"

J-

O;

1

=O

-D
D~

en
m
o

a

THE RESPIRATORY FUNCTION
OF THE BLOOD

CAMBRIDGE UNIVERSITY PRESS

JLoillron : FETTER LANE, B.C.

C. F. CLAY, MANAGER

CEliinburgrj : 100, PRINCES STKKKT

ILonUon: H. K. LEWIS, 13<i. GOWER STREET, \\'.C.

WILLIAM WESLEY & SOX, 28, ESSEX STREET, STRAND

Berlin: A. ASHER AND CO.

lUtpjig: F. A. BROCKHAUS

£etn lotk: G. P. PUTNAM'S SONS

Sotnbnjj anft 9TaIrutta : MACMILLAN AND CO., LTD.

[Toronto: J. M. DEXT AXD SONS, LTD.
JTokuo: THE MARUZEX-KABl'SHIKI-KAISHA

ill rights r>' served

THE RESPIRATORY FUNCTION
OF THE BLOOD

>./

ll h

BY

JOSEPH BARCROFT, M.A., B.Sc., F.R.S.

FELLOW OF KING'S COLLEGE, CAMBRIDGE

CO
00

UJ

MARINE

BIOLOGICAL

LABORATORY

LIBRARY

WOODS HOLE, MASS.
W. H. 0. I.

CO

Cambridge :

at the University Press

1914

Camtmljgr

PRINTED BY JOHN CLAY, M.A.
AT THE UNIVERSITY PBE8S

TO THE

PROVOST AND FELLOWS OF KING'S COLLEGE

CAMBRIDGE

I DEDICATE THIS VOLUME

PREFACE

AT one time, which seems too long ago, most of my leisure was
*- spent in boats. In them I learned what little I know of
research, not of technique or of physiology, but of the qualities
essential to those who would venture beyond the visible horizon.

The story of my physiological " ventures " will be found in the
following pages. Sometimes I have sailed single handed, sometimes
I have been one of a crew, sometimes I have sent the ship's boat
on some expedition without me. Any merit which attaches to my
narrative lies in the fact that it is in some sense at first hand.
I have refrained from discussing subjects which I have not actually
touched, but which might fittingly have been included in a modern
account of the blood as a vehicle for oxygen. Such are the relation of
narcosis to oxygen-want and the properties of intracellular oxidative
enzymes. The omission of these and other important subjects has
made the choice of a title somewhat difficult. I should like to have
called the book, what it frankly is — a log; did not such a title
involve an air of flippancy quite out of place in the description of
the serious work of a man's life. I have therefore chosen a less
exact, though more comprehensive title.

After all, the pleasantest memories of a cruise are those of the
men with whom one has sailed. The debt which I owe to my
colleagues, whether older or younger than myself, will be evident
enough to any reader of the book. It leaves me well-nigh bankrupt
—a condition well-known to most sailors. But I owe another large
debt of gratitude to those who, as teachers, showed me the fascination
of physiology, to Dr Kimmins*, and especially to Dr Anderson f. At
a later stage I learned much from Dr Gaskell, Professor Langley and
Dr Haldane.

* Formerly science master at the Leys School now Chief Inspector of the
Educational Department of the London Comity Council.

t Formerly supervisor in physiology to King's College, now Master of Gonville
and Caius College.

viii Preface

There are occasions on which every sailor of the deep sea has to
ship a pilot. Mr A. V. Hill has brought me into those harbours
which are best approached through the, to me, unknown channels of
mathematics. T have to thank him and also Miss Dale for much
help with my proofs.

In the preparation of this book my acknowledgments are due to
many friends and others for allowing me to reproduce their photo-
graphs or illustrations — Mr C. G. Douglas (Figs. 115 — 119), Prof.
Durig(Figs. 127 a, 131 and 133), Dr Aggazzotti (Figs. 127 b and 132),
Dr Haldane and the Council of the Royal Society (Figs. 71, 72 and
73), Mr Hartridge and the Council of the Royal Society (Fig. 104),
Dr Lewis and the publishers of Heart (Fig. 105), Dr Krogh and
the publishers of the Skandinavisches Arch, fiir Physiologic (Figs.
98 and 99), Dr Rohde and the publishers of the Archiv fiir
experimentelle Pathologic mid Pharmakologie (Figs. 52 — 55),
Prof. Nuttall and the publishers of the Journal of Parasitology
(Fig. 2), and Dr Warburg (" Ueber Hemmung der Blausaurewirkung
in lebenden Zellen," H. S. Zeits. fiir pht/s. Chemie, Band 76, S. 331,
Strassburg, Karl J. Triibner), and to the Editor of the Journal
of Physiology for permitting the reproduction of many figures of
my own.

Lastly, my thanks are due to certain Corporations who have
given me grants for research — the Royal Society, the British
Association, the Fellows of King's College and the Commission for
the Study of the Biochemical Effects of High Climates and Solar
Radiation presided over by Prof. Pannwitz of Berlin.

J. B.

CAMBRIDGE,

December 1913.

CONTENTS
PART I

THE CHEMISTRY OF HAEMOGLOBIN

CHAP. PAGE

I. THE SPECIFIC OXYGEN CAPACITY OP BLOOD 1

II. THE DISSOCIATION CURVE OF HAEMOGLOBIN 15

III. THE EFFECT OF TEMPERATURE ON THE AFFINITY OF HAEMOGLOBIN

FOR OXYGEN . 27

IV. THE EFFECT OF ELECTROLYTES ON THE AFFINITY OF HAEMOGLOBIN

FOR OXYGEN 41

V. THE EFFECT OF ACID ON THE DISSOCIATION CURVE OF BLOOD . 53

PART II

THE PASSAGE OF OXYGEN TO AND FROM THE BLOOD

VI. THE CALL FOR OXYGEN BY THE TISSUES ... 73

VII. THE CALL FOR OXYGEN CONSIDERED AS A PHYSIOLOGICAL TEST . 107

VIII. THE METABOLISM OF THE BLOOD ITSELF .... 120
IX. THE REGULATION OF THE SUPPLY OF OXYGEN TO THE TISSUES 134

X. THE UNLOADING OF OXYGEN FROM THE BLOOD 157
XI. THE RATE OF EXCHANGE OF OXYGEN BETWEEN THE BLOOD AND THE

TISSUES . . . !?2

XII. THE ACQUISITION OF OXYGEN BY THE BLOOD IN THE LUNG . 182

XIII. THE ACQUISITION OF OXYGEN BY THE BLOOD IN THE LUNG (contimied) 194

x Contents

PART III

THE DISSOCIATION CURVE CONSIDERED AS AN "INDICATOR" OF
THE " REACTION " OF THE BLOOD

CHAP. PAGE

XIV. THE DISSOCIATION CURVE IN MAN . . . . . 218

XV. THE EFFECT OF DIET ON THE DISSOCIATION CURVE OF BLOOD . -I'll

XVI. THE EFFECT OF EXERCISE ON THE DISSOCIATION CURVE OF BLOOD . 236

XVII. THE EFFECT OF ALTITUDE ON THE DISSOCIATION CURVE OF THE

BLOOD . 243

XVIII. THE EFFECT OF ALTITUDE ON THAT OF EXERCISK . . 266

XIX. SOME CLINICAL ASPECTS . 282

APPENDIX I. ON METHODS 290

APPENDIX II. ON THE AGGREGATION OF HAEMOGLOBIN .... 314

APPENDIX III. ON THE HYDROGEN ION CONCENTRATIONS OF REDUCED BLOOD 316

APPENDIX IV. THE CONSTANT OF AMBARD 317

INDEX • 318

PLATE . ... to face p. 16

PART I

THE CHEMISTRY OF HAEMOGLOBIN
CHAPTER I

THE SPECIFIC OXYGEN CAPACITY OF BLOOD

WRITERS on historical subjects frequently allow themselves to

llllOU the

I'o put the matter in another way ; blood carries about 40 times
as much oxygen as the same volume of plasma. Therefore to convey
as much oxygen round the body as is carried by the blood, would in
the absence of haemoglobin demand 150 kilos of plasma, or perhaps
more. The contents of the vascular system would therefore amount
to twice the present weight of the body. The body would in short
be unable to cope with the weight of its own blood. The whole basis
of its economy therefore hinges upon the accidental possibility of the
occurrence and properties of haemoglobin.

Nor is there any chemical substance which exactly resembles
haemoglobin, though in some of the lower animals there are poor
imitations of it. But for its existence man might never have at-
tained any activity which the lobster does not possess, or had he
done so, it would have been with a body as minute as the fly's.

B. R. F. 1

x Content*

PART III

THE DISSOCIATION CURVE CONSIDERED AS AN "INDICATOR" OF
THE "REACTION" OF THE BLOOD

CHAP. PAGE

XIV. THE DISSOCIATION CURVE IN MAN . . . . .218

XV. THE EFFECT OF DIET ON THE DISSOCIATION CURVE OF BLOOD . L'liT

XVI. THE EFFECT OF EXERCISE ON THE DISSOCIATION CURVE OF BLOOD . 236

XVII. THE EFFECT OF ALTITUDE ON THE DISSOCIATION CURVE OF THE

RT.nnn

ERRATA.

p. 67 last line, for preparing read preferring.
p. 97 line 27, for Neumann read Neuman.
pp. 130 and 131, for Pikes read Pike's.
p. 259 et seq., for Chisholin read Chisolm.

?/ AV // Kxn

P- = 16, for - = - read =

p. 316 line 17, for Kx, read A'=10*.

PART T

THE CHEMISTRY OF HAEMOGLOBIN
CHAPTER I

THE SPECIFIC OXYGEN CAPACITY OF BLOOD

WRITERS on historical subjects frequently allow themselves to
speculate upon the momentous issues which follow from trivial
circumstances. A general stays to drink the health of his king in
some tavern, and the delay determines the result of a battle and
therewith also the reigning dynasty and thgJatutc religion. In like
manner I have sometimes indulged in the luxury of inquiring what
would have happened to the animal kingdom had it not been for
the accidental occurrence of the chemical substance haemoglobin.
Consider the respiration of muscle. It is among the most efficient of
machines. In warm-blooded animals muscle when working at its full
power uses up its own volume of oxygen in about ten minutes'1'. This
oxygen is carried to it by the haemoglobin of the blood — a substance
so rich in oxygen that a relatively small quantity of blood satisfies
the need of the muscle. Were it not for this red pigment some 200 c.c.
of fluid would have to be circulated through every gram of muscle in
ten minutes of time.

To put the matter in another way ; blood carries about 40 times
as much oxygen as the same volume of plasma. Therefore to convey
as much oxygen round the body as is carried by the blood, would in
the absence of haemoglobin demand 150 kilos of plasma, or perhaps
more. The contents of the vascular system would therefore amount
to twice the present weight of the body. The body would in short
be unable to cope with the weight of its own blood. The whole basis
of its economy therefore hinges upon the accidental possibility of the
occurrence and properties of haemoglobin.

Nor is there any chemical substance which exactly resembles
haemoglobin, though in some of the lower animals there are poor
imitations of it. But for its existence man might never have at-
tained any activity which the lobster does not possess, or had he
done so, it would have been with a body as minute as the fly's.

B. K. F. 1

2

Chapter I

Consider the question from the standpoint of evolution. If the
wing or the beak of a bird is too short for the highest efficiency it
can be modified in successive generations by a process of natural
selection until it attains the most suitable proportions. But in dealing
with chemical substances nature is set a very different task : either
they occur as stable compounds or they do not occur. Unless we are
to suppose that in times gone past there has been an infinite variety
of possible and impossible chemical combinations, we must suppose
that nature so far from adapting chemical substances to herself has
had to adapt herself to the chemical substances which exist. And
therefore I will adopt as a basis for the consideration of respiration
an investigation into the chemical nature of the substance haemo-
globin on which respiration, in the higher animals at least, depends.
I will begin therefore by considering the hard facts with which we
have to do in the chemistry and physics of the subject.

The first question which I would answer is, Is haemoglobin a
single substance ? In books on Chemistry it is usual to give con-
siderable prominence to the difference between a mixture and a
compound, and to prescribe obedience to the "law of combination
in definite proportions " as one of the conditions which a chemical
compound must satisfy. The table which is given below will show

Analysis of Oxy haemoglobin (2).

Animal

C

H

N

S

O

F

P

Author

Horse ..

54-75

6-98

17-35

0-42

20-12

0-38

Abderhalden

DOR ..

54-57

7-22

16-38

0-57

20-43

0-34

Jaquet

Cat

54-60

7-25

16-52

0-62

20-66

0-35

Abderhalden

Guinea-pig ...
Fowl

54-12

52-47

7-36
7-19

16-78
16-45

0-58
0-86

20-68
22-5

0-48
0-34

0-197

Hoppe-Seyler
Jaquet

Pie

54-17

7'38

16-23

0-66

21-36

0-43

Otto

54-60

7-25

17-43

0-48

19-60

0-40

Abderhalden

that this condition is very far from being fulfilled by haemoglobin,
for the discrepancies which appear between the various analyses are
too great to be accounted for by errors in the very accurate analytical
method of combustion.

On the other hand the uniformity of the results yielded by
accurate combustions depends upon the purity of the substance
which is being analysed. And the difficulty of obtaining haemo-
globin uncontaminated by other bodies, more especially by salts and

The specific oxygen capacity of blood

by water, is very great. We must however provisionally conclude that
haemoglobin differs in different species, and perhaps even in different
individuals. This difference may be explained, in part at all events,
by a difference in the globin portion of the molecule, which, on account
of its relatively great weight, plays a predominant part in the data of
chemical analysis, whilst it is but by-play in the study of haemoglobin
as a respiratory pigment. It would be enough for our purpose could
we show some constancy in the iron-containing portion of the molecule
and leave on one side the question of the exact composition of the
protein with which it is combined.

For this reason much laborious work has been done on the deter-
mination of the " specific oxygen capacity " of haemoglobin, that is,
on the number of cubic centimetres of loosely combined oxygen
which correspond to every gram of iron in the compound. Suppose
for instance that it appears, as the result of analysis, that 401 cubic
centimetres of oxygen correspond invariably to a gram of iron. It
would follow that haemoglobin obeyed the law of definite proportions
so far as oxygen and iron were concerned, and also that it obeyed
the law of combination in simple proportions. For, expressing this
ratio of the iron to the oxygen by weight, every 56 grams of iron
would correspond to 32 grams of oxygen. In other words, the
oxygen and the iron would be united in the proportion of one atom
of iron to two of oxygen.

The following table will however show that the history of the
subject provides us with but scanty hope of reaching this ideal (3).

Observer

Animal

Number
of Cases

Extreme
figures

Mean

Bohr

Dog

92

328 468

375

Tobiesen

Dog and Calf

17

378 429

388

Abrahamson

Ox

32

301 391

351

Pig

5

284 401

341

Bohr

Horse

9

379 426

411

Bornstein & Miillerf4) ...
Masing & Siebeck (5)

Cat
Man, Ox, Babbit

5

372—403

401
397

Butterfield(6)

Ox

3

389 — 395

397

Man

391

Man (diseased)

11

384 -409

399

The discrepancies between the various analyses amount in some
cases to one-third of the whole quantity of oxygen measured, and
some little consideration must be given to their interpretation.

1—2

4 Chapter I

Till recently there have been two schools of thought with regard
to the meaning of the figures given. Of these the first teaches that
the sources of analytical error are so great as to make more accurate
analysis impossible, whilst the second, represented chiefly by the late
Professor Bohr of Copenhagen, frankly admits that the want of uni-
formity is so great as to render untenable the idea of haemoglobin
as a simple substance. He explains the divergencies which we have
noted as being due to a mixture in different proportions of four
substances which he calls a, /3, 7 and S haemoglobins'7', each with a
different oxygen capacity from the others.

To the two schools mentioned above has now been added a third
which teaches that the combination of oxygen and haemoglobin is
not in the old-fashioned sense a chemical combination at all, but that
it is a manifestation of the physical phenomenon known as adsorption,
and that it therefore depends essentially on the surface conditions of
the molecules of oxygen and haemoglobin respectively. The pro-
perties of these surfaces may presumably be altered by all sorts of
variations in the collateral substances which are present in the
solution of the uniting molecules. The amount and nature of the
salts present for instance might be supposed to alter the charges on
the molecules, and in so doing to affect the amount of oxygen with
which a given quantity of haemoglobin would unite.

Within recent years, partly on account of the improvements in the
analytical methods both for oxygen and iron and partly on account
of the increased importance of the subject, it has become more and
more desirable that some re-investigation by direct methods of
this specific oxygen capacity of haemoglobin should take place. I
mean by methods in which the oxygen is measured as such, and the
iron as a salt of the metal, as opposed to the indirect spectro-photo-
metric methods which have given uniform, and apparently excellent
results in the hands of Butterfield (6). This investigation has lately been
undertaken by Peters (8) ; for the purpose of estimating the oxygen he
has used the differential method of blood analysis based upon the
observation of Haldane (9) that oxygen is eliminated from haemoglobin
quantitatively by potassium ferricyanide. The theoretical accuracy of
the method has been confirmed by the researches of Professor Franz
Miiller (10) in Berlin, while its practical details have been so far simpli-
fied that some half-dozen analyses can easily be performed with as
many cubic centimetres of blood in two hours. Peters therefore has
been at a great advantage as compared with his predecessors, whose

The specific oxygen capacity of blood .">

individual analyses, if theoretically somewhat more accurate, extended
over two or three days and entailed the use of large quantities of
haemoglobin. It was impossible in their case to obtain large numbers
of analyses which could be averaged and from which the errors could
be eliminated to some extent by statistical treatment.

In estimating the iron by titration with titanium advantage has
been taken of new methods of analysis which are much simpler and
more accurate than the older permanganate titrations.

The theory of the titanium method of estimating iron is repre-
sented by the following equation :

TiCl, + FeCl3 = TiCl4 + FeCl2.

The method has this great advantage over that of titration with
potassium permanganate, that it is not vitiated by the presence of
chlorides.

In practice defibrinated blood was centrifugalised, the corpuscles
washed twice in isotonic salt solution, and as much as possible of
the fluid was got rid of. The cream of corpuscles was laked by
the addition of twice its volume of dilute ammonia (4 c.c. of strong
ammonia per litre). This solution, which we shall call solution A,
was centrifugalised again to rid it of any corpuscles which did
not lake and of other debris. Portions of it were then measured
out from the same burette, both for the iron estimations and the
oxygen analyses. For the former 50 c.c. of the solution was evaporated
in a platinum crucible and carefully "ashed." It was at this point that
the nicety of the determinations really entered if accurate results were
to be obtained. The ashing must take place at a temperature which
is neither too hot nor too cold. It is best carried out in a " muffle
furnace." To quote Peters, " If the carbon is not completely burned
away iron will still remain which cannot be removed by boiling with
acids ; whereas on the other hand if the ash is heated to a high
temperature in removing the carbon the iron becomes insoluble. In
both cases the result will be a loss of iron." After the ashing is
complete the iron is dissolved up in strong hydrochloric acid, and a
trace of hydrogen peroxide is added to insure the complete oxidation
of the iron. The excess of hydrogen peroxide is subsequently boiled
off and the titration with titanous chloride is made, potassium sulpho-
cyanide being used as an indicator.

Some idea of the scale on which Peters' experiments were con-
ducted may be gleaned from an account of a single experiment. In
addition to making two or three analyses of iron such as have just been

6 Chapter I

described, he made from the same solution, run from the same burette,
sixteen oxygen estimations. These were divided into four groups :
the average of each group was taken, and the mean of these four
averages was taken. The greatest error which entered into Peters'
blood-gas analysis was doubtless in the measurement of 3 c.c. of
fluid from an ordinary 50 c.c. burette, the meniscus in the case of
haemoglobin solution being none too well defined. This error, serious
though it appears, is discounted by the fact that the 16 samples for
analysis were run consecutively out of the burette. There may be
a certain ruggedness in the individual figures, but the averages of the
groups of four are very close to one another, since a positive error
in one sample entails an equal negative error in the next : in the
aggregate 48 c.c. used there is no appreciable error as compared
with the 50 c.c. used for the iron analysis.

Peters' figures for the oxygen of a single experiment are as
follows :

Oxygen in 3 c.c. of solution A in c.c.

Group (1) Group (2) Group (3) Group (4)

•3667 -3667 -3576 -3527

•3449 -3374 -3640 -3510

•3482. -3439 -3638 -3614

•3455 -3517 '3455 -3537

Total 1-4053 1-3997 1-4309 1-4188
Mean -3513 -3499 -3577 -3547

Taking the average of these four means we get

•3518
•3499
•3577
•3547

Total 1-4136
Mean -3534 c.c.

When it is remembered that the whole of these operations, both iron
analyses and gas analyses, could be carried through successfully in a
day, it will be clear what an advance Peters has made by the use of
the modern technique both as regards the concordance of his figures
and the certainty with which he has been able to put them forward.
His figures for the volume of oxygen per gram of iron are as follows :

Ox 394, 401, 399

Sheep 387, 384

Pig 388

Dog 384

Average 391

The specific oxygen capacity of Mood 7

It is not very easy to discern the processes by which conviction
grows in the mind ; probably the mere inspection of the figures given
is sufficient to convince the reader that so far as the relation of the
respiratory oxygen to the iron of haemoglobin is concerned, these
quantities are related in the proportions of two atoms of oxygen to
one of iron. To me, who had the privilege of seeing Peters work
from week to week, conviction came in a slightly different way; it
developed as it were like the image on a photographic plate. As one
experimental difficulty after another was overcome, as one source of
error after another was weeded out, as the worker himself developed
in skill and in capacity, just so surely did the results which he ob-
tained approach the theoretical figure with greater certainty till at
the end when all the difficulties had been overcome and when Peters
himself had attained to the rank of a first-rate exponent of the
technique, I arrived at a stage of conviction in which I never
doubted, when he undertook an experiment that the result would be
between 385 and 405. Perhaps there could be no surer proof that
all thought of the wide differences between different kinds of haemo-
globin, alleged to exist by Bohr and others, had passed out of our
horizon, than the fact of our almost laughable concern at the end of
the work as to why the average figure was .391 and not 401. We, in
the laboratory, thought perhaps that Peters did not perform sufficient
experiments to obtain a true average, or that some trace of
methaemoglobin was always present or, most probably, that some
trifling error had crept into standardisation of the apparatus used.

The probability of the last source of error at least seemed
sufficiently great to warrant the initiation of a fresh research on
the subject, which was undertaken by Burn. Moreover on quite
general grounds it seemed desirable to undertake something of the
sort, for independence of the fallacies of a single experimental
procedure is of the essence of all sound experimental work.

The problem was to find a method of calibrating the differential
blood-gas apparatus (12' in such a way that the possible errors involved
in the method which Peters had used would be of a different character
from those involved in the new method.

The details of the differential apparatus will be found in the
Appendix ; the form used by Peters is shown in Fig. 1. All we need
say here is that the oxygen is liberated from the haemoglobin in one
of its two bottles. The pressures in these become unequal, and the
difference of pressure is indicated by the movement of the fluid in the

8

Chapter I

manometer. The measurement consists in reading the difference in
the level of the fluid surface on each side ; the problem is to turn
this difference of level read in mm. of clove oil into a measurement of
quantity in cubic centimetres of oxygen which has been liberated.

FIG. 1. — Form of differential apparatus used in Peters' and Burn's researches.

It may be shown mathematically that if x be the difference in
level of the fluid surfaces, V the volume of air in each bottle, a the
sectional area of the tubing, p the barometric pressure in mm. of
clove oil, and q the quantity of oxygen liberated, all the quantities
being expressed in mm.,

'-
»P

Now the quantities V, p and a are measurable, and V and a
may be measured once for all. It was by this method that Peters
calculated the relation of q to x in his experiment. There is no

The specific oxygen capacity of blood 9

doubt of course as to the theoretical accuracy of the calculation, but
it is evidently not impossible that when the greatest accuracy is
concerned there may be weak points about its application. We are
using fluids for instance which leave films on the glass and exert
vapour pressures and so forth, and it may be that the trifling-
constant errors involved might in the aggregate amount to one or
two per cent.

Burn therefore went back to quite another method of relating q
to x, which was tried, though without complete success, by Roberts
and myself. He first defined the conditions under which it might be
relied upon, after which he obtained with it results of remarkable
consistency. He liberated a known quantity of oxygen in the
apparatus and read off the difference in level; thus knowing both
q and x he found the constant K by which x must be multiplied to

(V
give q, and which should of course be equal to ( — \-a\.

The known quantity of oxygen was liberated from a known volume
of standardised hydrogen peroxide by potassium permanganate accord-
ing to the following equation :

K2Mn208 + 3H2SO4 = K2S04 + 2MnS04 + 3H20 + 50,
5O + 5H2O2 = 5H20 + 502.

So far as manipulative details are concerned the oxygen is libe-
rated in just the same way as it is from a haemoglobin solution by
ferricyanide. As the sources of manipulative error are the same in
each case, the estimation of K by the H202 method should give a
correct value of K for the ferricyanide method. The validity of the
method depends upon the assumption that the calculated quantity
of oxygen is released.

The following is the result of one series of experiments in which
Burn determined K for a certain apparatus, No. 19, by the two
methods.

K determined by H202

3-87

3-87

3-87

3-78

3-87

3-82

3-78

3-82

3-78

3-82

3-81

3-89

3-78

A" determined by measurement
of V, p, and a

3-72

Mean 3-83

The extreme results of the H2O3 method differ from one another
by 3 "/..

10 Chapter I

We may now tabulate the results of three of Burn's experiments.

Series

Apparatus

No. of
Exps.

Extreme results
by H202 method

A
Mean K by
H202

B

Mean A' by
formula V\p + a

A-B %

I

23

13

3-55—3-68

3-62

3-53

2-6

II

19

13

3-78—3-89

3-83

3-72

2-9

III

24

7

3-50—3-61

3-56

3-48

2-2

Mean 2-5

The chemical method of calibration therefore gives to K quite
constantly a somewhat higher value, for the different apparatuses
which were tried, than does the physical method. The mean differ-
ence is 2'5 %•

Let us now look at Peters' results in the light of this newly
acquired information. Had the constants of his apparatus (and the
apparatuses alluded to in the above table were among those used
by Peters) been obtained by the chemical method, his figures for the
specific oxygen capacity would all have come out 2%5 °/0 higher or
thereabouts, and would therefore have been as follows :

Specific oxygen capacity as
given by Peters

Ox

.. 394 401 399

404

Cat

.. 395

405

Sheep

387 384

397

Pig

388

398

Dog .

. 384

394

Average .

. 391

X.

401

Specific oxygen capacity

recalculated by chemical method

of calibrating apparatus

411 409
394

Theoretical figure 401

As a third method of calibrating the apparatus has been worked
out by Homiann which gives results identical with that of Burn (see
Journal of Physiology, vol. XLVII) it seems probable that Peters'
values as corrected by Burn are more trustworthy than his un-
corrected ones.

The figures arrived at by the chemical method of calibration give
an average which is so near to 400'8 that there can be no doubt
that the oxygen and iron are united in haemoglobin in the ratio of
two atoms of the former to one of the latter.

The propositions, therefore, (a) that the bloods of different animals
have fundamentally different oxygen capacities, and (6) that analyses
furnish any serious reason to doubt the validity of the oxygen to iron
being related as two atoms to one, have ceased to trouble my mind.

The specific oxygen capacity of blood 11

The above experiments were all made upon the blood of animals
in good health, or at all events not known to be in bad health.

Now that the physiological problem may be regarded as settled
the pathological ones remain, and become the more ripe for settle-
ment. These offer quite distinct subjects from that which we have
discussed, inasmuch as there may be intermediate bodies present
which are being worked up into haemoglobin in the blood, or degene-
rate substances such as methaemoglobin. One aspect of these
problems resolves itself into an investigation of whether the oxygen
capacity and the colour go hand in hand in the case of anaemic
patients — a most important consideration for the only real use of
the haemoglobinometer is as an index of the oxygen carrying power
of the blood. According to Morawitz (13) they do.

I will not enter upon a discussion of what really can only be
settled by accurate analysis.

In the meantime it is clear that in doubtful cases the haemo-
globinometer may be put on one side and the oxygen capacity of the
blood measured directly. The following is a case in which this has
been done. The oxygen capacity of the blood was observed systemic-
ally throughout the case by actual oxygen determinations, each of
which was carried out upon one-tenth of a cubic centimetre — about
two drops — of blood, with results which inspire complete confidence.
This direct procedure obviated all assumptions as regards specific
oxygen capacity of pathological blood. The case was one of so-called
biliary fever in the horse investigated by Nuttall and Strickland (U) in
the Quick Laboratory at Cambridge. The object was amongst other
things to determine the oxygen capacity of the blood whilst it was
suffering from the ravages of the blood parasite Nuttallia equi.
This parasite makes its home in the red blood corpuscles. The
stages of its development are shown in Fig. 2, the description of
which I quote from Nuttall and Strickland's paper.

"N. equi multiplies slowly and in the following manner :

(1) The minute piriform or oat-shaped parasite enters a fresh
corpuscle and (Fig. 2 ; 2, 3, 4, 5) grows in size, being slightly
amoeboid, with a general tendency to resume a pear shape. Definitely
amoeboid movements (6) are, however, only to be seen distinctly when
the parasite has attained a certain size. Judging from the form of the
chromatin masses, stages 7, 8, 9, 10 follow next. The rest of the cycle
has been continuously observed in the living parasite : the formation
and breaking-up of the cross-form, the scattering of the daughter
cells within the corpuscle, and their escape from the corpuscle — "

12

Chapter I

The liberation of the parasites is associated with a great breaking
down of the corpuscles. Clearly therefore the oxygen capacity of
the blood cannot be judged from corpuscle counts, firstly, because
of the large amount of haemoglobin in the plasma and, secondly,
because there is no assumption that the fresh corpuscles formed

stowing the usual mode

of multiplication
4 of

Nuttallia equi

in tbe circulating blood

FIG. 2. — Illustrating the usual cycle of development of N. equi in the circulating blood.

under the stress of such a disease contain the normal quantity of
haemoglobin. The oxygen capacity must be tested, if at all, either
by the direct test of oxygen measurement, or by the haemoglobino-
meter, the use of which depends upon the assumption of the very
point which we have declined above to discuss without further data.
I append the history of the case and its chart.

