a seventh. 
 
 It is thus somewhat remarkable that until extremely low baro- 
 metric pressures, such as under 100 mm., were reached, the deaths 
 of the animals from want of oxygen should have coincided so 
 closely with a threshold oxygen pressure in the inspired air. The 
 probable explanation of this has already been referred to in 
 Chapter VI. With fall of barometric pressure the rate of dif- 
 fusion in a gas increases rapidly, since the mean free path of 
 each molecule before it strikes another molecule is increased. As 
 a consequence, the oxygen molecules in the neighborhood of the 
 alveolar epithelium reach it more rapidly, so that when there is 
 scarcity of oxygen the blood can be more readily saturated to the 
 existing mean oxygen pressure in the alveoli, or to whatever 
 higher oxygen pressure can be produced by active secretion. The 
 excessive fall in alveolar oxygen pressure at low barometric pres- 
 sures is thus partially compensated. 
 
 An experiment which Paul Bert describes (p. 749 of his book) 
 would seem to confirm this explanation. A bird was placed in the 
 apparatus (Figure 93) and the pressure reduced to 220 mm., at 
 which the animal had severe symptoms of anoxaemia. The pres- 
 sure was then raised to normal, not with air, but with nitrogen. 
 The animal died almost at once, though the partial pressure of 
 oxygen was 6 per cent, and the alveolar oxygen pressure must have 
 been raised, owing to the greatly diminished proportion of aque- 
 ous vapor in the alveolar air at normal barometric pressure. 
 
 The importance of the CO 2 present in the air was not noticed 
 by Paul Bert. In all his experiments where the oxygen pressure 
 of the inspired air fell to about 3.5 per cent before death there 
 was also a considerable proportion of CO 2 in the inspired air. 
 This CO 2 must have stimulated the respiration greatly, in the 
 manner already explained so fully, thus diminishing the fall in 
 alveolar oxygen pressure. The presence of CO 2 tends to diminish 
 the percentage saturation of the haemoglobin in the arterial blood, 
 owing to the Bohr effect already referred to at length in Chapters 
 IV and VII, but there is the counterbalancing advantage that the 
 haemoglobin holds on less tightly to oxygen in the systemic 
 
RESPIRATION 363 
 
 capillaries. The excess of CO 2 has, however, another quite dis- 
 tinct effect in counterbalancing the effects of the low alveolar oxy- 
 gen pressure : for the circulation can increase, owing to thensthnu- 
 lus of anoxaemia, without the counteracting effect due to the 
 production of alkalosis through deficiency of CO 2 . In this way the 
 oxygen pressure in the systemic capillaries is kept considerably 
 higher than if there were no excess of CO 2 in the inspired air. 
 
 Other things being equal, the presence in the inspired air of 
 a moderate proportion of CO 2 diminishes the effects of oxygen 
 deficiency, as can easily be shown experimentally. The CO 2 , by 
 increasing the breathing, raises the percentage of oxygen in the 
 alveolar air; and a very small excess in the alveolar CO 2 pressure 
 is sufficient to produce a large effect on the breathing. There is 
 consequently a considerable increase in the alveolar oxygen pres- 
 sure. That, however, the effects of CO 2 in relieving anoxaemia 
 are not simply due to the increased oxygenation of the blood can 
 be shown most strikingly in CO poisoning. A given percentage 
 of CO is less poisonous when administered to an animal breathing 
 human expired air. As this does not raise the alveolar oxygen 
 pressure, the effect cannot be due to increased oxygenation of the 
 arterial blood, and must be put down to increase in the circulation 
 rate, and consequent better supply of oxygen to the tissues. Lor- 
 rain Smith and I found that excess of CO 2 has no effect in stimu- 
 lating oxygen secretion by the lungs. 
 
 Although Paul Bert had in reality proved quite conclusively 
 that the physiological effects of low atmospheric pressures depend 
 on the lowering of the oxygen pressure, the theory was promi- 
 nently brought forward by Mosso twenty years later that these 
 effects are due primarily to excessive loss of CO 2 from the body, 
 or "acapnia." Mosso imagined that as a physical consequence of 
 the low atmospheric pressure more CO 2 than usual is washed out 
 of the blood in the lungs, and that this is the cause of mountain 
 sickness. 1 His physical chemistry was completely at fault. If the 
 volume of air breathed did not alter, the partial pressure of CO 2 
 in the alveolar air would remain the same, and no more CO 2 would 
 be given off at low than at ordinary atmospheric pressure. Actu- 
 ally, however, there is an excessive loss of CO 2 at low atmospheric 
 pressure, and this is due to the increased breathing caused by the 
 anoxaemia. Moreover we can, for the reasons already explained, 
 mitigate the anoxaemia by adding a suitable proportion of CO 2 
 
 1 Mosso, Life of Man on the High Alps (translation), London, 1898. 
 
364 RESPIRATION 
 
 to the inspired air. Acapnia may thus be looked on as a contribu- 
 tary cause of the symptoms, so that at first sight there seems to be 
 some experimental support for Mosso's theory. The acapnia, 
 although most important, is, however, only a secondary result of 
 the lowered oxygen pressure. This aspect of the matter has become 
 clear only recently through the work of Kellas, Kennaway, and 
 myself (see Chapter VI), and independently along closely similar 
 lines by that of Yandell Henderson and Haggard. 2 
 
 Mosso held to his acapnia theory till the time of his death, and 
 it was quite in vain that I myself endeavored to persuade him that 
 Paul Bert was right. "Acapnia" became for a time to many 
 physiologists the same sort of ignis fatuus as "reduced alkaline 
 reserve" has been in recent years. In 1906, however, Zuntz and 
 his colleagues placed the main facts in true perspective in an ac- 
 count of investigations carried out at high altitudes in the Alps. 3 
 
 We must now consider acclimatization to high altitudes and 
 anoxaemia caused in other ways. Paul Bert in his book (pp. 336, 
 1105) describes and discusses acclimatization, though he had not 
 himself studied it experimentally. The evidence pointing to the 
 fact of acclimatization was clear. He suggested that the tissues 
 become gradually accustomed to a smaller supply of oxygen in 
 the blood, and perhaps become more economical in their use of 
 oxygen. He also, however, suggests that the oxygen capacity of 
 the blood may become increased at high altitudes; and this he 
 afterwards verified by actual examination of blood taken from 
 animals living at high altitudes. 4 
 
 In 1892 Viault showed that the number of red corpscles per 
 unit volume of blood is increased at high altitudes, and Miintz 
 that the percentage of iron is increased. Various subsequent ob- 
 servers-established clearly the fact that in animals and persons 
 living at high altitudes there is an increase in both the percentage 
 of haemoglobin and the number of blood corpuscles in the blood. 
 By far the most complete and accurate series of observations on 
 the increase in haemoglobin was that carried out in connection 
 with the Pike's Peak Expedition by Miss Fitz Gerald on persons 
 living permanently at different altitudes in the Rocky Mountains 
 and elsewhere in America. Figure 96 shows graphically the 
 average results obtained at different altitudes. 
 
 It will be seen from this figure that on an average the per- 
 
 2 Haggard and Henderson, Journ. Biol. Chem., XLIII, p. 15, 1920. 
 
 8 Zuntz, Loewy, Miiller, and Caspari, HofienkUma und Bergivanderungen, 1906. 
 
 4 Paul Bert, Comptes rendus, XCIV, p. 805, 1882. 
 
RESPIRATION 
 
 365 
 
 centage of haemoglobin varies inversely with the barometric 
 pressure, and that even quite small diminutions in barometric pres- 
 sure are effective in causing a rise in the haemoglobin percentage. 
 In different individuals, however, the effects on the haemoglobin 
 percentage of a given diminution in barometric pressure vary 
 
 CMWfUfe 
 
 7* 
 
 zx 
 
 IVOOO 
 
 rapoo 
 12,000 
 
 700 650 600 
 
 Figure 96. 
 
 Average haemoglobin percentages in persons living 
 permanently at different altitudes (FitzGerald). 
 
 considerably. Thus among the persons acclimatized on the 
 summit of Pike's Peak (barom. 453 mm.) the rise in haemoglobin 
 percentage varied from 13 to 53 per cent of the normal. The rate 
 at which the haemoglobin percentage rises when a person goes to 
 a high altitude varies also. In some persons the rise is very slow; 
 and in consequence of this some observers have failed to detect 
 any rise on going for a short time to a high altitude. 
 
 As the average rise in haemoglobin percentage is appreciable 
 with only small increases of altitude, one would expect to find 
 that with increase of atmospheric pressure above normal the 
 haemoglobin percentage would fall below the normal value at 
 sea level. That this is actually the case was shown for dogs and a 
 monkey by A. Bornstein, who kept the animals under atmospheric 
 pressure of about three atmospheres or 2,280 mm. in the Elbe 
 tunnel at Hamburg during its construction. 5 She found that the 
 
 'Adele Bornstein, P finger's Archw., 138, p. 609, 1911. 
 
366 RESPIRATION 
 
 haemoglobin percentage and number of red corpuscles fell about 
 20 per cent, and that there was no fall in the case of animals kept 
 in the tunnel at a place where the atmospheric pressure was not 
 increased. It appears, therefore, that the haemoglobin percentage 
 is regulated generally in relation to the oxygen pressure in the 
 arterial blood, and rises or falls according as this pressure is 
 diminished or increased. 
 
 It is easy to see what the physiological advantage will be, other 
 things being equal, of a rise in the haemoglobin percentage. As 
 the blood passes through the systemic capillaries, its oxygen pres- 
 sure will fall more slowly than usual. Hence although the arterial 
 oxygen pressure is considerably below normal, the venous oxygen 
 pressure will be much more nearly normal, so that the lowering 
 of the oxygen pressure in the tissues is diminshed. There may be 
 much more of available oxygen in the arterial blood at a high 
 altitude than at sea level, but this in itself avails nothing, since it 
 is the pressure, and not the quantity, of oxygen in the blood that 
 counts. To explain the beneficial effects of increased haemoglobin 
 percentage at high altitudes and in other conditions where chronic 
 arterial anoxaemia exists we must consider the effects of the 
 increased haemoglobin on the oxygen pressure in the tissues. At 
 the same time we must bear in mind the influence of increased 
 haemoglobin percentage in diminishing the CO 2 pressure, and 
 therefore the hydrogen ion concentration, in the tissues ; and this 
 brings us to a second factor in acclimatization. 
 
 In recent years it has gradually been shown more and more 
 clearly that at high altitudes the volume of air breathed is in- 
 creased and remains so after acclimatization. This was already 
 more or less evident from the measurements by Zuntz and his 
 colleagues of the volume of air breathed and respiratory exchange 
 at high altitudes, and, as mentioned in Chapter VI, was rendered 
 quite clear by the experiments of Boycott, Ogier Ward, and 
 myself on the alveolar air at low atmospheric pressures. We drew 
 the conclusion that the blood, apart from the CO 2 contained in it, 
 becomes less alkaline at low atmospheric pressures, so that less 
 CO 2 is needed to excite the respiratory center. This diminution in 
 the "fixed" alkalinity of the blood was already known through 
 titrations. Barcroft then found on the Peak of Teneriffe that in 
 spite of the lowered pressure of CO 2 in the arterial blood, the 
 dissociation curve of the oxyhaemoglobin of the blood in presence 
 of the alveolar CO 2 pressure remains sensibly normal. This also 
 pointed in the same direction. The phenomena did not, however, 
 
RESPIRATION 367 
 
 correspond with those accompanying excess of lactic acid in the 
 blood, and Ryffel was unable to find any such excess in the blood 
 or urine. Accordingly the conclusion was drawn by my colleagues 
 and myself after careful observations during the Pike's Peak 
 Expedition, that the diminution in available alkali in the blood 
 must be due to a lowering in the level of concentration to which 
 the kidneys regulate the fixed alkali in the blood. We thought 
 that the anoxaemia must influence the kidneys specifically in this 
 direction. 
 
 The Anglo-American Pike's Peak Expedition 6 was planned 
 with the special object of studying acclimatization to the oxygen 
 deficiency of the air at high altitudes. We selected Pike's Peak 
 (14,100 feet) because it was possible, not only to get apparatus 
 and supplies to the summit easily by the cogwheel railway, but also 
 to live there without the disturbing effects of cold and hardship. 
 We were thus enabled to watch in ourselves the progress, which 
 was very striking, of acclimatization, and to observe the effects of 
 the rarefied air on the numerous unacclimatized persons who came 
 up. 
 
 It is evident that a simple increase in the breathing must 
 greatly diminish the arterial anoxaemia at high altitudes : for not 
 only will the alveolar oxygen pressure be increased, but in conse- 
 quence of excessive removal of CO 2 , the haemoglobin passing 
 through the lungs will combine more readily with oxygen, in 
 accordance with the discovery, already often alluded to, of Bohr 
 and his pupils. It might thus appear as if a simple increase in 
 breathing were the natural adaptive response to the anoxaemia 
 of high altitudes and other conditions. But, as already pointed 
 out, such a response is, except for a very short period, or to a very 
 limited extent, prevented, owing to the effect of the lowered CO 2 
 pressure in diminishing the breathing; and an increased circula- 
 tion rate (which would also tend to diminish the fall of oxygen 
 pressure in the tissues) is also prevented in the same way. More- 
 over the increase in percentage saturation of the haemoglobin in 
 the tissues is in any case of only limited advantage, since, owing 
 to the lowered CO 2 pressure, the haemoglobin holds on more 
 tightly to the oxygen. Nevertheless there will be some increase in 
 breathing and circulation rate; and this will represent a com- 
 promise between the effects of want of oxygen and of deficiency 
 
 8 Douglas, Haldane, Henderson, and Schneider, Phil. Trans. Roy. Soc., B, 
 203, 1913- 
 
368 
 
 RESPIRATION 
 
 DOUGLAS 
 
 40 
 36 
 32 
 
 28 
 24 
 
 gjj 
 
 COLORADO 
 SPRINGS 
 
 5 
 
 l/Afl 
 
 1IT 
 
 ^?F 
 
 . 
 
 
 
 P 
 
 ^I/C 
 
 
 COLO^/VDO 
 SPRINGS 
 
 
 A 
 
 r H 
 
 /y/i 
 
 KF/ 
 
 
 no 
 
 100 
 90 
 80 
 70 
 
 60 
 
 50 
 /in 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 v*. 
 
 ^^^^ 
 
 
 
 v 
 
 
 
 
 
 
 
 
 
 
 
 
 
 /^/ 
 
 >^ 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 g 
 
 s 
 
 
 ? 
 
 
 
 
 
 \fi 
 
 ^~ 
 
 ^ 
 
 ^ 
 
 
 
 -~ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^^ 
 
 
 
 
 "~^~~ 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 x-~" 
 
 
 ~\ 
 
 
 ^__ - 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 II 
 JULY 
 
 HA LOAN E 
 
 4 & 12 16 20 24 28 / 59 
 AUGUST SEPTEMBER. 
 
 Figure 97. 
 
 Pressure of COz and oxygen in alveolar air of three members of the Pike's 
 Peak Expedition at about sea level (Oxford and New Haven), at Colorado 
 Springs (6,000 feet), and on Pike's Peak (14,100 feet). Thick line = alveolar 
 COz pressure, and thin line = alveolar oxygen pressure. Interrupted lines = 
 normal alveolar COz and oxygen pressures at sea level. 
 
RESPIRATION 369 
 
 of CO 2 . It is during this condition that mountain sickness is pro- 
 duced. 
 
 In the course of a day or two, or of several days, the mountain 
 sickness passes off if the altitude is not too great ; but the breathing 
 is only slightly increased further, as we found on Pike's Peak 
 (Figure 97) by analyses of the alveolar air. Further light on 
 acclimatization was afterwards thrown by Hasselbalch and Lind- 
 hard 7 in a series of observations during which they remained for 
 a number of days in a steel chamber at reduced pressure. They 
 found by direct measurement that after acclimatization the hydro- 
 gen ion concentration of the blood is approximately normal, thus 
 confirming Barcroft's conclusions from observations of the dis- 
 sociation curve of the oxyhaemoglobin of the blood. They also 
 found that the excretion of ammonia in the urine is distinctly 
 diminished; and this led them to the conclusion that the very 
 slight acidosis which presumably causes the increased breathing 
 is due to diminished formation of ammonia in the body. 
 
 In a still more recent investigation 8 by Kellas, Kennaway, and 
 myself, we found that on exposure to a considerable diminution 
 of atmospheric pressure there is at once a very marked decrease 
 in the excretion of both acid and ammonia by the kidneys. The 
 urine may become actually alkaline to litmus. These observations 
 threw r a new and quite clear light on the increased breathing at 
 high altitudes. It became evident that the increased breathing is 
 primarily due simply to the stimulus of anoxaemia. This increased 
 breathing not only raises the alveolar oxygen pressure, but also 
 washes out an abnormal proportion of CO 2 and thus produces a 
 condition of slight alkalosis, to which the perfectly normal re- 
 sponse is a diminution of ammonia formation and in the acidity 
 of the urine, as explained in Chapter VIII. This response tends to 
 continue until the normal reaction of the blood is restored, owing 
 to reduction in the "available alkali" in the body. There is no 
 acidosis at any stage of the process ; the supposed acidosis is only 
 the compensation of an alkalosis. Nevertheless the process of 
 compensation is never quite complete. If it were so the excretion of 
 ammonia would return to its normal value on acclimatization, 
 whereas actually there is still, as shown by Hasselbalch and 
 Lindhard's observations, a slight but distinct diminution in am- 
 monia excretion. Moreover if the compensation were complete 
 
 7 Hasselbalch and Lindhard, Biochem. Zeitschr., 68, pp. 265 and 295, 1915; 
 and 74, pp. i and 48, 1916. 
 
 8 Haldane, Kellas, and Kennaway, Journ. of Physwl., LIU, p. 181, 1919. 
 
370 RESPIRATION 
 
 there would be no extra breathing caused by the immediate effect 
 of the anoxaemia. Actually there is still a slight amount of extra 
 breathing from this cause, since on raising the alveolar oxygen 
 pressure there is an immediate, though comparatively slight, rise 
 in the alveolar CO 2 pressure, as we found on Pike's Peak when a 
 mixture rich in oxygen was breathed in place of ordinary air. 
 The evident reason why the compensation does not become more 
 complete is that if it were made more complete the normal com- 
 position of the blood would be very seriously altered; and such 
 alterations tend to be resisted. The compensation thus represents a 
 compromise. 
 
 A similar interpretation of the apparent slight acidosis of high 
 altitudes was reached on independent grounds by Yandell Hender- 
 son, and published shortly before our paper appeared. 9 As already 
 mentioned in Chapter VIII, he and Haggard made the very 
 important discovery that with prolonged and very excessive 
 ventilation of the lungs (thus producing great alkalosis) the 
 available alkali or "alkaline reserve" of the blood diminishes 
 greatly. A similar diminution occurs at high altitudes, and Hen- 
 derson attributed it to the increased breathing produced by the 
 anoxaemia, and was thus the first to identify its true nature as a 
 compensatory response to the alkalosis produced by the increased 
 breathing. 
 