The specific oxygen capacity of blood 13

Nuttallia equi Horse III.

The Horse was inoculated subcutaneously with 20 c.c. of defibrinated blood taken from the jugular
vein of Horse II at mid-day on 2. vin. 1910. Horse III had previously been infected with Piroplasma
caballi on 22. vi. 1910, but had recovered (see Protocol of P. caballi Horse II). N. equi appeared in the
horse's blood on the 10th day, and it died at 5.30 p.m. on the 20th day after inoculation (21. vni. 1910).

Temp. °F.

Day M. E/

1 97-8 Inoculated.

2 98-2

3 98-2

4 97'2 Oxygen Leuco-

5 98-2 — capacity cytes

R.b.c ofhaemo- per

6 99 "0 — per c.mm. globin c.mm.*

7 98-2 — 8,888,000 0'227 8,640

8 98-2 8,872,000 0'230 12,900

9 97-4 8,968,000 0'222 7,120

10 100-4 — Parasites appeared 9,155,000 0'240 7,800

% r.b.c.

infected (S) (L) (D) (2—4) F

10 102-0 0-2 34 65 1 • •

11 103-0 3-0 23 75 1 • 0'4 9,004,000 10,650

104-2 2-0 43 50 4 2 • Horse weak.

12 104-2 Horse weak, jaundice; off 7,240,000 0-230 9,540

106-6 11-6 23 70 3 41 his food.

13 105-0 10-3 47 45 0'6 2 5 Very weak; haemoglobin- 5,390,000

104-6 6-8 50 43 2 32 uria; blood appears watery,

coagulation retarded.

14 103-0 8-4 56 38 0'2 4 1 Bather better; feeding better, 5,365,000 0'141 12,000

101-7 10-6 62 33 2 21 haemoglobinuria.

15 100-6 5-2 31 65 1 3 0-4 Ditto; urine smoky. 4,812,000 0'104 7,520

102-4 7-4 28 66 0'6 5 0-6

16 102-4 9-2 15 80 0-6 3 0'2 Very weak, especially hind 3,840,000 0-115 9,700

100-8 6-8 35 57 1 41 legs, which are stretched

apart.

17 103-0 8-0 19 75 1 3 0'2 Seems better, feeding; swell- 3,848,000 0'104 13,500

103-0 12-8 28 67 2 3 • ing on jaw (abscess).

18 103-0 8-6 25 69 1 4 0'3 Weak ; jaundice persists. 3,230,000 O'lOO 22,600

102-2 13-2 20 73 2 5 •

19 103-2 6-6 23 74 1 2 • Very weak, lying down. 3,120,000 0-066 33,600

101-0 4-8 20 75 1 4 • <-—

20 98-4 8-4 29 53 2 11 6 Helpless, urine smoky; leading™

found dead at 5.30 p.m. SS^of3

1 c.c. of
* Average number of counts = 5. blood in c.c.

The signs denote corpuscles containing

(S) = small and medium sized piriform or oval parasites.

(L) = large rounded parasites.

(D) = dividing and cross-forms.

(2-4) = two or more parasites.

F = free forms.

14

Chapter I

FIG. 3.— Chart of Nuttallia eifiti Horse III.

REFERENCES

(1) Chauveau and Kaufmann, Comptes Rendus de VAcad. Franc, cm, 1063.

(2) Miiller, Oppenheimeijs Handbuch der Biochemie, p. 672, 1908.

(3) For discussion of literature see Peters, Journal of Physiol. XLIV, p. 134, 1912.

(4) Archie fiir Anat. u. Physiol. 1907, p. 470.

(5) Deut. Arch, fiir Klin. Med. xcix, p. 130, 1910. Masig, Ibid, xcvm, p. 123, 1909.

(6) Zeitsch. f. physiol. Chem. XLII, p. 143, 1909.

(7) Bohr, Skand. Arch, m, p. 76.

(8) Peters, Journal of Physiol. XLIV, p. 131, 1912.

(9) Haldane, Ibid, xxn, p. 298, 1898 ; xxv, p. 295, 1900.

(10) Miiller, P ft tiger's Archiv, cm, p. 541, 1904.

(11) Barcroft and Burn, Journal of Physiol. XLV, p. 493, 1913.

(12) Barcroft and Roberts, Ibid, xxxix, p. 429, 1910.

(13) Morawitz and Itami, Detitsch. Arch.f. Klin. Med. c, p. 191.

(14) Nuttall and Strickland, Journal of Parasitology, v, p. 65, 1912.

CHAPTER II

THE DISSOCIATION CURVE OF HAEMOGLOBIN

IF a haemoglobin solution be shaken up with oxygen, the haemo-
globin unites with a definite quantity of the gas, which is in the
proportion of 32 grams of oxygen to each 56 grams of iron in the
haemoglobin. However much stress we may lay upon this fact — and
we cannot lay too much stress upon it — the most elementary con-
sideration of the haemoglobin in the circulation reveals the fact
that it is always united with less oxygen than the total amount
possible, frequently with much less and in any case with no precise
or invariable amount.

The next step therefore, if we are to regard oxyhaemoglobin as a
chemical compound, is to inquire whether we can reconcile the fact
that haemoglobin in the body unites now with more, now with less
oxygen, with the known laws of chemical action.

The most obvious law which might illuminate this problem is the
law of mass action. Our inquiry therefore resolves itself into this :
granted a solution which contains (1) oxygen, (2) oxyhaemoglobin,
and (3) reduced haemoglobin, does the amount of oxyhaemoglobin
depend upon the concentration of oxygen in the solution ?

The answer to this question can only be supplied by experiment.
The experiment is not a difficult one. It is easy to obtain solutions of
haemoglobin containing known concentrations of oxygen in solution,
for the concentration of oxygen depends directly upon the oxygen
pressure of the atmosphere with which the solution is in equilibrium.
If a be the volume of oxygen which is dissolved in 1 c.c. of the
solution at the temperature of the experiment and at 760 mm.
pressure, then the concentration of oxygen at any other pressure

p is p j—: . The experiment then will consist of exposing portions
of a haemoglobin solution to various atmospheres containing known

16

Chapter II

pressures of oxygen and subsequently determining the amounts of
oxy- and reduced haemoglobin in each sample after an equilibrium
has been established between the haemoglobin and the atmosphere.

Suppose we have five closed vessels each containing a small
quantity of haemoglobin solution and also at the same time oxygen
at the following pressures, namely 0, 10, 20, 40 and 100 mm. of

10

20

40

100

FIG. 4. — The numbers denote the pressure of oxygen in mm.

mercury. After the fluids had been shaken up thoroughly at 38° C.
the concentrations of oxygen would be

(1) (2) (3) (4) (5)

10

760

20

That is :

0 -00029

of oxygen in each c.c. of fluid.

760
00058

40;

•00116

•0029 c.c.

Now we must find out what proportion of the haemoglobin is
oxyhaemoglobin, and the following are figures such as we would
obtain :

Vessel

(1)
0°/0

(2)

55%

(3)
72°/0

(4)

(5)
92%

The best idea we can get of the relation of these numbers to one
another is to place the following picture before our eyes. Suppose
the haemoglobin in each case to be in a cylindrical tube and that
the oxy- and reduced haemoglobin could be separated from one
another, the former being red and sinking to the bottom and the
latter purple and rising to the top, wre should obtain five cylinders
as shown in Fig. 5 corresponding to the oxygen pressures in the five
tonometers.

B

•e

5

to cj

o"

§ 0-

•i *"

10

* i

<^s

S

O CNj

*53

c

03

a:

Q^

o

d

•e

o

to

o

•5

-5

*

^

o

.

"&»

,

S

o

^

c

4>

.

hs.

-S

_o

a:

g£

,s

^

>0

-o
o

o

o

&s

^

~

.s

0

cs

"^

•o

s

•^

o

as

o

53

O

s

?

Q

O

X

o

to

0

d

ci

03

10

CN

to

to
3

•5

a:

Q..

to

-^

si.

to

O

^

O

ts

c

<S

o

"Si

o

o

Q.

to

O

N!

5

Q.

«

"to

10

-S

to

^

"5.

•5

-5
.0.

0

5

%

o

5

53

-e
Si

H

0

-s:

O

Si

O

I
Q

I

O

10 20

40

100

FIG. 5.

Percentage saturation
with oxygen

100

FIG. 6.

mm. pressure

0 10 20 30 40 50 60 70 80 90 100

Oxygen pressure
FIG. 7. mm.

FIG. 5. — Percentage saturation of haemoglobin with oxygen at 37° C. corresponding to
oxygen pressures of 0, 10, 20, 40 and 100 mm. of oxygen respectively.

FIG. 6. — The same spaced with the oxygen pressure as the abscissa.

FIG. 7. — Dissociation curve representing the equilibrium between oxygen, oxyhaemoglobin
(red) and reduced haemoglobin (purple).

Tlte dissociation curve of haemoglobin 17

We are impelled to ask whether any definite relationship exists
between these quantities of oxyhaemoglobin and the oxygen pressures
to which they correspond. In doing so we must bear in mind that
we shall shortly be face to face with a very serious position, for the
very relationship which we are about to investigate is defined by
the law of mass action itself, and should the relation as found not
agree with that as prescribed, our theory must be abandoned and
some other explanation of the properties of haemoglobin must be
found. Let us introduce the element of pressure quantitatively by
spacing the cylinders apart from one another at distances which are
proportional to the concentrations (or pressures) of oxygen dissolved
in the solution, as is done in Fig. 6.

By joining the points which divide the blue and red portions of
the cylinders we obtain a curve which relates (1) the percentage of
the total haemoglobin which is oxyhaemoglobin to (2) the concentra-
tion of oxygen dissolved in the fluid at all concentrations of oxygen
between 0 and '0029 c.c., and consequently to all pressures of oxygen
between 0 and 100 mm. This curve representing the equilibrium
between oxygen and haemoglobin is called the dissociation curve of
oxyhaemoglobin. It is shown in Fig. 7.

Let us now turn from the observed properties of haemoglobin to
the other side of the question, namely the requirements of the law of
mass action. This law has been stated quantitatively by Guldberg
and Waage in the following terms :

" The velocity of chemical change is proportional to the product of
the active masses of the reacting substances." The chemical change
is conceived of as taking place in both directions simultaneously,
that is to say, oxyhaemoglobin is all the time being formed and
being broken up, and we therefore have these two changes taking-
place at the same time, (1) the formation, (2) the breakdown
of oxyhaemoglobin. Since these changes balance one another, the
whole system being in equilibrium, the velocities of the two changes
are equal. Taking first the formation of oxyhaemoglobin, the re-
acting substances are reduced haemoglobin and oxygen and the
velocity of their reaction is proportional, says the law, to the product
of their concentrations in the solution. If C0 be the concentration
of oxygen and CR of reduced haemoglobin, then the velocity of the
reaction is proportional to, or is equal to, a constant k multiplied by
the product of CR and C0, i.e. k (CR x C0). As regards the other
phase of the reaction — the breakdown of oxyhaemoglobin — there

B. R. F.

18

Chapter II

is but one active substance, namely oxyhaemoglobin ; let its con-
centration be Ca, then the velocity of the reaction is k', another
constant, multiplied by CH- Taking the two together, we have

The concentration of oxygen as we have already stated is p x — - ,

where p is the oxygen pressure to which the solution is subjected
and a the coefficient of solubility, therefore

replacing all the constants by one constant K,

KpCR = CH.

FIG. 8. — Rectangular hyperbola. The products of x and ?/, such as x1y1, x<,y.2 for

all points are equal.

If the concentrations of oxy- and reduced haemoglobin are re-
garded as percentages of the total concentration of haemoglobin,
then Ca =-- 100 - CR,

KpCR=IOO-CR,

CR 100

100

The dissociation curve of haemoglobin

19

If we write y for CR and x for ( p + ^ j , we obtain the relationship

xy = & constant, namely -~- . If we now construct a curve plotting y

vertically and x horizontally, and observing the condition that for
any point the product of x and y must be constant, we obtain a
curve known as a rectangular hyperbola (Fig. 8).

This curve does not at first sight relate the concentration of
reduced haemoglobin (€R) to the pressure p, but to something else,

namely p + -^ . It remains for us to find out what is the relation

A

Of CR tO p.

B'

O

100 .

A B

FIG. 9.

Consider the special case in which CR = 100, i.e. all the haemo-
globin is reduced haemoglobin,

^ 100
JR~ K '

The distance AB (Fig. 9) in that case = ™ Hence the distance

from BB' of any point on the hyperbola is a measure of the pressure.

We have therefore derived the following information from the law

of mass action. If the reaction is a chemical one involving single

2—2

20

Chapter II

molecules only, the relation of (1) the oxygen pressure to (2) the per-
centage of oxy- to total haemoglobin is capable of being represented
as a rectangular hyperbola, the origin of which is at once the point at
which there is no pressure of oxygen and no oxyhaemoglobin, and
the curve approximates to a line representing complete saturation.
If now we turn back to Fig. 7 we shall see that the curve which we
have shown as representing the relation of the pressure of oxygen
and the percentage of oxyhaemoglobin is identical with the curve we
have just given, and therefore the law of mass action is to this extent
satisfied.

Yet this satisfactory result was not reached without a struggle.
The ease with which it can be demonstrated at the present time
presents a very pleasant contrast to the tiresomeness of the path

FIG. 10. — Shows a series of rectangular hyperbolae, each with A as its origin,

and approximating to O.Y.

which had to be trod before the present point of vantage was
reached.

The history of the subject forms an interesting commentary upon
the psychology of research. The law of mass action was first quanti-
tatively applied to the reaction

by Hiifner'1', who quite unjustifiably assumed the applicability of
the law to the reaction in the form in which we have given it above
and obtained a curve very similar to the one which is represented
from entirely theoretical considerations.

The dissociation curve of haemoglobin

21

You can have any number of rectangular hyperbolae all of which
pass through the point A and approximate to OX, such as are
shown in Fig. 10: these all satisfy the condition stated above

n, 100

The difference between them lies in the value of K. Now Hiifner
assumed the correctness of the equation and set out to find the value
of K. This can be done from one point. He used a number of
samples of haemoglobin prepared in different ways, determined a
point for each, found the value of K, averaged these values and
produced a curve. But a nemesis awaited Hiifner. His speculations

80

70

60

30

20

10H

10 20 30 40 50 60 70 80 90 100

FIG. 11. — Dissociation curves of haemoglobin. H according to Hiifuer.

B according to Bohr.

fell in the hands of a physiologist of a diametrically opposite school.
Bohr (2) had inherited a tradition from the great laboratory of Ludwig
which, though it may carry its holders to excessive lengths, at least
forms a useful corrective to unjustifiable generalisations. Bohr's
motto was that every experiment had a value, nothing which was
obtained as the result of a test in the laboratory was set aside on
the ground of its inherent unlikelihood, of its failure to fit general
principles. Bohr therefore determined to map out the curve relating
the pressure of oxygen to the relative quantities of oxy- and reduced

22 Chapter II

haemoglobin point by point, irrespective of laws, and to find out
experimentally what it actually was like.

Fig. 11 shows that the actual curve determined point by point
differs fundamentally from Hiifner's hyperbola. The two cross,
therefore Bohr's cannot be a hyperbola, and further it has a
double contour and therefore is not a curve of the same order
at all.

At the present time it is scarcely necessary to dwell on the theory
which Bohr propounded to explain his curve. In two words it was
that when oxygen broke from oxyhaemoglobin, the haemoglobin
itself split into globin, and something rather like haematin which
he called haemochrome. Another explanation of this curve will be
alluded to in another place. The whole question took on a new
aspect when it became clear, as will appear hereafter, that the
affinity of haemoglobin for oxygen is profoundly influenced by the
nature of electrolytes (3) in the solution.

The bearing of this discovery was at once grasped by Ff. Roberts (4)
who suggested making a solution after the method of Bohr* and
then dialysing a portion of the solution to get rid of such residual
salts and traces of ether as might be in it ; then comparing the
dissociation curves obtained from the dialysed and undialysed
portions.

The experiment was performed on a solution obtained after
three days' dialysis in an aseptic dialyser; it was free from the
characteristic smell which always clung to our preparation of dog's
haemoglobin made by Bohr's method, it was quite neutral in
reaction ; Hardy kindly undertook to test its freedom from salts,
and showed that in saline concentration it was equivalent to a
•004 N solution of sodium chloride. The points on the dissociation
curves of the two solutions were then determined, and are shown in
Fig. 12. The round points are those of the dialysed solution, the
square ones are those of the undialysed solution. It is at once
apparent that the latter are in very close agreement with the curve
published by Bohr (denoted by II in the diagram), whilst the former
fall so nearly on the rectangular hyperbola (denoted by I) as to make
it in the highest degree probable that could the salts be entirely elimi-
nated the coincidence would be complete. Whilst in our opinion the
theory advanced by Bohr to explain his own curve may be set aside,

* I.e. precipitating the haemoglobin crystals with ether from centrifugalised
corpuscles, washing them, redissolving them and shaking out the ether.

The dissociation curve of haemoglobin

23

another theory also advanced to explain the same curve demands
consideration. It is the theory that oxygen is united to the haemo-
globin, not chemically in the ordinary sense of the word but by the
physical process known as adsorption, that is to say that small masses
of haemoglobin, each composed of a number of molecules, exist in or

10

20

30

40

50

60

70

80

100

FIG. 12. — Ordinates = percentage saturation of haemoglobin with oxygen. Abscissae =
tension of oxygen in mm. of mercury. Curve I = rectangular hyperbola. .XT=800.
Curve 11 = Bohr's dissociation curve of haemoglobin.

O Points determined from dialysed solution.
Q Points determined from undialysed solution.

out of the solution and that the surface charges on them attract mole-
cules of oxygen which adhere to the surface of the haemoglobin. The
phenomena of surface effects are destined to play so great a part in
the future of physiology that this theory, which is due to Wolfgang
Ostwald(5), is well worth the consideration of those interested in the

24 Chapter II

subject. Moreover it presents another very attractive feature. The
difference between oxyhaemoglobin and the body believed to be
isomeric with it, methaemoglobin, has always been somewhat of
a mystery. It would therefore be a relief to regard methaemo-
globin (6) as a chemical compound of oxygen and haemoglobin, whilst
oxyhaemoglobin was a physical combination of the two bodies.

Before dealing with the matter at greater length we should make
it clear that on the adsorption theory there is no reason why the two
bodies should be isomeric. The amount of oxygen which haemo-
globin should unite with has not a fixed limit on this theory. The
relationship which we have pointed out in Chapter I, that of the
oxygen and the haemoglobin being united in proportion to two
atoms of the one to one molecule of the other, would be a purely
accidental one with particular significance.

The adsorption theory of oxyhaemoglobin was originally put for-
ward merely as an explanation of the facts then known, but not in any
way as the result of experiments intended to show that it was more
likely to be correct than the chemical theory.

Since its advent, however, a very laborious research has been
carried out by Manchot'7' which must be considered, for if the
experimental evidence advanced by this worker can be accepted it
could be most easily interpreted along the lines of this adsorption
theory. The point which is fundamental to Manchot's work is
that the amount of oxygen with which 1 gram of haemoglobin can
unite is dependent upon the concentration of the haemoglobin
solution. Up to a certain point the more dilute the haemo-
globin solution the greater the quantity of oxygen it can unite
with proportionally. Manchot's method was somewhat as follows :—
(1) he reduced the blood used, till it no longer gave the oxyhaemo-
globin bands, (2) he then diluted it with various fluids, shook the
solutions with oxygen and measured the amount absorbed. The
experimental procedure failed to satisfy us on two counts : (1) we
do not regard the spectroscope as a satisfactory index of the initial
reduction, (2) no attempt was made to test whether the haemoglobin
at the end would yield, to a vacuum or to ferricyanide, the amounts
of oxygen alleged to have been taken up by adsorption. Burn (S) has
been at great pains to verify Manchot's results, having due regard to
the conditions just stated. He reduced the blood with the pump till
no more gas came off. He then determined the amount of gas which
1 c.c. of blood would absorb when shaken with air in the differential

The dissociation curve of haemoglobin 25

apparatus, and the amount which it would afterwards give out when
treated with potassium ferricyanide. When due allowance had been
made for the fact that "pumped-out blood" absorbs the carbonic-
acid of the air as well as the oxygen, and for other trifling physical
factors, Burn obtained the following figures :—

Degree of Dilution
of Blood

Oxygen capacity by
absorption method

Oxygen capacity by
ferricyanide method

Undiluted

17-6 °L

17-4°/n

5 times

17-6 °/0

17-5 °L

7 times

17-8 °/n

17-1 °L

Each figure represents the average of four determinations: clearly
dilution has no effect on the oxygen capacity.

In passing it is interesting to note that Burn was able to simulate
some of Manchot's results, when he used only the absorption method.

A caution must here be given against the attractiveness of
analogies between certain inorganic oxidations and that of haemo-
globin. There is perhaps no greater danger at the present time, to
one who would research in this particular subject, than the temptation
of trusting to analogies instead of to analysis.

Before leaving the point of analogies we must warn the reader
against one which is frequently put forward, namely, that of the
union between calcium oxide and carbonic acid. In this case if the
pressure and the amount of carbonic acid united to the calcium
oxide be plotted after the manner described above we get quite a
different type of curve. At a certain critical pressure the carbonic
acid begins to come off" and at that pressure it all comes off. The
essential difference between the two cases lies in the fact that, in the
case of the haemoglobin reaction the reacting substances are all in
solution, while in the case of the calcium carbonate the reaction is
between a solid and a gas and therefore involves a change of phase.
The number of phases being one greater, the number of degrees of
freedom is one less.

26 Chapter II

REFERENCES

(1) Hiifner, Arcli.f. Anat. und Physiol. v, p. 187, 1901.

(2) Bohr, Zentrlb.f. Physiol. xvn, p. 682.

(3) Barcroft and Camis, Journal of Physiol. xxxix, p. 118, 1909.

(4) Barcroft and Roberts, Ibid. p. 149.

(5) Wolfgang Ostwald, Biochem. Ctrlb. p. 386, 1908.

(6) Hasselbalch, Biochem. Zeitschr. xix, p. 435, 1909.

(7) Manchot, Liebig's Annal. d. Chem. p. 369, 1909 ; Biochem. Ztschr. XLIII, p. 438,

1912.

(8) Burn, Journal of Physiol. XLV, p. 482, 1913.

CHAPTER III

THE EFFECT OF TEMPERATURE ON THE AFFINITY
OF HAEMOGLOBIN FOR OXYGEN

IN perusing the last chapter one point will not have escaped
the reader. I may state it as follows. Let it be granted that the
dissociation curve of dialysed haemoglobin takes the form of a rect-
angular hyperbola, passing through the point A and approximating

FIG. 13.

to the line by OF. It is possible to draw any number of such curves.
Why should the curve of haemoglobin be this particular one, which is
represented by the thick line in Fig. 13, rather than any other, such
as those represented by the thin lines ?

Let us go back upon the argument which we have already used.

28 Chapter III

The equation which gave us this hyperbola was

For any value of p there can be only one value of CR so long-
as K, which we have already described as a constant, remains in-
variable.

To go back one stage further K itself involved three other
expressions which were also treated as constant — a, the amount
of oxygen which is dissolved in 1 c.c. of water at 760 mm., and k
and k', the velocity constants of the forward and backward phases

of the reaction

02^± HbO2.

Now in treating all these expressions as constants, as we have
already done, we have made the assumption that the temperature is
constant, for in reality they all vary with the temperature. Take
the simplest of them for instance, a; it is known that the amount
of oxygen dissolved in the solution varies inversely with the absolute
temperature. The changes of k and k' with the temperature are not
exactly known, but are certainly large.

But concerning the relation of K (which only involves the ratio of
k to k') to temperature a great deal is known. Our knowledge of the
subject is due chiefly to van 't Hoff and Arrhenius. So far from
being a chance affair the relation of this expression K to changes
of temperature depends upon such fundamental properties of the
substance as the amount of heat given out when haemoglobin and
oxygen unite. That they do unite with the evolution of heat is clear
from an application of the principle of Le Chatelier to the fact that
to heat oxyhaemoglobin tends to dissociate it. That is to say, if at
38° C. and at 20 mm. O2 pressure we have 71 % of the haemoglobin
as oxyhaemoglobin, then at a higher temperature, say 49° C., and the
same pressure a less percentage of the haemoglobin will be oxyhaemo-
globin and a greater quantity reduced haemoglobin.

The law of van 't Hoff(1) relating the values Kl and K« for any
temperatures Tl and Tz is

-q 2'a-Tx

Kz = K,e2 T^ ,

e being the base of the Napierian system of logarithms, and q the
heat evolved when one gram-molecule of haemoglobin unites with
one gram-molecule of oxygen.

Relation of temperature to dissociation

29

As soon as we have grasped this a whole world of facts is
before us.

Firstly then we must have a different hyperbola for every different
temperature. The one we have obtained is not a general expression
of the relationship of oxygen to haemoglobin, it is merely the relation-
ship at the particular temperature at which we chanced to do our
experiments, namely 38° C. Therefore the law of mass action has
become doubly exacting. Not only must the dissociation curve be
a certain shape, but it must be that shape at all temperatures, and
moreover the curves at all temperatures must be related according

FIG. 14.

to a certain law. Have we the means by which to test the truth
of the chemical conception in the light of these exacting require-
ments? Of the quantities in the above formula we have already
a value of KI corresponding to the temperature T1} viz. 38° C. One
could determine the value of K2 for some other temperature T.2, and
then if only — q were known one would have all the data for testing
whether K.2 as calculated was the same as that which was obtained by
experiment. But the truth is that we do not know q, or rather we
did not when the problem first attracted the attention of Hill (2) ;
it was therefore impossible to attack it in just this way. But though
it was impossible to make this test, it was possible to do something

30 Chapter III

less — namely this. Assuming the truth of the equation we have all
the data for the determination of q. Now if our assumption is
correct we may determine q not merely from the values of K at
these two temperatures, but from the values of K at numerous other
temperatures; and the test lies in the fact that the values of q
so determined should all be identical, they being independent deter-
minations of a definite physical constant — the heat of formation
of one gram-molecule of oxyhaemoglobin. Apart from the law of
mass action such an identity would be mere coincidence as unlikely
as if half a dozen church spires of different design taken at random
proved to be the same height. This test was applied; it was not
necessary to work out the whole curve in each case. Assuming
as a result of the law of mass action the hyperbolic nature of the
curve at each temperature, or, to put the matter in another way,
that K has a constant value for each curve, it was only necessary to
take one point at each temperature. The simple method of doing
this was to put a small quantity of haemoglobin solution into a flask
with a large quantity of an atmosphere containing but little oxygen,
and to determine the percentage saturation of the haemoglobin with
oxygen at several temperatures. The volume of gas must of necessity
be so large in comparison to that of the haemoglobin solution that
any reaction which takes place between the two may be regarded as
not altering the composition of the gas. Apart from the necessary
apparatus for blood gas analysis no further equipment was needed
for this experiment than a saucepan, a litre flask and a thermometer.
The saucepan lid was perforated to allow of the neck of the flask
protruding ; the flask was filled with nitrogen containing a small
quantity of oxygen ; about 10 c.c. of haemoglobin solution were put in.
The saucepan was filled with water which surrounded the flask, and
the lid was fixed immovably on the saucepan. At any temperature
it was a simple matter to establish an equilibrium between the
haemoglobin solution with its small volume and large surface and
the gas in the flask. When the equilibrium was once established the
apparatus was inverted. The solution ran down into the neck, where
it presented a very small surface to the atmosphere above it, and
therefore did not alter sensibly in percentage saturation whilst a
sample was being abstracted for analysis. Such an experiment
carried out at five different temperatures gave values as follows
for the percentage saturation :

Temperature 16° 24° 32° 38° 49°

Percentage saturation ... 92 71 37 18 6

Molecular heat of oxy haemoglobin 31

Are these figures such as to give a constant value for q ? The
answer is as follows. If q were 28,700 the following would be the
percentage saturation :

Temperature 16° 24° 32° 38° 49°

Percentage saturation... 90 71 41 22 6

There is no greater divergence here than can be accounted
for by the error of the method used. A similar experiment, at a
much higher oxygen pressure, with a haemoglobin solution made in
a different way but also dialysed, yielded the following results :

Temperature 16° 25° 32° 38°

\ Observed . 96 89 77 52

Percentage saturation jCaleulated;- 9? 89 u -4

The value for q used in this calculation was 27,700. It seems
clear then, that as far as our present experimental methods will
carry us, q has a constant value of approximately 28,000 gram-calories
over a range of temperature of 16° to 49°.

There is never the same feeling of exhilaration about testing
known phenomena by known laws as there is about breaking abso-
lutely fresh ground. Though the conception which we were testing
had passed triumphantly through a successful ordeal, the research
which we are describing took on a newr and extremely interesting
aspect at this point, for now that the bonajlde nature of q seemed to
be established it was clear that we had hit upon a new method of
attacking the molecular weight, and this aroused our enthusiasm the
more as the molecular weight determinations of haemoglobin were in
a very chaotic condition. It is true that, shortly before his death,
HUfner<3), in collaboration with Gansser, had made a determination by
a measurement of the osmotic pressure of haemoglobin. The haemo-
globin was placed in an osmometer which was permeable to salts;
the final osmotic pressure observed corresponded to a molecular
weight of about 16,000. In the performance of this experiment
Hiifner had assumed that the salts distributed themselves equally
on each side of the membrane. This assumption seems to be un-
justified, though the extent to which unequal distribution of the
salts would vitiate such a result has not been accurately ascertained.
Moreover others had attacked the same problem. Weymouth Reid (4)
and Roaf (5) had made similar experiments. The former arrived at
a value of 48,000, the latter at the conclusion that the value varied
within very wide limits according to the nature of the solvent used.

32 Chapter III

The reader may say at this point that we have ourselves proved
that the molecule of haemoglobin contains but one atom of iron and
is therefore the simplest possible. That assumption underlies the
hyperbolic form of the curve which is based upon the equation

For inasmuch as there are two atoms of oxygen to one of iron,
if the haemoglobin molecule contained two atoms of iron the equation
of the reaction would not be that just given but

using Hb to denote the simplest possible formula for haemoglobin.

Let us therefore inquire more particularly where we stand in this
matter. Let us see what we have assumed and what we have proved.
We have shown in Fig. 12 that the points obtained experimentally
from a dialysed haemoglobin solution are consistent with the view
that the dissociation curve is a hyperbola and, not as Bohr sup-
posed, inconsistent with that view ; moreover it must be remembered
that though the fluid in that experiment had a low electric conduc-
tivity it was not absolutely salt free. We have shown then that the
facts are consistent with the equation

Hb + 03^: HbO2.

But have we shown at that point that they are inconsistent with the
equation

Hb, + 2O2 = Hb2O4

or

As a matter of fact the points as determined would agree nearly as
well with a curve which was drawn on the basis of the second of these
equations though not on one which was based on the third. Let it be
clear however that we are not giving up the position which we estab-
lished in Chapter II, namely the probability of the oxygen being
chemically united to the haemoglobin.

I am indebted to Hill for calculating the curve (Fig. 15, Curve II)
which would be yielded by the equation

Hb2 + 2O2 = Hb204.

The points determined when compared with this curve appear to
be not altogether out of conformity with it and suggest that the
points are really between the two curves. And this is very likely
to be the case. For on our supposition it is clear that the actual

Molecular weight of haemoglobin

33

hyperbola could only be reached if all the salts were got rid of,
that is to say that the hyperbola is a theoretical limit which cannot
actually be attained, and similarly the unimolecular is a theoretical
limit of the aggregation of the molecules. In actual fact it is not
unlikely that the great majority of the haemoglobin molecules are
single molecules and that these are mixed with a certain number of
others. Therefore there is the possibility of considering what curve
would be given if the average number was n = 1'5. This curve agrees
with the. determined points just about as well as the hyperbola and
it is therefore not unreasonable to suppose that the actual solution,

30

40

50 60 70 80 90 100

FIG. 15. — Curves I, II and III represent dissociation curves in which the mean number of
molecules in each aggregate would be 1, 1-5 and 2 respectively. The points are those
experimentally determined.

which was equivalent to a '004 normal salt solution — and it is the
first traces of salt which would exercise the most profound influence
in causing the molecules to aggregate — contained some aggregated
molecules.

Such an hypothesis would at once hold out a hope of bringing our
own results into line with those of Roaf (5) whose general theme is
that the osmotic pressure of the protein, and therefore its molecular
weight, varies with the nature of the solvent in which the haemo-
globin is dissolved. As we shall have to discuss this matter more
particularly in the next chapter we shall leave it at this point.

B. R. F.

34 Chapter III

In the present chapter we have established a value for the heat
of molecular combination of oxyhaemogiobin : this involves the idea
that K is constant at any temperature but not the assumption that
haemoglobin is necessarily Hbj rather than Hb2 or Hb3, i.e. that
its molecular weight is approximately 16,700 rather than 33,400 or
50,100. We are therefore in this position ; osmotic pressure experi-
ments have yielded all of these values as well as many others, our
own results hitherto lean to the first, do not exclude the second but
seem to exclude the third. Can we use the heat of formation of
a gram-molecule of oxyhaemogiobin, HbnO2w, to decide the question?
We can. For it proved possible to determine the actual amount of
heat developed when one gram of haemoglobin unites with oxygen :
calling this quantity H, then the molecular weight of the molecule

We need hardly remind the reader that as compared with the
accuracy of chemical analysis this determination may be rather
rough. If, for instance, we obtain a value for the molecular weight
of say 13,000 or 20,000 we might in either case regard n as unity.
In practice the observation of the heat of combination of oxygen
with haemoglobin proved much simpler than we had expected. The
haemoglobin was reduced in a vacuum pump and allowed to run
directly from the pump into a cylindrical De war's flask, the surface
of the solution being protected from air by a coating of neutral olive
oil. A sample of haemoglobin solution was analysed and oxygen was
then bubbled through the fluid for a given time — about 5 minutes—
at the end of which a second sample was withdrawn and analysed.
The rise of temperature was observed with a simple Beckmann's
thermometer. The necessary corrections were made for the amount
of heat which the solution acquired if standing by itself, and so
forth, and when this had been done we arrived at the rise of tem-
perature in the haemoglobin solution as the result of the oxidation.
We have therefore two measurements: (1) the amount of heat
produced by the oxidation and (2) the amount of oxygen absorbed.
On the assumption, the correctness of which has been shown in
Chapter I, that each gram of haemoglobin unites Avith 1'34 c.c. of
oxygen we may calculate the number of grams of haemoglobin which
have been oxidised ; we therefore have all the data which we re-
quire. The result of our observations was that H, the quantity
of heat evolved when 1 gram of haemoglobin unites with oxygen, is

Molecular iveight of haemoglobin

35

1'85 calories*. We have indicated above that the molecular weight
of haemoglobin (Hb),t is to be found by dividing the quantity q by
H. The result of which is approximately 15,200 — a figure very near
to the theoretical figure 1(3,669. f

We have then by the above reasoning shown that the molecular
weight of (Hb)rt = 16,669, therefore Hb must either be that figure or
some submultiple of it. Thus if n was 2, then Hb would be 8335,

FIG. 16. — A, oxygen holder. B, warm bath for heating oxygen up to temperature of
haemoglobin. C, coil of metal tubing. D, chamber for measuring temperature
of oxygen. E, rubber pressure tubing. F, De war's flask containing haemoglobin.
G and //, Beckmaun's thermometers.

and so forth. Clearly this last alternative is an impossible one, for
it would involve the splitting of the atom of iron. The figure 16,669
containing 56 parts of iron is the smallest figure for the molecular
weight which would fit the facts of chemical analysis. The analyst
would admit a multiple of 16,669 but not a submultiple. We have
therefore no alternative but to regard n as unity, a result which

* This result is the mean of three experiments which gave values of 1'82, 1'9S,
1'75, respectively.

3—2

36

Chapter III

agrees with all we have as yet considered with regard to haemo-
globin.

We are now in a position, from the values of K which we have
determined above, and from the knowledge that dialysed haemoglobin
exists in single molecules, to calculate the percentage saturation of
haemoglobin with oxygen at any oxygen pressure and at any tem-
perature. Below we give a diagram of the dissociation curves of
oxyhaemoglobin at the temperatures 16°, 25°, 32°, 38° and 49° C.

100

10

80

90

100

FIG. 17.— Dissociation curves I, II, III, IV and V correspond to 16°, 25°, 32°, 38° and 49° C.
respectively. Oxygen pressure plotted horizontally, percentage of reduced haemo-
globin vertically downwards.

In order to arrive at some sort of knowledge, as to how fast the
forward and backward reactions of the oxidation and reduction of
haemoglobin go on, and as to how far these rates are affected by
temperature, some further experiments were done by Barcroft and

Rate of reduction of haemoglobin

37

Hill. Nitrogen was bubbled through a suitably protected solution of
haemoglobin at a constant rate but at two different temperatures.
The rate at which the haemoglobin was oxidised was noted in each
case. The experiment was first performed as follows. We started
at a temperature of 18° C. with the blood fully saturated. The
course of the experiment may be followed by reference to Figs. 18
and 19 ; the former shows the apparatus used, the latter the results
obtained. As regards the apparatus, the haemoglobin solution was
placed in the cylinder A. Nitrogen was passed at a constant rate in

/Vitroofn, from,
Marnotte'sBottU M

f/t-trootn, to

i Bottle N

. JL--.

FIG. 18. — Apparatus for the determination of the rate of oxidation of haemoglobin.

the direction of the arrows, first through a wash-bottle containing
an alkaline solution of pyrogallol, then through the haemoglobin
solution ; the whole apparatus was immersed in a water-bath and in
that was kept at the required temperature. Samples of haemo-
globin, which were small relative to the whole volume of haemoglobin
used, could be abstracted from A at any time through a tube let in
through the cork for the determination of the percentage saturation
with oxygen. At any time in the experiment the inlet and outlet
tubes for the nitrogen could be clamped and the whole apparatus

38

Chapter III

warmed up rapidly by replacing the water in the bath by water at
a higher temperature. The flow of nitrogen then was resumed at
the altered temperature.

Nitrogen was allowed to run for 35 minutes, by which time the
haemoglobin was slightly reduced. An analysis showed that 6 °/0 of
the haemoglobin was reduced haemoglobin and 94 % oxyhaemo-
globin. The point at which we have arrived is represented as D in
Fig. 19. There is possibly an error of about 2 °/0 in this measurement
either way, i.e. the percentage saturation may have been 92% or
96 °/0 (d or d'\ The nitrogen was now stopped, the temperature of the
bath was raised to 38° C. (the time which the process took is omitted

100
90
80
70
60
50

B C

40
30
20

ia"c

\

10 20 30 40 50 60 70 80 90 100

FIG. 19. — Curve representing the calculated degree of dissociation at any time of haemo-
globin which was reduced by a stream of nitrogen bubbled at a uniform rate.
Percentage saturation plotted vertically, time in minutes horizontally. The points
represent actual determinations, A — D at 18° C., D — G at 38° C.

from the diagram) and the nitrogen was restarted. The haemoglobin
now reduced very rapidly. After this three other determinations
were made as follows :

Point on curve D

Time (minutes) measured from D 0

Percentage saturation of haemoglobin ... 94 %

E
7-5

77%

F
23

60%

G
53

26%

Thus whilst at 18° C. 35 minutes had been required for the reduction
of the haemoglobin from 100 % to 94 % saturation, at 38° C. it only
required 7'5 minutes to reduce it from 94 °/0 to 77 °/0. Perhaps
the best comparison of the times necessary to produce a given
reduction can be obtained by extrapolating the curve DEFG back-
wards to B, in which case AC represents the time necessary to

Rate of reduction of haemoglobin

39

produce the reduction from 100 to 94 % at 18°C., whilst BC repre-
sents the time necessary for the same reduction at 38° (J. ; the former
is 35 minutes the latter about 2'5 or at most 3 minutes. It is usual
to compare the rates at which chemical changes take place, not
over a range of 20° C. as in this case, but over a range of ten degrees.
If the times necessary to produce a given change at temperatures
20° C. apart be in the ratio 35 : 3, the corresponding ratio for a
difference of 10° C. will be the square root of the ratio of 35 : 3,
namely 3'4 : 1, i.e. the coefficient is increased between 3 and 4-fold
by a rise of temperature of 10° C.

A few lines back I spoke of extrapolating the curve DEFG. Let
me say a word or two in justification of this phrase. I do not mean

100
90
80
70
60
50
40
30

20

10

20

30

40

50 60

70

80

90

100 110 120

FIG. 20. — Curve representing the calculated degree of dissociation at any time in a
haemoglobin solution through which nitrogen bubbled at a uniform rate. Percentage
saturation plotted vertically, time in minutes horizontally. Points represent actual
determinations.

merely producing the curve backwards by its general appearance,
though for the present purpose no practical error would be involved
by this procedure. I wish to make it clear that the line DEFG is a
definite curve with a definite equation. The reduction of haemoglobin
by nitrogen, if it takes place as a chemical breakdown, must obey laws
which prescribe a definite form to the curve relating the percentage
saturation to the time. The equation relating these two quantities
has been worked out by Hill, and DEFG as drawn in the figure is
such a curve. The coincidence of the points as determined experi-
mentally with the curve is sufficiently good. The test would however
have been more crucial had the points been more numerous.

40 Chapter III

Hill and I therefore performed another experiment from this
point of view, i.e. to see whether the actual points determined
followed a curve prescribed by the laws of mass action. We varied
the conditions by making- the haemoglobin in a different way, making
the rate of reduction much slower and determining six points instead
of four. The result worked out as follows :

Temperature 38° C.

Points A B C D E F

Time from beginning of Exp 0 14-5 32 55 80 114 mins.

Percentage saturation observed ... 92'5 83 71 64 55 46

Ditto calculated 92-5 82 72-5 63 51-5 46

The correspondence of the determined points with the theoretical
values is equally complete under these altered circumstances.

There seems to be scarcely any limit to the possibilities which
haemoglobin offers to the student of colloids : a colloid with a simple
chemical reaction which can be treated quantitatively with little
difficulty is of very rare occurrence. There is little doubt, if haemo-
globin could be made and kept with ease, that it would be a medium
for an enormous volume of illuminating work.

REFERENCES

(1) Nernst, Theoretical Chemistry.

(2) Barcroft and Hill, Journal of Physiology, xxxrx, p. 411.

(3) Hiifner and Gansser, Archie f. Anat. und Physiol. p. 209, 1907.

(4) Weymouth Reid, Journal of Physiology, xxxm, p. 12, 1905-6.

(5) Roaf, Proc. Physiol Soc. p. 1, 1909.

CHAPTER IV

THE EFFECT OF ELECTROLYTES ON THE AFFINITY OF
HAEMOGLOBIN FOR OXYGEN

UP to the present we have made no attempt to treat of the work,
which has recently been carried out on the affinity of haemoglobin for
oxygen, on any sort of chronological basis, and now, having arrived at
the fourth chapter we have also arrived at the point at which much
of the work discussed in the preceding pages really began, and from
which it proceeded always from the point of greater complication to
the point of greater simplicity. It was the constant effort to unravel
a knot which gave us the threads from which ultimately to weave a
fabric.

What we wished to do may sound simple. In practice it has
proved to be quite the reverse. It was to calculate the oxygen
pressure in the capillary circulation.

We were therefore face to face with the problem of whether it
was possible from the percentage saturation with oxygen of a given
sample of blood to calculate from the dissociation curve of blood the
oxygen pressure to which the blood was exposed. This was our first
practical introduction to the dissociation curve. When we consulted
the literature of the subject we found that each paper brought our
ideas into a state of greater complexity. The authors who had
written most recently on the subject were Zuntz'1', Loewy(2), Bohr(3),
and Franz Miiller(4). Zuntz and Loewy had made the important con-
tribution that apparently no two samples of blood, whether from man
or beast, could be relied upon to have the same dissociation curve.
Their view had been supported by Bornstein and M tiller who worked
on cat's blood, whilst Bohr and his colleagues had stated that this
was in large part due to the fact that the dissociation curve is very
sensitive to the presence of carbonic acid ; their observations were
discounted or neglected by many authors. Mtiller, however, had also

42 Chapter IV

shown that, even taking this fact into consideration, the blood of the
dog had a smaller affinity for oxygen than had that of the horse.

Camis(5) and I determined to perform some experiments to test
the accuracy of the above observations : this did not seem difficult
in view of the fact that we had at hand the differential method
of blood gas analysis — a method which offered the prospect of
making in a few days analyses that would previously have taken
many weeks.

Like those of the proverbial golfer who strikes the ball with un-
erring accuracy on the occasion of his first visit to the links and then
only, so our very early efforts seemed to augur speedy success — in
our case the construction of a uniform dissociation curve for the blood
of various animals. When we came to the blood of man however we
could never make the dissociation curve agree with that of the cat or
the rabbit. We went back to it time after time ; the result was the
same, human blood took up less oxygen at low tension than did that
of the cat or rabbit. It was then clear that we had found our way
into the morass in which our predecessors had already floundered so
hopelessly, and our newer and more certain methods instead of saving
us from their embarrassments had only made the uncertainty of our
position more certain. There was no overlapping of the curves, no
confusion of the points — human blood had a smaller affinity for
oxygen than cat's blood.

We then entered upon six months of research which became
weekly more and more depressing. Occasional flickers of light ap-
peared— they turned out to be but will-o'-the-wisps — each of which
beckoned us to a more inevitable disappointment than the last. It is
easy to discuss the merits of apparatus, to gauge the help given by
this or that method, yet as I look back upon this period of gloom, I
discern more and more clearly as time goes on that the determining
factors which made this research fruitful did not come from methods
or from apparatus, but from the unfailing zeal and good will of my
colleague Camis. His goodness of heart rose superior to the trials
caused not only by his own disappointments but by mine.

We thought to simplify the issue by ceasing to work with blood
and substituting haemoglobin. The simplest way of making this
substitution was to lake the blood. In order to do this efficiently we
added an equal part of dilute ammonia solution such as we used in
our gas analysis apparatus, which solution was made by adding 4 c.c.
of strong ammonia to a litre of distilled water.

Effect of alkali

43

When we left off working with blood and commenced working
with solutions of haemoglobin, the discrepancies were quite as great
as, if not greater than, they had been with blood, and intellectually the
same difficulties presented a much greater embarrassment because
they were much more paradoxical. With regard to blood, it might
be said with some reason that the haemoglobin was modified by the
fact of its being enclosed in the red blood corpuscles. Indeed, many
suggestions might be put forward as to how this could happen. The
surface charge of the corpuscle, if it has one, might increase or
decrease the affinity of a corpuscle as a whole for oxygen. The
surface membrane of the corpuscle might in different cases present

%100

20 40 60 80 100mm.

FIG. 21. — Comparison of points obtained from blood laked with dilute ammonia with
dissociation curves of Bohr and Hiifner. Oxygen pressure horizontal. Percentage
saturation vertical.

different obstacles to the diffusion of the oxygen within its walls.
The existence of points or particles in a fluid is known to have a
considerable influence on the accumulation of gas in their vicinity.
All these unknown factors might be advanced with regard to the
corpuscle.

When, however, we laked the blood, and so obtained a solution
of haemoglobin, free from corpuscles, it was not possible to apply
any of these considerations, and therefore the fact that no two
solutions of haemoglobin presented identical affinities for oxygen was
much more difficult of explanation.

44

Chapter IV

Moreover the curves obtained did not agree with those either of
Bohr or of Hiifner and they certainly were not in any rectangular
hyperbola.

We thought it best to get rid of the proteids of the serum which
might in some way adsorb oxygen. So we dissolved the haemo-
globin in Ringer's solution. The only trace of uniformity in our
result seemed to be that it differed from all our other results even
as they had done from one another.

The curve which we obtained was not of any particular mathe-
matical form which we could discern. It even differed from that of
a suspension of washed corpuscles in Ringer's solution.

7o100

60

20

20 4O bO 00 100mm.

FIG. 22. — Comparison of points obtained from a solution of haemoglobin made alkaline
with ammonium carbonate, with the dissociation curves of Bohr and Hiifner. Oxygen
pressure horizontal. Percentage saturation vertical.

Lastly we made solutions of haemoglobin in different ways.

Our great effort was to obtain the maximum degree of concen-
tration of the haemoglobin. The importance of this point had been
rendered evident by the earlier work of Hiifner, whose solutions of
haemoglobin had been so dilute that the oxygen physically dissolved
in the solution bore a considerable ratio to the total oxygen present.
To this fact we then attributed the discrepancies to be found in
Hiifner's results. On looking back now we can see that it merely
served to obscure discrepancies of a much more deep-seated nature, of

Effect of alkali

45

which Hiifner himself appears to have been oblivious, but which were,
at the time of which we were speaking, only too evident in our own
experiments. It seems certain that Hiifner's solutions of haemo-
globin, which he regarded as identical, must have differed from one
another in their affinities for oxygen as much, if not more, than
our own.

To get the maximum amount of haemoglobin into solution wre
found it advisable to add a little ammonium carbonate ; but we were
always assailed with the same result, namely, that no two solutions
gave identical dissociation curves.

% 1 00

80

60

40

20

0

7

20 40 60 80 100mm.

FIG. 23. — Comparison of points on haemoglobin solution prepared by Bohr's method,
with Bohr's dissociation curve. Oxygen pressure horizontal. Percentage saturation
vertical.

One afternoon Camis, Hill and myself, over the laboratory tea,
fell to discussing a subject which had frequently been discussed
before, namely Bohr's haemochrome theory. We had long since given
up any attempt to harmonise our curves with those of others, but
the question in our mind was whether with different constants the
most recent curve which we had obtained could be made to agree
with the general mathematical expression of the theory. Camis read
out the positions of the points.

Tension of oxygen 12-5 15 -5 31 45 72

Percentage saturation... 29 40 66 77 -5 90-5

46

Chapter IV

Hill dotted them in in pencil on Bohr's curve and showed me the
figure ; it was as shown in Fig. 23.

The solution like most of those made previously had been made
by Bohr's method, quoted in the last chapter ; in this particular case
however we had omitted, whether by intention or forgetfulness
I cannot now remember, to add any ammonium carbonate.

100

90

80

70

60

50

40

30

20

10

10

20

30

40

Iv

50

70

80

90

100

FIG. 24. — Ordinate = percentage saturation of haemoglobin. Abscissa = tension of oxygen
in mm. of mercury.

I. Dissociation curve of haemoglobin dissolved in water.
II- „ „ „ » -7°/0NaCl.

III. „ „ -9°/0KCl.

Eectangle surrounding point — magnitude of experimental error. Temperature 37 — 38° C.

We could only attribute the divergence of our previous curves
from that of Bohr to the ammonium carbonate which had been
added somewhat casually. We therefore dispensed with this and

Effect of salts 47

then we were able time after time to obtain solutions which were
identical with that figured by Bohr for haemoglobin.

To test the effect of salts in a more systematic way we made
solutions of haemoglobin, divided them into various portions, added
various salts to the portions and compared their dissociation curves
with that of the original haemoglobin as prepared by Bohr's method.

In Fig. 24 are given the dissociation curves of oxyhaemoglobin
in the solutions of electrolytes which are of greatest physiological
importance, namely sodium chloride and potassium chloride isotonic
with one another. We shall reserve the consideration of some
others till later. From this point constructive work on the dis-
sociation curve of blood commenced. In the red blood corpuscle
haemoglobin is dissolved in a solution of salts which differs very
widely in different members of the animal kingdom. But it is
evident that the first step towards building up the curve of blood
from that of haemoglobin is to investigate a haemoglobin solution
dissolved in the actual salts which are present in the red blood
corpuscles. But here a fresh complication arose, for the salts differ
in the corpuscles of each different species. We performed the follow-
ing experiment : a solution of haemoglobin was divided into two
portions, (a) and (6). To the portion (a) were added the salts of the
human red corpuscles as determined by the old, though still classical,
analysis of Schmidt""1. To the portion (6) were added the salts of
the dog's red blood corpuscles as determined by the recent, and ex-
cellent, analysis of Abderhalden'7'. The dissociation curves of these
two solutions were determined in the presence of approximately
40 mm. pressure of carbonic acid, and gave the results depicted in
Figs. 25 and 26.

It was at least evident that the two curves were quite different,
that the difference could by no possibility be explained by experi-
mental errors, and that one had arrived at at least one tangible
reason why the blood of one animal should differ from that of
another ; but even then it seemed that we were far from a complete
solution of the whole problem before us. It was not unlikely that
any number of other reasons might conspire to add or detract from
the differences we had observed, and therefore our next move wras to
make as faithful a comparison as possible between the curves which
we had just observed, and the actual curves which could be obtained
from human and canine bloods respectively when exposed to the
same conditions as the haemoglobin solutions (a) and (b). We hoped

48

Chapter IV

that by finding out what discrepancies still existed, to shed some light
on the factors which remained for investigation. Great then was our
excitement and delight when we found that point after point on the
curve of human blood lay also on the curve of the haemoglobin

100

120

130

140

150

160

80

90

100

FIG. 25. Ordinate = percentage saturation. Abscissa = tension of 02 in mm. of mercury.

Dissociation curve of human blood at 40 mm. tension of CO2. • = determination for
human blood. o = ditto for haemoglobin solution with salts of human red blood
corpuscles. The area drawn round the point is the experimental error in each case.
Temperature 37 — 38° C. More recent observations have shown that the portion of
the curve below 15mm. pressure, while traversing the shaded areas, approaches the
base line more gradually.

solution (a), whilst the corresponding points for dog's blood adhered
with equal fidelity to the curve obtained from solution (6). The
points in question are also given in the figure, and the reader will
observe that only in one point on each curve does there seem to be

The blood of different species

49

a slight tendency for the actual points obtained from blood to diverge
a little from the curve obtained from the haemoglobin solution. That
is just where the curve bends round at the top. To this we must add
that more recently the curve of the same human subject, Dr Camis him-
self, has been redetermined at Pisa by more exact methods than were

TOO

110

90

80

70

60

50

40

30

20

10

ffi

L/L

- o/

tr

LU

10

20

'

30

40 50 60 70 80 90 100

FIG. 26. — Ordinate = percentage saturation of haemoglobin. Abscissa = tension of 02 in
mm. of mercury. Dissociation curve of dog's blood at 40 mm. tension C00. Dotted
line = dissociation curve of human blood (see Fig. 25). •= determination for dog's
blood. o = ditto for haemoglobin solution with salts of dog's red blood corpuscle.
The area drawn round the point is the experimental error in each case. Temperature
37_38° C.

then at our disposal. These methods enabled us to work accurately
at pressures of oxygen under 10 mm. The Pisa curve seems to differ
from that given above to a trivial though appreciable extent. The
difference, however, is quite negligible as compared with the dif-
ferences which exist between his blood and that of a dog, and it is

B. R. F. 4

50 Chapter IV

highly probable that it depends upon the method employed. The
following are the data of the curves obtained for man and the dog :

Blood at approximately 40 mm., tension of C02 and 37 — 39° C.

Abscissa 10 15 20 25 30 35 40 45 50 60 80 100 160 mm. 02

° 21'5 45'5 59'5 69 75'5 80 83'5 86 88'5 91'5 93'5 97 °/°
7 11-5 18'5 20'5 32'5 41'5 50'5 60'5 69 80 88'5 90 ~°/°

It is scarcely possible to leave the subject of the action of salts
upon haemoglobin without saying what there is to be said on this
question of how they act.

This has to some extent been foreshadowed in the previous
chapter. The theory which we wish to discuss is that, while the
actual union of the oxygen and the haemoglobin is a chemical action,
the quantity and character of the salts present cause the molecules
of haemoglobin to adhere to each other in varying degrees. Any
alteration in the degree of aggregation of the molecules would by the
laws of mass action entail an alteration in the dissociation curve.

First of all let us discuss the question of whether there is any
independent reason for supposing that the presence of salts does
produce such an aggregation of the molecules.

The evidence on this head is two-fold :

(1) Roaf showed that the osmotic pressure of haemoglobin in the
presence of salts, i.e. in blood laked with water or with very dilute
sodium bicarbonate, corresponded to aggregates of about two mole-
cules. Hiifner and Gansser with pure haemoglobin obtained results
corresponding to one molecule.

(2) Mines has shown that a dialysed solution of haemoglobin is
precipitated by minute quantities of trivaleht ions in just the same
way as it is by acids, and precipitation is only one stage further than
aggregation.

The colloidal nature of haemoglobin*.

Although it is certain that the union between haemoglobin and
oxygen is a true chemical combination and not a mere surface ad-
sorption, the colloidal condition of haemoglobin remains of great
importance as a factor influencing the equilibrium between haemo-
globin and oxygen at relatively low oxygen tensions. For although
the oxygen capacity of haemoglobin must be independent of the state
* I am indebted to Mr Mines for the following note.

Haemoglobin as a colloid 51

of dispersity of the substance, the extent to which oxyhaemoglobin
will dissociate in the presence of low tensions of oxygen may con-
ceivably be largely influenced by this factor.

Solutions of haemoglobin however finely dispersed will remain
colloidal, for the haemoglobin molecule is of such dimensions that it
presents surface and itself constitutes a colloidal particle.

As regards the production of actual visible agglutination or pre-
cipitation of haemoglobin by electrolytes, haemoglobin may be said
to present the characters of a negatively charged emulsoid or lyophil
colloid. It has been shown that a number of negative suspensoid
colloids have their electric charge affected, and thus are precipitated,
quite as readily by complex trivalent kations (such as Co(NH3)6'") as
by simple trivalent kations (such as La"', Nd'", &c.), while emulsoids,
many of which have their charge affected readily by simple trivalent
kations, are almost uninfluenced by complex trivalent kations. The
latter is the case with oxyhaemoglobin. A carefully dialysed solution
of the pigment is precipitated by a moderate concentration of a salt
of lanthanum or cerium (cerous) while no visible effect is produced
by the addition of luteo-cobalt chloride even in large concentration.

Oxyhaemoglobin also shares a property possessed by some sus-
pensoids and some emulsoids, namely, extreme sensitiveness towards
the hydrogen ion. A low concentration of acid suffices to precipitate
oxyhaemoglobin.

It is to be expected that a change in the hydrogen ion con-
centration in the direction of increased acidity, insufficient to cause
visible agglutination, will yet cause some aggregation of the molecules
of haemoglobin. The fact that gross aggregation is readily obtained
by acid lends support to the idea that the effects produced upon the
dissociation curves of oxyhaemoglobin by still smaller amounts of
acid (Bohr, Barcroft and Orbeli) are due to a minor degree of aggre-
gation of the molecules.

If this hypothesis is a true one, it is to be expected that of
the simple kations, exclusive of the hydrogen ion, those of different
valency will produce very different effects. Thus a certain small
concentration of a salt yielding simple trivalent ions (for instance
lanthanum nitrate) will produce an effect on the dissociation curve
of oxyhaemoglobin which will be equalled only by a concentration of
a magnesium salt some hundreds or perhaps thousands of times as
great, while of such salts as NaCl a relatively enormous concentration
will be required. Further we should expect to find phosphates and

4—2

52 Chapter IV

citrates very active in opposing the changes produced by polyvalent
kations.

We are now face to face with the influence of " reaction " on the
dissociation curve of blood.

REFERENCES

(1) Zuntz and Loewy, Archie f. Anat. u. Physiol. p. 166, 1904.

(2) Loewy, Ibid. p. 246.

(3) Bohr, Hasselbalch and Krogh, Skand. Arch. f. Physiol. xvi, p. 412.

(4) Bornstein and Miiller, Archie f. Anat. u. Physiol. p. 478, 1907.

(5) Barcroft and Camis, Journal of Physiol. xxxix, p. 118.

(6) Schmidt, Halliburton's Text-book of Physiol. and Pat hoi. p. 265.