 It is evident that the compensatory change in the available 
 alkali of the blood and whole body tends to make increased breath- 
 ing possible with a minimum stimulus from actual anoxaemia. 
 The anoxaemia tends, therefore, to be relieved. In other words a 
 process tending to acclimatization has occurred. It will be noted 
 that the phenomena have been interpreted on what is usually 
 called a teleological basis, though no conscious adaptation of 
 means to end is implied, but only a tendency of the living body to 
 maintain its normal standards. The justification for this mode of 
 interpretation, and the demonstration that it constitutes the neces- 
 sary scientific basis of physiology, will be postponed to the next 
 chapter. 
 
 In connection with the Pike's Peak expedition Miss FitzGerald 
 carried out a large series of investigations of the alveolar air of 
 persons living permanently, and therefore fully acclimatized, in 
 towns and villages at different altitudes in or near the Rocky 
 Mountains. At a later date further observations were made at 
 
 9 Yandell Henderson, Science, May 8, 1919; and Haggard and Henderson, 
 Journ. BioL Chem., XLIII, p. 15, 1920. 
 
RESPIRATION 
 
 371 
 
 lower altitudes in South Carolina. 10 The average results are 
 shown in Figure 98. The results for men and women are given 
 separately, as men have a higher average alveolar CO 2 pressure 
 than women, as mentioned in Chapter II. It will be seen that 
 within the limits of atmospheric pressure investigated, the aver- 
 
 Cas pressure Altitude 
 
 mm. of 800 75O 70O 65O 60O 550 SOO 45O 4QO 350 ZOO 2 SO 20O >n Ft 
 
 /so 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 
 30JDOO 
 29.000 
 28OOO 
 27.000 
 26DOO 
 25000 
 24,000. 
 7-5,000 
 22,000 
 21,000 
 20,000 
 19.OOO 
 18,000 
 17.000 
 16,000 
 15.000 
 14,000 
 13,000 
 12,000 
 11,000 
 10,000 
 9,000 
 
 6,000 
 5,000 
 4,000 
 3.000 
 2,000 
 
 t.ooo 
 o 
 
 140 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 - 
 
 J30 
 120 
 //O 
 
 too 
 
 90 
 80 
 70 
 
 60 
 50 
 40 
 . 30 
 20 
 IO 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 /< 
 
 I 
 
 
 - 
 
 
 
 
 
 
 
 
 
 
 
 / " 
 
 
 
 - 
 
 A/veo/ar 
 
 
 
 
 
 
 
 
 L 
 
 / 
 
 
 
 
 
 pressure *^ 
 
 & 
 
 t 
 
 
 
 
 
 // 
 
 
 
 
 - 
 
 
 
 
 ^X. ^N 
 
 ^K 
 
 
 
 
 
 * 
 
 
 
 
 
 
 
 
 
 X 
 
 \0 
 
 
 
 7 
 
 
 
 
 
 
 
 
 
 
 
 
 5 
 
 /^ 
 
 f 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 \ 
 
 x 
 
 
 
 
 
 
 At 
 
 veoJa 
 
 r* 
 
 
 
 / 
 
 ? 
 
 
 x - 
 
 N X 
 
 
 
 
 
 /n re/at/on 
 to 
 
 o o 
 
 D*- 
 
 4 
 
 
 
 
 N 
 
 : \> 
 
 
 
 
 Off 
 
 P 
 
 nospfn 
 -essu) 
 
 >r/c 
 "e 
 
 , 
 
 r 
 
 o 
 
 
 "^ 
 
 
 
 
 >^s 
 
 X N 
 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 --- 
 
 ^?i 
 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 3OO ISO 700 650 600 S5O 5OO 4SO 4OO 3SO SOO 2SO 20O 
 Atmospheric pressure in mm. of mercury. 
 
 Figure 98. 
 Alveolar gas pressures in relation to barometric pressure or altitude. 
 
 age alveolar CO 2 and oxygen pressures fall proportionally to 
 the atmospheric pressure. To judge from these results the al- 
 veolar oxygen pressure at the height of 24,600 feet reached by the 
 
 10 FitzGerald, Phil. Trans. Roy. Soc., B, 203, p. 351 ', and Proc. Roy. Soc., B, 
 88, p. 248. 
 
372 RESPIRATION 
 
 Duke of Abbruzzi's expedition would only be about 31 mm., and 
 the CO 2 pressure about 2 1 mm. The figures, according to a form- 
 ula of Henderson, 11 would be oxygen 38 mm., and CO 2 15 mm. 
 
 Acclimatization would be a very incomplete process if it de- 
 pended solely on the increased breathing observed at high alti- 
 tudes. In spite of increased breathing and coincident increased 
 saturation of the arterial blood owing to the alkalosis produced, 
 there is at first very distinct cyanosis when persons first go to a 
 high altitude. On Pike's Peak this was very striking, though in 
 different persons the degree of cyanosis varied greatly. The fact 
 that there was so much cyanosis although the mean alveolar oxy- 
 gen pressure was about 50 mm. sufficient in presence of the 
 lowered alveolar CO 2 pressure to saturate the haemoglobin of 
 average human blood to 85 per cent or more is now explicable 
 by the fact that, as explained in Chapter VII, the oxygen pressure 
 of the mixed arterial blood is very appreciably below that of 
 the mixed alveolar air, and particularly at lowered atmospheric 
 pressure. The cyanosis disappears, however, after a day or two, 
 or sometimes longer, of mountain sickness; and in persons who 
 have reached the high altitude by gradual stages, as in the Him- 
 alayas, there may, apparently, be little or no cyanosis, and certainly 
 no mountain sickness. Among the party of four Europeans with 
 the Duke of the Abbruzzi, who gradually reached a height of 
 24,600 feet in the Himalayas, there were no signs of mountain 
 sickness or undue exhaustion at any stage. In the account of the 
 expedition the conclusion was even drawn that "rarefaction of 
 the air, under ordinary conditions of high mountains, to the 
 limits reached by man at the present day (a barometric pressure 
 of 12.28 inches or 312 mm.) does not produce mountain sick- 
 ness." 12 Mountain sickness, and its accompaniments were con- 
 sidered to be "in reality phenomena of fatigue." The writer of 
 this account was not aware of the fact that mountain sickness is 
 easily produced in unacclimatized persons without any fatigue, 
 and occurs quite readily in persons sitting in a steel chamber or 
 going by train to a high altitude. 
 
 We may contrast the experience of the Duke of Abbruzzi's 
 party with that of Hasselbalch and Lindhard in their steel 
 chamber. 13 They started altogether unacclimatized, from the sea- 
 level air pressure of Copenhagen, and only reduced the pressure 
 
 n Y. Henderson, Journ. BioL Chem., XLIII, p. 29, 1920. 
 
 12 Filipo de Filippi, Karakouram and, Western Himalaya, London, 1912. 
 
 18 Hasselbalch and Lindhard, Biochem. Zeitschr., 8, p. 295, 1915- 
 
RESPIRATION 373 
 
 to 520 mm., corresponding to a height of 11,000 feet; but after a 
 few hours they became so seriously affected by mountain sickness, 
 with alarming cyanosis, intolerable headache, and feelings of 
 asphyxia during the night, that they had to raise the pressure to 
 584 mm. (about 7,000 feet). Those ascending Pike's Peak started 
 from a height of about 6,000 feet and were thus partially acclima- 
 tized; otherwise their symptoms would doubtless have been more 
 marked than they actually were. 
 
 In Chapter IX the quantitative evidence has already been given 
 that at high altitudes after acclimatization the lungs actively 
 secrete oxygen inwards even during rest, and that were it not so 
 the immunity from symptoms of mountain sickness among ac- 
 climatized persons would be totally unintelligible. It only remains 
 to discuss here some special points with regard to oxygen secre- 
 tion. 
 
 The fact that some time is needed before oxygen secretion is 
 effectively established at a high altitude, accords exactly with the 
 fact that it takes a man some time to get his lungs and other parts 
 of his body into good physiological training for heavy muscular 
 exertion. As was pointed out in Chapter IX there is now very 
 clear evidence that in persons who are in good training oxygen 
 secretion by the lungs plays a very important part, whereas in 
 persons not in training any secretion evoked by muscular work is 
 so feeble as to be quite ineffective. Both at high altitudes and in 
 training for muscular exertion the power of secretion develops 
 with use; and development occurs in exactly the same manner 
 with the exercise of all other physiological functions. At high 
 altitudes the stimulus to secretion originates in consequence of the 
 imperfectly saturated condition of the arterial blood; and al- 
 though after acclimatization is established the saturation of the 
 arterial blood with oxygen becomes less incomplete, yet part of 
 the incompleteness must remain; otherwise there would be no 
 stimulus to oxygen secretion. In this connection it should be noted 
 that the arterial oxygen pressure given by the carbon monoxide 
 method is the average oxygen pressure of the blood leaving the 
 alveoli, and not the oxygen pressure of the mixed arterial blood. 
 The latter value is undoubtedly a good deal lower for the reason 
 already explained. 
 
 It has for long been well known to mountaineers that persons 
 who are in good physical training for hard work are far less 
 susceptible to mountain sickness and the other characteristic effects 
 of high altitudes than those who are not in training. This fact is 
 
374 RESPIRATION 
 
 the origin of the common and quite erroneous opinion that 
 mountain sickness is due simply to exhaustion and has nothing to 
 do with barometric pressure. It now seems probable that in so 
 far as acclimatization is due simply to increased power of oxygen 
 secretion good physical training in heavy exertion will do as 
 much as continued exposure to the high altitude. As we have 
 already seen, however, acclimatization consists not merely in 
 increased power of oxygen secretion, but also in increased haemo- 
 globin percentage and diminution in the available alkali in the 
 blood and tissues so as to permit of increased breathing without 
 the development of alkalosis. It takes time to bring about these 
 changes, and they are not brought about by training for muscular 
 work. The increased haemoglobin, though it was the first acclima- 
 tization change to be discovered, is probably of relatively minor 
 importance, inasmuch as recovery from mountain sickness and 
 related conditions commonly occur before there is any noticeable 
 change in the haemoglobin percentage. The diminution in avail- 
 able alkali seems to be much more important, but the process is 
 evidently a rather slow one. This is readily intelligible when one 
 considers the amount of alkali that has, apparently, to be got rid 
 of, partly by excretion through the kidneys, and partly through 
 suspension of formation of ammonia inside the body. Possibly 
 this part of acclimatization might be greatly hastened by the 
 administration of ammonium chloride, the striking effects of 
 which on the blood reaction were described in Chapter VIII. 
 
 The question of acclimatization has assumed new interest, owing 
 to the recent great extension of the use of aeroplanes at high 
 altitudes. The great advantage of good physical training seems 
 evident in this connection. At the same time it also seems evident 
 that only a limited amount of acclimatization can be produced 
 either by physical training or by intermittent exposures in aero- 
 planes to low atmospheric pressure. The limitation was distinctly 
 evident in the experiments, mentioned in Chapter IX, on the 
 degree of acclimatization produced by intermittent exposures at 
 low pressures. 
 
 We must now discuss the symptoms of balloonists and other 
 airmen at very great altitudes, and the means of averting these 
 symptoms. Enormous heights can easily be reached by balloons; 
 and quite recently, in consequence of the great improvements 
 during the war in the construction of aeroplanes and their engines, 
 a height nearly as great as those reached in balloons has been 
 reached in aeroplanes. The limitation in the heights to which men 
 
RESPIRATION 375 
 
 have hitherto been able to go is due entirely to the physiological 
 effects of the reduced oxygen pressure and the quite evident im- 
 perfections of the apparatus used for overcoming these effects. 
 
 Hot-air balloons were devised by the brothers Montgolfier, and 
 first used at Paris in 1783. Shortly afterwards the well-known 
 French physicist Charles invented the hydrogen balloon and made 
 the first ascent in 1785, reaching a height of 13,000 feet. Higher 
 ascents were soon after made, and in 1804 another Frenchman, 
 Robertson, reached about 26,000 feet and was greatly affected. 
 In the same year Gay-Lussac went to about 23,000 feet, but only 
 noticed slight effects. It seemed pretty evident that the limit of 
 safety was about 25,000 feet, but until 1875 no balloonist seems to 
 have been actually killed by asphyxiation due to the rarefied air. 
 
 In 1862 the well-known meteorologist Glaisher and the bal- 
 loonist Coxwell made a famous very high ascent from Wolver- 
 hampton; and Glaisher's account of the symptoms observed was 
 very full and valuable. 14 In 48 minutes they had reached a height 
 at which the barometer stood at 10.8 inches (274 mm.). Glaisher 
 found that after this he could no longer read his thermometer or 
 even his watch. His last reading of the barometer was 9.75 inches 
 (248 mm.), which he estimated as corresponding to 29,000 feet. 15 
 He then found that his arms and legs were paralyzed, and then 
 his neck also, so that he could not hold up his head. He could still 
 vaguely see Coxwell, who had climbed up to free the rope of the 
 valve, this having got tangled, owing to rotation of the balloon. 
 He tried to speak, but could not, and then suddenly he became 
 blind. He says, "I was still completely conscious, and my brain 
 was as active as in writing these lines." Then suddenly he lost all 
 consciousness and appears to have been unconscious for about 
 seven minutes, during which Coxwell had fortunately succeeded in 
 stopping the ascent of the balloon and bringing it down again for 
 a considerable distance. During Glaisher's return to consciousness 
 he first heard the words "temperature" and "observation," but 
 without seeing anything. Then he began to see his instruments 
 vaguely, and then other objects, and finally was able to take up 
 his pencil and continue his observations. The barometer was then 
 ii T /2 inches (292 mm.). Coxwell had never lost consciousness. 
 He climbed down with great difficulty. Seeing Glaisher's condition 
 he tried to pull the valve rope, but found that his own arms were 
 now paralyzed. He then, with great presence of mind, got hold 
 
 14 Glaisher, Travels in the Air, London, 1871. 
 
 15 It is somewhat doubtful whether the aneroid barometer was correct. 
 
376 RESPIRATION 
 
 of the rope with his teeth, and so succeeded in opening the valve 
 and turning the balloon downwards. By his presence of mind 
 and determination he saved both Glaisher's life and his own. 
 
 The next very high ascent was made by the three French sci- 
 entists Croce-Spinelli, Sivel, and Tissandier in 1875, and re- 
 sulted in the death of the two former. This tragic occurrence 
 revealed in a very clear manner the insidiousness of the onset of 
 dangerous anoxaemia, and the absolute necessity for taking the 
 most efficient means of guarding against it at very high altitudes. 
 Croce-Spinelli and Sivel had tried the effects of oxygen in Paul 
 Bert's steel chamber, as well as during a previous ascent to about 
 25,000 feet. They were thus familiar with its effects. The balloon 
 was therefore provided with bags of oxygen. Paul Bert, who was 
 away from Paris at the time, had, however, written to them that 
 the bags provided were too small to last for more than a short 
 period. There was not time, however, to get larger ones, and for 
 this reason they decided not to begin using the oxygen till they 
 felt themselves really in need of it. They reached a height of 
 about 24,600 feet with the barometer at 300 mm. and the balloon 
 no longer rising. At this point Sivel asked both his companions 
 whether they would go higher, and on receiving their assent cut 
 the strings of three bags of sand used as ballast. Figure 99 repre- 
 sents the appearance of the car of the balloon at this point. In 
 Tissandier's notebook there was the entry "1.25, T = 10, 
 B = 300. Sivel throws ballast. Sivel throws ballast." The writing 
 was scarcely legible, and the repetition of the words was charac- 
 teristic of the symptoms of anoxaemia. The balloon then rose 
 rapidly. Tissandier relates that he tried to take up the mouthpiece 
 of the oxygen tube, but his arms would not move. Nevertheless 
 he had no sense of the danger, but felt happy that they were 
 rising. He saw the barometer passing 290 and then 280 and wished 
 to call out that they were at 8,000 meters, but his voice was 
 paralyzed, and immediately afterwards he lost consciousness and 
 did not wake up till about forty minutes later. 
 
 The balloon was then descending rapidly and he noted that 
 the barometer was at 315. His companions were still unconscious. 
 He let go some ballast, and shortly afterwards Croce-Spinelli 
 woke up and let go more, including the aspirator. He then became 
 unconscious again. The balloon must have gone up, and he did 
 not wake up again till an hour and a quarter later. The balloon 
 was then at about 20,000 feet and falling very rapidly. Both Sivel 
 and Croce-Spinelli were dead. Tissandier had great difficulty in 
 
Figure 99. 
 
 Sivel, Tissandier, and Croce-Spinelli in the car of 
 the Zenith. Sivel preparing to cut the strings of the 
 ballast bags at 300 mm. barometric pressure. Croce- 
 Spinelli with the bubbling arrangement for breathing 
 oxygen in his hand. Tissandier reading the barometer. 
 The oxygen bags are seen above the car, and the re- 
 versible aspirator fixed to the basket work. 
 
RESPIRATION 
 
 377 
 
 letting go the anchor and landing safely, but succeeded. Figure 
 100 indicates diagrammatically the course of the balloon. The 
 maximum height was given by an automatic recorder. 
 
 9000 
 
 3000 
 
 2000 
 
 MMW/////,'. y////////////^^^^ 
 
 L0ire ri.-.! ' C<ron 
 
 Chotfouroux INOHC 
 
 Figure 100. 
 Diagram of the voyage of the Zenith, April 15, 1875. 
 
 It was clear that all three had been paralyzed before they tried 
 to breathe the oxygen. Doubtless they were all convinced that 
 they felt all right and in full possession of all their faculties. The 
 feeling of self-confidence seems always to be present in conditions 
 of gradually advancing anoxaemia. I have experienced it myself, 
 not only in steel chambers, but also in experimental CO poisoning; 
 and the conviction that one is fully competent is still present in 
 spite of the knowledge that this conviction may be a gross illusion. 
 A man who is grossly intoxicated by alcohol has just the same 
 
378 RESPIRATION 
 
 insane confidence that he is all right. At very high altitudes in 
 balloons or aeroplanes it is imperative that oxygen should be 
 breathed continuously. 
 
 For about twenty years after the accident just described no 
 further very high ascents in balloons seem to have been attempted. 
 The next high ascents were made in Germany, starting with an 
 ascent by Berson and Gross to 26,000 feet in 1894. Berson alone 
 then reached a height of 30,000 feet; and finally in 1901 Berson 
 and Siiring reached about 36,000 feet (n,ooo meters), with a 
 barometric pressure of 1 80 mm. In all these ascents oxygen was 
 used, without which they would have been quite impossible ; but 
 at the end of the last ascent both Berson and Siiring became un- 
 conscious, though fortunately not before the former had pulled 
 the valve rope and thus turned the balloon downwards. Berson 
 had the cooperation of the Austrian physiologist, von Schrotter, 
 and the latter in his book describes not only the ascents, but 
 various preliminary experiments in a steel chamber and experi- 
 mental ascents in which he made physiological observations. Von 
 Schrotter had thoroughly grasped Paul Bert's work and was not 
 misled by the mistaken opposition of some physiologists to the 
 oxygen theory. 16 
 
 Berson and Siiring used steel oxygen cylinders from which a 
 constant stream of oxygen came to them through a tube which 
 they could hold in the mouth. The cylinders were a great improve- 
 ment on the bags used by Croce-Spinelli and his companions, but 
 in other respects the arrangement was very imperfect, as von 
 Schrotter pointed out. With any increase of breathing the volume 
 of oxygen supplied became insufficient, so that only a mixture of 
 air and oxygen was breathed, the air being taken in through the 
 nose or by opening the mouth. Moreover it required constant 
 attention to inspire through the mouth, even if the supply of 
 oxygen was adequate. It was no wonder, therefore, that first 
 Siiring and then Berson was overcome. 
 