(7) Abderhalden, Lehrbuch d. Physiol. Chem. Zweite Aufl. pp. 732 et seq.

CHAPTER V

THE EFFECT OF ACID ON THE DISSOCIATION CURVE

OF BLOOD

BEFORE entering into a discussion as to whether the influence
of " reaction " on the affinity of blood for oxygen is different in
nature or only in degree from that of salts, I will state the facts
which must form the basis of any such discussion.

Doubtless the affinity of haemoglobin for oxygen is very sensitive
to small changes in acidity or alkalinity. This statement unfortu-
nately does not rest on experimental determinations of the influence
of minute quantities of acid or alkali on the dissociation curve of an
otherwise pure solution of haemoglobin, for no such determinations
exist. In the case of acid it remains to be seen whether such an
experiment can be carried out without the precipitation of the
haemoglobin. What information there is on the subject, then, is
derived from experiments on blood and not on haemoglobin.

In two words, the effect of increased acidity of the solution is to
lessen the effective concentration of the oxygen in the solution which
contains the haemoglobin. To give a concrete instance, that of my
own blood. The presence of 40 mm. pressure of carbonic acid halves
the effective concentration of oxygen. In order therefore to pro-
duce a given degree of saturation in the presence of 40 mm. CO2
pressure, the concentration of oxygen must be about twice the
value which would have sufficed in the absence of the carbonic
acid.

The concentration of free carbonic acid in the solution, necessary
for the production of a measurable change in the affinity of haemo-
globin for oxygen, is very small, and amounts to something of the
order of one part in one hundred million. This fact is directly
deducible from the data furnished by Bohr, Hasselbalch and Krogh (1),

54 Chapter V

who were the discoverers of the influence of carbonic acid on the
dissociation curve.

Carbonic acid however is not specific in its effect. Lactic acid
acts in precisely the same way (2) ; so also do mixtures of lactic and
carbonic acids.

Having made the rough statement that the effect of acidity is to
reduce the effective concentration of oxygen, I must now state more
precisely (1) what I mean by "reaction," and (2) what I mean by
reduction of the effective concentration of oxygen.

It seems that the influence of acids depends on the change in
hydrogen ion concentration which they cause in the blood : and this
appears to be different from the change of hydrogen ion concentration
which they would cause in water or salt solution.

For instance, mineral acids in equal molecular concentration are
generally much more powerful than acids such as lactic, acetic or
carbonic. They cause a much greater concentration of hydrogen
ions. Yet when they are added to serum the change in the hydrogen
ion concentration which they produce is nearly the same. This change
has been measured by the method of titration with an indicator which
changes colour gradually over a considerable range of hydrogen ion
concentrations.

Added to normal salt solution the following strengths of acid
would produce equal changes in the hydrogen ion. concentration :

Lactic Hydrochloric Formic Acetic

M/200 = M/300 M/80

M/110 = M/200 M/130 M/50

M/50 M/170

M/10 M/130

But the relative quantities of these acids which produce equal
changes in hydrogen ion concentration in serum are

Lactic Hydrochloric Acetic C02 Hydrogen ion cone.

M/200 M/200 10-6'8

M/100 M/120 10-H-1
M/100 weaker than M/100

•2 °/0 = exposure to atmosphere of 100 % 10-4'«

These results were obtained by titration with the indicator
referred to above.

Now so far as the affinity of the haemoglobin for oxygen is con-
cerned, the same change is wrought by increasing the concentration

Effect of acid on dissociation curve 55

of lactic acid to '2 °/0 as by exposing the blood to 100 % CO.,, a change
which in each case increases the hydrogen ion concentration of the
plasma from lO"75 to lO"4'5.

Similarly, blood made up with lactic acid to M/120 has about the
same affinity for oxygen as blood made up to M/130 with hydrochloric
acid, and so forth.

The affinities of the samples of blood for oxygen were not com-
pared in these cases by means of the dissociation curve, but by another
method, that of bubbling a uniform stream of nitrogen through the
blood and testing the rate at which it became reduced, i.e. the time
taken to attain a given degree of reduction. Allusion has already
been made to this procedure. The method was adapted somewhat
for the purpose by Mathison (3), who performed the experiments with
which I am now dealing. It is described in Chapter XI. Here I am
discussing the results.

This method no doubt expresses dynamically the facts concerning
the relation of haemoglobin to oxygen, just as they are expressed
statically by the dissociation curve. It is very desirable however that
actual determinations should be carried out in which the hydrogen ion
estimations are measured with the gas chain battery, and the affinity
of the haemoglobin for oxygen by means of the dissociation curve.
This is a very difficult matter, but the work which has recently been
published from Hasselbalch's (4) laboratory offers a bright promise in
this respect.

In the meantime, let me give as examples the data of the two
cases I have mentioned. From them it will be seen (1) that the
addition of acids greatly accelerates the reduction of the blood;
(2) that the concentrations of the acids necessary to produce ap-
proximately equal effects are those in which they produce equal
increments in the hydrogen ion concentration.

From Fig. 27 it will be seen that the times taken, by blood contain-
ing *2 °/0 lactic acid, to be reduced by nitrogen, and by blood without
lactic acid to be reduced by 100% CO2, are indistinguishable, and are
about one-tenth as great as the times necessary for the same reduction
without the acids; i.e. the blood is reduced from 100 °/0 saturation to
80 °/0 saturation, in about 2 minutes with the acids present, and about
22 minutes with the acids absent.

From Fig. 28 it will be seen that normal blood is reduced under
the conditions of the experiment (in which the bubbling was more
rapid and the temperature higher than in the preceding one) from

56

Chapter V

100 % to 35 % in 15 minutes, while when it contained M/130 HC1
and M/120 lactic acid the same reduction was effected in about 6 '5
and 7 '5 minutes respectively.

,0 1 0 2Q 40 60

100

80

60

40

20

100

80

60

40

20

• Normui Blood

"u?

0-2% (

FIG. 27.

N
M

E
HCL

lood

10

20
FIG. 28.

40

60

FIGS. 27 and 28. — Katio of reduction of blood with a uniform stream of oxygen free gas.
Percentage saturation vertically. Time in mins. horizontal. Fig. 27. — Comparison
on normal blood, with blood containing lactic acid and with blood reduced by carbonic
acid instead of nitrogen. Fig. 28. — Shows the approximate equality of M/130 HC1
and M/120 lactic acid in their effects.

Having made it clear that the sense in which I use the word
" reaction " is the strict one, namely the hydrogen ion concentration,
I will leave this point and pass to the consideration of the effective
concentration of oxygen.

Effect of acid on dissociation curve

57

The simplest meaning which would attach to the words "increased
hydrogen ion concentration reduces the effective concentration of
oxygen " would be as follows : In order to obtain a given percentage
saturation of the blood with oxygen in the presence of the various
concentrations of acid, the pressure of oxygen with which the blood
would be in equilibrium would require to be higher in proportion to
the increase in the H-ion concentration : this proportion also should
be the same all along the curve.

•O05 -010 -015 -O20 -025 -030 -035 -040 -O45

PRESSURE OF CO IN PERCENTAGE OF ONE ATMOSPHERE.

FIG. 29. — Dissociation curves of CO-haemoglobin in absence of oxygen, at 38° C. and
with various pressures of C0.2.

o Blood of Douglas. • Blood of Haldane.

For instance, my blood is 45 D/0 saturated at 10 mm. oxygen
pressure in the absence of C02, in the presence of 40 mm. CO2 it is
45 % saturated at 24 mm. 02 : now if the relation which we have

Chapter V

defined be true the oxygen pressures on the two curves (0 CO2 and
40 mm. C02) for any given percentage saturation should be in the
ratio of 10 to 24. In other words the effect of the CO2 is to make
24 mm. oxygen pressure behave like 10 mm., and x mm. pressure
behave like x x ££.

Here we find ourselves facing the question of whether the effect
of " reaction " is of a fundamentally different character from that of
salts. The effect of salts was to produce a curve essentially different
from that which would be obtained in their absence. The dissociation
curve of pure haemoglobin is a hyperbola, that of haemoglobin in

100

90
80
70
60
50
40
30
20
10

0 10 20 30 40 50 60 70 80 90 100

FIG. 30. — Ordinate = percentage saturation with oxygen. Abscissa = oxygen pressure in
mm. The abscissa corresponding to any percentage saturation on Curve I is 11/20,
on Curve II 16/20, and on Curve III 27/20, of the abscissa corresponding to the same
percentage saturation on the dissociation curve of my blood at 40 mm. CO2 pressure.
Determinations: • 3mm. O 20 mm. ® 90mm. CO2 pressure.

salt solution is something else, namely an S-shaped curve. But if
I have correctly described the effect of " reaction," it does not affect
the essential character of the curve, but merely the scale on which
it is drawn. This conception is fundamentally different from that
with which I finished Chapter IV. There was no other suggestion
in it than that the effects of acids and salts might turn out to be
the same.

The question is one which must be put to the test of experiment.
The effect of acid, as I have described it, was suggested to Douglas

Effect of acid on aggregation 59

and Halclane by the curves which they obtained for the dissociation of
carboxyhaemoglobin into reduced haemoglobin and CO, in the presence
of different CO2 pressures. Clearly they approximate to this ideal.

It seemed desirable to redetermine the corresponding points for
oxyhaemoglobin for the purpose of ascertaining whether there were
any difference between the curves at various CO2 pressures other
than the scale on which they were drawn. To this matter Poulton
and I addressed ourselves. We made determinations of points on
the dissociation curve in the presence of 3, 20 and 90 mm. C02
respectively. These points are shown in Fig. 30. The curves in this
figure are all derived from my dissociation curve at 40 mm. pressure.
For any percentage saturation the pressure on the line corresponding
to 3 mm. C02 line is ^, that on the 20 mm. line is ^, and on the
90 mm. line is f£ of the pressure on my 40 mm. line.

Not until Poulton and I arrived at 90 mm. C02 was there any
suggestion of a discrepancy. The discrepancy that occurred then
may be due to experimental error, for the determinations could just
be included within the extreme limits of such errors. Nevertheless
in view of the fact that there is a similar disposition of the points
on the corresponding curve for CO haemoglobin (Fig. 29), and also
that a similar tendency is evident in some of the curves obtained
on Monte Rosa, I am inclined to think that at C02 pressures of
about 100 mm. the dissociation curve does become slightly more
curved than it would be were it a mere horizontal expansion of the
curves obtained at lower pressures *.

This slight divergence at a very high concentration of CO2 does
not seem to invalidate the general truth of the conception that the
main action of increased hydrogen ion concentration is to reduce
the effective concentration of oxygen in precisely the way we have
described. One of the most unbending of physical laws (Boyle's law )
bends when pushed to extremes. The laws of ionic dissociation do the
same, and it could scarcely be expected that our present case should
be more rigid than these. The important point is that over ranges
of CO2 pressure which are not extreme the curves appear to differ
merely in the scale on which they are drawn. This, as I have indi-
cated, suggests that in the presence of considerable quantities of salt
small changes in hydrogen ion concentration do not appreciably alter
the degree of aggregation of the molecules.

Now that I have stated the main facts about the relation of salts

* See Appendix II.

60 Chapter V

and acids to haemoglobin it may be profitable to give a brief sketch
of an attempt which has recently been made to find a physical expla-
nation of these facts.

Any such explanation starts from the proven fact that the dis-
sociation curve is a rectangular hyperbola, representing a simple
reaction

This reaction involves single molecules only, and has a definite
equilibrium constant.

The equation for this curve may be written

y Kx

100 ~= 1 + Kx '

where y = the percentage saturation of the haemoglobin with oxygen,
x = oxygen pressure, and K is the equilibrium constant of the curve.

This curve, as I have said, assumes that the molecules of haemo-
globin are single ones. Hill <7) amplified this equation with the object
of ascertaining the shapes of the dissociation curves which would be
obtained on the assumption that the molecules of haemoglobin fall
into aggregates as the result of the addition of salts. At the end
of Chapter IV I pointed out that there was independent reason to
suppose that this was the case.

If on the average there are n molecules in each aggregate, the

equation becomes

y Kxn

100 ~1 +Kxn'

The solution is conceived of as containing haemoglobin in the various
degrees of molecular aggregation, the reactions of which would take
place in the following way :

Hb + O2^

Hb3 + 3O2 ^± Hb3O6, &c., &c.

The equation however has certain limitations of which the only one
to be considered, so long as we are confining our attention to oxy-
haemoglobin, is that there is no considerable quantity of intermediate
oxides such as Hb,02 in the solution (8'. The theory does not preclude
the formation of such substances. It is highly probable that Hb202
for instance is a stage on the way to Hb204. But to make the
equation hold, the suboxides can only exist as stages in the reaction ;
in short the assumption is that if two molecules of Hb,O2 met, they

Effect of salts on aggregation

(51

would resolve themselves into Hb2 and Hb204 as being the more stable
arrangement.

And now does the theory fit the facts ?

FIG. 31. — Dissociation curves of haemoglobin solutions in NaoHPO4, NaHC03, KC1 and

Kxn
NaCl. • Points actually determined. Curves drawn from formula z//100 = - — j^—n-

Ordinate = percentage saturation with oxygen. Abscissa = oxygen pressure in mm.

First let me consider the simplest cases. Below is a table of the
values of K and n which correspond to the various solutions of
haemoglobin which have been investigated.

62

Chapter V

K
n

N
ioo(

90
80
70
60
50
40
30
20
10
0

i2HP04 NaHC03 KC1 NaCl (NH4)2C03 (

•0453 -0155 -00204 -0062 -006257
1-670 1-835 2-49 1-778 3-19

) 10 20 30 40 50 60 70 80

R

INGEF

s So

.UTIO

W^^,

jj— — ~~

_-

-X

6

0

100
90
80
70
60
50
40
30
20
10
n

/

/*

/

7

/ 1

Jx

00

10

20

30

40

50

/

<^_

— Ji

/

90

1

NH

412^

^3

80

V

70

^•-~

— -—

60

/

^

50

r

40
30

10

0

/

BLOC

D

/

/

/

'X

1

/

2

Distilled water
dialysed

•125

10

20

30

40

50

Fie;. 32. — Dissociation curves of haemoglobin solutions in Ringer's fluid and (NH4).>CO.,,
and of blood. x Points actually determined. Curves drawn from formula
A'r"

?//100 = - — jf—n'
pressure in mm.

Ordinate = percentage saturation with oxygen. Abscissa = oxygen

Aggregation theory 63

Whether or no the curves drawn from these data fit the points as
determined may best be seen by actually drawing the curves, and
placing the points in relation to them.

This is done in Figs. 31 and 32.

To pass to a more complicated case — that of a solution of haemo-
globin in a mixture of salts, namely Ringer's fluid: taking n = ^'l 11
and K ='00427, a curve is produced which satisfies the properties
of haemoglobin in Ringer's solution, as well as does the freehand
curve of Barcroft and Camis(9).

From Ringer's solution we may pass to the still more complicated
problem which is presented by blood.

Here we have the advantage, in some cases at all events, of work-
ing with curves upon which much more time and labour have been
spent than have been claimed by the solutions of haemoglobin.

Some forty or more determinations have been made on the normal
blood of Mr C. G. Douglas <10) of St John's College, Oxford, some by
myself, some by Haldane and Douglas ; and as our methods differed
somewhat in detail, the fact that we arrived at the same result proves
that Douglas' normal curve has been determined with as much
accuracy as our present methods will admit of. Yet in spite of
the great number of determinations which have been made, the
curve as drawn freehand through the points is extremely close to
that calculated from the equation, with the values ^='000212,
n=2'5. This curve and the corresponding points are given in
Fig. 33. The curve differs very much from those of the solution
given just previously, inasmuch as the S-shape is much more evident.
This difference is due to the fact that the curve is determined in
the presence of 40 mm. C02 pressure, which was the pressure of that
gas in Douglas' alveolar air and presumably in his body generally.

Each test which we apply, and which is met satisfactorily, leads us
on to another. We were drawn into the discussion which has formed
the subjects of this present chapter and the last by the discovery that
the bloods of different animals had not identical affinities for oxygen.
The theory we are testing must therefore not only hold true for the
blood of a particular person, but it must also hold good for the bloods
of different species.

It has been tested in the case of man, sheep, dog and cat, and
seems to be true for them all.

But the differences which have been found between different
species apply to a less extent to different individuals, at all events to

64

Chapter V

different individuals of the human race. Probably my own blood
has been studied in greater detail than that of any person except
Douglas. It is quite certain that my blood differs from his. It is
also certain that with a slightly different value for K a curve for my

100

90

80

70

60

50

40

30

20

10

10

20

30

40

50

60

70

80

90

100

90

80

70

60

50

40

30

20

10

10 20 30 40 50 60 70 80 90

FIG. 33.— Douglas' blood and Barcroft's blood exposed to 40 mm. C02 pressure. • Points

Kxn
actually determined. Curves drawn from formula ?//100= n . Ordinate =

percentage saturation. Abscissa = oxygen pressure in mm.

blood can be calculated from Hill's equation, which fits the points
which have been determined (Fig. 33).

The normal dissociation curves for Douglas' blood and of my
own have been determined in the presence of CO2 as I have said

Aggregation theory

65

above. The amount of CO2 present has been in each case that
present in the alveolar air. We arrive at a very interesting result,
however, by studying the curves of the same blood with different
quantities of carbonic acid. In a couple of words it is this : the
equations for all such curves seem to* differ from one another in
the value of K only, n remaining constant for the series. This is but
another way of saying that the series consists of a single curve drawn
to different horizontal scales and can be derived from the statement
that the effect of the CO* is to lessen the effective concentration

of oxygen.

100i—

90 100

0 10 20 30 40 00 GO 70 80

FIG. 34.— Dissociation curves of Barcroft's blood. Exposed to 0, 3, 20, 40 and 90 mm. C02 .
Ordinate = percentage saturation. Abscissa = oxygen pressure.

In Fig. 34 n is 2*5 and the values of K for the various CO2
pressures concerned are

CO, pressure in mm 0 3 20 40 90

K -00258 -00130 -000505 -000292 -000135

The fact that n remains constant throughout the whole series of
curves places the theory under discussion upon a wholly different
stratum of probability.

Up to this point we have considered K and n as being merely
mathematical constants though it is true they were arrived at by a
physical process of reasoning ; all that we have claimed for them is
that by suitably changing them we can reproduce the curves obtained
by analysis. Now however we are face to face with the fact that one

* See Appendix II.

B. R. F. 5

6(5 Chapter V

I

of these remains constant over a large series of curves. It is very
improbable that such a constancy is merely fortuitous.

The classical example of this principle was first put forward by
Laplace with reference to the direction of the orbits of the members
of the solar system. In its up-to-date form it is as follows : "...the
tale of the asteroids has now approached five hundred and out of this
huge number of independent planetary bodies there is not a single
dissentient in the direction of its motions. Without any exception
however they all perform their revolutions in the same direction as
the sun rotates at the centre. When this great host is considered
the numerical strength of the argument " (that the arrangement is
referable to a physical cause and is not purely fortuitous) "would
require about 150 figures for expression*." Even before any of the
asteroids were discovered, Laplace considered the argument a strong
enough one to justify the nebular hypothesis.

It may be urged that the analogy is not sound, for it looks like
a comparison of something qualitative with something quantitative.
The planet must go round either clockwise or counter-clockwise,
whereas all that we can say of n is that it remains constant within
the limits of 2'45 and 2'5o over all the curves which have been deter-
mined for human blood. Closer than that we cannot determine it.
Even so, the solar system will not fail us for an example. The orbits
of the seven planets lie in the same plane within 9° of arc. The
chance against this taking place without a physical explanation is
about 10,000,000 : 1. This was Kant's argument in favour of the
nebular hypothesis.

We can calculate the value of n — on any one of the curves of which
we have been speaking — to within about four or five per cent, and
within these limits it is the same for all. I leave it to some
mathematician to say what the chances may be of n being the same
within four per cent, in a dozen curves, when, if it were a perfectly
fortuitous mathematical expression, it might be anything between
zero and infinity in any given case ; but I have probably said enough
to convince the reader that since n remains so constant it is probably
the expression of some definite physical fact.

It is clear that any theory which applies to the formation of

oxyhaemoglobin must also apply to CO-haemoglobin. Therefore the

equation may be put to the further test of applying it to the parallel

data with regard to the dissociation of carboxyhaemoglobin in the

* Quoted from the Earth's Beginning by Sir Robert S. Ball, 1909, p. 316.

Aggregation theory

67

presence of various pressures of carbon dioxide. It is not a little
startling to find that the curves given in Fig. 34 for oxyhaemoglobin
are almost superposable upon those in Fig. 29 given for CO-haemo-
globin by Douglas, Haldane, J. S. and Haldane, J. B.(11) *. In the figure
the curves or some at least of them are drawn in freehand.

100

90

80

70

60

50

40

30

20

10

•08

x OC

« 19 ri.m.COj
* 42ri.m.CO,

o

O79rji.rn.CO;
• A

•16

•24

FIG. 35. — Dissociation curves of CO -haemoglobin.
A'.r«

•32 -40

Curves drawn from formula

j//100=

1 + Kx"

Ordinate = percentage saturation with CO. Abscissa = CO pressure

in mm. Points indicated are the determinations of Haldane and Douglas.

The points determined by these observers have been transcribed
from their figure as faithfully as possible and are shown in Fig. 35.
The figure is drawn to a somewhat different scale from theirs for
the purpose of comparing the oxygen dissociation curves with the

* The aggregation theory has been modified by these authors in certain respects.
My reasons for prepariiijf the original form of the theory are given in the Biochemical
Journal, vn, p. 481.

5—2

68 Chapter V

CO-haemoglobin curves. It is so chosen that the middle point of the
CO curve (Fig. 35, .4) at 42 mm. C02 pressure coincides on paper with
the middle point for the corresponding curve of oxyhaemoglobin.
I then tested whether curves drawn from Hill's equation

fitted the points, with the result which may be seen in Fig. 35. Not
only do the lines fit the points absolutely in the case of each of the
four curves given, but as in the case of oxyhaemoglobin the lines are
all obtained with the same constant value of n, 2*5, and with a change
merely in the value of K. The 42 mm. C02 curve therefore not only
coincides with the similar curve for oxyhaemoglobin at the point of
50 per cent, saturation, but is the same identical curve with the same
value for n. The identity does not stop here, for the curves for other
C02 pressures are identical in Figs. 34 and 35. In fact Fig. 35 is the
exact counterpart of Fig. 34 but for the trifling difference caused by
the disparity between Douglas' blood and my own.

The aggregation hypothesis forms a complete explanation of the
facts that are known concerning the reactions

Hb + 02 ^=± Hb02
and Hb + CO ^ HbCO.

The more complicated case remains for consideration — that of
the reaction

Hb02 + CO ^± HbCO + O2.

There are several remarkable facts about this reaction, any one of
which might prove upsetting to a general theoretical explanation.

(1) The reaction, unlike those which we have discussed, is repre-
sented by a rectangular hyperbola.

(2) This hyperbola, unlike the curves of haemoglobin in the
presence of oxygen, or of CO separately, is almost unaffected by
acids and by salts.

(3) While the curve of the reaction

Hb02 + CO 5^ HbCO + O2

is a rectangular hyperbola in the presence of an ample supply of
O2 and CO, it ceases to be so when there is insufficient CO and O2
present to saturate the Hb, i.e. when there is a considerable quantity
of reduced haemoglobin present. Under these circumstances the
entire form of the curve changes in the way which is shown in
Fig. 36<u>.

Aggregation theory

69

Perhaps the most telling feature of the aggregation theory is its
adequacy to meet these unexpected and peculiar requirements.

At first sight it wholly failed to do so. If two assumptions be
made, the matter, as Hill has shown (8), becomes quite clear. Of
these one has already been made : it is the instability of unsaturated

, . affinity for CO .
compounds ; the other is that the ratio ffi .•; — ^ — ^ is even greater

in the case of the half-saturated than in the case of the completely
unsaturated compounds. With regard to the first assumption it
may be said that (1) no unsaturated oxide has ever been isolated

100
90
80
70
60
50
40
30
20
10

\

10

12

14

16

18

20

FIG. 36. — Heavy line represents dissociation curve representing partition of haemoglobin
between 02 and CO at 38° C. CO pressure = 0-00854 °/0 of an atmosphere throughout.
Light line represents the corresponding hyperbola.

(2) there is no transitional spectrum between that of oxy- and of
reduced haemoglobin.

It remains only to state the numerical relation between K and
the carbonic acid pressure. It will not surprise the reader to hear
that this relation is similar in the cases of CO-haemoglobin and
oxyhaeinoglobin. The two fall on the same curve, the difference
lying only in the scale on which the curve is plotted. The temptation
to speculate on the form of this curve is considerable, but till a greater
number of points are forthcoming the temptation must be repressed.
It must suffice to say that it is not a parabola — the curve which

70

Chapter V

would occur if the influence of the CO2 were simply to cause a greater
or less degree of adsorption of the haemoglobin. It is of interest
to record that the influence of CO2 is as evident at low temperatures
as at high ones. In this respect there is a sharp contrast between the
effect of CO2 and also of the acids engendered by oxygen want, and
the effect of oxygen want itself (a vacuum), on the reaction

Hb + 02 ^±: HbO.2.

•0025

•0020

•0015

•0010

•0005

•0004

•0003-
•00020
-00010

Too

FIG. 37. — Ordiuate = K. Abscissa = C02 pressure in mm. • Oxyhaemoglobin. x CO-
haemoglobin. (Left hand = oxyhaemoglobin, right hand = CO-haemoglobin.)

The reduction of the haemoglobin by the acids goes on both at 37 °C.
and at room temperature with equal activity, but the reduction of
haemoglobin by a vacuum is so slow at low temperatures that one is
tempted to conceive of the presence of acids rather than of absence

Aggregation theory

71

of oxygen as being the factor which determines the reduction of
blood in cold-blooded animals.

In the opening sentences of this book I alluded to the specula-
tions which I sometimes allow myself. I drew a contrast between
morphology and biochemistry in their relation to natural selection.
The fundamental basis of natural selection is variation, but it would

100

0 10 20 30 40 50

0 10 20 30 40 50

Fio. 38. — Dissociation curves of I, haemoglobin in -9 % KC1 in presence of 25 mm. C02 at
15° C. ; II, ditto in absence of CO,; III, blood at 37° C. exposed to 3 mm. C02;
IV, ditto exposed to 20 mm. CO,.

seem that the chemical properties are fixed. They are immutable
properties of the substance. The reader will not have read the
preceding chapters without discovering for himself that a way out
of this apparent impasse has been discovered. At all events in the
case of the reaction Hb + 0,, 5^ Hb02 this is so.

The velocity of this reaction is dependent to some extent upon

72 Chapter V

temperature, that is so much to the good, but the temperature of
the body is roughly speaking the same throughout. It has to suit
itself to a thousand reactions. But in virtue of the fact that the
haematin is allied to a protein the rate of reaction is influenced by
the degree of aggregation of the protein elements, and this may
"vary" with the concentration and nature of the electrolytes in
which they are placed and these electrolytes may differ from place
to place according as is most suitable for the chemical actions which
depend upon them. The same considerations may be applied to
the local effects of acids and alkalis. Here then is a possible basis
for biochemical "variation." The dog has salts of one type in his
corpuscle, the pig salts of another ; the dog's oxyhaemoglobin breaks
up therefore at one velocity, the pig's at another. This is no doubt
as much a matter of natural selection as are the morphological features
of the animals in question. But what is true of the reaction haemo-
globin is probably true of those of other colloids ; we recognise it in
the case of haemoglobin because haemoglobin has a simple reaction
with oxygen which we can investigate.

But I must cease from speculating and turn to the next portion
of my subject, namely the consideration of haemoglobin as a vehicle
for the transport of oxygen in the body.

REFERENCES

(1) Bohr, Hasselbalch and Krogh, Skand. Arch.f. Physiol. xvi, p. 390, 1907.

(2) Barcroft and Orbeli, Journal of Physiol. XLI, p. 353, 1910 ; Barcroft, Ibid.

XLII, p. 44, 1911.

(3) Mathison, Ibid. XLIII, p. 347, 1911.

(4) Hasselbalch, Skand. Arch, xxvn, p. 13, 1912.

(5) Douglas, Haldane and Haldane, Journal of Physiol. XLIV, p. 237, 1912.

(6) Barcroft and Poulton, Ibid. XLVI. ; Proc. Physiol. Soc. Feb. 1913.

(7) Hill, Journal of Physiol. XL. ; Proc. Physiol. Soc. p. iv, 1910.

(8) Hill, Biochemical Journal, vn, p. 471, 1913 ; Barcroft, Ibid. p. 481.

(9) Barcroft and Camis, Journal of Physiol. xxxix, p. 118, 1909.

(10) Barcroft, Ibid. XLII, p. 44, 1911.

(11) Douglas, Haldane and Haldane, Ibid. XLVI, p. 275, 1912.

(12) Barcroft and King, Ibid, xxxix, p. 375, 1909.

PART II

THE PASSAGE OF OXYGEN TO AND FROM THE BLOOD

CHAPTER VI

THE CALL FOR OXYGEN BY THE TISSUES

THE classical work of Pfliiger(1) on the combustion of living
material settled for all time, it seems to me, the logical order in
which the constituent processes of respiration should be treated.

The issue before Pfliiger may be stated in a few words. Is the
quantity of oxygen taken up by the cell conditioned primarily by the
needs of the cell, or by the supply of oxygen ? The answer was clear,
the cell takes what it needs and leaves the rest. Respiration there-
fore should be considered in the following sequence. Firstly the
call for oxygen, secondly the mechanism by which the call elicits a
response, the immediate response consisting in the carriage of oxygen
to the tissues by the blood and its transference from the blood to the
cell. Thirdly in the background you have the further mechanism by
which the blood acquires its oxygen. It is not the habit of writers
on respiration to adopt this order, quite the contrary, but their
reason for placing pulmonary respiration in the foreground of the
picture is a purely practical one — pulmonary respiration is more
evident, both to the eye and to the understanding. I imagine they
will not quarrel with me if I make an attempt to treat the matter
in what appears to be its logical sequence and make some estimate
of the call which the blood has to meet, before entering into a dis-
cussion of how the call is to be met.