 In one of the ascents by Berson and von Schrotter liquid air 
 was tried for the first time. It failed, partly because there was 
 no proper means of gasifying as much of the liquid as they re- 
 quired, and partly because the oxygen percentage in the gasified 
 liquid air was not high enough. Cailletet had, however, already 
 indicated a method of controlling the gasification, and this method 
 in an improved form was extensively used by the Germans during 
 
 18 Von Schrotter, Der Sauerstoft in cLer Prophylaxte und Therapie der Luft- 
 druckerkrankungen, 1906. 
 
 
RESPIRATION 
 
 379 
 
 the war for instance in the very high flights needed for bombing 
 London. It is of course necessary to use liquid oxygen. Simple 
 liquid air would evidently be quite useless ; but if ordinary liquid 
 air is allowed to evaporate for a sufficient time the nitrogen dis- 
 tills off, leaving a residue very rich in oxygen. It was this residue 
 that was employed by von Schrotter and Berson. 
 
 To improve upon the simple tube hitherto used, von Schrotter 
 strongly recommended the use of a face piece, and figures the 
 first form used. The face piece covers both mouth and nose, and 
 the oxygen passes into it through a tube in a constant stream. 
 This arrangement was introduced for aeroplanes before the war, 
 and is now extensively used. The airman can inspire or expire 
 air freely, but always receives a certain amount of oxygen, and has 
 not to think of his breathing. The amount of oxygen, whether from 
 a steel cylinder or from a Dewar flask of liquid oxygen, can be 
 adjusted according to the height, but it is simpler to arrange for 
 a constant supply which is sufficient, or more than sufficient, up 
 to a certain height. About half the oxygen is wasted, as it reaches 
 the face piece during expiration. This waste can be prevented by 
 an arrangement similar to that already described (Figure 49) in 
 connection with the administration of oxygen to patients. Priest- 
 ley and I found in steel-chamber experiments that with this ar- 
 rangement about I liter a minute (measured at sea-level pressure) 
 was sufficient up to a height of 28,000 feet during rest ; but at least 
 2 liters were needed for such exertions as an aeroplane observer 
 or pilot has to make. With the light steel cylinders or large Dewar 
 flasks now in use the waste of oxygen with the ordinary arrange- 
 ment of mask does not greatly matter, however. 
 
 A height as great as Berson and Siiring reached in a balloon 
 has quite recently (March, 1920) been reported as reached in an 
 aeroplane by Major Schroeder of the American Army Air Service, 
 who, however, also became unconscious, and had a very narrow 
 escape. How it was that the oxygen supply became insufficient in 
 this remarkable ascent has not yet been reported. 
 
 The heights hitherto attained represent by no means the limit 
 which Paul Bert's experiments on animals indicated when pure 
 oxygen is breathed. All that is shown by them is that the oxygen 
 supply was insufficient. At 36,000 feet a man breathing pure 
 oxygen would be quite unaffected by the altitude. The barometric 
 pressure is about 1 80 mm. In the alveolar air there would be a 
 pressure of 47 mm. of aqueous vapor and 40 mm. of CO 2 - Hence 
 (by difference) there would be 93 mm. of oxygen pressure; and in 
 
380 RESPIRATION 
 
 the rarefied air this would certainly suffice to saturate the arterial 
 blood to the same extent as at sea level. At 140 mm. of barometric 
 pressure there would still be at least 53 mm. of alveolar oxygen 
 pressure; and it is probable that marked symptoms of oxygen 
 shortage would only begin to appear at pressures below this. At 
 100 mm. they would become urgent in unacclimatized persons. 
 At 80 mm. Paul Bert's animals were at the point of death. 
 
 It is difficult to see how the addition of CO 2 to the inspired 
 oxygen could be of any service, although at moderate diminutions 
 of pressure CO 2 is of considerable service, as already pointed 
 out. When pure oxygen is breathed it is impossible to raise the 
 alveolar CO 2 pressure without lowering the alveolar oxygen pres- 
 sure; and at very low barometric pressures every millimeter of 
 alveolar oxygen pressure counts. Moreover rise of alveolar CO 2 
 pressure would, on account of the Bohr effect, tend of itself to 
 diminish the percentage saturation of the arterial blood with 
 oxygen and thus counteract any advantage gained by increased 
 rate of circulation. Aggazotti has shown 17 that when animals are 
 placed in oxygen containing a considerable percentage of CO 2 
 they are capable of withstanding extremely low pressures; but 
 the same was found by Paul Bert when the atmosphere was one 
 of pure oxygen. Aggazotti himself reached the very low pressure 
 of 120 mm. in a steel chamber while breathing oxygen with CO 2 
 added. 
 
 To make it safe to go much above 30,000 feet it would be 
 necessary to have an apparatus which made it certain that the 
 wearer always breathed pure oxygen, or at any rate oxygen not 
 mixed with any other gas than CO 2 . An ordinary mine-rescue 
 apparatus with the usual constant oxygen supply of about 2 liters 
 per minute (measured at ordinary atmospheric pressure) would 
 secure this result with a very moderate expenditure of oxygen. 
 Care would, however, be necessary to insure that both the purifier 
 and the oxygen supply worked properly at the low temperature 
 and pressure met with at very high altitudes. With a larger con- 
 sumption of oxygen an apparatus could be made to work safely 
 without a purifier. If it were required to go much above 40,000 
 feet, and to a barometric pressure below 130 mm., it would be 
 necessary to inclose the airman in an air-tight dress, somewhat 
 similar to a diving dress, but capable of resisting an internal pres- 
 sure of say 130 mm. of mercury. This dress would be so arranged 
 
 "Aggazotti, Arch. ital. de Biologie, XLVI, 1905. 
 
RESPIRATION 381 
 
 that even in a complete vacuum the contained oxygen would still 
 have a pressure of 1 30 mm. There would then be no physiological 
 limit to the height attainable. 
 
 The problem of going to very high altitudes with an oxygen 
 apparatus is similar to that of using a self-contained breathing 
 apparatus in mine air which is either intensely poisonous from 
 the presence of CO or H 2 S, or contains little or no oxygen. 
 This problem has been solved successfully, so that teams of 
 miners have worked daily for weeks or months at places a long 
 distance from where there was any oxygen in the air. The same 
 care as is needed and actually taken in the case of the mining 
 apparatus is even more necessary in the case of airmen at great 
 altitudes, but, owing to prevailing ignorance, has not yet been 
 applied. At 36,000 feet, for instance, with the barometric pres- 
 sure at a quarter the normal, an airman breathing pure oxygen 
 would be much nearer danger if, owing to some accident, he took 
 several breaths of the surrounding air, than a miner using a self- 
 contained breathing apparatus would be if he took several breaths 
 of an atmosphere of fire damp. The miner would have in his 
 lungs to start with a pressure of 700 mm. of oxygen, whereas the 
 airman would have only about 90 mm. To the airman at very 
 high altitudes it is therefore specially necessary to have an ap- 
 paratus which is perfect in its action and is used with all the 
 precautions which our existing physiological knowledge shows 
 to be necessary. 
 
CHAPTER XIV 
 General Conclusions. 
 
 ON looking back at the results reached in successive chapters of 
 this book certain points of general physiological significance 
 emerge. The present chapter will be devoted to their discussion. 
 
 It is evident that within the limits of health the breathing 
 represents the lung ventilation required to keep the reaction 
 and the pressure of oxygen in the blood supplying the re- 
 spiratory center constant within certain narrow limits, and that 
 the breathing increases or diminishes in accordance with the 
 quantity of air needed to produce this effect. The "chemical" and 
 "nervous" stimuli acting on the respiratory center cooperate in 
 bringing about the constancy. The circulation is, in the main, 
 similarly regulated so as to maintain a normal reaction and 
 oxygen pressure in each of the various organs, although other 
 factors may also determine the local circulation rate to some 
 extent. 
 
 The quantity of respired air required to keep the arterial blood 
 normal varies with the very variable consumption of oxygen and 
 output of carbonic acid by the whole of the living tissues. In 
 different individual parts of the body the variations in consump- 
 tion of oxygen and output of carbonic acid are still more striking; 
 and meeting these variations there are equally striking variations 
 in the local circulation rates. 
 
 What is regulated by the breathing and circulation is not pri- 
 marily the consumption of oxygen and formation of carbonic acid, 
 but the partial pressures, or diffusion pressures, of these sub- 
 stances. If their diffusion pressures become more than slightly 
 abnormal the result is, not a mere slowing or quickening of physio- 
 logical activity, but totally abnormal activity and abnormal change 
 in structure. What is immediately effected is the maintenance of 
 these pressures. The supply of oxygen and removal of carbonic 
 acid are such as to keep them approximately steady. We have also 
 seen that it is simply as an acid that carbonic acid is of physio- 
 logical importance, so that in reality a normal reaction, or normal 
 diffusion pressure of hydrogen and hydroxyl ions, and not merely 
 a normal diffusion pressure of carbonic acid, is maintained. 
 
RESPIRATION 383 
 
 After Harvey's discovery of the circulation and Lavoisier's 
 discoveries with regard to respiratory exchange and animal heat, 
 many physiologists looked upon circulation and breathing as 
 processes which primarily determine and regulate tissue activity. 
 We can trace this, for instance, in the physiological ideas of 
 Descartes and Liebig, and in ideas still to some extent prevalent 
 as to the causes of respiratory exchange, secretion, and growth. 
 Closer examination has shown that breathing and circulation are 
 responses to tissue activity, and do not primarily determine it. 
 
 Another tendency has been to regard the nervous system as the 
 primary autonomous regulator of breathing and circulation. The 
 evidence brought forward above has shown, however, that the 
 regulative influence of the nervous system is not autonomous, 
 but dependent on conditions of environment determined mainly 
 by varying tissue activity. 
 
 In his "Le9ons sur les phenomenes de la vie" (p. 121) Claude 
 Bernard drew the conclusion that "all the vital mechanisms, 
 varied as they are, have only one object, that of preserving con- 
 stant the conditions of life in the internal environment" (the 
 blood). No more pregnant sentence was ever framed by a physi- 
 ologist, and the long series of investigations described in the 
 present book may be regarded as an attempt to follow out in 
 regard to blood reaction and oxygen supply the line which 
 Bernard indicated. Physiological activities can in one sense be 
 summed up in the "preservation of the conditions of life in the 
 internal environment," with consequent maintenance of normal 
 structure. In another sense, however, physiological activity is 
 constantly disturbing the internal environment. What is actually 
 maintained is a dynamic balance between the disturbing and 
 restorative activities. The order displayed in this dynamic balance 
 is the order of biology. 
 
 In view more particularly of Paul Bert's experimental demon- 
 stration that the physiological action of gases dissolved in the 
 blood depends on the pressures which they exert in the surround- 
 ing atmosphere that is to say on their vapor pressures we may 
 conclude that it is the diffusion pressures of substances dissolved 
 in the blood that correspond to Bernard's "conditions of life." This 
 definition includes temperature : for diffusion pressure, other 
 things being equal, varies as the absolute temperature and indeed 
 gives us our measure of temperature, since the expansion of gases 
 or liquids, by which we measure temperature, depends on increase 
 of diffusion pressure. 
 
384 RESPIRATION 
 
 It is a familiar fact that, apart from the contained gases, the 
 composition of blood plasma is extremely constant. The varied 
 experiments initiated by Ringer and carried forward by many 
 other observers indicate directly the physiological importance of 
 the various salts or their ions which are present in blood plasma, 
 and render intelligible the exactitude with which their concentra- 
 tions are regulated by the kidneys. The facts collected in the pres- 
 ent book show that also as regards hydrogen and hydroxyl ions 
 and free oxygen the composition of the blood plasma in contact 
 with any particular part of the tissues is, and must be, very con- 
 stant, and is kept so by regulation of breathing, circulation, kidney 
 excretion, and other physiological activities. Thus oxygen and 
 hydrogen and hydroxyl ions take their place in a strict quantita- 
 tive sense beside the salts, proteins, sugar, etc., which help to 
 make up Bernard's "conditions of life." 
 
 We also now know that what is called the osmotic pressure of 
 blood plasma is so constant that the existing methods of measur- 
 ing it by depression of freezing point or vapor pressure are too 
 coarse for the detection of such differences as are constantly oc- 
 curring during life and evoking the ordinary physiological re- 
 sponses of the kidneys and other organs. Osmotic pressure de- 
 pends, however, as already mentioned (Chapter VIII) on the 
 difference between the diffusion pressure of a solvent in a solution 
 and in the pure solvent. It is thus in reality the diffusion pressure 
 of water in the blood that is maintained so constant. The diffusion 
 pressure of water can thus be placed in the same category as that 
 of other substances among Bernard's "conditions of life." The 
 experiments of Priestley and myself 1 on the excretion of water 
 by the kidneys show that the regulation by the kidneys of the 
 diffusion pressure of water in the blood is comparable in its ex- 
 treme delicacy to the regulation of blood reaction. 
 
 As a general rule salts, water, and various other substances 
 present in blood plasma are to only a very small extent used up 
 by or given off from the tissues. Hence in the case of most 
 tissues it would require only a very slow circulation to keep the 
 concentrations of these substances constant in the blood, pro- 
 vided that the temperature was constant. If, however, the circu- 
 lation were much slower than it is, and if this were rendered 
 possible by the provision in the blood of much greater capacity 
 for carrying oxygen and CO 2 as easily dissociable compounds, 
 
 1 Haldane and Priestley, Journ. of Physiol., L, p. 296, 1916; Priestley, Ibid,.* 
 L, p. 304, 1916. 
 
RESPIRATION 385 
 
 the even regulation of temperature in the body would apparently 
 become impossible, and in other ways the physiological inter- 
 connection between different parts of the body would be less close 
 and rapid. 
 
 Although water and salts are by ordinary measurements neither 
 absorbed by nor given off from most living tissues, it is evident 
 that this only means that passage of them into the tissues is bal- 
 anced by passage outwards. A liquid, like a gas, consists of mole- 
 cules in rapid movement and diffusing in all directions. We can- 
 not follow the movements of individual molecules, and can only 
 detect gain or loss when either the relative proportions of different 
 kinds of molecules alter, or the total number increases or dimin- 
 ishes. When as many molecules or ions of any one substance 
 are passing in as are passing out there appears to be neither ab- 
 sorption nor giving off of the substance. Nevertheless there is 
 continuous molecular or ionic exchange, and the blood is in con- 
 stant and active physiological connection with the surrounding 
 tissues. As is shown by the immediate effects of altering the 
 diffusion pressure of salts, water, or other blood constituents, the 
 exchange of molecules continues during life, whether a tissue is 
 "active" or "resting." In reality there is constant physiological 
 activity, and the conventional sharp distinction between conditions 
 of rest and activity is extremely misleading. 
 
 From the standpoint of physical chemistry life depends upon 
 the maintenance of a balance of molecular exchanges between the 
 tissue elements and their environments. If the balance is disturbed, 
 so that, for instance, too many or too few water molecules or 
 potassium, calcium, or sodium ions are passing from the blood 
 to the tissues or vice versa, life is imperiled. The case is exactly 
 similar with oxygen molecules, or with hydrogen and hydroxyl 
 ions. If the oxygen diffusion pressure in the plasma falls so low 
 that the proportion of oxygen molecules passing in is abnormally 
 low as compared with that passing out there is physiological dis- 
 turbance; and similarly, as shown in Chapter XII, when too much 
 oxygen is passing inwards. 
 
 Hitherto the supply of oxygen has not been regarded from this 
 standpoint. It has been generally assumed that the oxygen mole- 
 cules are all passing in one direction and that an irreversible re- 
 action occurs in the living tissues by which oxygen is fixed so that 
 no free oxygen molecules are returned to the environment. The 
 facts indicating the great importance of a certain definite dif- 
 fusion pressure of oxygen in the immediate environment of the 
 
3 86 RESPIRATION 
 
 tissue elements are inconsistent with this view. The experimental 
 evidence shows that we must place the diffusion pressures of oxy- 
 gen and carbonic acid in exactly the same category as the diffusion 
 pressures of water, salts, and other dissolved constituents of blood 
 plasma. This means that oxygen molecules are constantly passing 
 both outwards and inwards, although in ordinary tissues more 
 are passing inwards. It is only in oxygen-secreting tissues that we 
 find that on one side of the secreting membrane oxygen molecules 
 are passing more readily outwards, and only, so far as yet known, 
 in the green parts of plants and in the presence of light that free 
 oxygen is on all sides passing more readily outwards from living 
 tissue elements than inwards. But even in green plants, as Paul 
 Bert showed, a considerable diffusion pressure of oxygen is neces- 
 sary for life. 
 
 We can thus compare living structures to dissociable chemical 
 molecules and particularly molecules which, like haemoglobin, 
 form molecular compounds only capable of existing in so far as 
 rate of loss is balanced by rate of gain. We must, however, assume 
 that the dissociation and association are taking place simultane- 
 ously in many different directions, corresponding to the many 
 different substances present in the blood plasma and necessary 
 for life. We have also to remember that although the individual 
 tissue elements are all in connection, direct or indirect, with the 
 blood plasma, they are also in connection with one another, and 
 that this implies additional conditions of stability in connection 
 with which molecular or ionic gains and losses are balanced 
 against one another. 
 
 It is clear that the stability in respect of one kind of molecular 
 gain or loss determines the stability in respect of others. Thus a 
 small deficiency of oxygen molecules, or a small excess of hydro- 
 gen ions, in the blood plasma, disturbs the equilibrium of the 
 receptor elements in the respiratory center and leads to the extra 
 molecular discharges which show themselves in increased activity 
 of the center. Disturbances in other directions of the composition 
 of the blood plasma have similar results, though the receptors 
 are specially sensitive to changes in reaction or deficiency in oxy- 
 gen pressure. We can interpret similarly the mode of action of 
 various stimuli acting on living tissues, including what, for want 
 of more intimate knowledge, we call mechanical stimuli. Hence we 
 are led to the conception of a living organism as the seat of a vast 
 system of mutually dependent reversible chemical reactions. For 
 irreversible chemical reactions physiology has but little use. 
 
RESPIRATION 387 
 
 The mechanistic interpretation of life fails to take account of 
 the mutual dependence throughout a living organism of these 
 reactions. When we remove any part of the organism from its 
 physiological connection with its environment including the other 
 parts, we at the same time necessarily alter its reactions and the 
 stability of its living structure. Hence we cannot investigate an 
 organism as we investigate the parts of a machine by taking them 
 apart and ascertaining the properties and structure of each sepa- 
 rate part. The same criticism applies to what may be called the 
 ''hormone" theory of the interconnection between the parts of an 
 organism. On this theory the interconnection is brought about 
 through the existence of special chemical messengers, or "hor- 
 mones," produced in minute quantities by each organ, and bring- 
 ing about specific excitatory effects, resulting in coordinated 
 action. The hormone theory, like the mechanistic theory, tacitly 
 assumes that, apart from the influence of hormones, and of the 
 central nervous system, each part of an organism leads an inde- 
 pendent existence. The truth is that every substance which enters 
 into the life processes of any part of an organism is as much a 
 hormone as any other such substance. Water, for instance, is the 
 most abundant constituent of the body, and a very minute excess 
 in the diffusion pressure of water in the blood excites very striking 
 reaction in the kidneys. This minute excess seems, therefore, to 
 act as a hormone, just as a minute deficiency in alkalinity or in 
 oxygen pressure acts as a hormone to the respiratory center. 
 Since, however, water, hydrogen and hydroxyl ions, and oxygen 
 are influencing the body continuously, the conception of them as 
 hormones, acting only occasionally, is quite misleading. The 
 physiological interconnection between different parts of the body 
 is continuously in existence and far more intimate than is assumed 
 by either the ordinary mechanistic theory or the hormone theory. 
 