The present chapter will deal with the following theme: "There
is no instance in which it can be proved that an organ increases its
activity, under physiological conditions, without also increasing its
demand for oxygen."

The importance of the principle that increased activity of an
organ entails a call for oxygen has been self-evident ever since the
days of Ludwig. The reason why the work has not been carried out
on an extensive scale till recently is because the old methods were
quite inadequate.

Chapter VI

Skeletal 'muscle. Nevertheless I cannot pass over the bold attempt
made by Chauveau and Kaufmaim (2). My reasons for drawing atten-
tion to their work on the levator labii superioris and the masseter
muscles of the horse depend less upon the results which they obtained
(some of which do not altogether inspire confidence), than on the grasp
which they had of the problem. Their work was conceived along
physiological lines ; their idea was to determine the gaseous exchange
of the muscles with the least possible abnormality in the conditions
of the animal. They took no elaborate precautions against clotting
of the blood ; they simply had recourse to the horse as an animal
whose blood did not readily clot. They wanted a considerable
quantity of blood, for their samples for analysis had to be of the
order of lOOc.c. each. They therefore chose a smallish muscle and
one belonging to an animal so large that the bleeding entailed was
not felt by the animal. They record their surgical operation as
being so simple that the animal did not cease chewing its oats while
they were at work : thus they reaped the double advantage that the
muscles which they were studying had (a) a normal stimulus and
(6) a metabolism which was unhampered by anaesthetics whether in
rest or activity. There is one final point in which one could wish that
other workers had been able to follow Chauveau and Kaufmann —
they made an attempt to measure the degree of activity which was
induced in terms of absolute units of energy. For this purpose they
made measurements both (a) of the work done, by attaching a weight
to the muscle, and also (6) of the heat given out by the muscle during
its contractions. Nothing in short could have been more complete
than the scheme of their research. It has been a loss to science that
the actual number of experiments performed was small and that the
complete scheme was not carried through in any one experiment.

The following is an example of the figures which they obtained :

Extent of Gaseous Exchange in the Leeator Labii Superioris
of the horse in c.c. per gram of muscle per minute.

Best

Activity

Oxygen absorbed

COo given out

Oxygen absorbed

CO-2 given out

1

2
3

0-0032
0-0079
0-0028

0-0019
0-0058
0-0026

0-054
0-014
0-010

0-063

0-018
0-013

The call for oxygen by muscle

75

The above figures leave no room for doubt that the quantity
of oxygen used increases during activity.

Recent work on Skeletal Muscle. Quite recently the inquiry has
been pushed a good deal further by Verzar'3', who investigated the
time relations of the oxidation as compared with the muscular
activity. His technique was very different from that of Chauveau
and Kaufmann ; armed with the modern methods by which it is
possible to work on minimal quantities of blood, he made a muscle
nerve preparation of the cat's gastrocnemius muscle. The general
relations of the dissection will be seen in the accompanying figure.

FIG. 39. — G = gastrocnemius muscle. Sc- sciatic nerve. A = artery. Vf= femoral vein.
Fs = saphenous vein. P = pipette. M"=myograph.

The veins which contributed to the femoral vein below the
saphenous were all tied with the exception of that which came
from the gastrocnemius muscle. A cannula was introduced into the
saphenous vein, and when a sample of blood was required the clip
was removed from the saphenous vein and placed on the femoral
vein above the junction of the two vessels. The blood was collected
into a graduated 1 c.c. pipette so that the time taken for 1 c.c. to flow
through the muscle might be measured. The blood was prevented
from clotting by the intravenous injection of hirudin. The muscle
was thrown into contraction by the application of an electrical
stimulus to the sciatic nerve. The muscle lifted a weight, doing
about 70 grm. cm. of work at the beginning of each tetanus.

The following diagrams with the figures on which they are based
show quite distinctly the time relations of the tetanus and the call
for oxygen. The latter takes place :

76

Chapter VI

(1) During the contraction. The adequacy of the oxygen supply
depends upon the rate of blood flow in the muscle and this in turn
depends upon the pressure in the blood vessels.

(2) After the contraction. The response to the call is in every
case at its maximal value after the tetanus passes off, which shows
that the call for oxygen continues for some time after the actual
work is performed which the oxidation is designed to meet.

fIC. 2

500 sec

FIG. 40.— Oxygen used by gastrocnemius muscle. Ordinate = c.c. per gram. Abscissa =
seconds. The dotted vertical lines signify points at which the kymograph was stopped.

C.c. Oo necessary
to saturate
1 c.c. blood

Percentage
saturation
of oxygen

O2 pressure in
blood mm. Hg

Rate of
blood-flow
c.c. permin.

C.c. Oo used by muscl

per min.

_ gm.

Exp.

Venous

Arterial

Venous

Arterial

Venous

Arterial

per —
mm.

16

Normal

•057

•013

67

93

48

100

4-30

•190

•00670

Tetanus

•096

45

37

•67

•056

•00196

61" later

•163

7

11

3-00

•450

•01580

73" ,,

•136

•007

22

96

27

100

2-14

•276

•00960

Tetanus

•158

10

14

•75

•114

•00397

80" later

•156

11

15

2-61

•390

•01360

17

Normal

•046

•006

62

95

45

100

1-54

•062

•00320

Tetanus

•084 (?)

31

32

2-00

•144

•00770

57" later

•061

50

39

4-45

•245

•01250

45-5" „

•056

•007

54

94

41

100

2-07

•102

•00520

Tetanus

•061

50

39

1-62

•088

•00450

50" later

•069

43

36

3-52

•215

•01090

18

Normal

•055

•012

69

93

49

100

1-16

•050

•00303

Tetanus

•085

53

41

2-40

•178

•01080

15" later

•068

62

45

6-00

•336

•02030

11"
J-L »

•071

60

44

3-52

•208

•01260

11" „

•076

•Oil

58

94

43

2-40

•154

•00935

146" „

•072

•007

60

96

44

100

•90

•059

•00358

Total

oxygen

capacity of Wt. of
1 c.c. blood muscle

(c.c. Oo) gin.

•175

28 '6

•121 19-7

•179 16-5

The call for oxygen by muscle 77

Verzar's results have reformed our view as to the terms in which
we should express the increased oxidation which results from in-
creased activity. In treating of Chauveau and Kaufmann's work
we followed their method of stating their result when we said that
the " coefficient of oxidation " was increased say thirty -fold. This is
clearly an inadequate statement of the case. At a certain point in
the train of events the oxidation was increased to that extent. Really
we should ask, How much extra oxygen was used by the muscle as
the result of such and such a piece of work ? In order to answer
this question we must, it is true, consider the events taking place
during the work, but just as necessary is it to consider those taking-
place after it.

The third of the three experiments quoted above gives the most
satisfactory data which we have at the present time for relating the
functional activity of the muscle and the oxidation taking place
within it. The duration of this experiment was long enough to allow
the oxygen consumption to return almost to its original level. We
can therefore, by calculating the whole increase in the oxygen used
during the period of the experiment, obtain a minimal value for that
required — a value which is probably not very far from the true one.
Of the experiment in question one can say (1) that as the result of
the stimulus given, which lasted about 25 seconds, the muscle showed
increased oxygen intake for 220 seconds at least, and (2) that in this
time it used up 753 c.c. of oxygen as against '260 which it would have
used up at its normal rate ; therefore the stimulus which was given
was responsible for at least '5 c.c. of oxygen used by the muscle.

The fact that the increased oxygen consumption of the muscle
survived the increased functional activity must be viewed in conjunc-
tion with the recent work of Hill'4', who investigated the relation in
time of the functional activity of amphibian striated muscle and the
heat-evolution of the same.

Hill's method was as follows : he compared the curve of de-
flection of a galvanometer registering thermo-electrically the rise of
temperature of a live muscle when stimulated, with the curve of
deflection given by the rise of temperature due to electrical warming
of the same muscle after death. He found that the curve of de-
flection (coming back to the base line in 4 or 5 minutes by reason
of heat-loss) was the same for a muscle in nitrogen as for a muscle
warmed after death by a very short tetanising current : there is
therefore in nitrogen no heat-production except in the few moments
immediately following an excitation. In O2 however the curve of

78

Chapter VI

deflection of the live muscle continually diverged from the control
curve due to electrical warming for a short period. The only possible
explanation of this is that heat is being produced by the muscle in
O.2 for long periods after the contraction is over. For examples of
this see Figs. 41 and 42. The fall of the deflection is due to heat-loss
and, where there is no delayed heat-production, is a simple ex-
ponential curve. From the control curve of electrical warming it is
possible to calculate the coefficient of heat-loss in the case of any
particular thermopile used, and from this to ascertain, in the case of
the live muscle in 02, the true curve of heat-production.

160

SO 3C be SV 6c yc fc

FIG. 41. — Curves of galvanometer deflection for live muscle stimulated, and for dead
warmed (control). Also calculated curve of true heat-production of live muscle,
i.e. curve corrected for heat-loss.

It was found that approximately as much more heat is produced
in the 4 or 5 minutes following a single shock in O2 as was produced
in the first few moments after excitation occurred. There is there-
fore, but only in the presence of 02, a very large recovery heat-
production lasting for some minutes after the contraction is over,
which recovery heat-production one can scarce but associate with
the oxidative removal of fatigue products (lactic acid, Fletcher and
Hopkins*12').

It was found moreover that any process, as e.g. a previous tetanus,
which uses up the O2 existing already in the muscle delays or abolishes
the recovery heat-production. Hill concluded that oxygen is used,

The call for oxygen by muscle

79

and the delayed heat-production occurs in recovery processes : and
that these processes cannot occur in the absence of oxygen. He
argued moreover that his results were in favour, neither of the
hypothesis of intra-molecular oxygen, nor of the idea that oxygen is
used as required in the simple processes of energy liberation. He
suggested in fact that his results were in favour of the old view that
oxygen is used largely in the processes whereby the molecular

30

FIG. 42. — Galvanometer deflection for rise of temperature of muscle excited in nitrogen,
and later in oxygen, and finally warmed when dead (control). Note that the curve of the
deflection for the live muscle in nitrogen very nearly coincides with the control curve,
and that the curve for the live muscle placed in oxygen after nitrogen is considerably
displaced to the right.

machine — like a steam-engine charging an accumulator — builds up
bodies containing considerable amounts of potential energy which
(like the accumulator) can be discharged whenever desired on sub-
jecting the tissues to appropriate stimuli : that oxygen is used in
maintaining the activity, the state of potential energy, of the
organism, and is therefore largely used after activity has occurred
in * preparation for the next period of activity*.

The construction which is to be put on these experiments is that
* I am indebted to Hill for the above statement of his results.

80 Chapter VI

the chemical activity, whether expressed as heat or as the call for
oxygen, is not merely something which accompanies the contraction
of the muscle, but the contraction sets going a chain of chemical
events which are necessary for the restitution of the muscle into its
former state ; that the increased functional activity is responsible
for the increased oxidation is certain.

In the light of what has since been published one is inclined to
wish that the actual work done by Verzar on muscles had been of a
more definite character. It consisted of lifting a weight and keeping
it (or failing to keep it) at the height to which it had been lifted.
Clearly this is a very complex process. The first portion of it is readily
to be expressed in grm.-cms., the second is somewhat illusory and at
once brings us to the brink of controversy. We must avoid the
use of vague terms such as statical work. In some further, as yet
unpublished, experiments Hill has shown that the heat produced by
a sartorius muscle of the frog in maintaining for 1 second a tension
of 1 gram weight in 1 centimetre of muscle length is, including
recovery processes, about 25 x 10~6 gram-calories. This exceedingly
large number corresponds to the oxidation of 6 x 10~9 grams of
carbohydrate, which would correspond to an amount of O2 used,
4*4 x 10~6 c.c. In Verzar's experiment there was 0*5 c.c. of O2 used,
as a result of 25 sec. stimulus, or an average amount of 0'02 c.c.
per sec. In a muscle 5 cm. long Hill's number should give about
22 x 10~6 c.c., so that if his result is comparable with Verzar's, the
gastrocnemius used by Verzar would have exerted a tension of
1000 grams weight in an isometric contraction. It is a pity that
Verzar's experiment was not conducted isometrically and the ten-
sion exerted expressed in absolute units. In all future work on the
subject, isometric contractions, as Hill has repeatedly urged, should
be used. In any case it is very striking that the tension that could
be exerted, according to Hill's figures, is exactly of the right order
of magnitude. The gastrocnemius preparation of the cat, as used by
Verzar, could certainly lift about 1000 grams weight. It would be
of the greatest interest to ascertain exactly the amount of O2 used
by a muscle in maintaining a tension, per second, per gram weight
of tension maintained, per centimetre of muscle length. As Hill has
urged, the tension exerted and not the work done is the fundamental
quantity in the muscle: and therefore the O2 used in maintaining
unit tension for unit time on a muscle of unit length is the funda-
mental unit of oxidation.

The Heart The pioneer work on this subject was carried out by

The call for oxygen by the heart

81

Dixon and myself (5) : it was a simple investigation into the question
we are discussing, namely whether the oxygen consumption and the
carbonic acid production of the heart varied with the functional
activity. I call it the pioneer work because in the light of the
beautiful researches which have since been performed on the subject
by Rohde(6) and his colleagues in Heidelberg and by Evans (7) at
University College, London, it seems now as I read it over but " poor
stuff" ; nevertheless, I remember, we were not a little proud of it at
the time it was done, for to tell the truth it tested our powers to the
uttermost and I can only claim for it what Dr Johnson claimed for
the preaching of women, "Sir a woman's preaching is like a dog's
walking on its hind legs, it is not well done but you are surprised to
find it done at all." However it did prove the point at issue in a
primitive sort of way. Since many of the methods of altering the
functional activity which we chose have not been tested in the more
finished work of our successors and as the points which we raised
are those which have been elaborated by them, it forms a suitable
introduction to the later work.

The first point about the work in which it falls short of the ideal is
that it is not strictly quantitative — by which we mean that while the
gas measurements were quantitative there were no other records of
the changes in activity than the obvious alterations shown by the
graphic records of the heart's contraction. There is therefore no
means of judging of the relation between the functional activity of
the heart and the gaseous exchange, other than a comparison of the
figures for the blood gas exchange with the tracing.

This comparison we therefore proceed to give.

(a) The first case is that of cardiac augmentation with adrenalin.

Oxygen used per gram per min.
Period I Before adrenalin
Period II After adrenalin

°'C'

I II

PIG. 43.— Showing two periods, preceding and succeeding injection of adrenalin.

B. R. F. 6

82 Chapter VI

(b) Before and after pilocarpine.

Gas per gram per min.

Oxygen
Period used

I Before pilocarpine 0-050
II After pilocarpine 0-010

I II

FIG. 44. — Pilocarpine injected between periods I and II.

C02

given out

0-048 c.c.
0-003 c.c.

(c) Effect of atropine following on that of pilocarpine.

II

in

IV

FIG. 45. — Eecord of puppy's heart. Upstroke = systole. Period I is normal. Period II
shows the effect after injecting 5 rngs. of pilocarpine and Period III after 20 mgs.
Periods IV and V show the recovery of the heart after two successive doses of
atropine.

Gas per gram per rnin.

Period

I

II

III

IV

V

Oxygen

No drug '033

After 1 c.c. 0-5 °/0 pilocarpine -014

„ 1 c.c. 2 °/0 pilocarpine '009

,, 2 c.c. 2 °/0 atropine -015

•021

3 c.c. 2 °/0 atropine

CO.,
•041
•036
•003
•005
•008

The call for oxygen by the heart 83

(d) The effect of potassium chloride on a feebly beating heart.

Gas per gram per min.

Period Oxygen Carbonic acid

I Before KC1 -030 -061

II After „ -027 -048
„ Later -020 -020

Later . -012 -007

II

FIG. 46.— Before and after KC1.

(e) The effect of barium chloride following upon potassium
chloride (the same heart as above).

Oxygen per gram per min.

Fig. 46. Period II, under influence of KC1 -012

Fig. 47. Period I, after injection of BaCl2 -040

(f) The effect of a further dose of barium chloride which
induces contraction.

Gas per gram per min.

Period I

Period II..

Oxygen
•040
•030

CO,
•010
•013

FIG. 47. — Showing the effect of barium chloride.

6—2

84

Chapter VI

Effect of chloroform.

FIG. 48.— Upper tracing represents record of puppy's heart. Lower tracing = blood pressure
of the perfusing animal. Period I — normal. Period II shows the effect of injecting
20 minims of CHC13 water. The signal mark represents the time during which the
sample of blood was taken.

Gas per min.

Period I. Before injection of chloroform water
Period II. After injection of chloroform water

(K) Stimulation of the vagus.

Oxygen
3-0 c.c.
0-37 c.c.

Carbonic acid

8-8 c.c.
1-9 c.c.

III

FIG. 49. — Record of the movements of the heart of a small cat perfused from the
circulation of a large cat. Upstroke Asystole. Period I = normal. In Period II
the signal mark represents the time of vagus stimulation (coil at 10 cms.). The
third period corresponds to the after effect and in this period the third sample of
blood was taken.

The call for oxygen by the heart

85

Gas per gram per min.

'- ' s^

Oxygen Carbonic acid

Period I. Before stimulation 0'014 c.c. 0-038 c.c.

,, II. During ,, 0-009 c.c. 0-005 c.c.

,, III. After ,, 0-022 c.c. 0015 c.c.

It is sufficiently evident from the examples which we have given
that obvious changes in the activity of the heart run, roughly speaking,
part passu with changes in the call for oxygen.

There are two remaining points in the research which may here
be considered and which I am inclined to emphasise more strongly
than I did at the time, in view of the fact that they have both since
been confirmed by the work of Evans (7).

i II

FIG. 50. — Record of puppy's heart. Period I shows the condition during the incompetence
of the aortic semi-lunar valves (i.e. the isometric condition). Period II shows the
recovery after a tube had been introduced into the left ventricle (i.e. the isotonic
condition).

The first deals with a case in which the change in the activity of
the heart is not evident. It is distended with blood to such an
extent that it cannot contract. The condition of the heart may be
expressed in more than one way ; in our paper we stated that it was
undergoing isometric contractions : in the light of the more recent
work of Rohde((i) and of Hill(S), to which we shall shortly refer, we
would say that each beat expended itself simply in a change of
tension and not in a change of form.

The question then that was thus accidentally forced upon us was
whether, in the absence of evident contractions, there was increased
oxidation. The following is our description of this experiment: —
" The cavities on the left side of the heart were much enlarged.
The heart was endeavouring to contract but the resistance to the

86

Chapter VI

outflow of blood (i.e. the pressure in the aorta) being greater than
the force of contraction, it was performing a series of approximately
isometric movements. At the point where the tracing changes its
character the resistance to the outflow was abolished by introducing
a tube through the wall of the left auricle into the left ventricle
so that now the heart could drive its blood up the tube at each
contraction. In the first period the heart bore some resemblance to
an enlarged heart with incompetent semi-lunar valves : this is
characterised by the rapid pulse and the failure to produce effective
contractions. In the isotonic period the rhythm had returned to

FIG. 51.

its normal rate and the contractions were well marked. The oxygen
consumption in the first period was '174 c.c. per minute, in the
second "169. Another point of interest arises in that the slightly
greater oxygen consumption was associated with a reduced blood-
flow. In the isometric period the blood passing through the coronary
system was 1*9 c.c. per minute, in the isotonic it was 2'6."

The other point to which we would refer is the apparently much
greater irregularity of the figures for carbonic acid than of those for
oxygen. This irregularity is due to a number of causes, such for
instance as the solubility of the gas in the blood and any change in
the acidity (hydrogen ion concentration) of the heart itself, for it

The call for oxygen by the heart 87

has been made clear by Fletcher (9) that an evolution of CO2 follows
from a production of lactic acid in the frog's gastrocnemius.

Dixon and I came to the conclusion that the irregularity was
only apparent and that in reality the changes in C02 production
lagged somewhat in time after the changes in the oxygen intake;
this "hysteresis" of C02 was also apparent in the kidney (10) and
the intestine (11), and as we said above it has been confirmed in the
heart by Evans.

Evans was interested in the effect of increased activity, not only
on the call for oxygen, but also on the respiratory quotient of the beat-
ing heart. In a calculation of the ratio of the * ,, - by the

O2 taken in

organ, this carbonic acid " hysteresis " proved fatal to the application
to his problem of the method Dixon and I had employed. To obtain
valid data it was necessary to integrate the whole of the oxygen
taken in and of the CO2 given out over a long time (20 minutes).
For this purpose Evans had recourse to an ingenious device which
likely enough will prove useful for other purposes. He made a
heart-lung preparation something similar to that used by Stolnikoff
for measurement of the output of the heart. He then circulated air
through the lungs by an artificial respiration apparatus, and measured
the amount of oxygen taken up and of CO2 given out. After an
allowance had been made for the gaseous exchange of the lungs
themselves, these data furnished the respiratory quotient of the heart.

One of the merits of this preparation consists in the ease with
which the work performed by the heart can be varied. Taking the
work as being practically (the output) x (the pressure against which
the heart is working), the work may be increased in either of two
ways: (1) by feeding the right auricle with more blood, and (2) by
increasing the resistance in the aorta. Evans obtained an absolute
rise in the oxygen intake and in the CO2 output when extra work was
thrown upon the heart, but if the heart were in good condition he
obtained a fall in the gaseous exchange per kilogram-metre of work
performed and also a fall in the respiratory quotient. Judged as a
machine the efficiency of the heart was extremely low — from two to
ten per cent.

At the risk of going outside the title of my book I cannot refrain
from an allusion to the work of Rohde, despite the fact that he worked
on an excised heart perfused with Locke's solution ; for on the work
of Rohde I am an enthusiast. His technique may be considered

88

Chapter VI

under two headings, (a) his method for gauging the functional
activity, and (b) that for measuring the oxygen.

The research was performed upon the excised heart of the cat in
every case. A rubber balloon was introduced into the cavity of the
ventricle, and the balloon was distended with water at a known
pressure. Thus it was possible (1) to estimate the pressure against
which the heart was working, and (2) by taking a tracing to obtain
a record of (a) the rise of pressure during systole, and (b) the rate of
the pulse.

17cm

A

E

CJ

O

FIG. 52. — Eohde's arrangement for measurement of the functional activity of the

warm-blooded heart.

H — rubber balloon in left ventricle. A= burette for measuring constant pressure. B =
pressure bottle. C = burette for measurement of capacity of balloon during isometric
contractions. D = quick-silver valve. F-M = Frank's " Feder manometer." W =
wall of incubation chamber. Bore of tubing = 0'6 — 0'7 cm.

The apparatus is sufficiently described by the figure, from which it
will be seen that it is possible, by appropriately turning the taps,

(a) to make the heart beat against a known pressure isotonically, or

(b) by preventing the fluid from leaving the rubber bag to make the
beat isometric and that with any given volume of fluid in the bag.

The technique for measuring the oxygen was also very different
from that used by us. In the first place the heart was perfused not

The call for oxygen by the heart

89

with blood but with oxygenated Locke's solution. Here again little
explanation is needed beyond the statement of the general principle :
the Locke's solution circulated round a closed circuit in which was
placed among other things (1) the heart which deoxidised the solution,
and (2) a chamber from which the fluid could take up another stock
of oxygen before returning to the heart ; the supply of oxygen in
this chamber was kept up by the admission of fresh quantities of the
gas from a reservoir, and the quantity of the gas so admitted was
measured. In this way it was possible to find out how much gas the
heart was using over considerable periods of time and thus to rid the
calculations of momentary fluctuations. The method appears to give
reliable results for time intervals of ten minutes and over.

Of the results which Rohde obtained I will first refer to that of
a heart which, at the same initial pressure, was made to beat both
isotonically and isometrically.

Period of
observation

Kind of
contraction

Initial pressure
cm. water

Oxygen
used

Pulse rate

10.5 —10.20 a.m.

isotonic

43

7-20

138

10.25—10.40 a.m.

isometric

43

7-80

138

11.00—11.20 a.m.

isotonic

43

7'55

138

The more interesting and important part of the work may be best
gathered from the account of one of the 177 experiments which
Rohde performed.

The functional activity of the heart was altered by altering the
pressure in the pressure bottle.

Fig. 53 shows a series of tracings from a heart. The heart was
beating isometrically ; the tracings represent the changes of pressure
as registered by Frank's manometer. Take the first portion which
shows heart beats, what information does it give ? It tells (1) the
frequency of the beats, which is 159 per minute, (2) the initial pressure,
which is 35 mm. of mercury, (3) the maximal increase of pressure, or
as Rohde calls it the pulse pressure, which is 97 mm. In the other
tracings each one of these is changed. The question then is, upon
which of these does the oxygen used by the heart depend? The
answer is, that it does not vary directly with any one, but with the
product of the frequency and the maximal increase of pressure.
To put the thing in another way, if Q is the oxygen used, T the

90

Chapter VI

maximal increase of pressure in the beat and N the frequency,

Q

is a constant quantity.

Isometric curves from left ventricle of a surviving cat's heart

during a 2 hours' research.

Time in seconds.

A quarter of an hour after commencement of circulation with Locke's solution.

Half an hour ditto.

One hour ditto.

One and a quarter hours ditto.
Commencement of rise in pressure.

2 hours ditto.

FIG. 53.

The call for oxygen by the heart

91

The following data taken from the experiment in question
illustrate this point : —

(Q)

(N)

(T)

f\

Period

Initial
pressure

Oxygen used
per min.

Frequency
of pulse

Maximal
pressure

V
NT

9.50—10.10

35

9-45

159

97

613 x 10-6

10.10—10.30

59

10-45

165

106

597 „

10.30—10.50

35 9-75

168

98

593 „

It is perhaps even more clearly shown by the following chart : —

(/?afc

of P

;/se;

Pr

•ssur

of f.

u/se)

0 -

Cons

unpt

'on

l/j

.

-----

1

1

1 —

pres

sure

of pu

fee

r
•
i

numt

er of

pulse

beats

Time in periods of 20 minutes

FIG. 54. — Note the correspondence between the increments of oxygen used and of
the rate and maximal pressure of the pulse.

To the work of Hill(8) on the heat given out by frog's muscle, I
have already referred ; I must return again to it, for he came to the
same conclusion about the heat given out as Rohde had done about
oxygen taken in, namely that with any one contraction it varied
directly with the maximal tension set up in the muscle.

For any heart, the law that ^^ is constant, has of course its

limitations :—

(1) If subjected to very high pressures the amount of oxygen
required begins to go up relatively, in other words the heart is losing
its efficiency, it is being over taxed.

92 Chapter VI

(2) In the natural process of death the same change takes place.

(3) In the case of hearts treated with drugs, numerous depressing
drugs — chloral hydrate, atropine, KCN, muscarine and veratrine — also

cause an increase in the ratio ~~ .

On the other hand, strophanthin, adrenalin and — from the
physiologist's standpoint most important of all — vagus stimulation

produce their changes without any alteration in the ratio

The significance of this latter conclusion is very great, from the
point of view of the theory of inhibition, for it has been held by
Gaskell and others that the essential factor in the anabolic process
was a storage of oxygen.

Now when the vagus is stimulated the absolute quantity of
oxygen used goes down, but so does the number of contractions. One

. Consumption of Oxygen
in cmm.

number of beats
Maximal pulse pressure

f t t t

Muscarine little much Adrenalin
, , A,tropip ,

FIG. 55.

might express the above theory of inhibition as follows. Relatively
to the functional activity of the heart there should be more oxygen
taken up during vagus stimulation than at other times ; the experi-
ments of Rohde seem to show that this is not so.

We are tempted to enlarge upon the future that we see before
the type of experiment which we have been describing ; this however
we will leave to Rohde himself and must now pass to another series
of experiments conceived on the same lines.

The Kidney. Next to the contractile organs which we have just
studied, probably the best attested case of functional activity going
hand in hand with oxygen consumption is that of the kidney. There is,
so far as I know, no case where the kidney does work in which there
is not also an increased oxygen consumption by the organ. Now it
is necessary here to use the words " does work " in a strictly physical
sense, and not in the loose and general way which lends itself to

The call for oxygen by the kidney 93

all manner of false deductions. There is only one method by which
a measurement of the work can be made, viz. by the use of the
second law of thermodynamics and the laws of osmotic pressure.
Whenever a given solution (e.g. the blood), containing any number
of substances dissolved, is separated into two or more separate parts
of different concentrations (e.g. blood and urine), then work is done :
and this work is always positive and can never be negative. It
always requires the liberation of free energy outside to effect the
separation of a solution into two different solutions.

In order to avoid confusion we must emphasise that actually and
commercially (so to speak), to carry out the separation of the several
bodies, far more work will probably be done than is theoretically
necessary. In the same way the theoretical minimum work in foot-
pounds, which it is necessary to do in order to carry bricks up a hill,
is given by the weight of the bricks in pounds multiplied by their
vertical rise in feet : in practice, however, it will inevitably be the
case that very much more work than this will actually be done in
carting the bricks up the hill, depending on the state of the road, the
wind, the friction of the wheels, and the training of the horse, or the
internal friction of the engine which carries them up. But, in spite
of that, the only general and valuable estimate of the work to be done

is the product

(weight of bricks) x (vertical rise in feet),

for the actual amount of work expended depends entirely on the
method adopted, and the mechanism by which the work is done—
and that of course we do not in general know. The secretion of
urine may be regarded as the separation of one fluid, the blood, into
two fluids of the same total volume, the concentrated blood and the
urine. The actual energy used in carrying out the process of secret-
ing a given sample of urine we cannot calculate, until we know the
inner mechanism by which secretion is performed. All we can do is
to calculate, from the "molecular concentrations" of the several salts
in the urine and the blood, a certain quantity W, which is the minimum
work which would have to be impressed on the blood in order to
separate it into concentrated blood and urine. W is always positive
for every conceivable change, as will be shown below, and is obtained
on the hypothesis of reversible changes being carried out in concen-
trating the blood by means of semi-permeable membranes : there are
many processes by which the separation can be carried out, but the
second law of thermodynamics asserts that whatever be the process,

94 Chapter VI

provided it be reversible, the work done in accomplishing it must
be the same quantity W, while if the process be not reversible the
work done will be greater than TF.

Whatever then be the mechanism by which the secretion takes
place, we may assert definitely, and beyond the possibility of error,
that free energy to at least the value W must have been provided,
presumably at the expense of oxidative processes.

How then can we calculate W, the least work which must be done
in the secretion of a volume F litres of urine ?

Let the blood be assumed to contain the several constituents
AU A-2, ... An at molecular concentrations* c1? c2, ... cn, and the urine
secreted from this blood to contain the same bodies in concentrations
GI, Co', . . . cn'. Then the minimum work which it is necessary to do to
separate this urine from the blood is

W=VRT

c' c'

' log -- - (C/ - d) + C2' log -- - (C-2 - C,) + . . . ,

G\ ^2

where R is the gas constant (approximately two calories) and T the
absolute temperature.

£

Now each of the terms c'log -- (c' — c), which may be written

C

log - - ( 1 — ; ) , can be shown mathematically to be positive

__ c \ c j J

for all values of c' and c. Hence whatever type of urine has been
secreted W is the sum of a series of positive terms, and is therefore
positive. Every constituent A, therefore, whose concentration is
greater or less in the urine than in the blood, has added its quota
to the total minimum work it is necessary to do to secrete that
urine.

It has been assumed that the minimum work which is necessary
to separate a given urine can be calculated merely from the freezing
points of urine and blood. This is absolutely fallacious. These lower-
ings of freezing points give no clue as to the value of the expression
for W given above : they merely tell us the value of the two ex-
pressions

C\ ~r C% T • • • T Cji,

and Ci+c-2 + ...+ cn',

and therefore of the part of the expression for W given by

* Concentrations reckoned in gr. mols. per litre.

The call for oxygen by the kidney

Of the remaining terms in the expression for W,

95

the freezing points give us absolutely no evidence. Their use for this
purpose involves a very serious error, for these neglected numbers
may be relatively very large : for example, if AI represents urea,
d' the concentration of the urea in the urine is very large, while d
the concentration in the blood is extremely small. Hence the first
term

e/ log ^

is very large and cannot be neglected. Similarly if we consider some
body occurring largely in the blood, and not so largely in the urine, c'
is small while c is large, so that

VRTc'log-

may be finite and negative. In order therefore to calculate W we
can only gain satisfactory results if we know the concentrations in
blood and urine of each separate important constituent. Without
this information, the results deduced from freezing-point measure-
ments are completely fallacious.

V

large 6oc(y of
6 food

P,

P

V

B,, (B,,)

FIG. 56. — Bn — Bn represents the membrane used for concentrating the substance An,
Bn initial position. (BH) final position.

[An actual reversible process by which the separation might be carried out is as
follows : (It should be noted that I am not proposing in any way a hypothesis of
kidney secretion, but ani merely describing a method by which it is possible to
calculate the minimum work which it is necessary to do in order to separate the
urine).

Let us suppose a membrane Bl permeable to all the dissolved bodies except A1 :
then the substance A: which, in the urine in volume V is at concentration c/, and
which was initially in the blood at concentration cx, must have occupied a volume

96 Chapter VI

V — . Hence if we use the membrane B^ to concentrate this body from a volume

c\

c '
V - - to a volume F, we have to do work against its osmotic pressure. On the

ci

membrane we have two osmotic pressures working, one each side, viz. : the fixed
osmotic pressure p1 of A1 in the original solution, and the gradually increasing
osmotic pressure p of A1 in the gradually concentrating solution. The total work
done is

vc-

f r' \ f c /•' / /-\ /•'

-Pi (F--r} + pdv = -p, V - 1 i --, ) + Vc'RT log - ,
V c ) ]v c\cj 5 c'

according to the gas laws: or finally, since pv-nRT, where n is the number of

c'
molecules of gas considered, putting p-^V - - Vc'RT, we find the total work to be

C

= - V (c,' - c,) RT+ Vc{ R T log ^

ci

Using now a membrane B.2 permeable to all the bodies except A2, we find we
have to do work

- V (c,' -cJRT+ VcJ R T log - ,

C1

in concentrating Az; &c.

The total work W is the sum of all these terms, each of which must be positive*.]

The considerations which have just been applied to the problem
preclude us therefore from drawing any comparison between the
gaseous exchange of the kidney and the so-called work as calculated
only from the freezing points of the blood and urine. The freezing
points give us positive assistance in another direction, however, for
they show that as between a kidney which is not secreting and one
which is secreting the latter must be doing positive work in virtue
of its secretion, unless the urine and the plasma are of the same
crystalloid concentration for each salt separately. The urine secreted
in response to such diuretics as sodium sulphate and urea is not of the
same saline concentration as the plasma, as shown either by actual
determinations of the salts, or by the depression of the freezing points.
Since the gaseous exchange of the kidney has been determined in
many of these experiments on sodium sulphate and urea diuresis, we
may proceed to see whether there is a call for oxygen in such cases.

Let us be clear, however, how far we are relying on our freezing-
point determinations; we are merely using them as an indicator
to show that if the freezing points of the blood and the urine are
different, the saline concentrations of one or more salts cannot be
identical in the two fluids, and therefore work has been done. The

* The above discussion of the work performed by the kidney has been con-
tributed by Mr A. V. Hill.

The call for oxygen by the tissues

97

opposite case may of course exist, that in which the freezing points
are the same, but in which nevertheless there must have been work
done because the molecular concentrations of the salts, though
jointly the same in each fluid, are severally different.

Data of four experiments on the work of the kidney.

Kidney secreting not at
all or to a trifling extent

Kidney secreting

Oxygen used

Urine

Oxygen used

Urine

Qualitative evidence of work

fO-57 c.c.
\0-64 c.c.

0-00 c.c.
0-07 c.c.

2-95 c.c.
1-14 c.c.

0-85 c.c.
0-34 c.c.

AB = 0-o4, Ar=l'41
AB = 0'61, Ar=ri7

1-66 c.c.

0-00 c.c.

5-58 c.c.

1-53 c.c.

AB = 0-62, Ar=l'08

Analysis of urine

0-04 c.c.

0-03 c.c.

0-05 c.c.

0-04 c.c.

0-09 c.c.

0-09 c.c.

2-1 c.c.

2-8 c.c.

Sulphates as ]
Urea = 0-28%

NaCl = 0-39 °L

i = l"25° I0

i = 0-25°/0

Surgically, experiments of this character entailed a good deal of
difficult operating, which was carried out in the first two cases cited
above by Brodie, and if one followed the example of the surgical
text-books one would call it " Brodie's operation." These operations
were performed by him on dogs — the second two experiments having
been upon cats. The operative procedure has not been materially
altered in the numerous operations which have since been per-
formed in the Cambridge Laboratory by Straub, Knowlton, Winfield,
Neumann and myself, on the kidneys and suprarenals of cats and
rabbits. The problem was to perform the necessary manipulations
upon the kidney without having the experiment upset by the surgical
shock involved in the exposure and manipulation of the intestines.
Its solution lay in the fact that the shock might be obviated by
getting rid of the intestines. The rectum was therefore tied in two
places and cut between the ligatures; the same was done to the
following structures, in order — the inferior mesenteric artery, the
oesophagus, the coeliac axis, the superior mesenteric artery, and
the portal vein. Of course any minor vessels must also be ligatured ;
by this procedure the whole of the intestines, spleen, pancreas and
stomach may be removed, and with a little practice the blood pressure
at the end will be as high or higher than at the beginning of the
experiment. The animals were always anaesthetised, usually with
urethane. The blood from the kidneys was withdrawn through a
cannula placed in the vena cava, below the entrance of the renal

B. R. F.

98 Chapter VI

veins. It was then easy to divert the blood from these vessels into
the cannula, and measure the rate of flow as it emerged.

The Salivary Glands. The evidence which exists with regard to
the submaxillary gland is of the same general character as that which
we have given for the kidney. In this case also the deficiency lies on
the side of quantitative proof of the amount of work done. It is
known however that saliva is always poorer in salts than is plasma,
though the two differ when the secretion is less rapid than when
it is slow. We can therefore affirm positively that whenever the
gland secretes it must do positive work, inasmuch as the secretion
differs in its salts from the plasma, though we are unable to state
precisely even the minimal value of such work.

Of the salivary glands the submaxillary is that on which most work
has been done.

The gross fact is that stimulation of the chorda tympani produces
an increase in the metabolism of the gland. This could be shown
even in the old days of the blood-gas pump. The following are some
data from the dog.

Gaseous exchange in c.c. per minute (u).

Besting Active chorda

Exp. O2 C02 02 C02 Saliva secreted per mm.

11(1) 0-32 0-20 1-20 1'58 1'6 c.c.

,,(2) 0-30 0-20 0-93 0'60 1'6 c.c.

,,(3) 0-29 0-20 0-59 0'75 1-3 c.c.

IV 0-12 0-17 0-25 0-25 0-9 c.c.

V 0-12 0-12 0-54 0-60 I'O c.c.

VI 0-56 0-60 1-06 1-86 1-0 c.c.

In the few experiments which have been performed the amount
of oxygen used by the gland per cubic centimetre of saliva is more
uniform than might have been expected.

Kate of flow of saliva
Exp. per min. in c.c. Oxygen per c.c. of saliva

1 2-1 0-48
3-2 0-45
2-5 0-43

2 1-8 0-66
1-55 0-60
1-2 0-5

3 0-37 1-05
1-05 1-00

4 1-4 0-50
2-1 0-50
2-5 0-44

The call for oxygen by the tissues

99

It must however be remembered that these results were obtained
by methods which are now quite out of date ; moreover the presence
of CHC13 in the blood may have caused some inaccuracy unsuspected
at the time.

The more recent developments of the discussion of the relation
of activity to oxygen used, as in the case of striped muscle, have

250-TTT300

350 400 45O 500 550 600

FIG. 57. — Ordinate = volume of oxygen used in c.c. per minute. Abscissa = time in seconds.
Upper signal = duration of flow of saliva. Lower signal = duration of chorda stimula-
tion. I, short stimulation. II, long stimulation. Ill, fatigued gland.

gone to show that the oxygen is not used entirely at the time of the
secretion, but that it continues to be used for some little time after-
wards. This is not only so in the case of stimulation of the chorda
tympani, but it is also the case when saliva is elicited by the injection
of adrenalin. The experiments have been done on the cat.

Fig. 57 I, II and III shows the relation in time between the flow

7—2

100

Chapter VI

of saliva and the call for oxygen which is initiated by stimulation
of the chorda tympani.

No. I (Fig. 57) gives the rate of oxygen consumption which follows a
short stimulus, 20 seconds, while No. II shows the consumption follow-
ing a stimulus of ten times the duration. Not only does the oxygen
consumption long outlast the stimulus in each case, but it does so
for a much longer time in the case of the longer stimulus. This is

FIG. 58. — a, b, c, oxygen used in active as compared with resting glands in individual
exps. d, line represents mean rate of salivary secretion in c.c. per minute in the
three Exps. S — Sybase line for saliva. Black area — oxygen used by the active as
compared with the resting gland. 0 — 0 = oxygen base line.

probably a fatigue effect to some extent, for in No. Ill, in which the
animal was in bad condition as a result of surgical shock, the call for
oxygen only manifests itself very slowly and has scarcely passed its
maximum at the right of the figure, that is to say almost six minutes
after the stimulus has ceased.

Nor is it only as the result of chorda tympani stimulation that
we have this prolonged call. In the cat it follows upon stimulation

The call for oxygen by the tissues 101

of the sympathetic as well, at least on the accepted theory that
adrenalin action is sympathetic secretion.

Fig. 58 shows the oxygen consumption which results from injection
of 1 c.c. of Yotroo' adrenalin into the jugular vein of a cat in three
experiments. These ran an almost even course, so that it seemed
allowable to average out the results. Their mean is given at the
bottom of the figure. From it we can not only see that the oxygen
use outlasts the secretion of saliva by some considerable time, but we
can also calculate with a fair degree of accuracy the quantity of
oxygen which is used as the result of a certain quantity of salivary
secretion.

The Pancreas. Sufficient has probably been said about the sub-
maxillary to show that a great deal of knowledge has been gained
about it within the last ten years ; nevertheless this has only sufficed
really to open the door for a more exact investigation of the whole
problem. We have done enough to convince our readers that the
call for oxygen follows the functional activity of the gland, but we
want a much more rigid definition of the latter. Of the other organs
which have formed the subject of research, the pancreas'15', the liver'16'
and the suprarenal (17) glands, there is little which need be said; in the
case of the two former, experiments have been performed which clearly
increase the activity of the glands, but we have no measurements in
units of the increase.

The chief interest of these experiments perhaps lies in the very
diverse stimuli which are used for the purpose of bringing about the
increase. This is an important matter, because it is clear that if
functional activity, when produced by stimuli of the most diverse
kinds, evokes a call for oxygen, we are on a much safer footing in
supposing that the oxygen want is the direct result of the functional
activity.

So far as the pancreas was concerned the stimulus used was of
course secretin. The experiments were performed in collaboration
with Prof. Starling shortly after he and Prof. Bayliss discovered
the mechanism of pancreatic secretion. The experiments consisted
in isolating the tail of the pancreas in the dog from the rest of the
organ ; this may be done without upsetting the circulation. A cannula
is so placed that the blood from this portion may be collected with
ease, its rate of flow being measured at the same time and its gases
analysed. The experiments were done in the early days, before the
introduction of the differential method of blood-gas analysis, and

102

Chapter VI

whilst the ferricyanide method was still on its trial. We therefore
made a great point of checking the results obtained by the ferri-
cyanide method by those obtained with the blood-gas pump. After
arterial and venous samples had been collected for the determination

Volume of
blood emerging
from gland ---10

Water lost by
blood

0

4 Minutes

of the metabolism of the resting gland, an injection of secretin was
made, and when the juice flowed another set of samples was taken.
We made no allowance for the concentration of the blood as none
seemed to be necessary. Perhaps as we have not referred to this

The call for oxygen by the tissues 103

matter before we may therefore explain more particularly what
we mean.

When a gland secretes, water is taken from the blood. The result
is that the corpuscles in the venous blood are more numerous than
in the arterial blood per cubic centimetre. To find out how much
oxygen the blood has lost in its passage through the gland, it
is necessary to know the sum total of the oxygen lost by each
corpuscle ; therefore the amount of oxygen in the venous blood must
be subtracted not from the amount of oxygen in the same volume
of arterial blood, but from the amount of oxygen in a volume of
arterial blood which contains the same number of corpuscles (or the
same amount of haemoglobin) as are contained in the venous blood
collected.

In the case of the submaxillary gland exact measurements have
been made. The following figure shows the relation of the volume of
venous blood passing through the gland during stimulation of the
chorda tympani, to the volume of the saliva secreted and to the
amount of water lost by the blood. It will be seen that the two
latter are almost equal in amount ; the slight excess of the water
lost by the blood over the saliva secreted is to be accounted for no
doubt as lymph (18). To obtain the volume of arterial blood from
which a given volume of venous blood was derived one multiplies the
volume of saliva by I'l and adds it to the volume of the venous blood.

If the conditions in the pancreas and the kidney be considered in
the same way, it will be found, for experiments of the order of accuracy
of those with which we have been dealing, that the volume of pan-
creatic juice or urine is so small as compared with that of the blood
that no correction need be introduced for the concentration.

The following results for the oxygen used during rest and activity
were obtained by the ferricyanide method :

Oxygen absorbed per min.

Besting pancreas Active pancreas Response to injection

Exp. 1. -49 c.c. 1-71 c.c.1

•60c.c. 3-50C.C.J g°C

Exp. 2. -25 c.c. -51 c.c.)

•34 c.c.} g°°d

Exp. 3. -12 c.c. -28e.c/)

2-1 c.c. in 8'
•11 c.c. -34 c.c.J

Exp. 4. -08 c.c. -33 c.c.^l

•06 c.c. -29 c.c./

•9 c.c. in 5'

104 Chapter VI

With the blood-gas pump :

Oxygen absorbed per min.

Besting pancreas Active pancreas

•40 c.c. -53 c.c.

•23 c.c. -31 c.c.

The Liver. Lastly we come to an organ which is perhaps more
obscure than any of which we have yet treated — the liver. It is
possible to express the mechanical work performed by muscle for it
is doing a definite thing ; it is possible in the case of the kidney to
lay down lines for calculating at least the minimum work done by
that organ ; but who shall express in units what the liver is about.
Its functions are so manifold and in many cases so ill understood, and
the evidence of those functions is so difficult to estimate even when
they are understood, that at present there seems to be but little hope
of getting any accurate notions of its work. All that we can do is to
attempt to excite it by what we may regard as its normal stimulus,
namely the presence of food in the intestine. It seems at least a fair
assumption that the liver will increase in activity during digestion.

But the estimation of the oxygen used by the liver offers a very
difficult surgical problem ; fortunately it fell into the hands of a
skilful operator in the person of Shore, and it proved possible to
attain reliable results in a very considerable percentage of the
experiments.

The blood had to be collected from

(1) an artery,

(2) the portal vein,

(3) the hepatic vein,

in such a way that one could obtain one's samples and measure the
rate of flow in each of the two last vessels, without upsetting the
vascular conditions of the liver.

This is perhaps scarcely the place to describe the operative pro-
cedure in detail ; in a few words however the blood which runs into
the inferior vena cava from all organs except the liver is conveyed in
a hirudinised animal round to the superior vena cava. The vena cava
inferior is then tied above the renal veins and a cannula inserted
just above the ligature. By pinching the vena cava in the region of
the diaphragm, the blood from the liver may be collected by the
cannula. For the collection of the portal blood a cannula is intro-
duced into the splenic vein. Into this cannula the blood may be

The call for oxygen by the tissues

105

diverted. The greatest care must be taken to insure that the blood
is collected at the pressure at which it normally is in the vein.

We performed two series of experiments. The determinations
were made as a rule about noon ; in the first series the animals were
given a clear day without food having been fed the evening before
that, there being therefore about 42 hours from the time of the last
meal till that of the experiment. In the second series the animal,
a cat in every case, was fed the evening before the experiment, i.e.
about 18 hours previously. In neither case was much food found in
the intestines and therefore it is not surprising that the difference
in the metabolism of the abdominal viscera which drain into the
portal system was trifling. But the metabolism of the liver was
markedly higher in the case of the fed animals than in the case of
the unfed ones, as the following figures show :

Animals fasting

Animals fed

^ ^N

Oxygen used per gram per mil

A

Exp.

/ —
Portal organs

Liver

1

•012 c.c.

—

2

•013 c.c.

•005 c.c.

3

•012 c.c.

•005 c.c.

4

•Oil c.c.

•007 c.c.

5

•013 c.c.

•017 c.c.

6

•013 c.c.

•012 c.c.

7

•008 c.c.

•018 c.c.

Average

•012 c.c.

•Oil c.c.

Oxygen used per gram per mil

x

Exp.

/•• —
Portal organs

Liver

t

•034 c.c.

3 <

•Oil c.c.

(

•045 c.c.

9

•018 c.c.

•050 c.c.

10

•018 c.c.

•030 c.c.

f

•013 c.c.

•034 c.c.

\

•015 c.c.

•029 c.c.

12

•016 c.c.

•024 c.c.

•015 c.c.

•035 c.c.

In our discussion of the call for oxygen we have reviewed the
activity of many organs of the body, muscle, heart, kidney, secreting
glands and absorbing epithelium ; these organs are excited by the
most diverse forms of stimulus, electrical stimuli, hormones, drugs,
&c., and evince their activity by doing work of the most diverse
kinds ; in one respect only do they resemble one another, namely that
in no organ excited by any form of stimulus can it be shown that
positive work is done without the blood supply having to respond to
a call for oxygen.

106 Chapter VI

REFERENCES

(1) Pfliiger, "Ueber die physiologische Verbrennung in den lebendigen Organ-

ismen," Pftiiger's Arch, x, p. 350, 1875.

(2) Chauveau and Kanfmann, Comptes Rendus de VAcad. Frang. cm, p. 1063 and

civ, p. 1765.

(3) Verzar, Journal of Physiol. XLIII, p. 243, 1912.

(4) Hill, Ibid. XLVI, p. 28, 1913.

(5) Barcroft and Dixon, Ibid, xxxv, p. 182, 1907.

(6) Rohde, Arch.f. Exp. Path, und Pharmacol. Bd. 68, p. 401, 1912.

(7) Evans, Journal of Physiol. XLV, pp. 230, 231, 1912.

(8) Hill, Ibid. XLII, pp. 40 et seq., 1911.

(9) Fletcher, Ibid, xxm, p. 10. Fletcher and Hopkins, Ibid, xxxv, p. 301.

(10) Barcroft and Brodie, Ibid, xxxm, p. 67, 1905.

(11) Brodie and Vogt, Ibid. XL, p. 167, 1910.

(12) Fletcher and Hopkins, Ibid, xxxv, p. 247, 1907.

(13) Barcroft and Piper, Ibid. XLIV, p. 359, 1912.

(14) Barcroft, Ibid, xxvn, p. 31, 1901.

(15) Barcroft and Starling, Ibid, xxxi, p. 491, 1904.

(16) Barcroft and Shore, Ibid. XLIV, p. 296, 1912.

(17) Neuman, Ibid. XLIV, p. 188, 1912.

(18) Asher (for discussion in relation to lymph flow), Biochem. Zeitschr. xiv, p. 75,

1908.

CHAPTER VII

THE CALL FOR OXYGEN CONSIDERED AS A PHYSIOLOGICAL TEST

AN advantage of the assurance that every increase in the activity
of the cell means an instant call for oxygen lies in the fact that it
furnishes a method of deciding whether in certain cases there is or is
not increased activity on the part of the cell.

The most obvious instance in point is furnished by the kidney.
When Ringer's solution is injected into the blood of a cat or a rabbit,
there is an immediate increase in the amount of urine secreted. Yet
so far as may be judged from the nature of the secretion there is no
adequate reason to suppose either that there is or that there is not
increased activity on the part of the cells of the kidney. It might
be very plausibly supposed on the one hand that the mere fact of
increased flow of urine was an index of increased cellular activity—
a view which I myself held till a few years ago. On the other hand
in this particular case the urine secreted is, so far as its crystalline
constituents are concerned, of the same composition as the plasma.
Therefore it is possible theoretically for the urine to be excreted as
the result of some change in the vascular conditions, the energy
necessary for the filtration being supplied by the heart.

We* at once asked ourselves, Is there a call for oxygen? The
answer is sufficiently shown by the following experiment, a chart
of which is given in Fig. 60.

The result is clear, there is no call for oxygen, no evidence of
work done by the cells as would have been the case if for instance
a solution of sodium sulphate had been injected, or some drug which
essentially altered the composition of the urine. This fact is sufficiently
shown by the following chart in which the oxygen used by the gland
and the volume of urine secreted are shown both during a diuresis
caused by sodium sulphate and one caused by Ringer's solution.

* Dr Hermann Straub and myself.

108

Chapter VII

We need not however stop at this point. Nothing could be more
unsatisfying than to prove that such a diuresis was due to mechanical

FIG. 60. — Line = oxygen consumption. Black area = urine excreted.

o-i -

005-

FIG. 61. — Line = oxygen consumption. Black area = urine excreted.

causes without making any effort to see what the mechanical causes
at work might be.

The call for oxygen considered as a physiological test 109

In the experiment which I have quoted a number of changes
took place in the vascular conditions, any one of which might easily
have had an effect on the flow of urine, all of which may have con-
spired in this matter. There was for instance an increase in the
general arterial pressure, an increase in the rate of flow, a decrease
in the viscosity of the blood and a decrease in the concentration of
proteins in the plasma. We can proceed to eliminate these one by
one and see whether a diuresis follows, a diuresis which preserves its
characteristic feature of taking place without any call for oxygen.

OH

Girpuscfe Blood

I I

Corpusde

FIG. 62.— Rabbit. FIG. 63.— Rabbit.

Oxygen and urine plotted as in Fig. 60.

The change in the blood-flow and the general arterial pressure
may be eliminated together.

The method of performing the experiment which is least up-
setting in every way is to suspend, in the Ringer's solution, red blood
corpuscles, then to remove a certain quantity of blood and replace it
by the suspension of corpuscles. If this is done, one gets a very
considerable diuresis, the vascular conditions of the kidney remain
practically unaltered, the composition of the blood remains unaltered
as regards salts, and there is no tendency to anaemia. One factor only
has been introduced, the plasma is less concentrated in protein : yet
a copious flow of urine is at once set up ; the question faces one,

110 Chapter VII

Is this or is it not due to increased activity on the part of cells of
the kidney ? *

The record of such an experiment is given in Fig. 62. The animal
which was the subject of the experiment was secreting 0'05 c.c. of
urine per minute, its blood pressure was 95 mm. of mercury ; 22 cubic
centimetres of blood were taken out and 25 c.c. of Ringer's solution
wrere injected : the secretion at once rose to 0'4 c.c. per minute (see the
notch on the record of the diuresis at 12.23), the arterial pressure being
then but 52 mm. The suspension of corpuscles (25 c.c.) was put into
the jugular vein, the arterial pressure at once rose to 84mm., and the
diuresis reached the very large figure (for a rabbit) of 2'35 c.c. per
minute. The rate of flow of blood through the kidneys was slightly
slower in the diuretic period than before it, being 1 c.c. in 4'3 seconds
as opposed to 1 c.c. in 2'9 sees. The urine attained a value of 0'95 °/0
of chlorides during the diuresis, and the oxygen taken in presented
scarcely any variation.

During
Before diuresis — * , After

Oxygen taken in per gram of kidney per minute 0-104 c.c. 0-108 0-105 0-09

Similar results were obtained in a second experiment of the same
nature (Fig. 63). In it the oxygen taken in before the diuresis was
O'll c.c. per gram per minute, during the flow O'lO.

At this point something may be said about the theory of caffeine
diuresis. Up to the present time two theories have been put forward
to explain it, (1) that the caffeine acts as a specific stimulant to the
kidney cells and (2) that it acts by causing vaso-dilatation, accom-
panied to some extent by a paralysis of the hypothetical reabsorptive
mechanism of the tubules.

The idea that mere dilution of the protein constituents of the
plasma could cause a copious diuresis without any activity on the
part of the cells of the kidney was to us so interesting that we felt
bound to pursue it further. There seemed to be two possible ways
of explaining it. The first of these was that owing to the decreased
viscosity of the blood the pressure in the capillaries was greater than
formerly, the arterioles not damping the pressure to their normal
extent, and secondly, an explanation might be found by expanding
a conception put forward some years ago by Starling (2). To explain
the fact that the flow of urine normally stops when the arterial
pressure is abnormally reduced, Starling pointed out that the proteins

* The main points in the following discussion on the kidney have been confirmed
by the independent work of Prof. Tangl of Bnda-Pest.

The call for oxygen considered as a physiological test 111

in the blood exercised an osmotic pressure of about 25 — 30 mm. of
mercury. In so far as filtration could account for the flow of urine,
filtration could only take place and therefore urine could only flow
when the capillary blood pressure was greater than 25 mm., since the
proteins do not go through the wall of the glomerulus. In short the
available pressure for filtration is the difference between the capillary
blood pressure and the osmotic pressure of the proteins of the plasma.
A necessary corollary to this clearly is that if you dilute the proteins
and therefore lower their osmotic pressure you increase, other things
being equal, the available pressure for filtration.

The issue between these two explanations was taken up by

o-H

0-05-

&>

.5

Rabbit.

I

Caffeine

"T ' i

900

Caffeine

FIG. 64.

talfeine. Canine.

FIG. 65.

Oxygen and urine plotted as in Fig. 60.

Knowlton(3). His experiments may be summed up very shortly ;
always obtaining a diuresis which produced no work on the part of
the kidney. He showed :

(1) That if gelatine was put in the Ringer's fluid in such quantity
that osmotic pressure of the gelatine was approximately equal to that
of the proteins of the plasma, while the viscosity of the solution was
approximately that of defibrinated blood, relatively little diuresis
was produced. This is seen in the chart given in Fig. 66. Three
injections were made :

(a) Ringer's solution without gelatine.

(b) Ringer's solution and 5 °/0 gelatine.

(c) Ringer's solution with gelatine.

112

Chapter VII

The oxygen used was the same in each of the three cases ; there-
fore any difference in the quantity of urine was not due to a different
degree of activity in the kidney. Yet in the second case there was
almost no diuresis whilst in the first and third there was copious
diuresis.

(2) In control experiments starch in place of gelatine was added
to the Ringer's solution. The addition of soluble starch produces no

SALINE

GELATINE SALINE
SALINE

SALINE STARCH
SALINE

SALINE

STARCH
SALINE

FIG. 66. — Urine = black area. Blood pressure = broken line. The black line just below
the blood pressure shows the oxygen consumption by the kidney. Abscissae = time in
hours.

increase in the osmotic pressure of Ringer's solution but the fluid
has twice the viscosity of blood : yet when injections of Ringer's
solution with and without starch were made alternately the diuresis
was almost the same in every case.

(3) A 5 per cent, solution of gum acacia has almost no greater
viscosity than has the saline solution, but on the other hand it has a
considerable osmotic pressure though not quite equal to that of the
plasma of the gelatine. The result was the same as that of injection

The call for oxygen considered as a physiological test 113

of the gelatine though corresponding to the slightly lower osmotic
pressure ; there was just a little greater diuresis than in the case of
the starch, though still the diuresis was but small as compared with
that of the Ringer's solution which contained no gum acacia.

•*.

3
cc.

2 -

cc

cc

III 10

GUM c«l IMF

ACACIA S*LINE

II.3O

1 2.. 00

GUM
ACACIA

SALINE

SULPHATE

GELATINE
SULPHATE

SULPHATE

FIG. 07. — Urine and blood-pressure plotted
as in Fig. 66.

FIG. 68. — Urine, blood-pressure and
blood-flow through the kidney
plotted as in Fig. 66.

The following are the data of the osmotic pressure :

Solution Temp.

5 °/0 gelatine in normal saline 37° C.