 In the case of chemical compounds which we ordinarily regard 
 as being stable in their existing environment, and not in a constant 
 state of association and dissociation, it is well known that the 
 particular nature of one of the atomic linkings may make a great 
 difference to the others. Thus the general properties of an or- 
 ganic compound may be greatly changed when a hydrogen atom 
 is replaced by a chlorine atom or a methyl radicle. We have also 
 seen in Chapter IV how in oxyhaemoglobin the affinity of the 
 haemochromogen part of the molecule for oxygen is affected by 
 changes in environment affecting primarily another part of the 
 molecule. From the point of view of our present chemical knowl- 
 
3 88 RESPIRATION 
 
 edge there is thus nothing new in principle in the fact, character- 
 istic of physiological reactions, that any particular reaction is de- 
 pendent upon the whole life of an organism. Nevertheless it is just 
 here that we strike the dividing line between the physical sciences 
 and biology. 
 
 A physiological reaction, when we examine it closely, is always 
 found to depend on a vast number of conditions of structure and 
 environment. It is true that under "normal conditions" the same 
 stimulus will produce the same reaction again and again; but 
 when we inquire what normal conditions represent we find some- 
 thing which is indefinitely complex from the physical and chemi- 
 cal standpoint. We have only to alter slightly the diffusion pres- 
 sure of one or other of the many substances, only partially known, 
 in the blood plasma, in order to obtain a quite different reaction. 
 For instance a given fall in the diffusion pressure of oxygen fails 
 to excite the respiratory center if the hydrogen ion concentration 
 of the blood is very slightly below normal; and if the calcium 
 ion concentration were a little above or below normal there would 
 doubtless also be an abnormal result. The presence of a trace of 
 ether or morphia, or probably of numerous other substances, 
 affects the center in a similar manner. The excitability of a tissue 
 to any given physical or chemical stimulus may thus vary in- 
 definitely under slightly different conditions. 
 
 If we attempt to investigate physiological phenomena from the 
 standpoint merely of physics or chemistry, we are thus at once 
 landed in confusion. In investigating ordinary physical or chemi- 
 cal phenomena, we can examine one by one the parts or units 
 we are dealing with and ascertain their properties, so that from 
 the empirical knowledge thus gained we can predict what will 
 result when they act on one another. In other words we can give 
 physical and chemical explanations of their mutual action. But 
 when we attempt to do this as regards the actions on one another 
 of the parts of an organism, or of the organism and its environ- 
 ment, we are met by the difficulty that we cannot ascertain the 
 structures and properties of any of the separate parts, since their 
 structures and properties actually depend on the existing physio- 
 logical relations of the parts and environment to one another. The 
 relativity of the phenomena confronts us at every turn in the 
 attempt to reach physical and chemical explanations of physio- 
 logical reactions. 
 
 Up to a certain point we can, it is true, understand living organ- 
 isms mechanically. We can, for instance, weigh and measure them 
 
RESPIRATION 389 
 
 and their parts, and investigate their mechanical and chemical 
 properties. This enables us to predict certain points in their be- 
 havior, as shown, for instance, in Chapters IV and V. But when we 
 look more closely it becomes quite evident that the knowledge 
 we gain from mere physical and chemical examination hardly 
 touches any fundamental physiological problem. We cannot es- 
 cape from the relativity of the phenomena we are dealing with. 
 
 The only way of real advance in biology lies in taking as our 
 starting point, not the separated parts of an organism and its 
 environment, but the whole organism in its actual relation to 
 environment, and defining the parts and activities in this whole in 
 terms implying their existing relationships to the other parts and 
 activities. We can do this in virtue of the fundamental fact, which 
 is the foundation of biological science, that the structural details, 
 activities, and environment of organisms tend to be maintained. 
 This maintenance is perfectly evident amid all the vicissitudes of 
 a living organism and the constant apparent exchange of material 
 between organism and environment. It is as if an organism al- 
 ways remembered its proper structure and activities ; and in repro- 
 duction organic "memory," as Hering figuratively called it, 2 is 
 transmitted from generation to generation in a manner for which 
 facts hitherto observed in the inorganic world seem to present no 
 analogy. We can discover and define more and more clearly by 
 investigation these abiding details of structure and activity, dis- 
 tinguishing accidental appearances from what is really main- 
 tained; and this process of progressive definition is the work of 
 the biological sciences. 
 
 If we look back on the general outcome of the investigations 
 summarized in this book, it is evident that the progress made has 
 consisted in distinguishing underlying identity of activity amid 
 superficial appearances which at first sight present confusion. In 
 the second and third chapters it was shown that behind the ir- 
 regularities of ordinary breathing the mean pressure of CO 2 in 
 the alveolar air is maintained steady within narrow limits for 
 each individual ; and in a later chapter it was more definitely 
 shown that this implies a similar steadiness in the CO 2 pressure of 
 the respiratory center. In Chapter VIII this conclusion is widened 
 by the evidence that CO 2 pressure is only important as an index 
 of blood reaction, and that it is blood reaction, and not mere pres- 
 sure of CO 2 , that is kept so constant by the breathing. In Chap- 
 
 2 E. Hering, Memory as a Generalized, Function of Organized Matter (1870). 
 English Translation, Chicago, 1913. 
 
390 RESPIRATION 
 
 ters VI, VII, and IX it is shown that there is similar maintenance 
 of the pressure of oxygen in the blood, and in Chapter X evidence 
 is collected that the circulation is so regulated as to keep both the 
 oxygen pressure and the reaction very nearly steady in each part 
 of the body. Chapter XIII deals with the manner in which the 
 body adapts itself to an abnormal atmosphere in accordance with 
 the principles laid down in preceding chapters. 
 
 It is thus with the dominant fact that in various definite re- 
 spects the internal environment of the living body tends to be 
 maintained very steady that the investigations brought together 
 in the preceding chapters have mainly dealt. This dominant fact 
 is what makes a scientific treatment possible in actual practice, 
 and furnishes us with principles by means of which we can predict 
 physiological responses and at the same time gain a practical con- 
 trol of the living body, such as is required in medicine and sur- 
 gery. 
 
 When we find that a certain characteristic structure and internal 
 environment exists within a living organism, we have discovered 
 what at first sight appears to be a fact capable of definition, though 
 not of explanation, in physical and chemical terms. Thus the 
 "normal" diffusion pressures of substances present in the blood 
 are simply diffusion pressures which we can measure and define, 
 one by one, in ordinary physical and chemical terms. But when 
 something occurs which tends to alter one of the diffusion pres- 
 sures, or to disturb the structure, we realize more fully the real 
 nature of what is maintained in a living organism : for the altera- 
 tion is not entirely prevented, but met by active readjustment of 
 such a character that what we easily recognize as organic identity 
 is maintained. If, for instance, the oxygen pressure in the air 
 inspired is lowered, a quantitatively corresponding lowering in 
 the oxygen pressure of the blood passing through the tissues is 
 prevented by increased breathing, oxygen secretion by the alveolar 
 epithelium, and rise in the haemoglobin percentage. At the same 
 time other disturbances which would naturally result from these 
 changes are met by diminution in the "available" alkali in the 
 blood, increase in blood volume, and so on. A widespread re- 
 adjustment of physiological activities and of blood composition 
 has thus occurred, but with the result that the more fundamental 
 diffusion pressures of oxygen, hydrogen and hydroxyl ions, etc., 
 have altered only very little, and in this slight alteration they have 
 held together as a whole. The oxygen pressure, for instance, is 
 not restored at the expense of hydrogen-ion pressure or excessive 
 
RESPIRATION 391 
 
 work of the heart or lungs. What is maintained in the tissue en- 
 vironment is oxygen pressure in its organic relations. The rela- 
 tivity to one another of the phenomena of life stands out clear in 
 this maintenance of organic identity. 
 
 In the course of biological investigation we meet on all hands 
 with similar examples of maintenance and reestablishment of 
 organic identity ; and the existence of this actively-maintained 
 identity is the scientific basis of practical medicine and surgery. 
 But for the fact that functional as well as structural compensation 
 is constantly occurring, not only under ordinary physiological 
 conditions, but also in cases of injury by disease or accident, and 
 that by observation and experiment we can learn to understand, 
 predict, and aid it, physicians and surgeons would be absolutely 
 helpless. Neither scientific biology nor scientific medicine could 
 be based on the ordinary working hypotheses of physics and chem- 
 istry, since these hypotheses furnish no sufficient means of under- 
 standing and predicting biological phenomena. Biologists, physi- 
 cians, and surgeons are not, and never will be, simply chemists 
 and physicists. 
 
 In physiology we are always dealing with responses to immedi- 
 ate stimuli ; but the responses are evidently determined in relation 
 to the maintenance of organic identity. They are organic responses, 
 and are simply rendered unintelligible when by the common con- 
 fusion in thought running through so much of the present teaching 
 of physiology they are represented as examples of mechanical 
 determination. Such expressions as the "mechanism" of respira- 
 tion, or secretion, or of maintenance of the internal environment 
 generally, are examples of this confusion. On closer examination 
 all the assumed mechanical reactions turn out to be expressions of 
 the organic maintenance which is the subject matter of the bio- 
 logical sciences. 
 
 Biology must take as its fundamental working hypothesis the 
 assumption that the organic identity of a living organism actively 
 maintains itself in the midst of changing external appearances. 
 This identity is not physical identity nor identity of form or 
 chemical composition, but something which we can perceive and 
 trace by exact quantitative investigation just as readily and ex- 
 actly as we perceive and trace physical identity in what we inter- 
 pret as the inorganic world. The science which traces this organic 
 identity is biology. Anatomy or morphology traces it as regards 
 structure, and physiology as regards activity. But since organic 
 structure is only the outcome or expression of ordered activity, 
 
392 RESPIRATION 
 
 and organic activity only the activity which expresses itself in 
 organic structure, the two branches of biology are in reality one, 
 and we may look forward to a time when the present wholly 
 artificial and sterilizing separation of them will disappear along 
 with the disappearance of the mechanistic theory to which the 
 separation is due. 
 
 The true scientific procedure of biology is different from that 
 of the physical sciences. In physics and chemistry the procedure 
 employed is to ascertain the properties of the separate units of 
 matter and energy with which it is assumed that these sciences 
 deal. Thus from the properties and movements of the parts of a 
 material system such as a machine we can predict its behavior 
 and can design and control machinery. From the properties and 
 movements of the molecules in a given quantity of gas we can 
 predict its behavior. From the properties of the atoms of carbon 
 and other elements we can predict the existence and many of the 
 properties of carbonaceous and other compounds. But we cannot 
 predict in this way the behavior of a living organism. The re- 
 lationships, for instance, into which the carbon atoms as inter- 
 preted by chemistry enter within living organisms show them- 
 selves to be too complex and changeable, so that, apart from the 
 biological method of treatment, we should be totally at a loss. In 
 the physical sciences we are looking at collections of units, each of 
 which is looked at from the outside. In biology we are looking at 
 each unit from the inside, and biological results afford abundant 
 justification for this method of looking at them. 
 
 It may appear at first sight as if the biological method were 
 unscientific, and the claim may be made that it ought to be, and 
 ultimately must be, possible to advance in biology by the method 
 of the physical sciences. This claim must now be examined care- 
 fully. 
 
 The reason why the physical or chemical method of treatment 
 is so unsatisfactory in biology is that in connection with living 
 organisms the properties of the parts show peculiarities which we 
 do not meet with in what we distinguish as the inorganic world. 
 Let us take the case of nitrogen atoms. When nitrogen is pres- 
 ent as a gas at an ordinary temperature the properties of its mole- 
 cules seem to be very simple for all practical purposes. The mole- 
 cules simply repel one another when they meet, or when they 
 encounter molecules of other gases; and the kinetic theory of 
 gases, based on this simple assumption, enables us to predict with 
 the greatest accuracy the behavior at ordinary temperatures and 
 
RESPIRATION 393 
 
 pressures of a mass of gaseous nitrogen or of a mixture of nitrogen 
 with another gas. But if we raise the temperature sufficiently, and 
 hydrogen or oxygen is present, the nitrogen combines with it, form- 
 ing ammonia or oxides of nitrogen. The properties of nitrogen 
 have thus shown themselves to be more complex than the simple 
 kinetic theory of gases assumed. But from the atomic theory as ap- 
 plied in chemistry, and the theory of valencies, we can still predict 
 more or less successfully the composition of the compounds formed 
 by the nitrogen. Most of their special properties have to be ascer- 
 tained by experiment; but once ascertained they can be used for 
 the purpose of predicting how these compounds will behave un- 
 der quite new conditions. It is exactly the same when we come to 
 the complex proteins and other organic nitrogenous compounds 
 which can be separated from the bodies of organisms. So long as 
 they are separated from living organisms we can investigate 
 them just as we investigate other chemical compounds, and they 
 present no real obstacle to such investigation. 
 
 The obstacle appears whenever the assumed chemical mole- 
 cules are participating in the life of an organism. Their proper- 
 ties seem then to become fluid and dependent from moment to 
 moment on the position of each molecule relatively to multitudes 
 of other molecules of the most diverse kinds. We consequently 
 cannot trace the individual molecules, and cannot tell whether or 
 how they are in combination with other molecules. They seem to 
 develop a quite indefinite potentiality of exhibiting unsuspected 
 properties. 
 
 Now this fact shows us clearly that the simple atoms and mole- 
 cules of physics and chemistry are only a sensuous illusion : for, 
 behind the supposed simplicity, indefinite potentialities are hid- 
 den and actually show themselves in connection with the phenom- 
 ena of life. The properties and activities of what we call atoms 
 or molecules are in reality a function of their relations to 
 other atoms and molecules ; and this fact, which is not at once evi- 
 dent in what we call the inorganic world, becomes perfectly evi- 
 dent in biological phenomena. Organic individuality is something 
 very evident to our perception, and has thus the same claim to 
 reality as inorganic structure; but, from a purely physical and 
 chemical point of view, living structure and activity constitute not 
 merely a molecular flux like that of a river or a flame, but an 
 altogether undefinable flux undefinable because we cannot define 
 the molecular changes. It is biological and not physical or chemi- 
 cal structure and activity which biological investigation enables 
 
394 RESPIRATION 
 
 us to define more and more accurately and fully. Only when an 
 organism is dead do we seem to have before us a physical and 
 chemical complex. 
 
 Those who insist that physiological activity must in reality be 
 physical and chemical change have to answer a previous ques- 
 tion as to the justification for the assumption of physical and 
 chemical reality. The molecules, atoms, and electrons of the 
 physical sciences seem real enough so long as we confine ourselves 
 to the superficial aspect of reality which is dealt with by the 
 physical sciences; but it is the same reality that is dealt with by 
 biology, and we reach a different interpretation of it through the 
 study of biological phenomena. In this interpretation the self- 
 existent individuality of atoms and molecules fades away in rela- 
 tivity. 
 
 The modern world has become so accustomed to the material- 
 istic assumption which identifies the mechanical interpretation of 
 reality with actual reality that in spite of the existence of biology, 
 psychology, ethics, religion, and philosophy it is difficult at pres- 
 ent to obtain even a hearing for the view that physical reality rep- 
 resents no more than superficial sensuous appearance. By the help 
 of various makeshift hypotheses such as those of vitalism or 
 animism, the real philosophical problem as to the ultimate validity 
 of the physical interpretation of reality has been evaded for the 
 time. But these evasions cannot satisfy us, and the problem comes 
 up in a clear-cut and definite form in connection with the relation 
 of biology to physics and chemistry. The facts dealt with in the 
 latter sciences present us with one interpretation of "reality," or 
 "nature," and those dealt with by the former present us with a 
 different one. 
 
 Which of the two interpretations corresponds more closely to 
 actual reality? There appears to me to be no doubt that the biolog- 
 ical interpretation does. The progress of the physical sciences 
 has taught us that the gases, liquids, and solids which to super- 
 ficial examination appeared to be continuous and inert substances 
 are not only discrete but made up of molecules in continuous 
 relative movement, and, in the case at least of solids and liquids, 
 continuously affecting one another's movements and properties. 
 We now know also that atoms themselves are systems of still more 
 elementary units moving relatively to one another at enormous 
 velocities, and that in chemical combination, and even in solution 
 or what we call simple mechanical interaction, these systems are 
 modified, as shown, by electrical phenomena. The chemist can 
 
RESPIRATION 395 
 
 determine with great apparent accuracy the proportions of hydro- 
 chloric acid and water in an aqueous solution. In actual fact there 
 may be practically no hydrochloric acid molecules present and far 
 fewer simple molecules of water than would appear from the 
 analysis, since the molecules are partly ionized and partly com- 
 bined with one another in various forms. Consequently the re- 
 sults of the analysis represent only a "practical" convention, how- 
 ever useful this convention may be. In reality the properties 
 of both the conventional hydrochloric acid and the conven- 
 tional water depend on the particular conditions existing in the 
 solution. But the inquiry can be, and has been, pushed still further. 
 At first sight it seems as if, in whatever way the molecules of 
 water and hydrochloric acid may be split up or combined, the 
 mass present is something independent of changeable relations. 
 But here, again, the progress of physical science has indicated 
 that even the mass of what is present depends upon relative move- 
 ment, and finally that absolute movement in empty space is a 
 conception to which no experimentally verifiable meaning can be 
 attached. 
 
 We only deceive ourselves when we imagine that in physical 
 and chemical investigation we are free of relativity. Behind 
 all the superficial appearances of a "real" physical world, rela- 
 tivity finally appears; but in biological phenomena the relativity 
 is always evident and prominent, and precludes the possibility of 
 even a conventional physical and chemical interpretation of 
 the observed facts. In frankly accepting relativity, and framing 
 her interpretations on a principle based upon it, biology comes a 
 step nearer to actual reality than the physical sciences. 
 
 It has come to be popularly believed that if we knew enough 
 of the physics and chemistry of what occurs in a living organism 
 biological interpretation could be reduced to physical and chemi- 
 cal interpretation. Though the attempts to give physical and 
 chemical interpretations of biological phenomena have never been 
 successful, and their failure in detail is becoming more and more 
 evident with the progress of both physiological and physical in- 
 vestigation, labored endeavors are still made to teach physiology 
 and represent the growing body of physiological knowledge in 
 physical and chemical terms. The investigations described in the 
 present book illustrate the fruitlessness of these attempts. In the 
 phenomena connected with breathing we are everywhere dealing 
 with organic regulation in other words with the manifestations 
 amid superficial changes which at first sight puzzle and confuse 
 
396 RESPIRATION 
 
 us, of organic identity. It is the same in connection with the phe- 
 nomena of circulation, excretion, absorption, and other physio- 
 logical activities. I wish to claim very definitely that in dealing 
 with biological phenomena and putting her questions to Nature, 
 biology must use her own working hypothesis, and not those of the 
 physical sciences. 
 