5 °/0 gum acacia in normal saline 37° C.

3 °/0 soluble starch in normal saline 37° C.

Osmotic pressure
of colloid

23 mm. Hg

12 mm. Hg

2 mm. Hg

These figures for gelatine and for gum acacia are considerably
lower than the figures given by Moore and Roaf. This is to be
explained in part by the fact that they were dissolved in Ringer's
solution instead of distilled water and thus the "solution aggregates"
of the colloids were altered.

B. K. F. 8

114 Chapter VII

The comparative viscosity of the solutions used was determined by
measuring the time of flow of a given quantity through a tube of small
bore. The apparatus used was kindly lent by Mr W. B. Hardy. It con-
sisted of a capillary tube and arrangements for causing flow at constant
temperature and under constant pressure. At a temperature of 37° C.
and pressure of 37 mm. of water, the results were as follow :

Comparative viscosity from average time of flow.

Distilled water 1 min. 5 sec.

Gum acacia (5 °/0) in saline ... ... 2 ,, 22 ,,

Defibrinated blood ... ... ... ... 5 ,,

Gelatine (5 °/0) in saline ... ... ... 6 ,,

Soluble starch (3 °/0) in saline 10 ,, 20 ,,

It is clear therefore that the efficiency of starch, gum acacia
and gelatine in counteracting the effect of the Ringer's solution is
in proportion to their osmotic pressures and not in proportion to
their viscosities. Nor can this effect be attributed to changes in the
rate of flow of blood through the kidney, to the general arterial
pressure or to the activity of the cells.

But if further proof of the mechanical nature of Ringer diuresis
be needed it may be found in the fact that these characteristic
actions of gelatine and starch can only be obtained on those forms of
diuresis which are unaccompanied by a call for oxygen. Salt solution
seems frequently to act in just the same way ; in such cases what we
have just said about Ringer's solution might with equal truth have
been said of a solution of sodium chloride. Sometimes however
sodium chloride may cause a secretion by the tubules as well as
its mechanical diuresis.

We have already shown that diuresis obtained by sodium sulphate
is not of this character. There is increased work done by the kidney
as evidenced both by the call for oxygen and by the fact that the urine
is, or may be, almost a pure solution of sodium sulphate.

The presence of gelatine in the sodium sulphate solution injected
does not counteract the action of the sodium sulphate or at least it
does so only to a trifling extent, for although sodium sulphate has a
specific secretory action it must of course have a salt action as well.

The information which has been recounted in the preceding pages
appeared to be in conflict with a statement which is often put
forward as being one of the fundamental facts with regard to urinary
secretion, namely that a kidney, the artery of which has been clamped,
does not secrete urine for a long time. Winfield and I determined

The call for oxygen considered as a physiological test 115

therefore to test the matter for ourselves and find out if possible the
limits within which this statement is true. It is clear at the outset
that, while it may be perfectly true that such a kidney does not
secrete, it is quite a different matter to suppose that it cannot be
made to do so by appropriate methods.

We first tested the matter with Ringer's fluid and with sodium
chloride on the one hand, these being diuretics which can act
mechanically and elicit no call for oxygen. In contradistinction to
these we used caffeine, this being as we have already seen a diuretic
which acts solely in virtue of its stimulating effect upon the epithelium
of the kidney and has no salt effect. The result was perfectly clear
after asphyxiation of the cells of the kidney by clamping ; the sodium
chloride produced a copious diuresis, the caffeine on the other hand
produced no diuresis whatever.

The following data are those of three experiments in which
diuresis was evoked, in the case of Exps. 1 and 2 by injection of
Ringer's fluid, in the case of Exp. 3 by injection of 10 c.c. of 5 per
cent, sodium chloride, the animal in each case being a rabbit.

Exp.

Normal
flow of
urine
c.c. per min.

Normal
blood-
flow
c.c. per min.

Saline diuresis
before
clamping
renal artery

Corresponding
blood-flow

Length of
time of
clamping
artery

Second
saline
diuresis

Corre-
sponding
blood-flow

If-)

o-i

1-75 c.c.

10 sees.

1*3 c.c.

_

1(6)
2

0-1

0-4

24

2-2 c.c.

3-8 c.c.

24 c.c.

5 min.
10 min.

1*3 c.c.
1-3 c.c.

9 c.c.

3

0-05

48

1-3 c.c.

63 c.c.

18 min.

0-9 c.c.

24 c.c.

It is clear that any diminution which takes place in the diuresis
is amply accounted for by the less satisfactory vascular condition,
and in any case there is in each experiment a good diuresis after the
artery has been clamped for periods varying from 10 seconds to
18 minutes.

On the other hand the following experiment will show how different
is the effect of clamping upon the diuresis obtained by caffeine :

Duration of
clamping

Diuresis caused
by '01 grm. of
caff. sod. ben.

Diuresis caused by
10 c.c. 5 °/0 NaCl

Left kidney

20 minutes

0-00 c.c.

0-4 c.c.

Eight kidney ...

not clamped

0'5 C.C.

1-2 c.c.

8—2

116 Chapter VII

The statement then that the kidney does not secrete after being
clamped is subject to certain limitations. If the word "secrete"
is used in the specialised sense it is true, but it can excrete if the
mechanical conditions are so altered as to promote increased
filtration.

Another problem very similar to that which we have just studied,
in which the " call for oxygen " has been used as a test of the
activity of the cells, is furnished by the work of Brodie and Vogt.
The question was, Does the passage of water through the cells of the
intestine involve the activity of the intestinal epithelium ?

First let me consider cases in which the fluid is passing from the
lumen of the intestine to the blood, i.e. cases of absorption from the
intestine.

When physiological saline is placed in the intestine and thence
finds its way into the blood, work is not necessarily performed in
order to satisfy the energy conditions of the system. In practice
of course there is likely to be some, just as some expenditure of
energy is necessary to drive even the most facile bicycle along the
most level and perfect track. The absorption of distilled water is
still more interesting for then work might actually be done on the
system. There is therefore no a priori reason for supposing that
the absorption of these fluids from the intestine either is or is not
a process involving the activity of the cells.

The simple way of putting the matter to the test was to deter-
mine the effect on the oxygen used up by the intestine. In practice
the experiment is far from simple, but in the hands of a physiologist
who probably has no superior alive at the present time in the type of
manipulative procedure involved, the experiments were satisfactorily
carried out. The result was an increase in the oxygen consumption,
which clearly showed that the absorption of water was an active
process. The following are the data of a couple of experiments from
Brodie and Vogt's paper.

1. Absorption of Physiological Saline.

11.25. Operation completed.

11.45. B.F1. = 46-875 c.c. per min.= 0-433 c.c. per grm. per min. B.P. =95.

11.49. Blood samples taken, Stage I.

11.55. Injection of 100 c.c. of warm sodium chloride solution, 0'93 °/0.

11.58. Rate of absorption 0'4 c.c. per min.

12.03. ,, ,, I'O c.c. per min.

12.06. B.Fl. = 46-875 c.c. per min. =0'433 c.c. per grm. per min. B.P. =130.

The call for oxygen considered as a physiological test 117

12.08. Blood samples taken, Stage II.

12.11. Rate of absorption 1-97 c.c. per min.

12.13. ,, ,, 2-0 c.c. per min.

12.24. B.F1. = 28-85 c.c. per min. = 0-267 c.c. per grm. per min. B.P. =92.

12.27. Blood samples taken, Stage III.

12.28. Rate of absorption 1-9 c.c. per min.
12-36. ,, ,, 2-0 c.c. per min.
12.45. ,, ,, 0-83 c.c. per min.

12.52. B.F1. = 35 -72 c.c. per min. =0-331 c.c. per grm. per min. B.P. = 85.

12.53. Blood samples taken, Stage IV.

12.54. Rate of absorption O'O c.c. per min.

No fluid was recovered from the intestines. Weight of intestines 108 grams.

The analysis of the blood gases gave :

Stage I Stage II Stage III Stage IV

Time 11.49 12.08 12.27 12.53

O2 absorbed by grm. per min 0-0087 0-0194 0-0210 0-0068

Blood-flow „ „ 0-433 0-433 0-267 0-331

Rate of absorption c.c. per min O'O 1'97 1*9 O'O

Time after injection 13' 32' 58'

In percentages of the value in the resting organ :

0., 100 224 241 78

B~F1 100 100 62 76

2. Absorption of Distilled Water.
In this group we have one experiment.

Dog. Intestines washed out with 300 c.c. warm saline solution. Artificial respiration.
In this experiment the rate of absorption was followed.

11.10. Operation completed.

11.48. B.F1. =35-71 c.c. per min. =0-331 c.c. per grm. per min. B.P. = 105.

11.50. Blood samples taken, Stage I.

11.55. Slow injection of 100 c.c. of distilled water at 37° C.

11.59. Commence to determine rate of absorption.

12.00. Absorbed 2 c.c. =2 c.c. per min.

12.01. „ 2-5 c.c. =2-5 c.c. per min.

12.02. „ 3-5 c.c. = 3-5 c.c. per min.
1204. ,, 2-6 c.c. = 1-3 c.c. per min.

12.05. B.F1. = 35'71 c.c. per min. =0-331 c.c. per grm. per min. B.P. = 75.

12.06. Blood samples taken, Stage II.
12.09. Absorbed 12 c*e. = 2-4 c.c. per min.

12.11. ,, 4-3 c.c. =2-15 c.c. per min.
12.14. ,, 6-0 c.c. = 2-0 c.c. per min.

12.21. Second set of observations of rate of absorption.

12.24. Absorbed 3'0 c.c. = 1-0 c.c. per min.

12.25. B.F1. =25 c.c. per min. = 0-232 c.c. per grm. per min. B.P. =76.

12.26. Blood samples taken, Stage III.

118 Chapter VII

12.29. Absorbed 3-2 c.c. = 0'64 c.c. per min.

12.31. ,, 1-7 c.c. = 0-85 c.c. per min.

12.34. ,, 2-5 c.c. =0-83 c.c. per min.

12.37. ,, 4-0 c.c. = 1-33 c.c. per min.

12.43. ,, 7'1 c.c. =1-18 c.c. per min.

12.53. B.F1. =30 c.c. per min. =0-278 c.c. per grm. per min. B.P.=87.

12.55. Blood samples taken, Stage IV.

1.02. Experiment stopped.

13 c.c. of fluid were recovered from the intestine. It was alkaline in reaction and gave
a fair precipitate with silver nitrate. It did not contain any appreciable amount of mucin.
Weight of the intestines 108 grams.

Stage I Stage II Stage III Stage IV

Time 11.50* 12.06 12.26 12.55

O2 per grm. per min 0-0056 0-0177 0-0120 0'0121

B.F1. per grm. per min 0-331 0-331 0-232 0-278

Blood-pressure 105 75 90 87

Kate of absorption 0-0 2-4 0-64 1-2

Time after injection 11' 31' 60'

* Injection at 11.55.

Or as percentages of the values for the resting organs :

02 100 316-1 214-3 216-1

Blood Flow 100 100 70-1 84-0

An endeavour was made to test the further question whether the
absorption of peptone involved activity of the cells; in this Halli-
burton (5) and Miss Cullis took part. Their experiments showed that
such an activity took place in the experiments, it is not however clear
that the activity was greater than during the absorption of a corre-
sponding quantity of the normal saline in which the peptone was
dissolved.

We may now pass to the other phase of the question, namely the
investigation of what happens when the fluid passes from the blood
to the lumen of the intestine. This of course is a matter of great
interest to the pharmacologist, and inasmuch as it was performed
from his point of view, the presence of a strong solution of magnesium
sulphate in the intestine was used for the purpose of evoking a flow
of fluid. The result of the experiment seemed to show that in this
case the activity of the cells was not invoked but that the action of
the magnesium sulphate was a purely mechanical one.

Dog of 14 kilos weight. Intestines washed out with warm saline. Artificial respiration
maintained throughout the experiment.

12.42. Operation completed.
1.09. B.F1. =25 c.c. per min. =0-357 c.c. per grm. per min. B.P. — 135.

The call for oxygen considered as a physiological test 119

1.10. Blood samples collected, Stage I.

1.18. Injected 50 c.c. MySOi solution (4-697 %).

1.25. Increase of fluid in intestine 0'5 c.c. per min.

1.32. ,, ,, ,, 0-5 c.c. per min. No movements to be seen.

1.39. B.F1. = 27"8 c.c. per min. = 0'397 c.c. per grm. per min. B.P. =110.

1.40. Blood samples taken, Stage II.

1.46. 15 c.c. of fluid removed from the intestine.

1.51. Increase of intestine volume 0'25 c.c. per min.
2 09. ,, ,, ,, 0*01 c.c. per min.

2.10. B.F1. =27-8 c.c. per min. =0'397 c.c. per grm. per min. B.P. = 105.

2.11. Blood samples taken, Stage III.
2.15. Animal killed.

64 c.c. of fluid were recovered from the intestine. The fluid was very viscid and had
much mucin in it. Alkaline in reaction and slightly blood stained. It contained
2-48 °/0 MgS04. The fluid removed at 1.46 was clear and contained 4-04 °/0 MgS04.
The intestine weighed 70 grms. The mucous membrane was very distinctly injected
throughout especially over all Peyer's patches. It was covered with thick mucus. The
solution injected contained 2-35 grm. MgS04. That withdrawn 0-61 grm. and that
recovered at the end of the experiment 1-59 grm. This leaves 0-15 grm. unaccounted for.
This probably was partly absorbed, partly adherent to the mucous membrane although
the intestine was washed out and the sulphate estimated in the washings. The minimum
amount of water secreted into the intestine was 29 c.c.

The samples on analysis gave the following results :

Stage I Stage II Stage III

Time 1.10* 1.40 2.11

02per grm. per min. ... 0-0191 0-0203 0-0166 c.c.

Blood Flow „ ... 0-357 0-397 0-397 c.c.

Time after injection ... 22' 53'

* Injection at 1.18.
Or in percentages :

02 100 106 87

Blood Flow . 100 111 111

REFERENCES

(1) Barcroft and Straub, Journal of Physiol. xn, p. 145, 1910.

(2) Starling, Ibid, xxiv, p. 317, 1899.

(3) Knowlton. Ibid. XLIII, p. 219, 191*1.

(4) Brodie and Vogt, Ibid. XL, p. 135, 1910.

(5) Brodie, Cullis and Halliburton, Ibid. XL, p. 173, 1910.

CHAPTER VIII

THE METABOLISM OF THE BLOOD ITSELF

ONE of the subtler problems in physiology, and one which has
made a considerable appeal to some minds, is that of the extent to
which the blood itself can be regarded as a living organ of the body.
In some works blood is described as one of the " connective tissues,"
the essential difference between it and for instance cartilage being
that the matrix is fluid instead of solid. This question brings us back
to a statement of what we regard as the essential criteria of life.
Without indulging in any general statement on the subject it is fairly
obvious that any tissue which we regard as alive must have a meta-
bolism of its own, and for that reason various workers have from time
to time tried to estimate the amount of oxygen used up by blood
per minute and the amount of carbonic acid given off by it. If such
a metabolism were proved to exist, it would be natural to discuss
the extent to which the plasma, the red corpuscles and the white
corpuscles respectively participated in it.

The early work on this subject was of course wholly vitiated by
ignorance of the growth of microorganisms, the so-called metabolism
of the blood being simply evidence of bacterial action.

Within the last few years, the problem has been taken up afresh
by two workers, Warburg (1) and Morawitz'2', both at that time working
at Heidelberg, though in different laboratories. As will be seen they
worked from different points of view.

Before describing their work in greater detail there is one point
which should be made clear. The oxidative processes in the blood, in
so far as they exist, are of two quite different natures which must
not be confused. In the first place there may be substances which
are readily oxidisable, which have found their way into the blood
in (small quantities from the tissues and which take up a certain

The metabolism of the Hood itself 121

quantity of oxygen ; these incomplete products of metabolism have
been studied by Pfliiger and more recently by Krogh and others* ;
once oxidised they are done with ; they have nothing to do with
the subject which we are about to discuss. In the second place we
have a call for oxygen which may be attributed to the life of the
corpuscles of the blood itself. This is a steady oxidation which goes
on from hour to hour.

As regards the life of the red corpuscles a natural line of inquiry
was that adopted by Warburg, namely to get some idea of the im-
portance of the nucleus to the metabolism of the corpuscle. He
therefore compared the metabolism of healthy human blood with that
of avian blood.

The method was very simple. The blood was obtained aseptic-
ally (as shown by cultures in bouillon), centrifugalised in sterile salt
solution and received in two glass bottles of 3c.c. each. In these
bottles were some glass beads. The corpuscles and the salt solution
were thoroughly mixed by shaking, the oxygen in the blood of the
one was estimated at once, the other was incubated for a given time.

The following results were obtained for human blood : the degree
of reduction is obtained by dividing the difference between the
oxygen in the two samples by the oxygen in the original blood, and
then multiplying by one hundred.

No. 1 No. 2

Duration of incubation ... 3 15 17 20 hours

Degree of reduction 3 % 5% 9% 10%

In rabbit's blood the rate of reduction is considerably greater.

Duration of incubation ... 2| 2| 8 7 8£ 12^ hours

Degree of reduction 8% 8 °/0 22% 19% 31% 29%

As compared with these it will be seen that the metabolism of the
nucleated corpuscles of the goose is very great, and is maintained at
a fairly even level.

Duration of incubation ... 150 102 150 minutes

Fluid in which corpuscles) Kinger without

are suspended . . . j Plasma NaHC03 Friedentl

Reduction per hour 33% 32% 33%

* Whilst this book has been in the press a very instructive paper on this subject
has appeared by Evans and Starling, Jo urn. of Physiol., XLVI.

122 Chapter VIII

These observations were so controlled as to make it clear that the
effects were not in any way due to leucocytes.

The reaction just described between oxygen and the nucleated
corpuscles of the goose has been used by Warburg as a typical re-
action by which to test the action of narcotics upon living matter.
Having determined the rate at which reduction takes place in
the manner we have described he went on to determine the rate
at which it is inhibited by various substances. The results of a
single set of determinations will show the general plan of the
research.

Suppose the problem be to determine the effects of various
strengths of potassium cyanide upon living matter; the corpuscles
were suspended in '9 °/0 NaCl solution containing the required
strength of potassium cyanide and incubated for 5 hours (in the
case of other organic substances usually 2 hours). The degree of
reduction which was found would then be expressed as a percentage
of what it would have been had the potassium cyanide not been
present.

The following figure will show the elegance of the results which
may be obtained by this method.

In contradistinction to the small amount of metabolism which
normal blood exhibits, a most interesting fact was discovered by
Morawitz(2), namely that the blood of anaemic animals had a very
considerable metabolism.

The technique is simple. Healthy rabbits were given daily in-
jections of phenylhydrazine hydrochloride until they became anaemic.
When the haemoglobin content had reached about 20 % of its normal
value, the blood was withdrawn from the carotid or other artery as
the result of an aseptic operation, defibrinated by shaking with glass
beads and thoroughly shaken with oxygen. One portion of the
blood was analysed at once in the Barcroft-Haldane apparatus, the
remainder, about 3 c.c., was placed in a glass bottle from which
air was excluded, and kept in a water-bath at known temperature
and for a known time. The oxygen present was then analysed. An
example will perhaps make the method most clear.

Rabbit No. 3. Made anaemic by injections of phenylhydrazine
between 10 May and 4 June. Haemoglobin value sank to 18%.

Bled and killed on June 4. Maximal oxygen capacity of 1 c.c. of
blood 0*043 c.c. After two hours at room temperature — oxygen in
blood nil.

The metabolism of the blood itself

123

Morawitz performed numerous control experiments, and the series
yielded the following data :

(1) That the oxidation is a function of the corpuscles, since it
is just as evident in a suspension of them which has been removed
from the plasma and washed three times over in normal salt solution.

(2) That it has an optimum temperature equal roughly to that
of the body.

(3) That it cannot be accounted for by nucleated corpuscles,
either red or white.

(4) That it is in short due to the young unnucleated red cor-

80 1 —

70

^.60

1 1 5°

I!40

u

^ 30

20

10

0'01 0-O2 0-03 0-04 0-05 0-06 0-07 0-08 0-09 0-1

M/iooo KCN (M =65)
FIG. 69.

puscles which are present in large numbers in the blood of the
anaemic rabbit.

The work was continued by Itami(3), who followed the whole
course of numerous cases of anaemia produced experimentally both
in dogs and rabbits. The anaemia was in some cases post-haemor-
rhagic, in some cases induced by phenylhydrazine. The technique
of the blood-gas analyses was the same as that of Morawitz.

Let us consider some abstracts of typical experiments. The first
shows the influence of continuous bleeding. It is evident from the
last column but one that as more and more new corpuscles were
produced the oxygen consumption of the blood itself became greater
and greater in amount, corresponding to the increased number of
young corpuscles present.

124

Chapter VIII
Rabbit, Male, 2900 grams.

Maxim. O., per cent.

Red

Hbby

Oo

Date
after
bleeding

blood
corpuscles

Nucleated
elements

Haldane's
haemoglobin-
ometer

Estimated
from
haemoglo-

Observed

after
5 hours

Oo

used" up

Bleeding

(days)

Millions

per cent.

binometer

per cent.

per cent.

c.c.

reading

1

5-21

6500

82

15-2

\15-4
j 15 '3

14-2
14-0

7-8

20

3

4-79

3250

74

13-7

U4-1
|13'9

12-2|
12 -0(

13-6

20

5

4-45

4850

68

12-6

112-8
/12-9

10-1

10-0

21-1

30

7

4-06

5000

62

12-5

Jll-9

/12-0

3-7
4-0

60-0

35

9

3-29

6500

54

10-0

110-2

(10-3

2-5
2-6

75-0

—

The second experiment we quote shows the effect of a single
violent haemorrhage after which observations were made from day to
day. On the first and third days of the experiment 600 and 400 c.c.
of blood were removed from a dog of 20 kilos, which may therefore
be supposed to have started with less than a kilo of blood. From
the last column but one it will be seen that the intensity of the
oxidation in the blood and therewith the rate of regeneration of
red blood corpuscles reached its maximum about the eighth day
and then gradually diminished.

Dog, 20 kilos.

Red

Hbby

Maxim. O., per cent.

0.,

Date
after
bleeding

blood
corpuscles

Nucleated
elements

Haldane's
haemoglobin-
ometer

after
5 hours

02

used up

Bleeding

Millions

per cent.

Estimated

Observed

per cent.

per cent.

c.c.

1

6-93

14000

110

20-4

(20-4
J20-2

18-81
19-0)

17-0

600

3

5-03

10000

78

14-4

114-7
(14-6

11-3)
11-2J

22-8

400

5

4-35

20000

68

12-6

(12-8
1 12 -6

4-7)
4-8)

62-7

—

8

4-18

19500

66

12-2

(12-4
(12-5

4-3)

4-2[

65-3

—

12

4-42

17500

76

14-1

(14-3
"(14-2

10-4)
10 -5 [

26-0

—

15

5-26

14000

80

14-8

J14-9

12-21

18-1

(14-8

12'lj

In the third experiment given here he again controls the question
of whether the increased oxidation in the blood can be attributed
to the nucleated elements. For this purpose he divides the blood

The metabolism of the blood itself

125

into two portions. One of these, B, is rapidly defibrinated by vigorous
shaking, the other, A, is defibrinated very slowly. In the latter case
blood-counts showed that there were three times as many nucleated
corpuscles as in the former, yet the percentage of reduction was not
very different in the two cases.

Experiment on dog.

Red

Hbby

Maxim. O2 per cent.

O.7

blood Haldane's
corpuscles haemoglobin-
ometer

Nucleated
elements

after
5 hours

02
used up

Millions

per cent.

Estimated

Observed

per cent.

per cent.

B

3-32

50

6000

9-3

J 9-2

| 9-4

5-0(

4-8l

46

A

3-96

54

18500

10-0

uo-i

10-3

5'0| -2.q

4-7

Experiment on rabbit.

B

2-60

40

2250

7-4

7-6

1-5)

80

A

2-73

41

7500

7-6

J7-7

\ P7 rj

53!

87

(1-1

A short while ago I came across an instance of the type of blood
which Morawitz describes. The observation is not without interest
in itself and it may be useful as a warning to some. I will therefore
describe it.

In the Biochemical Journal Moore and Wilson (4) published some
observations on the alkalinity of the blood in cancer. It occurred to
me that such changes in reaction could be observed by exposing the
blood to a known oxygen pressure and observing whether the haemo-
globin became more or less saturated with oxygen than would
normally have been the case.

Dr Hopkins was kind enough to give me the blood of numerous
rats, which he was killing at the time and which had grown sarcomata.
The result of the experiment seemed interesting beyond my ex-
pectation, for the rats with sarcomata possessed blood which reduced
to a much greater extent than did the blood of normal rats. Indeed
it was possible to draw a very presentable curve showing that the

... „ reduced haemoglobin ,.

quantity 01 , , . — x 100 or percentage reduction ot

total haemoglobin

the blood exposed to 17mm. oxygen pressure followed the weights

126

Chapter VIII

of the tumours. The interpretation of this would have been — the
bigger the tumour the more acid the blood. My suspicion was
aroused by the great degree of reduction which some of the blood
underwent. This caused me to do control experiments in which I
treated some of the same blood with air instead of 17 mm. of oxygen ;
these control experiments led us to suspect that the blood reduced
itself.

A second series of experiments were undertaken for the object
of testing this point. The technique was simple. The rats were
killed by cutting the throat. The blood was whipped with feathers,
thoroughly saturated with air and placed in a 1 c.c. glass syringe.
Care was taken to prevent air entering the syringe which was placed
in an incubator at 37° C. for f hour. The blood was then expelled
into the bottle of a differential apparatus under ammonia. The
percentage saturation was measured in the usual way, and as a known
quantity of blood was used the total oxygen capacity was also
obtained.

Many of the samples showed a great degree of reduction. In
the series there were two classes of animals, those with tumours of
5 grams and under and those with tumours of 13 grams and over.
The blood which reduced itself to a great degree during the
45 minutes incubation came in each case from the rats with large
tumours, but the degree of reduction followed the total oxygen
capacity of the blood more closely than the weight of the tumour.
Roughly speaking the more anaemic the rat the greater was the
amount of reduction in the blood. The following table gives the data.

Number of rat

71

71 A

72 A

73

73 A

74

74 A

75

Control

sheep s blood

Weight of tumour (grams)

1

0 +

23

13

34

1

5

14

—

Percentage saturation of'i
blood after | hour incu- >
bation )

85

89

49

0

64

84

80

66

98

a _>-- f c.c. of O.j per c.c. of \
8,'S blood

•14

•111

•070

•044

•088

•134

•132

•069

—

S? a ] Gower-Haldanehae-f
O § ( moglobinometer ... (

75

60

38

24

47

78

77

38

—

Fig. 70 represents these data in the order of the haemoglobin
value showing how close is the correspondence between the degree
of anaemia and the power of the blood to eat up its own oxygen.

The metabolism of the blood itself

127

It is therefore necessary to control many researches by testing
for self-reduction in the blood. With this warning let me pass to a
wholly different subject for which this property of young corpuscles
has been used as a test.

Armed with the discovery that freshly formed red cells have an
appreciable metabolism, Morawitz and Masing(5) sought to solve a
problem which has been something of a stumbling-block to physio-
logists, namely whether or no there is an increased formation of red
blood corpuscles at high altitudes.

When they attacked the problem it was capable of statement as
follows : It was known that life at increased altitude caused an
increase in the number of red blood corpuscles per cubic millimetre

100

FIG. 70. — Shows the degree of anaemia and the degree of self-reduction of the blood in
each of eight rats which had sarcomatous tumours. o= Percentage saturation of
blood after incubation. •= Haemoglobin value (100 corresponds to '185 c.c. oxygen
per c.c. of blood). The figures along the abscissa denote the rats.

of blood taken from the finger or ear. This increase was accounted
for in three different ways by different authors.

(1) That there was an actual increase in the formation of red
corpuscles.

(2) That there was a diminution in the plasma leading to a
greater concentration of the corpuscles.

(3) That the corpuscles became unevenly distributed, being in-
creased relatively to the plasma in the peripheral blood and decreased
relatively to the fluid elements of the blood in the viscera.

It is not easy to arrive at a judgment as to whether or no there
is an increased formation of haemoglobin in man. The most obvious
path along which to seek a solution is that of measuring the total
blood volume by the carbon monoxide method, either as described
by Haldane(6) or by Plesch(7). Indeed measurements of this sort have

128

Chapter VIII

been made since the experiments of Morawitz and Masing which we
are about to describe. Douglas (8) in Tenerifte obtained the following
data upon my own blood.

Date

Hb°/0

Oxygen capacity
in c. c. per c. c.
of blood

Total oxygen
capacity of blood
in body

Grams of
haemoglobin

Volume
of blood

Cambridge.

Feb.

5

6
21

Mean for.

March 21

„ 23

25

100
100
100

100

103
105
113-5

•185
•185
•185

•185

1017
1053
1025

1032

140

5500

146

5700

142

5550

143

5583

Orotava (sea level}.

995
1000

•191
•194
•210

144
157

5220
4770

Canadas (7000 feet}.

March 29

108

•200

April 2

108

•200

950

142

4750

„ 4

—

—

1075

—

—

„ 11

Ill

•205

940

145

4580

The last two columns indicate that in my own case there was no
increase in the amount of haemoglobin, but from some cause or other
a concentration of the blood.

To return however to Morawitz and Masing, they showed that an
appreciable rise took place in the metabolism of the blood itself, as
the result of bleeding to the extent of 400 c.c. To put this in another
way, if the body is called upon suddenly to make haemoglobin to the
extent of 8 % of the haemoglobin in the body, the formation of young
corpuscles with high metabolism will be sufficient to be appreciated
by their desire for oxygen. The experimental procedure was to
shake the sterile blood, incubate it for 5 hours, at the end of which
time the oxygen in the blood was measured by a gas analysis apparatus.
Normal blood would have lost less than '01 c.c. of oxygen per c.c. of
blood, while that from the patient after bleeding would have lost more
than *01 c.c. of oxygen.