 The organic regulation which we find everywhere in a living 
 organism does not represent something imposed from without on 
 the processes occurring in the organism, but is simply a natural 
 expression of the reality which is present. It is Nature we are 
 studying in biology, not a special "vital force" or other super- 
 natural influence. But the biologist must be free to interpret 
 Nature in his own way; and it is Nature as Hippocrates saw her, 
 and not as Democritus saw her, that he sees and cannot help 
 seeing. Organic regulation, maintenance, and reproduction are 
 nothing but the expression of this biological Nature. 
 
 The universal acceptance among biologists of the doctrine of 
 evolution has often been assumed to carry with it the corollary that 
 life has arisen out of inorganic conditions ; and in this way a short 
 cut has been made to the conclusion that biology must in ultimate 
 analysis be nothing but physics and chemistry. This reasoning 
 cannot be justified. Even in the simplest forms of life it is still 
 unmistakably life that we are dealing with ; and if we succeed in 
 tracing life to yet simpler forms we shall still find life, so that the 
 "inorganic conditions" into which we have traced life will ap- 
 pear to be something very different from inorganic conditions 
 as we now represent them to ourselves. 
 
 We can see, and particularly clearly in the case of higher or- 
 ganisms, that the life of each organism is an association of the 
 lives of more elementary organisms, each of which shows its full 
 being only in the life of the whole, but is also more or less capable 
 of independent existence. It is by the separation and subsequent 
 full development of these more elementary organisms that re- 
 production is brought about. The life of a higher organism has 
 been said to be the "sum" of the lives of its constituent cells. Such 
 an expression is, however, misleading: for a cell apart from its 
 particular place in the living body, or the particular environment 
 which exists there, behaves very differently from the same cell 
 in its proper place. It thus cannot be physiologically defined apart 
 from its place in the whole organism. The organism as a whole is 
 no less real because it includes in its life the lives of individual 
 cells, and each cell, as shown very clearly in connection with the 
 
RESPIRATION 397 
 
 facts first discovered by Mendel with regard to reproduction, in- 
 cludes the lives of still more elementary centers of life. The same 
 reasoning applies, of course, to communities of what appear at 
 first to be quite separate organisms. An organism separated from 
 its kind is an artificial abstraction, just as is an organism separated 
 in other ways from its environment. 
 
 Although such processes as respiration, circulation, secretion, 
 absorption, and various forms of nervous activity, occur inde- 
 pendently of consciousness, many bodily activities are accompa- 
 nied by consciousness. Muscular exertion, for instance, is for the 
 most part consciously determined, and as muscular activity de- 
 termines breathing, and in other ways the breathing is determined 
 by conscious activity and under direct conscious control, it is 
 necessary to refer to the relation of conscious to unconscious 
 bodily activity. 
 
 We can interpret unconscious physiological activity from the 
 biological standpoint which has hitherto served us in the interpre- 
 tation of breathing, circulation, etc. ; but it is different with con- 
 scious activity. In perception we are aware of what we interpret 
 as "objective reality/' and voluntary actions are quite evidently 
 determined by this awareness. The awareness signifies that in 
 perception, as distinguished from a simple physiological reaction, 
 the reaction is not simply definable as occurring at a certain 
 moment or within a certain definite time, but involves also past 
 and future times, as well as surrounding space. When I see my 
 pen now, I see it as a material structure which has existed and 
 will continue to exist. I also see it as being in relation to many 
 other things not at the moment visible in the physiological sense. 
 The light in which I see it is not merely that of an electric lamp 
 but of all my other experience. When I write with the pen the 
 movements of my muscles are determined by the actual presence 
 to me of innumerable past, present, and anticipated future events 
 in both my own individual history and that of mankind. The past 
 events are not simply past and done with, like events interpreted 
 physically or biologically, but they, and not their mere effects, 
 are still present and active. What I have experienced before, what, 
 for instance, I have read of Hippocrates, or Johannes Muller, or 
 Claude Bernard, or Paul Bert, is still taking on fresh meanings in 
 my mind and directly determining my action now. The same is 
 true of all I have absorbed of the common spiritual heritage and 
 anticipations for the future of my country or of mankind. Actual 
 memory is no mere organic memory. I am living and acting in a 
 
398 RESPIRATION 
 
 spiritual world for which separation, not merely in space, but 
 also in time, has none of the meaning which it possesses for the 
 world interpreted physically or biologically. Along the years 
 and across the oceans action and reaction are direct in this spirit- 
 ual world. 
 
 It is evident that in conscious activity we are face to face with 
 facts that neither physical nor biological hypotheses are capable 
 of interpreting. Yet conscious activity manifests itself in connec- 
 tion with the same beings that seem also to live and breathe as 
 mere organisms, or to consist of nitrogen, hydrogen, oxygen, 
 carbon, and other atoms leading a wild and undefinable dance. 
 In presence of the evidence of life we cannot rest satisfied with the 
 physical and chemical interpretation of these beings ; but simi- 
 larly in the presence of conscious activity we cannot rest satisfied 
 with the biological interpretation. Biological phenomena show 
 us that the physical interpretation of the universe is only an im- 
 perfect preliminary interpretation for which all that can be said 
 is that it is of essential practical use in the absence of fuller 
 knowledge. But the facts relating to perception and volition show 
 us that the biological interpretation is also no more than a prac- 
 tical makeshift. As mathematicians, physicists, chemists, biolo- 
 gists, we are only "practical" men, though we often take our 
 practical working hypotheses for representations of actual re- 
 ality. We do so by unconsciously neglecting for the time a great 
 part of the facts to be explained in particular the facts that our 
 world not only includes living organisms, but is a known world 
 and a world of spiritual values. In reality our sciences are only 
 making use of abstractions of a limited practical value. 3 
 
 In conscious activity the self-conserving and species-conserving 
 organic activities of living organisms take on a new and far wider 
 interpretation. Mere organic self-conservation appears now as 
 conservation of a system of consciously realized interests; and 
 social interests assume a commanding position as compared with 
 individual interests. In so far as bodily interests are carried out 
 consciously, therefore, the physiological interpretation of them 
 recedes into the background ; and this is still more true of the 
 physical and chemical interpretation. 
 
 In the preceding chapters, I have attempted to justify the 
 physiological interpretation of unconscious bodily activities by 
 pointing out how breathing, circulation, etc., are manifestations 
 
 3 A fuller discussion of this point of view will be found in my book "Mech- 
 anism, Life and Personality," New Edition, 1921. 
 
RESPIRATION 399 
 
 of the maintenance of organic identity. Up to a certain point one 
 can apply the same reasoning to conscious activities by showing 
 how exquisitely dependent they are from moment to moment on 
 the integrity of normal "conditions of life" in the internal en- 
 vironment, and how they play their part in maintaining this in- 
 tegrity in accordance with Claude Bernard's conception. But such 
 treatment of them is wholly insufficient, since they evidently par- 
 ticipate in that spiritual world to which reference has already 
 been made. Hence they cannot be described in terms of the work- 
 ing hypotheses of biology, and attempts to describe them ade- 
 quately in such terms are merely childish. A fortiori they cannot 
 be described in physical terms. 
 
 Perception and volition are often referred to as processes occur- 
 ring in the cerebral hemispheres as a result of physical impulses 
 communicated along sensory nerves from outside. For certain 
 limited practical purposes this is a useful view to take of them. 
 But, as already pointed out, perception and volition as such are 
 not capable of description as events occurring at a certain time and 
 place, since from their very nature they include other times and 
 places, and may be said to be creative of time and space. The 
 working conception under which we attempt to describe them as 
 events occurring here and now is totally inadequate; and in so 
 far as we express them in terms of this conception we reduce them 
 to mere abstractions. By a process of abstraction we can observe 
 in ourselves and interpret as mere physiological or even physical 
 events our perceptions and voluntary actions. These observations 
 constitute an important part of our existing practical knowledge, 
 but they belong to physics or physiology, and not to psychology, 
 since in making them we deliberately leave out of account all that 
 is characteristic of conscious activity. 
 
 To those who argue that all our conscious activities are depend- 
 ent on physical conditions, the reply is that "physical conditions" 
 are in ultimate analysis only imperfect abstractions. If once we re- 
 gard them as anything more, we are plunged into all the difficul- 
 ties which modern philosophy since Descartes has been continu- 
 ously and successfully grappling with. The universe is a spiritual 
 universe, and not a dualistic universe of matter and mind. 
 
 This book is concerned with physiology and not psychology. 
 I have claimed for physiology its rightful practical sphere in 
 distinction from that of physics and chemistry. But we have 
 reached a limit to the sphere of physiology when we come to deal 
 with conscious activity. 
 
APPENDIX 
 
 THIS appendix contains a description and discussion of several special 
 methods of blood examination associated with my name, together with 
 modifications introduced by myself and others since the methods were 
 originally described. Methods of gas analysis are not included, since 
 these are collected in my book "Methods of Air -Analysis." 
 
 Until a few years ago the gases present in the easily dissociable and 
 free state in blood were universally determined by means of the mercurial 
 vacuum pump, which had been gradually perfected by Lothar Meyer, 
 Ludwig, Pfliiger, and others, while Leonard Hill had considerably 
 simplified it for ordinary uses. It required, however, an inconveniently 
 large amount of blood and was also not very accurate, since even when 
 large volumes of blood were used errors due to gas adhering to the glass 
 could not be avoided. The presence of these errors was clearly shown by 
 the fact that the amount of nitrogen apparently obtained from the blood 
 was not only variable, but much greater than the amount which the blood 
 was capable of dissolving. The excess of nitrogen could be calculated as 
 due to contamination with air from the pump; but this correction was 
 not very satisfactory, since gas must also be left in the pump at the end 
 of the operation of pumping. The discovery which I made in 1897, that 
 oxygen or CO can be liberated quantitatively from oxyhaemoglobin or 
 CO -haemoglobin by ferricyanide, 1 made it possible to dispense with the 
 blood pump and greatly simplify blood-gas determination and increase 
 its accuracy. With the new method Lorrain Smith and I found also that 
 the oxygen capacity of blood varies exactly as its coloring power, so that 
 the oxygen capacity can be determined colorimetrically. The methods 
 now to be described are based partly on the ferricyanide reaction and 
 partly on the colorimetry of blood. 
 
 A. Determination of Oxygen Capacity of Blood Haemoglobin 
 
 by Ferricyanide 
 
 The following method of determining very accurately the oxygen 
 capacity of the haemoglobin in blood or a solution of haemoglobin was 
 first fully described in ipoo. 2 Although the oxygen capacity can be de- 
 termined with much smaller quantities of blood by the apparatus de- 
 scribed below, it seems useful to describe also the earlier method, as it 
 
 ^aldane, Journ. of Physiol., XXII, p. 298, 1898. 
 2 Haldane, Journ. of Physwl., XXV, p. 295, 1900. 
 
RESPIRATION 
 
 401 
 
 can be carried out with very simple apparatus, easily put together in any 
 laboratory, and suitable not only for exact research, but for use by 
 students. The chemical facts on which the method is based have already 
 been referred to in Chapter IV. 
 
 The apparatus is shown in Figure 101 and the process is as follows. 
 Twenty cc. of the oxalated or defibrinated blood thoroughly saturated 
 with air by rotating it in a large flask, are measured out from a pipette 
 into the bottle A, which has a capacity of about 120 cc. 
 
 Figure 101. 
 
 Apparatus for determining the oxygen capacity of haemo- 
 globin in blood. 
 
 As it is important to avoid blowing expired air into the bottle, the 
 last drops of blood are expelled from the pipette by closing the top and 
 warming the bulb with the hand. In filling the pipette, care must also be 
 taken that the corpuscles have not had time to begin to subside in the 
 vessel from which the pipette is filled. Thirty cc. are then added of a 
 solution prepared by diluting ordinary strong ammonia solution, (sp. 
 
402 RESPIRATION 
 
 gr. 0.88) with distilled water to i/25oth, and the mixture shaken. The 
 ammonia solution prevents CO 2 from coming off and also lakes the 
 blood. Unless the blood is laked, the ferricyanide cannot act on the 
 haemoglobin, since the corpuscle walls are impermeable to ferricyanide. 
 About 4 cc. of a saturated solution of potassium ferricyanide are then 
 poured into the small tube B (the length of which should slightly exceed 
 the size of the bottle) and placed upright in A. The rubber stopper, which 
 is provided, as shown, with a bent glass tube connected with the burette 
 by stout rubber tubing of about i mm. bore, is then firmly inserted, 
 and the bottle placed in the vessel of water C, the temperature of 
 which should be as nearly as possible that of the room and of the 
 liquid in the bottle. If the stopper is not heavy enough to sink the bottle, 
 the latter should be weighted. By opening to the outside the three-way 
 tap (or a T tube and clip) on the burette, and raising the leveling tube, 
 which is held by a spring clamp, the water in the burette is brought to 
 a level close to the top. The tap or T tube is then closed to the outside, 
 and the reading of the burette (which should be graduated to .05 cc., and 
 read to .01 cc.) taken after careful leveling, as soon as the temperature 
 has become constant, as shown by the constancy of the reading. Mean- 
 while the water gauge (which has a bore of about 2 mm.) attached to the 
 temperature and pressure-control tube is accurately adjusted to a defi- 
 nite mark. This is easily accomplished by sliding the rubber backwards 
 or forwards on the narrow glass tube D. The control tube is an ordinary 
 test tube containing some mercury to sink it. 
 
 As soon as the reading of the burette is constant, the bottle is tilted so 
 as to upset B, and is shaken as long as the gas is evolved. During this 
 operation B should be repeatedly emptied, as otherwise the oxygen dis- 
 solved in its liquid might not be completely given off. When the evolution 
 of gas has ceased, the bottle is replaced in the water. If, as is probable, 
 the very sensitive pressure gauge indicates an alteration in the tempera- 
 ture of the water, cold water from a tap, or else warmed water, is added 
 till the original temperature has been reestablished, and the reading of 
 the burette noted as soon as it is constant. The bottle is again shaken, 
 etc., to make sure that the result is constant; and usually about fifteen 
 minutes will be needed to complete the operations. The temperature of 
 the water in the jacket of the burette 3 and the reading of the barometer 
 are now taken, and the oxygen evolved is reduced to its dry volume at o 
 and 760 mm. A table can be used for the reduction, and one is given in 
 Methods of Air Analysis. 
 
 * The jacketing of the burette may be omitted, in which case the thermometer 
 should be suspended with its bulb close to the upper part of the burette. 
 
RESPIRATION 403 
 
 To calculate the oxygen evolved from 100 cc. of blood, allowance must 
 be made for the fact that a 20 cc. pipette does not deliver 20 cc. of blood, 
 but only about 19.6 cc. The actual amount of shortage can easily be 
 determined by weighing. A further slight correction is needed on ac- 
 count of the fact that the air in the bottle at the end of the operation is 
 richer in oxygen than at the beginning, so that, as oxygen is about a 
 third more soluble than air, slightly more gas will be in solution. With a 
 bottle of 120 cc. capacity and 20 per cent of oxygen in the blood, the air 
 in the bottle will evidently contain about 26 per cent of oxygen, so that, 
 assuming that the coefficients of absorption of oxygen and nitrogen in 
 the 54 cc. of liquid in the bottle are nearly the same as in water, the 
 correction will amount at 15 to .03 cc. in the reading of the burette, if 
 the oxygen capacity is normal, or 0.75 per cent of the oxygen given off. 
 
 In order to make quite sure that no oxyhaemoglobin remains in the 
 solution owing to a reshrinkage of corpuscles on adding the ferricyanide, 
 and consequent escape of some of the oxyhaemoglobin from the action 
 of the ferricyanide, the liquid in the bottle can afterwards be examined 
 as follows. Part of it is diluted with 0.8 per cent salt solution, shaken up 
 in a test tube with expired air so as to render the solution just acid, and 
 examined spectroscopically. Any trace of oxyhaemoglobin left in incom- 
 pletely laked corpuscles is shown by the presence of the characteristic 
 absorption bands. These are completely absent if only methaemoglobin 
 is present, as ought to be the case. If they are present the result will be 
 too low, and the experiment must be repeated with saponin added. 
 
 If the blood is saturated with CO instead of oxygen the reaction is 
 slower, but gives precisely the same result. The correction for physically 
 dissolved gas is, however, scarcely appreciable, as CO is very little more 
 soluble than air. If the blood has begun to decompose, owing to bacterial 
 action, the result will of course be too low, and this can easily be detected, 
 because of the fact that each successive reading of the burette will be 
 lower, owing to the disappearance of oxygen. 4 There is no appreciable 
 error, owing to the tension of ammonia vapor in the air ; and the method 
 is one of extreme accuracy and certainty. Different determinations ought 
 not to differ by more than i/2ooth of the quantity measured. On com- 
 paring the results with those from the pump, after allowance in the case 
 of the pump for oxygen in simple solution in the blood, or adhering to 
 
 4 Under certain abnormal conditions even fresh mammalian blood, as Douglas 
 (Journ. of PAysiol., p. 453, 1910) has shown, may in presence of the ferricyanide 
 absorb an appreciable amount of oxygen before a determination is complete : in 
 which case the quicker method described below is greatly preferable. An appreciable 
 absorption can also be detected in normal fresh human blood left for an hour or 
 two in the apparatus. 
 
404 RESPIRATION 
 
 the glass in the pump, I obtained the following results, using a large- 
 sized Bohr pump with every precaution. 
 
 VOLUMES OF OXYGEN PER 
 
 ioo VOLUMES OF BLOOD 
 
 
 By blood By ferricyanide 
 pump method 
 
 Defibrinated ox blood 
 
 24.38 
 
 24.43 
 
 
 i 
 
 24-35 
 
 Oxalated 
 
 20.36 
 
 20.47 
 
 
 
 20.57 
 
 
 
 22.2O 
 
 Oxalated 
 
 22.40 
 
 
 Average 
 
 
 22.33 
 
 22.38 22.39 
 
 B. Determination of Oxygen Capacity of Blood Haemoglobin 
 by Haemoglobinometer 
 
 Colorimetric methods of estimating the relative concentrations of 
 haemoglobin in blood have been used for long; and in 1878 the late 
 Sir William Gowers introduced his well-known and extremely con- 
 venient "haemoglobinometer" for clinical purposes. 5 In this apparatus 
 there are two tubes A and B (Figure 102) of equal diameter; A is 
 sealed and contains picrocarmine jelly of such strength and composition 
 that when 20 cubic millimeters of normal human blood are diluted with 
 water in the tube B to the mark ioo, the tints of the liquid in the two 
 tubes are the same. If the blood contains abnormally little or much 
 haemoglobin, the quantity of water required to produce the tint of the 
 standard picrocarmine solution will be correspondingly less or more, 
 so that the percentage of the normal proportion of haemoglobin can be 
 read off on the tube. The diameter of the tubes and strength of the 
 picrocarmine or haemoglobin solution are so chosen that any variation 
 from the normal strength can be perceived with the maximum of readi- 
 ness. A solution much stronger or weaker would not be suitable. The 
 design is thus not only extremely convenient, but also thoroughly cor- 
 rect in principle. 
 
 When it was discovered that the coloring power and oxygen capacity 
 of haemoglobin are strictly proportional to one another it became evident 
 
 5 Gowers, Trans. Clinical Soc., XII, p. 64, 1878. 
 
RESPIRATION 
 
 405 
 
 that the Gowers haemoglobinometer could be made a very exact instru- 
 ment for determining the oxygen capacity of blood, and could also be 
 improved in other respects. I introduced the necessary improvements in 
 1 90 1. 6 For the picrocarmine solution there is substituted a i per cent 
 solution of blood with an oxygen capacity of 18.5 cc. per 100 cc. of blood, 
 since the average of a number of normal men showed that this is the 
 average oxygen capacity for men. To make this solution keep its coloring 
 
 Figure 102. 
 
 Gowers-Haldane Haemoglobinometer 
 
 A Glass tube containing blood solution of standard tint. 
 B Graduated tube. 
 C Rubber stand for tubes A and B. 
 D Capillary pipette and suction tube ; wires for cleaning 
 
 the pipette are supplied. 
 E Bottle with pipette stopper. 
 F Glass tube holding 6 lancets. 
 G Tube and cap for fixing over ordinary gas burners. 
 
 power it is saturated with CO, and sealed up with only CO, and no oxy- 
 gen, in the empty space above the blood solution. Hoppe Seyler had 
 already found that a strong solution of CO-haemoglobin retains its 
 coloring power. This is also true for a dilute solution ; and the standard 
 haemoglobinometer tubes filled and sealed twenty years ago have re- 
 mained absolutely unaltered in color. 
 
 * Haldane, Journ. of Physwl., XXVI, p. 497, 1901. The instrument is made 
 by Hawksley, Wigmore Street, London, W. 
 
406 RESPIRATION 
 
 One defect of the picrocarmine tubes arose from the fact that the picro- 
 carmine is not completely stable, so that after a time its color alters. But 
 even the original standard was somewhat indefinite, depending as it 
 did on the particular percentage of haemoglobin in the sample of normal 
 blood with which it was standardized. Another defect depended on the 
 fact that the colors of the blood and picrocarmine solution are not the 
 same spectrally. In consequence of this a color match with one quality 
 of light is no longer a match with a different quality of light. Thus in 
 ordinary artificial light the reading of the instrument is quite different 
 from that in average daylight ; and in different qualities of daylight, and 
 with different observers, the match differs. The same defect exists in 
 various later forms of haemoglobinometer, where colored glass or colored 
 paper is used as a standard. By using CO haemoglobin as the standard 
 solution, and saturating the blood under examination with CO or coal 
 gas these defects are avoided. 
 
 To avoid errors due to inequality in the diameters of the tubes, each 
 tube has first of all two marks placed on it the first at the level when 
 .2 cc. of water are introduced into a dry tube, and the second at the 
 level given by 2 cc. The distance between these two marks must corre- 
 spond exactly in the standard tube and measuring tube and this must be 
 borne in mind if either tube gets broken and has to be replaced. The 
 20 cubic millimeter pipette is also standardized by weighing on a deli- 
 cate balance. 
 
 To make a determination, some water is first introduced into the 
 measuring tube. Twenty cmm. of blood from a prick in the finger or 
 ear are then measured into this water from the dry pipette. The blood 
 sinks, and the pipette is rinsed out with some of the water standing above 
 the blood. Some coal gas or CO is then run into the upper part of the 
 measuring tube through narrow rubber or glass tubing, and the top of 
 the tube promptly closed with the finger. With the thumb of the same 
 hand on the lower end of the tube the latter is then inverted several 
 times so as to saturate the haemoglobin completely with CO, but without 
 warming the contents of the tube. The finger can then be slid off the 
 open end of the tube without the slightest loss of liquid. More water 
 is now added by means of the dropping pipette until the tints appear 
 equal. When this point is reached the level is read off after a short inter- 
 val to allow liquid to run down. Another drop is introduced, and then 
 another, until the tints appear unequal again; and the mean of the 
 readings giving equality is taken as showing the required percentage. 
 This indicates the oxygen capacity of the haemoglobin in percentages of 
 18.5 cc. of oxygen capacity per 100 cc. of blood. 
 
 In judging of equality in tint the tubes are held up before a window 
 
RESPIRATION 407 
 
 or an opal shade covering a gas flame or electric lamp. At every observa- 
 tion the tubes are transposed. This is essential since it will be found 
 that in all probability the tint of one tube will appear deeper when it is 
 held on one side than when on the opposite side. If, for instance, the 
 tubes are nearly equal in depth of color they will appear equal when one 
 tube is on the right or left side, but not vice versa. A slight inequality 
 of this kind is rather a help to accuracy, as probably only one reading 
 will give equality on both sides. With careful work any error in a de- 
 termination should not exceed 0.5 per cent. The method is thus one of 
 great accuracy. 
 
 It is often loosely assumed that colorimetric estimations are uncertain. 
 This is certainly not the case if they are properly carried out, with ap- 
 preciation of the precautions needed to avoid the errors referred to 
 above, of physiological origin. Another common misconception is that a 
 uniform colored surface is necessary, and that, as a tube does not give 
 this, a method such as that just described must be inaccurate. The 
 surfaces need not be uniform, provided they are similar to one another, 
 as in the case of two similar tubes. 
 
 The correctness of a Gowers-Haldane haemoglobinometer can be 
 checked at any time by the ferricyanide method described under A or 
 C. Another check on the correctness of the standard solution is that it 
 must have practically the same pink tint as fresh blood saturated with 
 CO. If there has been any defect in filling, the standard tube will appear 
 yellower. With a proper standard tube one can tell at once by the 
 absence or presence of yellow color whether a patient's blood is free from 
 methaemoglobin or other abnormal blood pigments. 
 
 For ordinary clinical work it is convenient to work ordinarily with a 
 picrocarmine standard tube, and only occasionally ascertain the correc- 
 tion necessary with this standard. The correction can easily be made by 
 comparing the results for the same person and time with the two tubes. 
 
 C. Determination of Oxygen and Carbon Dioxide in Blood by 
 Ferricyanide and Acid 
 
 As mentioned in Chapter IV, a method, based on the use of ferri- 
 cyanide, was described in a paper by Mr. Barcroft and myself in I9O2. 7 
 The principle of this method is that, without permitting any previous 
 contact of the blood with air, the oxygen of a small measured volume of 
 blood is liberated by ferricyanide in a closed vessel, and the pressure 
 produced by the liberation measured without any alteration being allowed 
 in the volume of gas in the vessel. The CO 2 is then similarly liberated by 
 
 7 Barcroft and Haldane, Journ. of Physiol., XXVIII, p. 232, 1902. 
 
408 RESPIRATION 
 
 acid, and its pressure measured. When certain corrections are made, it 
 is then possible to estimate either the total oxygen and total CO 2 , or the 
 combined oxygen and combined CO 2 in the blood. The gas is measured 
 by the increase of pressure at constant volume, and not by the increase 
 of volume at constant pressure. Theoretically, either method is correct, 
 in accordance with Boyle's Law ; but as Barcroft required a method for 
 dealing with very small quantities of blood, and a very delicate pressure- 
 gauge was needed in any case, it seemed simpler to graduate the pres- 
 sure gauge in millimeters, and keep the gas at constant volume, retain- 
 ing, however, the control vessel, as in the original form of apparatus. 
 I therefore designed the apparatus as it was originally figured in our 
 paper, and the tests we made gave very satisfactory results so far as 
 they went. 
 
 One defect of the apparatus described in the previous section is that 
 a considerable time is needed to reach temperature equilibrium and to 
 shake out all the extra free oxygen from the blood solution. The latter 
 defect would apply still more to an apparatus in which CO 2 had to be 
 shaken out. In the new apparatus the volume of liquid was therefore 
 greatly diminished, and the relative volume of air to blood solution 
 greatly increased; and this was also rendered advisable owing to the 
 fact that nearly as much CO 2 remains in solution in the liquid as is 
 present in an equal volume of air. The increased volume of air had, 
 however, the disadvantages, first that the pressure of ammonia in the 
 air introduced an appreciable source of error, and secondly that much 
 more care was needed as to temperature equilibrium in the blood vessel 
 and control vessel. A further source of error was slight variation in 
 capillarity at different levels in the gauges of the blood vessel and 
 control vessel. In spite of all improvements in this apparatus and the 
 methods of using it, there appears to be a range of error with it of at 
 least 2 per cent of the quantity measured, even when the error due to 
 ammonia vapor is completely eliminated. 
 
 The apparatus was rendered much more convenient, though also less 
 easy to make or repair, by Brodie. 8 It was also simplified by Barcroft; 
 who named his modification the "differential" apparatus. 9 Barcroft con- 
 nects the gauges of the blood vessel and control vessel, so that there is 
 only one manometer instead of two, and estimates the gas given off from 
 the readings of this compound gauge. With this construction the ap- 
 paratus works at neither constant volume nor constant pressure, so that 
 the gas given off cannot be correctly deduced from the mere readings of 
 the gauges. He therefore calibrates the apparatus empirically with the 
 
 8 Brodie, Journ. of Physiol., XXXIX, p. 391, 1910. 
 
 9 Described fully in Barcroft's book, The Respiratory Functions of the Blood. 
 
RESPIRATION 409 
 
 help of the oxygen liberated from a titrated solution of hydrogen perox- 
 ide. But this is a rather serious complication, and even if the calibration is 
 correctly made it can only apply correctly at a certain barometric pressure 
 and would not be quite valid over the variations of barometric pressure 
 ordinarily met with. I cannot, therefore, regard this plan as satisfactory 
 for some kinds of exact work. On the other hand this objection 
 does not apply where the empirical calibration is not needed, as in 
 determinations of the percentage saturation of haemoglobin with oxy- 
 gen for instance in investigating dissociation curves of oxyhaemo- 
 globin. Barcroft has also devised a small model, for which only o.i cc. 
 of blood is required. 
 
 A very different form of the ferricyanide method has recently been 
 introduced by Yandell Henderson and Smith. 10 The blood (i cc.) is 
 introduced (under ammonia solution without contact with air, just as 
 in the Barcroft-Haldane method) into the bottom of a diffusion tube 
 of about 12 cc. capacity. This tube is provided with a 3-way tap at the 
 bottom end and a thin rubber stopper at the top, and is graduated for a 
 short distance from the top. A fine hypodermic needle is then thrust 
 through the rubber to equalize the pressure inside and outside of the 
 tube, the needle withdrawn, and the blood and ammonia solution mixed 
 so as to lake the blood. Ferricyanide solution is then injected through the 
 stopper, and the tube rotated for five minutes so that the whole excess 
 of free oxygen diffuses out into the air of the tube. The tube is then 
 inverted and the stopper removed under water so that the pressure inside 
 and outside the tube is equalized. The volume of gas in the tube is read 
 off ; and finally nearly the whole of this gas is drawn into a Haldane gas- 
 analysis apparatus, and the oxygen percentage determined. From the 
 increased oxygen percentage of this gas as compared with air, and the 
 volume of gas in the tube, the oxygen given off by the blood can easily 
 be calculated. The CO 2 in the blood is estimated similarly; and both 
 oxygen and CO 2 can be estimated in the same sample of blood. This 
 method seems to be about as accurate as the Barcroft-Haldane method, 
 and to be easier for those familiar with accurate gas analysis. It appears 
 to be specially suitable for comparisons of the arterial and venous blood 
 in animals ; and evidently any CO in blood can be estimated conveniently 
 by this method, which also has the advantage that corrections for physical 
 solution of gases are greatly reduced. 
 
 Still another method is to use the Van Slyke vacuum apparatus in 
 connection with ferricyanide. 11 This, however, involves the various 
 
 10 Yandell Henderson and Smith, Journ. of Bwl. Chem., XXXIII, p. 39, 1918. 
 
 11 Van Slyke, Journ. of Bwl. Chem., XXX, p. 347, 1917 ; and XXXIII, p. 127, 
 1918 
 
4 io RESPIRATION 
 
 sources of error connected with the use of a vacuum pump, or necessitates 
 analysis of the gas obtained from the blood. 
 
 Until recently we have used at Oxford the Brodie modification of the 
 Barcroft-Haldane apparatus. As, however, the range of error with this 
 apparatus has been about 2 per cent, I have quite recently devised a 
 new apparatus, with a view especially to more accurate determinations 
 of the oxygen in human arterial blood, and of dissociation curves. 12 
 With this apparatus it is possible to reach an accuracy as great as with 
 the original ferricyanide apparatus i.e., to within 0.5 per cent of the 
 oxygen capacity of the blood. This new apparatus will therefore be de- 
 scribed in full. On account of the present difficulty and expense in getting 
 glass apparatus made, it was designed so that it could if necessary be 
 put together in a laboratory from easily obtainable parts, just as in the 
 case of the original apparatus. 
 
 When blood from a blood vessel is used, a glass syringe with solid 
 glass piston is employed for obtaining the sample. This method was first 
 applied to human arteries by Hiirter, and developed by Stadie and others. 
 Professor Meakins, with whom I have been associated in work on human 
 blood gases, employs the following procedure. A very small quantity of 
 finely powdered potassium oxalate is introduced into the bottom of the 
 syringe. The piston is then introduced and a little liquid paraffin drawn 
 in, and as much as possible expelled again with the syringe pointing 
 upwards so as not to expel the oxalate. After disinfection of the skin the 
 needle (previously sterilized) is introduced into the radial artery or other 
 vessel, and about 5 cc. or more of blood withdrawn, a compress and 
 bandage being afterwards applied over the place for an hour if the 
 vessel was an artery. The needle is then removed and washed, and the 
 blood transferred (with the syringe pointing upwards) through a rubber 
 connection into a graduated pipette holding more than 2 cc. From this 
 pipette an exactly measured quantity of about 2 cc. is introduced beneath 
 the sodium carbonate or ammonia solution in the blood-gas flask. At the 
 end of the operation about 0.5 cc. remains in the pipette, so that none 
 of the blood has come in contact with air. 
 
 The apparatus is shown in Figure 103. In principle it is similar to that 
 shown in Figure 101, but designed for small quantities of blood and for 
 determining CO 2 . The blood is received in one of the small flasks shown, 
 while the other is for temperature control. Each has a capacity of about 
 20 cc. The procedure differs according as it is desired to determine the 
 oxygen or the CO 2 of the blood. In the former case the first step is to 
 measure 2 cc. of a i per cent solution of dried sodium carbonate into 
 one of the two small flasks (about 20 cc. capacity) shown and add a 
 
 "Haldane, Journ. of Pathol. and. Bacterial., XXIII, p. 443, 1920. 
 
RESPIRATION 
 
 411 
 
 small quantity of saponin on the point of a penknife. Exactly 2 cc., or at 
 any rate an exactly determined volume, of the blood is then measured 
 out from the pipette into the flask beneath the sodium carbonate solution. 
 The flask is then firmly corked and completely immersed beyond the 
 cork in the bath alongside the other (control) flask until the temperature 
 
 Figure 103. 
 Apparatus for blood-gas analysis. 
 
 of the air in the flask becomes completely steady. The flasks are con- 
 nected, as shown, by means of thick-walled rubber tubing of about 2 mm. 
 bore with the two gauges and gas burette fixed on the wooden stand. The 
 glass connections, taps, and gauges are also of 2 mm. bore, and so ar- 
 ranged that the connections of the two flasks are of equal volume. The 
 burette itself consists of an ordinary i cc. dropping pipette divided to 
 .01 cc., and therefore capable of being read to .002 cc. The correctness 
 of the graduation can easily be tested by weighing the water delivered by 
 it. The taps are at first left open to air, but are turned after a few 
 minutes so that the flasks communicate only with the gauges and burette ; 
 and the leveling tubes are previously adjusted so that the gauge levels 
 
4 I2 RESPIRATION 
 
 are at the zero marks and the burette level is at a convenient distance 
 below zero. The gauges are then carefully observed, and the water in 
 the bath is occasionally stirred by blowing air through it. It will be 
 found that when both the gauges are exactly adjusted they do not keep 
 even when left to themselves until at least ten minutes after the blood 
 flask has been placed in the bath. The alterations are compensated by 
 means of the leveling tubes; and when the gauges have come steady, 
 or only move together, the burette is read off exactly. The confining 
 liquid is distilled water containing a small quantity of bile-salts which 
 make the readings more certain and sensitive. 
 
 The blood flask is now agitated for two or three minutes in order 
 that the blood may take up all the gas it is capable of taking. At the 
 same time it is laked by the saponin. In the process of agitation the flask 
 is never removed from the bath. It is held by the neck with forceps or 
 something else interposed to shield it from the warmth of the fingers. 
 The gauges are now again adjusted, and, after they are quite steady, 
 which should be the case almost at once, the burette is again read off. The 
 difference between the two readings gives the gas absorbed by the blood 
 from the air. From this we can calculate the volume of oxygen absorbed 
 by the haemoglobin. 
 
 The first step in the calculation is to reduce the gas absorbed to its 
 dry volume at o and 760 mm. and calculate its volume per 100 cc. of 
 blood. For this purpose the barometer is read and the temperature given 
 by a thermometer (not shown in the figure) fixed on the front of the 
 stand, with the bulb close to the upper part of the burette. It is evident 
 that what is required is not the temperature of the bath or connections, 
 but that of the burette. The reduction is easily made with the help of a 
 table with factors for correction, such as that at page 60 (second edi- 
 tion) of my book on Methods of Air Analysis. 
 
 We have now to calculate how much of the gas absorbed has simply 
 gone into physical solution. Blood in the living body is saturated with 
 nitrogen at the partial pressure of the nitrogen in the alveolar air. 
 Allowing for the aqueous vapor present, this partial pressure is about 
 75 per cent of the existing atmospheric pressure. The coefficient of ab- 
 sorption of nitrogen in blood at 38C is .on, according to Bohr's de- 
 termination. Hence at ordinary atmospheric pressure there will be .83 cc. 
 of nitrogen (at o and 760 mm.) in solution in 100 cc. of blood. The 
 blood in the flask will become saturated at about 15 with nitrogen at a 
 partial pressure of about 78 per cent of an atmosphere ; and, as the co- 
 efficient of absorption is .016, about 1.25 cc. of nitrogen will be in 
 solution per 100 cc. of blood saturated with air at 15. Thus 100 cc. of 
 blood will take up .42 cc. of extra nitrogen on saturation. 
 
RESPIRATION 413 
 
 To calculate how much extra oxygen the blood will take up in simple 
 solution, we must know the partial pressure of oxygen at which the blood 
 taken from the living body is saturated, and this can be deduced pretty 
 accurately from the percentage saturation of the haemoglobin and the 
 dissociation curve of oxyhaemoglobin in human blood. Now it was found 
 by Meakins and Davies 13 that the haemoglobin of normal human arterial 
 blood is about 95 per cent saturated, which corresponds to an oxygen 
 pressure of 1 1 per cent of an atmosphere, or 84 mm. The coefficient of 
 absorption of oxygen in blood at 38 is .022. Hence there will be .24 cc. 
 of oxygen in simple solution in 100 cc. of arterial blood. At 15 the co- 
 efficient of absorption is .031 and at ordinary atmospheric pressures the 
 partial pressure of oxygen in the bottle will be 20.5 per cent of an 
 atmosphere. Hence .63 cc. of oxygen will be in solution in 100 cc. of 
 blood saturated with air at 15, and the extra oxygen taken up in solu- 
 tion will be .39 cc. Thus the total extra gas taken up in solution will be 
 .42 + .39 = .81 cc. in 100 cc. of blood, and only the balance of the 
 proportion actually taken up in the blood flask will go to saturate the 
 haemoglobin. Hence if the temperature of the water bath is 15 the 
 allowance for gas in simple solution will be .81 cc. 
 
 If the bath is above or below 15 this allowance will be a little less or 
 greater, and a calculation shows that for each degree above or below 15, 
 between the temperatures of 20 and 10, the allowance will have to be 
 diminished or increased by .038 cc. 
 
 An example will make the calculation of the percentage saturation of 
 the haemoglobin clear. Let us suppose that 2.15 cc. of arterial blood have 
 been delivered into the flask and the constant reading of the burette 
 after temperature equilibrium had been obtained was .072 cc., and after 
 agitating the blood .030. Thus 0.042 cc. of gas had been absorbed from 
 2.15 cc. of blood, or 1.95 cc. from 100 cc. The temperature was 14 and 
 the barometer 755 mm. Hence the factor for reduction to dry gas at o 
 and 760 mm. was 0.930. Therefore the dry gas at standard pressure and 
 temperature was 1.81 cc. The temperature of the bath was 13. Hence 
 .81 + .08 = .89 cc. went into physical solution, so that 0.92 cc. of oxygen 
 was absorbed by the haemoglobin. 
 
 To determine the percentage saturation of the haemoglobin it is 
 necessary to know the total oxygen capacity of the haemoglobin ; and this 
 can now be determined directly. To the tube passing through the stopper 
 of the blood flask there is attached a loop of wire into which a small 
 tube of thin glass can be inserted. In the tube is placed .25 cc. of satu- 
 rated ferricyanide solution and the flask closed and reinserted in the 
 water bath till temperature equilibrium is reached. The burette is again 
 
 "Meakins and Davies, Journ. of Patkol, and Bacter., XXIII, p. 451, 1920. 
 
414 RESPIRATION 
 
 read off, and the flask turned up so as to let the ferricyanide flow into 
 the blood solution. Before doing this, however, the blood solution should 
 be observed to make sure that it is perfectly laked and transparent; 
 otherwise more saponin must be added. The flask is now agitated as 
 long as gas continues to come off as shown by the movements of the 
 gauge. This will take three or four minutes. The burette is again read 
 off, which gives the volume of oxygen given off. This is reduced to dry 
 volume at o and 760 mm. and per 100 cc. of blood. 
 
 Let us suppose that the oxygen capacity of the haemoglobin in the 
 above example was 17.4 cc. per 100 cc. of blood. The percentage satura- 
 tion of the haemoglobin in the arterial blood was therefore 
 
 17.40 .92 
 
 100 x 94.7. 
 
 17.40 
 
 It is easier to determine the oxygen capacity by means of a Gowers- 
 Haldane haemoglobinometer, in which 100 per cent corresponds 
 to an oxygen capacity of 18.5. For this purpose a sample of the 
 blood drawn from the artery is used for the determination. In the above 
 example the oxygen capacity of 17.4 corresponds to 94 per cent on the 
 haemoglobinometer scale, and the range of error in carefully made 
 haemoglobinometer determinations is only about 0.5 per cent. The ac- 
 curacy of both methods is strikingly shown by the fact that in 36 determi- 
 nations by Meakins and Davies of the oxygen capacity of blood from 
 patients and healthy persons the maximum difference between the re- 
 sults by the haemoglobinometer and by the new method was under i per 
 cent of the oxygen capacity. 14 
 
 A haemoglobinometer can, of course, be exactly standardized by the 
 method just described. If the haemoglobinometer is used, it is unneces- 
 sary to use saponin or ferricyanide in determining the percentage satura- 
 tion of the haemoglobin in the sample of blood. The total available 
 oxygen in the sample of arterial blood is the oxygen combined with 
 haemoglobin plus the dissolved oxygen. This was, in the above example, 
 16.48 +.24 =16.72 cc. per 100 cc. of blood. 
 
 If, instead of being normal arterial blood, the sample was venous 
 blood, or arterial blood of abnormally low saturation with oxygen, the 
 calculation must be slightly modified, since less oxygen in simple solu- 
 tion is present in the sample. Thus if the blood turned out to be only 
 half saturated with oxygen the partial pressure of oxygen in the sample 
 would only be about 4 per cent of an atmosphere. Hence there would 
 only be .09 cc. of dissolved oxygen present, instead of .24 cc. This would 
 increase the correction at 15 for dissolved gas from .81 to .96 cc. a 
 difference which, however, affects the result but little. Ordinary varia- 
 
 14 Meakins and Davies, Journ. of Pathol. and, Bacter., XXIII, p. 454, 1920. 
 
RESPIRATION 415 
 
 tions of barometric pressure do not sensibly affect the correction, but at 
 high altitudes the correction must evidently be diminished in the pro- 
 portion of about o.i cc. for every 100 mm. of diminution in atmospheric 
 pressure. 
 
 If the blood is taken, not from the living body, but from a saturating 
 vessel, the gases dissolved physically must be calculated on the same 
 principle, allowing for their pressures in the vessel. 15 
 
 The method just described has been tested for accuracy in several 
 ways. In the first place it has been found that when blood fully saturated 
 with air at room temperature is placed under the sodium carbonate 
 solution in the ordinary way and then agitated after the gauges have 
 become steady, there is no sensible variation in the reading of the burette 
 afterwards. The constancy of the reading can be relied on to .002 cc. 
 with careful work. Hence the percentage saturation can be relied on to 
 0.5 per cent, or the oxygen capacity per 100 cc. of blood to o.i cc., if the 
 measuring pipettes are properly calibrated. This is as good a result as 
 could be obtained with 20 cc. of blood by means of the original ferri- 
 cyanide apparatus. The present method is therefore as exact as the 
 original one for determining the oxygen capacity of blood, but is quicker 
 and more convenient. By using sodium carbonate instead of ammonia 
 solution the errors due to diminution of the vapor pressures of ammonia 
 and water on mixing the blood with the solution are eliminated, while 
 the use of saponin, first introduced by C. G. Douglas, produces the laking 
 of the blood which is necessary in order to allow the ferricyanide to 
 act on the oxyhaemoglobin. The fact that, as has been found by Meakins 
 and Davies, haemoglobinometer estimations coincide within i per cent 
 with the results by this method furnishes further confirmatory evidence. 
 
 The new apparatus gives sharper results than the constant volume 
 method which Barcroft and I described in 1902. This is, I think, partly 
 due to the larger volume (2 cc.) of blood employed; partly to the fact 
 that the disturbance due to the use of ammonia solution is avoided and a 
 sharper index of temperature equilibrium is given by the two gauges 
 of the present apparatus ; and partly because the gauge levels are always 
 at the same place, whereas in the constant-volume apparatus the gauge 
 levels shift to places wide apart, so that small errors due to varying 
 capillarity of the gauge tubes are apt to tell. It is thus difficult, with the 
 constant-volume apparatus, to avoid errors within 2 per cent on either 
 side of the actual percentage saturation. 
 
 15 In the paper by Barcroft and myself where we first described the constant 
 volume blood-gas apparatus, the correction for gas in simple solution was un- 
 fortunately given incorrectly; and this doubtless accounts for the somewhat 
 distorted forms of the dissociation curves of oxyhaemoglobin in Barcroft's earlier 
 experiments on this subject. 
 
416 RESPIRATION 
 
 When it is desired to determine the CO 2 content of the blood the 
 procedure must be modified, as sodium carbonate cannot be used, and 
 2 cc. of blood would give too much CO 2 for the capacity of the burette, 
 apart from other causes of error. Therefore only about i cc. of blood 
 should be taken. This is delivered under 1.5 cc. of a solution of 4 parts 
 of ordinary strong ammonia solution (sp. gr. .88) to a liter of boiled 
 distilled water, and a trace of saponin added. To avoid the presence of 
 any carbonate in the ammonia the strong solution is first shaken up with 
 some unslaked lime and allowed to settle. The stock of dilute solution 
 is kept tightly corked. As soon as the ammonia solution is placed in the 
 flask, the latter is kept tightly corked until the blood is added, otherwise 
 a considerable amount of CO 2 may diffuse in and cause error. The blood 
 is shaken up to lake it, and .25 cc. of ferricyanide afterwards added to 
 liberate the oxygen, since if this were not done some oxygen might be 
 liberated by the acid. After all the liberated oxygen has been shaken off, 
 the small glass tube containing .25 cc. of 20 per cent solution of tartaric 
 acid is inserted and the burette read off after the gauges are steady. The 
 tartaric acid solution is then spilt into the blood solution and the flask 
 agitated under water till the CO 2 has completely ceased to come off, as 
 shown by the gauge. The burette is then adjusted and read off and the 
 volume of gas given off reduced to its dry volume at o and 760 mm. 
 and calculated per 100 cc. of blood. Part of the CO 2 however, remains 
 dissolved in the liquid in the flask, and must be allowed for. This liquid 
 is exactly the same as in the case of determination of CO 2 by means of 
 the constant- volume apparatus described by Barcroft and myself, so the 
 correction is made in a similar manner. At a temperature of 13 the 
 coefficient of absorption of CO 2 in this liquid was found to be i .00. Hence 
 if we know the total volume of the flask as compared with the volume 
 of gas in it when the liquid is also present, and the temperature of the 
 bath is 13, the total CO 2 liberated from the blood will be to the amount 
 shown by the burette as the total capacity of the flask to the volume of 
 gas in it when the liquid is also present. The capacity of the flask to the 
 cork is about 20 cc. Let us suppose that as determined by weighing with 
 the cork in place it is 20.5 cc., including the capacity of a piece of glass 
 tubing of about 4 mm. bore and two inches long which passes through 
 the cork. The volume of liquid in the flask is 3.0 cc. Hence if the tempera- 
 ture of the bath is 13 the total volume of CO 2 liberated is obtained by 
 
 20 ^ 
 multiplying the corrected volume actually read off by or adding 
 
 17 per cent. If the temperature is above or below 13 a fortieth must be 
 subtracted from or added to this addition, since the solubility of CO 2 
 diminishes by about a fortieth for each degree above 13, and increases 
 
RESPIRATION 417 
 
 similarly for each degree below 13. The glass tubing passing through 
 the cork is 4 mm. in bore in order to give room for the CO 2 given off 
 without its coming in contact appreciably with the rubber connecting 
 tubing. For determining CO 2 in blood it is better to use an ordinary 
 cork than a rubber stopper in the blood flask, as the rubber leads to a 
 slow absorption of CO 2 . 
 
 A further negative correction is required for any CO 2 present in the 
 solutions used, or absorbed from the air in the flask ; also for the small 
 error in the opposite direction owing to disappearance of ammonia 
 vapor from the air of the flask. The joint correction, which ought to be 
 very small, and may be either positive or negative, can be ascertained 
 by a blank experiment in which boiled distilled water in place of blood 
 is used in the flask. Or if the capacities of the two bottles are nearly 
 equal the blank experiment may be performed in one flask along with the 
 blood experiment in the other. In this way the correction is eliminated. 
 
 As shown by this method by Meakins and Davies, arterial blood gives 
 slightly more CO 2 than denbrinated blood at the same partial pressure 
 of CO 2 , as found in the experiments of Christiansen, Douglas, and 
 myself. 16 
 
 The following example illustrates the mode of calculation. The volume 
 of CO 2 given off from i.oo cc. of human arterial blood was 0.482 cc. as 
 read from the burette. Reduced to dry gas at o and 760, and calcu- 
 lated per 100 cc. of blood, this was 45.9 cc. The correction at 13 for the 
 CO 2 left in solution was 16.5 per cent, but as the temperature of the bath 
 was 15 the proper correction was 15.3 per cent. Hence the CO 2 con- 
 tained in 100 cc. of blood was 52.9 cc. A blank control experiment was 
 made simultaneously in the other flask, so there was no further correction. 
 
 In any cases where both the oxygen and CO 2 in a sample of blood are 
 required, it is better and quicker to make the determinations simul- 
 taneously in two different apparatus. 
 
 For the proper working of the apparatus it is essential that all the 
 joints, including the cork, should be absolutely tight. There is no 
 difficulty about this if the rubber tubing used is smooth and clean. To 
 test for tightness the burette should be read after the gauges are steady. 
 Positive or negative pressure is then produced for some time in the 
 apparatus by raising or lowering the leveling tubes. On readjusting the 
 gauges, the reading should be exactly the same as before, if the apparatus 
 is tight. If a leak exists it can soon be localized by putting pressure on 
 one part after another of the connections. 
 
 The apparatus can be put together without very much trouble, and if 
 
 16 Christiansen, Douglas, and Haldane, Journ. of PAysiol., XLVIII, p. 272, 
 1914. 
 
4 i8 RESPIRATION 
 
 three-way taps are not available T tubes may be substituted. Messrs. 
 Siebe Gorman & Co., 187 Westminster Bridge Road, London S. E., 
 supply it. 
 
 D. Colorimetric Determination of Percentage Saturation of 
 Haemoglobin with CO 
 
 This very convenient method is used in determining the oxygen pres- 
 sure of arterial blood, the total haemoglobin in the body, or the blood vol- 
 ume, as well as for investigations as to the properties of CO haemoglobin 
 and the phenomena of CO poisoning. It depends on the fact that a dilute 
 solution of CO haemoglobin has a pink color, quite different from the 
 yellow color of similarly diluted oxyhaemoglobin. 
 
 I originally used this color difference as an easy and delicate means 
 of recognizing the presence of CO in blood and roughly estimating the 
 saturation with CO; and I then thought that as it is impossible to 
 recognize by the difference of tint a difference of less than about 5 per 
 cent in the percentage saturation of haemoglobin with CO, the method 
 was at best a rough one. Various recent writers have fallen into the same 
 error. Further experience showed that with proper precautions the 
 method gives results of great accuracy. The following description is 
 taken almost verbatim from the account of the method given in 1912 by 
 Douglas and myself in our paper on oxygen secretion. 17 
 
 A solution of normal human blood (or blood from the animal experi- 
 mented on) is prepared of such strength as to correspond to about 0.5 
 per cent of the proportion of haemoglobin in standard human blood of 
 100 per cent strength by the Gowers-Haldane haemoglobinometer scale. 
 Two test tubes of equal bore of about 0.6 inch are selected, and into 
 each of these 5 cc. of the blood solution are measured with a pipette. 
 From a o.i per cent solution of carmine in ammoniacal distilled water 
 (this solution being kept in the dark in a cupboard) a dilute solution of 
 carmine in distilled water with a strength of tint about equal to or rather 
 greater than that of the blood solution is then prepared in a measuring 
 cylinder. The requisite amount of dilution (about one-twentieth of the 
 o.i per cent solution if the latter has been recently prepared) can easily 
 be estimated by the eye, and can be obtained at once, when experiments 
 are made daily, by diluting to a definite extent. A burette is filled with 
 the carmine solution, and another burette with water. The blood solution 
 in one of the test tubes is then saturated with CO by allowing coal gas to 
 run through the free part of the test tube, quickly closing the tube with 
 the thumb, and shaking the blood solution with the gas for a few seconds. 
 
 " Douglas and Haldane, Journ. of PAysioL, XLIV, p. 305, 1912. 
 
RESPIRATION 419 
 
 When looked at against the sky, the solution will now have a deep 
 purplish-pink tint, as compared with the brownish yellow of the normal 
 blood solution. The carmine is now added from the burette to the normal 
 blood solution until its tint is about equal in quality to that of the satu- 
 rated blood solution. It will then probably be found that the depth of 
 tint is too great in the tube containing the carmine. Water is then added 
 from the other burette until the depth of tint is equal, and if necessary 
 more carmine, until complete equality of both tint and depth of color is 
 obtained. In judging of this, the test tubes should be held up against the 
 sky, and it is absolutely necessary to change them repeatedly from side 
 to side ; otherwise gross error is certain. It will nearly always be found 
 that the right-hand tube appears a little yellower or pinker than the left- 
 hand one ; and a little deeper or less deep in color. This difference is in 
 reality a great help to accuracy. A point is first reached when the tubes 
 appear equal in tint or depth when held in one position, but unequal in 
 the other, and the end point, when the difference is the same on one side, 
 whichever tube is on that side, can be estimated with great delicacy. 
 The additions of carmine (or water) are continued until this point is 
 passed ; and if two successive additions both show equality, the mean of 
 the two readings is taken as correct. 
 
 To the carmine solution in the measuring cylinder a proportion of 
 water is now added equal to what had to be added from the water 
 burette to the carmine required to reach the end point of the titration. 
 The carmine solution is then ready for use. It will probably be found 
 that about 6 cc. of carmine are needed to reach the end point. The 
 amount required varies, however, according to the condition of the 
 strong carmine solution and the quality of the daylight. The carmine 
 solution is not stable, and it gradually becomes less deep in color, and 
 redder in tint than when first prepared. Hence the quantity of carmine 
 solution needed increases from month to month, and the extent to which 
 it has to be diluted for use diminishes. If the dilute solution is left for a 
 day or two exposed to light it becomes very markedly redder and more 
 dilute. 
 
 The titration of a blood sample is carried out as follows. One or two 
 drops of blood are needed, and are at once diluted with water. Half of 
 the dilute solution is poured into one of the two test tubes (always the 
 same one as that used for the saturated blood in standardizing the 
 carmine), and 5 cc. of the normal blood solution are measured with a 
 pipette into the other. Water is then allowed to drip from a tap into the 
 solution of the blood under examination until its depth of tint is about 
 equal to that of the normal solution. Carmine solution is now added to 
 the normal blood solution from the burette until the tints are equal, 
 
420 RESPIRATION 
 
 more water being also added to the other tube if necessary. The solution 
 under examination is then saturated with coal gas and the addition to 
 the normal blood solution of carmine is continued until the tints are 
 again equal. To illustrate the method of calculating the result we may 
 suppose that in the first result equality of tint was observed with 1.2 and 
 1.3 cc. of carmine, mean 1.25, and that in the second 6.4 and 6.8 cc. gave 
 equality, mean 6.6; the percentage saturation X is then given by the 
 result of the following proportion sum : 
 
 6.6 1.25 
 
 r r-: ; : : 100 : X 
 
 5 + 6.6 5 + 1.25 
 
 or, more simply, 
 
 6.25 6.6 
 
 100 x x = 35.1 per cent. 
 
 1.25 n.6 
 
 It is clear that the more carmine has already been added to the 
 normal blood solution the less effect on its tint will any further addition 
 have. Hence in approaching the point of equality only o.i cc. is added at 
 a time if not more than 2 cc. have already been added, whereas after 
 already adding 6 cc. it is useless to add less than about 0.4 cc. at a time. 
 
 The titration is repeated with the other half of the blood solution 
 for further safety, and it will be found that apart from accidents the 
 two results will nearly always agree within i per cent of the total satu- 
 ration. This accuracy is very surprising at first sight, since colorimetric 
 determinations have in general a rather bad reputation among chemists. 
 The carmine titration is also no ordinary colorimetric titration, but one 
 in which the quality, and not the density, of tint is estimated. We believe 
 that the bad results commonly obtained with "colorimeters" are due to 
 the two solutions being in some fixed position determined by the apparatus 
 used. An error of 10 per cent or more may easily occur from this cause. 
 Far more accurate results can be obtained with two ordinary test tubes 
 repeatedly transposed, as above described, than with complicated and 
 expensive colorimeters. 
 
 It will be found that the amount of carmine giving equality varies dis- 
 tinctly for different individuals. The proportional difference is, however, 
 the same at the two stages of the titration, so that the percentage result 
 obtained is the same. For the same individual the amount of carmine 
 needed varies, also, with different qualities of daylight, and is usually 
 less towards evening. This does not affect the percentage result, however, 
 provided that the two stages of the titration are completed by the same 
 light. 
 
 All these differences are due to the facts that the two solutions are 
 not spectrally identical; nor is the daylight at different times of day; 
 nor are the retinae of different persons equally sensitive to differences 
 
RESPIRATION 421 
 
 in any particular part of the spectrum; nor, finally, is any part of the 
 retina of one individual constant in its excitability for either white 
 light or colored light; the excitability of any one part being dependent 
 on side light falling on neighboring parts of the retina. The numerous 
 colorimeters, haemoglobinometers, etc., in which these sources of error 
 cannot be eliminated, are liable to very gross error, and appear to be 
 responsible for the discredit under which colorimetric methods suffer. 
 
 With ordinary artificial light the differences in tint between the various 
 solutions become almost invisible. The dimmest daylight is better than 
 ordinary artificial light. With blue spectacles, however, the differences 
 become very evident, and fairly good results can be obtained in the 
 titration if the carmine is made of the proper strength (very much 
 stronger) to suit the light. Daylight is, however, far better. 
 
 It is essential to accuracy with the carmine method that the carmine 
 solution should accurately match the standard blood solution in depth of 
 color. If the two do not correspond, it is easy enough to get a result: 
 for when the solution in one test tube is too deep in color it is only 
 necessary to incline the other in order to make its depth of tint appear 
 equal. The calculation of the percentage saturation becomes fallacious, 
 however, as is easily seen. One source of slight error in the titrations is 
 that a carmine solution which, when made up, exactly matches the blood 
 solution in depth, may, towards evening, be rather too strong, owing to 
 change in the light. This change can, however, be detected and rectified 
 very quickly, and attention would automatically be called to it by the 
 fact that considerably less carmine than before would suffice to produce 
 the tint of fully saturated blood solution. 
 
 A further source of possible fallacy depends on the liability of blood 
 solution to decomposition. It is essential that the blood should be fresh, 
 and diluted with clean water in a perfectly clean vessel. Solution which 
 has been kept more than a few hours is useless. It may show no methae- 
 moglobin band, and appear to be unaltered; but on saturating it with 
 CO it will probably no longer give the full pink color of undecomposed 
 haemoglobin, and its depth of color will also be found to be less than 
 before. It is thus mixed with colored decomposition products which make 
 it useless for titration. The tint on saturation with CO affords a far more 
 delicate index than spectroscopic examination of the freedom of a blood 
 solution from pigments other than haemoglobin. 
 
 When blood saturated, or partly saturated, with CO is diluted with 
 water, a small part of the CO must necessarily go into solution in the 
 water, as some dissociation of the CO-haemoglobin occurs. To demon- 
 strate this it is only necessary to saturate some blood with coal gas and 
 dilute some of it to 0.5 per cent with water. It will be seen at once that 
 
422 RESPIRATION 
 
 the diluted blood is distinctly less pink than some of the same solution re- 
 saturated with coal gas ; and on titration the blood which has been simply 
 diluted will be found to be not more than 88 or 89 per cent saturated. 
 The percentage dissociation can be calculated if we know the partial 
 pressure of CO corresponding to various percentage saturations of the 
 haemoglobin at room temperature, and also the coefficient of solubility 
 of CO. 
 
 In the case of human blood, half-saturation occurs at room tempera- 
 ture in presence of air with about .05 per cent of CO. Hence with 50 per 
 cent saturation of a blood solution saturated with air the partial pressure 
 of CO will be .05 per cent of an atmosphere. Now 100 cc. of water (and 
 presumably also of a very dilute blood solution) dissolves about 2.5 cc. 
 of CO from an atmosphere of pure CO at room temperature (i5C.), so 
 that at a partial pressure of .05 per cent it will dissolve 2.5 x .0005 
 .00125 cc. of CO, whereas 100 cc. of 0.5 per cent blood solution can take 
 up in chemical combination .0925 cc. of CO. Hence the proportion of the 
 haemoglobin dissociated is .00125 in .0925, or 1.35 per cent, so that if 
 50 per cent saturation were found by titration to be present we should 
 require to add 1.35 per cent to obtain the true result. By a similar 
 calculation we find that if the blood were found by titration to be 89 
 per cent saturated, we should have to add on u per cent in order to 
 obtain the true result, which would thus be 100 per cent. When human 
 blood is fully saturated with coal gas, the result actually found by titra- 
 tion, after dilution of the blood to 0.5 per cent, is 89 per cent, provided 
 the light is not bright. Hence the calculation agrees with the actual result. 
 In the brighter light of the middle of the day the result is, however, 2 or 
 3 per cent lower, even with a north light ; and on going outside so as to 
 increase the light, and avoid the absorption of actinic rays by window 
 glass, the result is still lower. This effect is due, as was pointed out by 
 Haldane and Lorrain Smith, to the action of actinic rays in dissociating 
 CO haemoglobin. The varying effect of light renders the carmine titration 
 with very high saturations of the blood with CO somewhat uncertain. 
 With low saturations, such as we have usually worked with, any error due 
 to this cause is trifling. We have at all times avoided bright light as far 
 as possible, and where it was necessary, as in the case of dissociation 
 curves, to titrate with high saturations of the blood, up to 80 or even 
 85 per cent, we have done the titrations by evening light. As an alterna- 
 tive, we might have used narrower test tubes and a greater concentration 
 of the blood solution, so as to diminish the correction for dissociation; 
 but it is easier to judge the tints accurately when ordinary test tubes are 
 employed, and comparatively few determinations were needed with very 
 high saturations. 
 
RESPIRATION 
 
 The following scale of corrections was used for human blood. 
 
 423 
 
 Observed -percentage 
 
 Correction 
 
 saturation 
 
 added 
 
 10 percent 
 
 0.15 
 
 20 " 
 
 0-35 
 
 30 
 
 0.6 
 
 40 
 
 0.9 
 
 50 
 
 1-35 
 
 60 " 
 
 2.0 
 
 70 " 
 
 3-i 
 
 80 " 
 
 5-4 
 
 89 
 
 II.O 
 
 For mouse blood the corrections used were 50 per cent higher, since 
 the partial pressure of CO required to produce a given saturation of the 
 blood with CO is about 50 per cent higher for mice than for men. 
 
 As already mentioned, the results of duplicate or triplicate titrations 
 of the same sample of blood agree very closely, the variation in the per- 
 centage saturation found hardly exceeding i per cent or 0.5 per cent 
 from the mean. When, as in determinations of arterial oxygen pressure, 
 two samples not differing much in percentage saturation are compared 
 successively with the same standard blood solution, the difference in their 
 percentage saturations with CO can be determined with corresponding 
 accuracy; for any errors due to imperfect preparation of the standard 
 solutions, or to the allowance for dissociation, will affect both results 
 equally. To determine the absolute range of the latter errors we made a 
 number of analyses of definite mixtures of normal blood with the same 
 blood saturated with coal gas. The coal gas contained about 7 per cent 
 of CO, and allowance was made for the small amount of CO present in 
 simple solution in the saturated blood. 
 
 The ox blood used for these mixtures was measured out from a pipette, 
 the blood being kept constantly stirred to prevent sedimentation of the 
 corpuscles. This method, though fairly accurate, is liable to slight 
 errors on account of variations in the quantity of blood which is left 
 adhering to the pipette. The following percentage saturations were ob- 
 tained on different occasions. The same carmine solution was used by 
 both observers. 
 
 In series (2) and (3) the mixtures were made with blood laked by 
 dilution to half, and were unknown to the observer. In (i) and (4) the 
 mixtures were made with whole blood, and were known to one observer. 
 In (4) each observer made up his own carmine solution. 
 
424 
 
 RESPIRATION 
 
 It will be seen that the maximum error was 2.0 per cent, this including 
 any error in making the blood mixtures and standardizing the carmine 
 solutions. With double determinations the error was considerably less. 
 
 Found. 
 
 (i) Calculated, 
 
 Found (3) Calculated (C 
 
 ..>.) 
 
 33-7 
 
 (33-9 (J.S.H.) 
 
 20.3 
 
 20.9 
 
 
 (34.0 (C.G.D.) 
 
 33-9 
 
 32.1 
 
 
 
 50.8 
 
 49.2 
 
 
 (76.0 (J. S. H.) 
 
 67.7 
 
 68.1 
 
 75-8 
 
 (76.8 (C.G.D.) 
 
 
 
 
 (75.2 (J.S.H.) 
 
 
 
 
 (4) 
 
 Found 
 
 
 (2) Calculated, 
 
 Found (/. S. H.*) Calculated 
 
 (C.G.D.-) (J. 
 
 S.H.) 
 
 25.4 
 
 26.8 1 1.2 
 
 10.3 
 
 11.4 
 
 33-9 
 
 338 25.2 
 
 26.0 
 
 27.1 
 
 50.8 
 
 52.3 50.5 
 
 5L9 
 
 52.5 
 
 677 
 
 68.3 80.8 
 
 79-9 
 
 81.4 
 
 E. Determination of Blood Volume in Man during Life by CO 
 
 Since CO is not oxidized or otherwise destroyed in the living body, 
 and since it forms a relatively very stable molecular compound with 
 haemoglobin, but with no other substance in the body, it is evident that 
 if we administer to an animal a known amount of CO, and then de- 
 termine the percentage saturation of the haemoglobin with CO and the 
 total CO capacity of a given volume of blood, we can determine the CO 
 capacity of the total blood in the body, and hence deduce also the blood 
 volume. The blood volume during life was first determined in this way 
 by Grehant and Quinquaud, 18 who used dogs for the purpose and em- 
 ployed the blood pump for the blood-gas analyses. In 1900 Lorrain 
 Smith and I introduced a much simpler method, easily applicable to 
 man; 19 and this method has been extensively used for physiological, 
 clinical, and pathological work, as mentioned in Chapter X. 
 
 The apparatus required for administering the CO to a man is shown 
 diagramatically in Figure 104. The subject breathes through a glass 
 mouthpiece A, the nose being clipped or held. The mouthpiece communi- 
 cates by ^-inch rubber tubing with a bladder or india-rubber bag B of 
 
 18 Grehant and Quinquaud, Journ. de I'anat. et de la -physwl., p. 564, 1882. 
 M Haldane and Lorrain Smith, Journ. of Physwl., XXV, p. 331, 1900. 
 
RESPIRATION 
 
 425 
 
 at least 2 liters capacity. Between the bag and mouthpiece there is inter- 
 posed a cylindrical metal vessel containing moist granulated soda lime 
 or other suitable absorbent to absorb CO 2 . The end of this vessel may be 
 made to screw on and off, with an air-tight rubber washer; or may be 
 made in two pieces, the outer of which slides over the inner, as shown 
 in the figure, the junction being made air tight with plasticine. The soda 
 
 Figure 104. 
 Apparatus for determining blood volume in man. 
 
 lime is kept in position by two circular pieces of wire gauze, one of which 
 is pushed into the end of the inner vessel, and the other into the end of 
 the outer vessel. Good soda lime can be made by stirring fresh slaked 
 lime in powder with a strong solution of caustic soda till the mixture 
 granulates, and then sifting off the fine powder and coarse lumps by 
 means of two sieves. Granulated caustic soda will also answer. There 
 should be no appreciable resistance to breathing, and one tin of soda 
 lime should last for several experiments. When the soda lime is spent it 
 ceases to heat, and the breathing begins to become increased, owing to 
 unabsorbed CO 2 . 
 
 The narrow graduated cylinder D is filled under water with CO, of 
 which a stock, prepared from formic and pure sulphuric acids, can be 
 kept in a large bottle. Just before the experiment, some of the CO is, by 
 turning the water tap E, driven out through the test tube and 3-way tap 
 F to the outside. In this way all the air is expelled up to the 3-way tap. 
 The water tap is then closed, and afterwards the 3-way tap. Oxygen from 
 a steel cylinder is now turned on through the tube C to displace CO 
 
426 RESPIRATION 
 
 from the tubing, which is then connected with the bag as shown in the 
 figure, and the bag filled pretty full with oxygen. Meanwhile the height 
 of the water in the cylinder is accurately read off, and the temperature 
 of the cylinder and barometric pressure noted. 
 
 The subject of the experiment now begins to breathe from the bag, 
 oxygen being supplied as required. The water tap is now slightly 
 opened, and the tap F turned so as to let CO as well as oxygen pass. The 
 required volume of CO is in this way very gradually driven in from the 
 measuring cylinder, about 20 cc. being passed in per minute. After the 
 CO has been passed in, the water tap is turned off, and the 3-way tap 
 turned so as to shut off the CO. The CO is absorbed from the bag very 
 rapidly and completely. The oxygen supply is continued for at least ten 
 minutes, after which the subject is allowed to absorb most of the oxygen 
 in the bag. About 1 5 minutes after the last of the CO has been given, a 
 drop or two of blood is taken and diluted for analysis by the carmine 
 method described above. At the same time the oxygen capacity of the 
 blood is determined in the ordinary way by the Gowers-Haldane haemo- 
 globinometer. For further certainty it is well to make both determina- 
 tions in duplicate. 
 
 As a little air always gets mixed with the CO, a sample of the CO in 
 the cylinder should be taken for analysis. It is usually sufficient to 
 determine the CO 2 (of which none should be present) and oxygen. From 
 the latter the proportion of air can be deduced. 
 
 Let us suppose that 150 cc. of CO were given, the temperature 12, 
 and the barometer 765 mm. ; also that there was 0.82 per cent of oxygen 
 in the CO, corresponding to 3.9 per cent of air. 150 cc. of gas saturated 
 with moisture would correspond to 142.5 cc. of dry gas at o and 760 mm. 
 But as 3.9 per cent of this was air, only 137 cc. of CO were administered. 
 Let us also suppose that the percentage oxygen capacity of the subject's 
 blood was 18.1 (98 per cent by the haemoglobinometer), and the per- 
 centage saturation with CO was 19.5. The total oxygen capacity or CO 
 
 capacity must have been 137 x =703 cc. ; the blood volume 703 x 
 
 i9-5 
 
 IOO 
 
 =3880 cc. If the subject's weight was 60 kilos this corresponds to 
 
 18.1 
 
 6.5 liters of blood to 100 kilos of body weight; and this result is usually 
 expressed as a blood volume of 6.5 per cent of the body weight. 
 
 In the original description of our method, we directed that the blood 
 sample should be taken within two or three minutes of the cessation of 
 administration of CO, as we assumed that by that time the CO would 
 be evenly distributed in the blood all over the body. The results from 
 samples taken three minutes after the first sample confirmed this as- 
 
RESPIRATION 427 
 
 sumption. When, however, Douglas and Boycott made a number of de- 
 terminations with a much larger bag which necessitated continuation of 
 the breathing for a considerable time after the CO had been given, they 
 obtained higher average results for the blood volume in man than 
 Lorrain Smith and I had got. Douglas and I therefore reinvestigated the 
 question as to how long the CO requires to distribute itself equally, and 
 found that when the samples were taken only two or three minutes after 
 cessation of the administration of CO the percentage saturations of the 
 blood were from 10 to 25 per cent higher than 15 minutes later. After 
 10 to 15 minutes, however, the saturation remained constant if the 
 subject continued to breathe from the bag. Our original experiments gave, 
 therefore, results for the blood volume which were too low probably by 
 about 25 per cent. The average blood volume in man by the CO method 
 is about 6.5 to 7 per cent of the body weight, and the total oxygen capacity 
 of the haemoglobin about i.i to 1.3 liters per 100 kilos of body weight. 
 
 It is probable that, as regards most of the circulating blood, mixture 
 with any added substance such as CO takes place very rapidly. In some 
 parts of the body, however, the circulation is so slow that a considerable 
 time is required for mixture. 
 
 Douglas, 20 and also Boycott and Douglas 21 applied the above de- 
 scribed CO method to animals, and took the opportunity of comparing the 
 results with those obtained by the older colorimetric method of Welcker, 
 which can only be applied after death. The series by Douglas showed an 
 average difference of 3 per cent, and that of Boycott and Douglas of 
 +5-5 per cent with the CO method as compared with the Welcker method. 
 It is evident, therefore, that no substance except haemoglobin combines 
 with CO. It must be remembered, however, that many of the muscles 
 contain some haemoglobin, and that by both methods this small fraction 
 of the total haemoglobin is estimated as if it belonged to the blood. 
 
 In using the CO method for human experiments it is necessary to 
 adjust the volume of CO administered to the patient's weight and prob- 
 able oxygen capacity, so that the percentage saturation of his haemo- 
 globin is not likely to rise above about 20; otherwise slight headache 
 may result. For persons of ordinary weight about 1 50 cc. of CO would 
 be suitable ; but in cases of pernicious anaemia or anaemia from loss of 
 blood, and in children or persons of low weight, far less CO should be 
 given. On the other hand in cases of polycythaema it may be necessary 
 to give 300 cc. or more in order to obtain a percentage saturation suffi- 
 cient for a satisfactory titration of the blood. As CO only leaves the 
 blood slowly when the percentage saturation is low, it is hardly neces- 
 sary, except in very exact experiments, to keep the patient breathing 
 from the bag after all the CO has been absorbed. 
 
 20 C. G. Douglas, Journ. Physiol., XXXIII, p. 493, 1906, and XL, p. 472, 1910. 
 n A. E. Boycott and C. G. Douglas, Journ. Path, and Bact., XIII, p. 256, 1909, 
 and A. E. Boycott, same Journal, XVI, p. 485, 1911. 
 

 THE UNIVERSITY LIBRARY 
 
 UNIVERSITY OF CALIFORNIA, SANTA CRUZ 
 
 SCIENCE LIBRARY 
 
 This book is due on the last DATE stamped below. 
 
 REC'D DEC 1 5 
 N.B.-HOLD 
 
 JUN9 1971 
 REC'D JUN 1 7 
 AUG 30 1972 
 
 APR 2? 1976 
 
 Rtli'U APR 2 8 
 MAY I 3 
 
 ^ 
 
 50m-4,'69(J7948s8)2477 
 
 AUG 2 1978 
 AUG i 1978 
 
 7 
 
 REC'D SEP 1 1 1978 
 NOV5 1978 
 
 REC'D OCT 2 5 1978 
 MAY 2 1980 
 
 MAY 2 5 1980 
 REC'D JUN 3 1980 
 
 JED FEB1? 1981 
 
 )EC 14*99 
 
 DEC 05 1993 
 
QP121.H26 Sci 
 
 3 2106 00259 4924