If then there were at high altitudes a formation of new blood,
erythrocytes corresponding in quantity to those in 400 c.c. of normal
blood, the blood would acquire the power of a considerable degree of

The metabolism of the blood itself

129

self reduction. The following table gives the results of Morawitz' and
Masing's experiments at Col d'Olen (10,000 feet) from which it will
be seen that they were able to discern no sudden increase in the
haemoglobin formation. The respiratory activity of the blood itself
was appreciably less in the case of "P.M." as the result of his climb
to Col d'Olen than in the result of losing 400 c.c. of blood, whilst in
the case of "E.M." there was no greater evidence of young corpuscles
on Monte Rosa than at Heidelberg.

I. Subject of research, P. M.

O2 capacity of blood

O2

in volume per cent.

dimi-

No. of

nution

Date

Hbin

per cent.

No. of red
corpuscles
per cub. mm.

white cor-
puscles per
cub mm.

in

volume
percent.

Freshly

After
5 hours

incubation

April 12

114

5,200,000

5200

21-1

20-2

0-9

\

„ 13

—

—

—

—

—

—

„ 15

110

5,100,000

5200

20-3

19-1

1-2

1 Bled 300 c.c.

July 29

115 (113)

(5,600,000)

—

21-5

20-6

0-9

f Heidelberg (115 m.)

Aug. 2

114

(5,400,000)

4000

21-3

20-3

1-0

„ 6

—

(5,400,000)

—

—

—

—

/

Aug. 13—17

—

—

—

—

—

—

Small journeys

Aug. 19

—

—

—

—

—

—

Climb to Col d'Olen

„ 21

116

( 5,300,000 I
|(4, 800, 000 M

4800

21-8

21-0

0-8

)

„ 23

,, 24

(115)
119

(5,000,000)

5,600,000

5000

22-2

21-3

0-9

[ Col d'Olen (3000m.)

„ 26

121

5,700,000

6500

22-1

22-2

0

„ 28

120

5,700,000

4500

21-9

21-0

0-9

/

Sept. 8

118

5,500,000

— .

—

—

—

) Bled 400 c.c.

,, 9

117

5,300,000

6400

21-6

20-1

1-5

j Heidelberg (115 m.)

II. Subject of research, E. M.

July 28

(108)

5,300,000

— •

20-4

19-5

0-9

|

„ 31

113

—

4100

20-7

20-2

0-5

f Heidelberg (115 m.)

Aug. 12

112

5,500,000

(6000)

—

—

—

)

Aug. 13—17

—

—

—

—

—

—

Small journeys

Aug. 19

—

—

—

—

—

—

Climb to Col d'Olen

„ 22

117

( 6,000,000 )
1(5,000,000) \

5700

22-3

21-5

0-8

\

„ 23
„ 25

118

(5,800,000)
6,100,000 {
(5,100,000) 1

5200

21-8

21-2

0-6

\- Col d'Olen (3000 m.)

,, 27

122

5,300,000

5300

23-0

22-2

0-8

When I was working at the Col d'Olen laboratory the question
of the increase of red corpuscles was the subject of research by
Cohnheim(9) and his collaborators. They arrived at the same

B. R. P.

130

Cluster VIII

conclusion as Morawitz, namely that in spite of all that has been said
and written on the subject they could detect no considerable increase in
the number of corpuscles or the amount of haemoglobin in the blood.
At the time we were at Col d'Olen, Haldane and his collaborators
Douglas, Yendell Henderson and Schneider (10), were investigating the
same problem on Pikes Peak (14,000 feet high). They went into the
matter with great care ; estimating (1) the number of corpuscles by
two different types of haematocytometer, Barker's and Thoma-Zeiss' ;
(2) the oxygen capacity of 1 c.c. of blood by a haemoglobinometer
and by direct determinations with ferricyanide ; (3) the total blood

-535K*

^':«s2kLJi

FIG. 71.— Pikes Peak observatory', Colorado, altitude 14,000 feet (Haldane,
Douglas, Henderson and Schneider).

volume and total oxygen capacity of the body by the CO method
already described. The following charts show clearly that the results
which they obtained for the various members of the party were very

uniform.

In each case there was a gradual rise in the total oxygen capacity
which reached its maximal value, only after some time, about three
weeks after the ascent, This perhaps is the essential point, for it
proves quite clearly that the body reacts to the altitude either by
producing increased quantities of haemoglobin or by retaining what
would otherwise be broken. This at least is a positive reaction.

The metabolism of the blood itself

131

The change in the blood volume is much less considerable, it is
also less constant in different individuals ; for instance, in the first
week Douglas' blood volume went down, Haldane's remained practi-
cally constant, those of Henderson and Schneider rose a trifle. The
factor which perhaps underwent the most constant change was the
percentage of haemoglobin in the blood. Cutting out the daily

DOUGLAS

OXFORD

130
120

oo:

-10.

o<o
o

SUMMIT OF PIKES PEAK

II

o ct
^ a-

hAutPj

OXK3RD

110
JOO

90

HALDANE

7 II

JULY

15 19 23 27 31 4 8 12 16 20 24 28 I J 9 13 17 21
AUGUST SEPTEMBER

\f>(\

^

>J— —

-0

1

1 — .

^.

a

>

•^

. - -

5^

^

V

' V

^

-

«r*

~ • •!

^-.

•v»

IrYi

/Vp^eA .-

•^

sfl

/v

-'

y-

'-

c

"•s.

^

^

;<«

on

-^ \/^*^.

hi

^

\

V

HENDERSON

I7n

^

^^^-

. — • — ^

s

v

)70

x

/

Xj

^^^

4

•^.

MO

X

A

r

_/ ^

**>

JV

C-K

V

^

>.

"•*-^,

,'

2

•^

— ^-

SCHNEIDER

I4<J

. — -

, — <

S^

pn

X

,^-
/\

fV

, ^

S

j —

s

~^-

f

^

£V

'

V

g

*s

T- •-

inn

H

?

— •

* ~

N»

S/

FIG. 12.

Abscissae and ordinates on the same scale in each case.

Ordinates represent percentages of the average values obtained before ascending the Peak
(Oxford and Colorado springs) on the particular subject.
-= percentage of haemoglobin.
= blood volume.
= total oxygen capacity or total amount of haemoglobin.

variations, which were considerable, this reached a maximal value of
about 120, in each case taking its original value as 100, but here again
it must be observed that the rise was but a gradual one.

A very complete series of results on the oxygen capacity of the
blood have been performed by Mr J. Richards (11), the manager of a

9—2

132

Chapter VIII

tin mine at Pazna in Bolivia situated at an altitude of 15,000 feet.
As coal gas was not available for the purpose of operating the
Haldane's haemoglobinometer, he used one containing picrocarmine
jelly which was standardised by Haldane before and after the series of
observations and found not to have varied more than 1 °/0. The read-
ings have been reduced to those of the Haldane's standard instrument.
The following was his itinerary :

Richards left Liverpool on Nov. 10

,, reached Buenos Ay res Dec. 5

,, started from Antofagasta Dec. 15

,, reached Tatoral Dec. 17

,, started from Tatoral Dec. 24

reached Pazna Dec. 24

Level in feet
sea

12,500
12,500
15,000

CI 25
WOVEMBER

t9

I 5 9

MARCH

13 17

a 7 II 15 IB 23 27 31 4- 8 IS 16 20 24 38 I 5 9 13 17 21
DECEMBER JANUARY FEBRUARY

FIG. 73.

= Mean percentage of haemoglobin at sea level.

= Actual percentage determined.

The chart shows the very definite rise in the haemoglobin-value
of the blood and that it did not reach its maximum till after two and
a half months' residence at the altitude of 15,000 feet.

What then is to be made of the very conflicting results which have
been obtained by different observers, each of unassailable position as
a physiological expert, and even of the same observer (Douglas) work-
ing at the different places but with the same method ?

For what it is worth, my view of the matter at the present time
is this. Col d'Olen (10,000 feet) and the Canadas (7000) are too low
to get much result, at all events in the short time that has been at
the disposal of the workers. In a fortnight at the Canadas the total
oxygen capacity of my body practically remained the same. In later

The metabolism of the blood itself 133

chapters in this book it will appear that in other respects I was also
unchanged — in fact the altitude had not as yet touched me. At
Col d'Olen even, changes such as diminution of alveolar CO2 pressure
are, in some individuals, very slight. This consideration will explain
why Douglas got negative results at Las Canadas, and why Cohnheim
and possibly Morawitz also got negative results at Col d'Olen, whilst
Zuntz, Haldane and others obtained positive results at higher altitudes.
It may of course be that in a longer time there would be well-
marked positive results at lower altitudes.

REFERENCES

(1) Warburg, Hoppe Seller's Zeitsch. LXVI, p. 305, 1910 ; Ibid. LXXVI, p. 331, 1912.

(2) Morawitz, Archie f. exp. Pathol. und Pharmacol. LX, p. 298.

(3) Itami and Morawitz, Deutsch. Arch.f. klin. Med. c, p. 191.

(4) Moore and Wilson, Biochem. Journal, i, p. 326 ; Mathison, this book, chap. v.

(5) Morawitz and Masing, Deutsche Med. Wochenschr. 1910, Nr. 8 ; Deutsche

Archie f. klin. Med.

(6) Haldane and Lorrain Smith, Journal of Physiol. xxv, p. 331, 1900.

(7) Plesch, Miinchener Med. Wochenschr. LVII, 8, p. 406.

(8) Douglas, Journal of Physiol. XL, p. 472, 1910.

(9) Cohnheim, Kreglinger, Tolber and Weber, Zeitsch. f. Physiol. Chem. LXXVIII,

p. 70, 1912.

(10) Douglas, Phil. Trans, of Roy. Soc. Ser. B, Vol. 203.

(11) Richai-ds, Supplement to Report of Pikes Peak Expedition. Phil. Trans, of

Roy. Soc. Ser. B, Vol. 203.

CHAPTER IX

THE REGULATION OF THE SUPPLY OF OXYGEN
TO THE TISSUES

FROM the consideration of the call of the tissues for oxygen, I pass
to the mechanism by which that call is met. The information about
the passage of oxygen from the blood to the organs may be considered
under two main headings: firstly the provision for bringing to the
organ the quantity of oxygen which is required, and secondly the
physical process by which the oxygen is transferred from the capillary
blood vessel to the cell. In the case of most organs of the body the
sole channel by which the organ receives its oxygen is its nutrient
artery. One organ, the liver, has two channels of supply; it is by
no means obvious whether the portal vein or the hepatic artery
provides the essential oxygen supply. I shall clear the ground by
first considering this question of the liver, then I shall pass to the
functional fluctuations in the oxygen supply of organs generally and
the mechanism by which these fluctuations are brought about.

Concerning the liver, then, let me quote two text-books of physio-
logy currently read by students : (1) "The liver has a double blood
supply : the portal vein which supplies a rich capillary anastomosis
round every liver cell and carries venous blood from the alimentary
canal, and the hepatic artery, which carries oxygenated arterial blood
and supplies chiefly the connective tissue surrounding the bile ducts
and blood vessels in the division between the lobules, known as
Glisson's capsule." (2) "The hepatic artery, the chief function of
which is to distribute blood for the nutrition of Glisson's capsule, the
walls of the ducts and blood vessels, and other parts of the liver."
No one reading these statements would have been surprised had
a quantitative investigation of the oxygen supply of the liver shown
that the hepatic artery brought but a trifling quantity of blood to
the liver and found oxygen merely for the accessory structures. Yet

The regulation of the supply of oxygen to the tissues 135

this proved to be exactly the reverse of what takes place. The fact
is (1) that the blood in the hepatic vein leaving the liver on the average
contains about as much oxygen as does that in the portal vein. The
comparison is shown in the following table :

Oxi/yen in 1 c.c. of blood entering the liver by the portal
vein and leaving it by the hepatic vein.

Mean of
Total 10 exps.

Portal vein -041 -049 -114 -038 -033 -033 -017 '032 -039 -094 -490 -049
Hepatic vein -042 -052 -088 "017 "074 -029 '000 -004 -056 -088 -450 -045

It is impossible to discuss the fate of the individual corpuscles on
their way through the liver. We only have the data furnished by a
knowledge of the initial and final conditions of the blood with regard
to the organ. When we speak of the oxygen which it receives from
the hepatic artery we mean the amount of oxygen in 1 c.c. of
the arterial blood minus the oxygen in 1 c.c. of blood from the
hepatic vein multiplied by the amount of blood entering the liver
from the artery per minute ; and similarly with regard to the portal
blood, the quantity of oxygen which the liver acquires from this
channel is regarded as the oxygen in 1 c.c. of portal blood minus that
in 1 c.c. of hepatic venous blood, multiplied by the number of c.c. of
blood entering the liver by the portal vein per minute. By the
addition of these two quantities (the oxygen acquired from the hepatic
artery and that acquired from the portal vein) we arrive at the total
quantity of oxygen used by the liver.

The following example which is typical will indicate what we
mean :

Oxygen in 1 c.c. of blood in hepatic artery -128 c.c.

„ ,, ,, portal vein -033 c.c.

,, ,, ,, hepatic vein -029 c.c.

Bate of flow of blood in hepatic artery 16 c.c. per minute

,, ,, portal vein 20 c.c. per minute

Oxygen taken from hepatic artery (-128- -029) x 16 = 1-58

portal vein (-033 - -029) x 20 -08

Oxygen used by liver per minute =1-60

In the majority of experiments the whole amount of oxygen
brought to the liver by the portal blood would not suffice for its
needs, whilst that brought by the hepatic artery would more than
suffice as the following figures show.

136 Chapter IX

Oxygen used

Oxygen in c.c. per minute.

Oxygen brought to liver by

by the liver Portal vein Hepatic artery

•35 -85 -47

•41 '47 '82

•77 1-85 '59

1-61 1-09 I'OO

•91 -66 3-00

1-76 -66 2-06

5-81 c.c. 5-58 c.c. 7'94 c.c.

Indeed the oxygen used by the liver is unaffected if the portal
vein be ligatured altogether and its blood diverted into the general
circulation without first passing through the liver. So much for what
we may call the "resting" liver; by that we mean the liver of an
animal which had not been fed for 36 hours before the experiment.

Important as is the arterial supply to the resting liver, it is much
more important in the case of the liver which is using up relatively
large quantities of oxygen. In animals which were fed 18 hours
before the experiment the quantities of oxygen brought to the liver
by the portal blood are of the same order as in those experiments
which we have already quoted ; the excess of oxygen brought comes
along the hepatic artery. From the following four experiments the
reader can gather the extent to which the liver would be starved of
oxygen did the hepatic artery perform no more important function
than that of supplying Glisson's capsule, and the walls of the blood
vessels and ducts.

Oxygen used

Oxygen in c.c. per minute.

Oxygen brought to liver by

by the liver Portal vein Hepatic artery
3-8 '28 3-53

2-17 '58 1-56

1-92 -78 3-67

1-72 2-25 3-02

9'61 c.c. 8-89 c.c. 11 '78 c.c.

The mechanism by which the arterial supply of blood is increased
during digestion is unknown, it may be to some extent local and
to some extent reflex. This at least is known from the work of
Burton-Opitz (2' : the branches of the hepatic artery are richly sup-
plied by vaso-constrictor nerves; these have been denied to the

The regulation of the supply of oxygen to the tissues 137

branches of the portal vein. We have therefore a complete scheme
before us. The regulation of the portal blood depends upon the
intestine and on the processes going on in the intestine. So far
as the liver is concerned the portal blood is the raw material on
which it works, but it has also a blood supply proper to itself for the
needs of its own metabolism.

Now to turn to the general question of the fluctuations which
occur in the blood supply of the organs of the body. We are not
concerned with such fluctuations as are brought about by the central
nervous system without reference to the needs of the organ for the
purpose of maintaining the general arterial pressure. These do not
touch the respiration of the organ. In Chapter V we have emphasised
the great variations which occur at different times in the need of the
organ for oxygen. Unless the tissue when at rest is to be flooded
with large quantities of unnecessary blood, the variations in its blood
supply must be of the same order as those of the oxygen required.
And for our present purpose we must regard the question of blood-
flow from the point of view of the oxygen itself.

Let us take the pancreas as an instance. It is excited to activity
by a stimulus which acts directly upon its cells ; the gland at once
requires four or five times as much oxygen as before : how is this to
be obtained ? Is there a reflex mechanism by means of which the
pancreas can beg a supply of blood from the vaso-motor centre?
We know of none. Does the secretion act on the vaso-inotor centre
itself ? We have no reason to suppose that its action is anything but
local. The broad question which is to be answered is this, Is there
evidence that each organ of the body is so far master of its own
metabolism that it can force the vascular system to give it the oxygen
which it requires ?

This is the sort of question which used to be put by Gaskell to his
class when I was a student ; since then the evidence on the subject
has increased many fold.

Let me start then by accepting (1) the conclusions of Severini(3),
Gaskell (4), and other more recent workers, in so far as they show that
certain products of metabolisms, to wit acids including C02 and lactic
acid, have the power of distending blood vessels, and (2) those of
Dale (5) and his colleagues to the effect that /3-iminazolylethylamine, a
body so closely bound up with the physiology of protoplasm that it is
liberated by the splitting off of C02 from histidine, will produce a
powerful dilatation of the vessels. So powerful indeed is the effect

138

Chapter IX

that the blood pressure of a monkey will fall to a little over half its
former value if but 0*0005 gram of the drug is injected intravenously.
How incredibly small a quantity of "/3-I" would suffice to produce
dilatation in an organ of a few grams weight, such as the monkey's
pancreas, were the drug produced in the organ itself.

Truly no great eifort of the imagination is necessary to conjure up
such a mechanism as I have outlined. Our first concern is to inquire
whether there actually exist cases of dilatation of which there is no
other satisfactory explanation than that of a local chemical action.
The chain of evidence appears to be most complete in the case of the
submaxillary gland.

In the cat stimulation of the cervical sympathetic nerve is known
to give a greater secretion of saliva in relation to the size of the
submaxillary gland than in other animals. The cervical sympathetic
also contains constrictor fibres running to the blood vessels of the

10

FIG. 74. — Monkey. A.C.E. Arm volume and blood pressure. Effect of 0-5 mgm. /3-I.

gland. Stimulation of this nerve produces in succession the following-
effects upon the blood-flow from the gland : (1) decreased, (2) in-
creased, and (3) decreased blood-flow. The second period or period
of dilatation comes immediately after the secretion of the saliva. In
Fig. 75 is shown the rate of flow of blood through the submaxillary
gland in a typical experiment. If we seek the cause of this dilata-
tion, which may be obtained in an even more striking fashion by the
injection of adrenalin, we shall see to start with that it cannot be
ascribed to mere loss of tone of the arterial wall, such as might be

The regulation of the supply of oxygen to the tissues 139

obtained by section of the cervical sympathetic ; for this constrictor
nerve is cut throughout the experiment, and was actually being-
stimulated at the very time that the dilatation was observed. Clearly
the dilatation takes place in spite of this stimulation ; the cause of
the dilatation, whatever it may be, is something which overrides the
excitation of the vaso-constrictor fibres.

Perhaps the most natural assumption to make is that there are
dilatator fibres mixed with the constrictor fibres in the sympathetic.
This assumption was made by Carlson and his colleagues, Greer, Becht
and McLean ((5), who, independently of us, observed the vaso-motor
phenomena which we have described. The second of the two papers
on the subject which appeared from his laboratory, in its closing
words states the issue as between true dilatator fibres, such as are

80 100 120 140

160 180

FIG. 75. — Ordinate = Rate of blood-flow through the submaxillary gland. | = Stimulation
of cervical sympathetic commences and lasts throughout. The figures represent
graduation of a tube along which the blood from the gland flowed. -0021 c.c. =1
graduation of tube. Abscissa— time in seconds.

assumed in the chorda tympani and local metabolic products. "If
there are some vaso-dilatator fibres... we should be able to show their
presence by cutting out the vaso-constrictors by chrysotoxine and
stimulating the vaso-dilatator apparatus by adrenalin, as chrysotoxine
seems to affect the vaso-constrictors, not the vaso-dilatators."

The same issue had been pointed out to me at a meeting of the
Physiological Society by Dale ; in effect it is this : Ergotoxine (or
chrysotoxine) does not paralyse the chorda tympani dilatator fibres,
but does paralyse the sympathetic constrictor fibres ; if therefore
there are fibres in the sympathetic which are functionally chorda
tympani fibres, then after the administration of ergotoxine and
paralysis of the constrictor fibres, stimulation of the sympathetic
will produce the same effect on a small scale as stimulation of the

140

Chapter IX

chorda tympani, i.e. dilatation. Now this is precisely what does not
take place. Administer ergotoxine in a suitable dose, stimulate the
cervical sympathetic either with an induction current or with adrenalin,
and you do not get dilatation; the effect of the excitation of the
nerve upon the vessels of the gland is nil. Nor is there any secretion
of saliva. The assumption that this dilatation is due to dilatator
fibres has therefore broken down. The rival theory holds the field.

To leave the matter at this point would be unsatisfying. It is
desirable that the theory of chemical, or as I prefer to call it

113.

FIG. 76.

•^^— = Arterial pressure.

— — — = Kate of flow of saliva in c.c. per minute.

• •••• = Bate of blood-flow through vessel of submaxillary drawn to one-tenth of the scale
of the saliva curve.

"functional," dilatation should stand not merely by default. There
should be actual proof that (1) the metabolism of the submaxillary
gland is increased when the sympathetic is stimulated, and (2) that
the actual metabolic products present are such as have the power
of dilating vessels.

What evidence, then, is there of increased activity of the tissues
in the gland when the sympathetic is stimulated ?

The answer can best be obtained when the sympathetic is stimu-
lated by means of intravenous injection of adrenalin, for when a

The regulation of the supply of oxygen to the tissues 141

faradic current is used, the issue is complicated by the fact that
the blood is almost entirely cut off from the gland at times, but with
stimulation by means of adrenalin all is in our favour ; we get a good
flow of saliva, a good dilatation, and the effect of the primary con-
striction, as far as rate of blood-flow is concerned, is largely counter-
balanced by the increased arterial pressure. The experiments were
performed on cats. A graduated tube was fixed into Wharton's duct ;
an observer registered the rate of flow of saliva on the kymograph
with a signal and tapping key. The blood from the gland was collected

FIG. 77. — A and B form continuous portions of a single tracing. From above downwards
(1) arterial pressure, (2) time of injection of adrenalin, (3) each mark of the signal sig-
nifies the flow of 0-01 c.c. of saliva, (4) blood signal marks correspond to O'l c.c.

in a cylindrical 1 c.c. pipette placed horizontally, the pipette was
graduated in tenths of a cubic centimetre and the moment at which
the blood meniscus passed each graduation was also recorded. The
general arterial pressure was registered from the femoral artery.
Samples of arterial blood for comparison with the blood from the
gland were obtained from the other femoral artery. The figure given
below shows the general time relations of the experiment, while we
have seen in Chapter VI that there is a great increase in the
metabolism of the gland as shown by the oxygen used up.

142 Chapter IX

The following are the data yielded by a couple of experiments on
this subject.

Oxygen used up in c.c. per grain of

gland per minute

, * >

Resting Secreting (adrenalin)

Exp. 1 0-028 0-052

Exp. 2 0-026 0-050

Speaking roughly the gland uses up about the same volume of
oxygen as it secretes of saliva.

The relation between the oxygen used by the gland and the flow
of blood and of saliva may be studied in rather greater detail in such
an experiment as that represented in Fig. 77.

The variations of oxygen consumption were estimated at different
times during and subsequent to the secretion of saliva produced by
an injection of adrenalin — the moment of taking the samples for
analysis, the rate of flow of the blood in each case, and the rate of
flow of the saliva are all shown in the figure. The quantities of
oxygen used were as follows. Taking the amount of oxygen used by
the resting gland (Sample I) as unity, the amounts used at different
times indicated in the figure are found by analysis to be

Sample II IV V VI

Eatio of oxygen used to that of resting gland ... 1-2 2-1 3'7 1*3

Moreover it is very interesting from our present point of view to
see that the heightened metabolism of the gland, like the heightened
blood-flow, outlasts the flow of saliva by a considerable time, and
outlasts the obvious direct vascular effects of the adrenalin by still
longer.

We have shown that there is dilatation of the arteries when
metabolism is induced by adrenalin, can we as a control experiment
prove that adrenalin produces constriction and constriction only when
there is not increased activity of the cells ? The following facts may
be cited in this connection. In the surviving gland, as in other per-
fused organs, the effect of drugs on the vessels survives the effect on
the organ itself. Adrenalin added to the saline which perfuses the
vessels of the submaxillary, a couple of hours after the death of the
animal, produces no secretion of saliva and protracted constriction of
the vessels ensues without the least vestige of dilatation ; this effect
may even be obtained on the following day.

The following example may be cited of a cat's submaxillary gland,
which was perfused directly through the submaxillary artery.

The regulation of the supply of oxygen to the tissues 143

Cat killed with chloroform about 11 a.m. perfused with warm Ringer's fluid.

11.48 30, 30, 30 drops per half minute.
11.50 Inject 1 c.c. pilocarpine.

44, 44, 45. Saliva.

37, 37. Saliva stops.

38 \

No saliva.

11.53
12.22
12.25
12.31
12.34
12.45
12.49
12.55
12.55-30 9, 10, 10, 10.

The same gland next day :
10.19 20, 18, 18, 16, 17. Inject adrenalin. 4, 3, 3, 4, 3, 3.

36
36
36
36
Inject 1 c.c. adrenalin.

No saliva.

FIG. 78. — Effect of injecting adrenalin on the rate of blood-flow of submaxillary gland.
A normal, B after clamping artery for 1 hour and 50 minutes.

We need not, however, have resort to artificial perfusions in order
to find out that adrenalin does not cause dilatation but constriction
when the gland is incapable of secreting. It is only necessary in the
living animal to clamp the submaxillary artery, wash out the gland
with cold saline, and leave it so for a couple of hours ; at the end of
this time the circulation of blood is reestablished. On the injection
into the arterial stream, close to the origin of the submaxillary artery,
of 0'5 c.c. of adrenalin*, we find that constriction only ensues

and that no saliva is secreted ; under normal circumstances saliva

* The adrenalin comes out from the gland with the blood whose rate of flow
is being measured, and so never reaches the general circulation.

144

Chapter IX

would have flowed freely, and the constriction would rapidly have
been overridden by dilatation.

It remains for us to see whether there are as a matter of fact bodies
secreted into the blood which are capable of dilating the vessels.
The bodies which are capable of causing dilatation may be divided
into two groups from our present point of view : (1) acids, (2) organic
bases, such as those isolated by Dale and Barger. The latter
act in extremely minute quantities, and nothing would appear more
probable than that such bodies should be produced in small quantities
in the breakdown of tissues. It is a commonplace that extract of
almost any organ produces such substances ; their very commonness

10

25 30 35 40 ~ 45 50 55 60

FIG. 79. — Ordinate^c.c. of CO., per 100 c.c. of blood. Abscissa=C02 pressure in mm.

has caused them to be largely overlooked in comparison with pressor
substances. It is certain that such substances may be obtained from
the submaxillary gland.

On the other hand, as regards the reaction of the blood, there are
special reasons why the fluid bathing the cells should be less alkaline
when the gland is secreting than when it is not doing so. The secretion
of the submaxillary is, of course, strongly alkaline. This means that
the bases are taken from the lymph and secreted whilst the acids are
thrown back into the blood. We can test the question experiment-
ally. We find that the blood in its passage through the gland does
actually change its hydrogen ion concentration, and it would there-
fore seem that the lymph in equilibrium Avith it cannot but dilate
the vessels.

The regulation of the supply of oxygen to the tissues 145

That the "chorda blood" is richer in acids as compared with
alkalis, than either the arterial blood or the venous blood, is doubly
manifest. The first method of demonstration is essentially that used
by Morawitz — namely, analysis shows that in equilibrium with
carbonic acid at a given pressure the chorda blood contains less
C02 than the ordinary venous blood, which in turn contains less
than the arterial blood, showing of course that the bases are more
completely saturated by other acids.

Fig. 79 shows the quantity of C02 contained in each of the three
samples of blood when in equilibrium with the pressures of CO2 at
which the tests were carried out. For a given CO2 pressure, say
41 mm., 100 c.c. of the chorda blood contains 36 c.c. of CO2, the resting
venous blood 39 c.c., whilst the arterial blood takes up 44 c.c.

The second method of showing that the chorda blood is the more
acid is by the effect on the affinity of its haemoglobin for oxygen.
The curves are shown in Fig. 88.

To summarise our case so far, stimulation of the submaxillary
gland by means of adrenalin furnishes us with a case of local dilata-
tion as the result of metabolic products, which dilatation can be shown
to be complete in every particular : the direct effect of the drug on
the vessels is constriction. This constriction is overridden when, and
only when, the gland secretes, in that case bodies which are known
to cause dilatation are demonstrably produced.

I do not think anyone can actually see a demonstration of the
phenomenon which I have just described without asking himself
whether this functional dilatation does or does not explain the whole
phenomenon of dilatation as observed in the salivary glands. In
other words, is it possible to demonstrate that the vaso-dilatation,
which follows upon stimulation of the chorda tympani, involves a
definite neuro-muscular vaso-dilatator mechanism ?

Claude Bernard supposed that when the chorda tympani was
stimulated three sets of fibres were thrown into action — the calorific,
the vaso-dilatator, and the secretory. The evidence of the calorific
was of the most direct kind, namely that heat was demonstrably
produced when the nerve was stimulated; the evidence of the vaso-
dilatator fibres was of the most direct kind, namely that the blood
stream was demonstrably accelerated; and the evidence of the secretory
fibres was of the most direct kind, namely that the saliva was seen
to flow.

The conception of the calorific fibres has long since gone; the

B. R. F. 10

146 Chapter IX

heat is due to the liberation of energy caused by the stimulation
of the secretory fibres ; the challenge has come to the dilatator fibres.
Are there really such things, or is the dilatation a " functional dila-
tation " ? This question must arise from general considerations and
from special considerations. Generally there is reason to believe that
a functional dilatation accompanies functional activity; particularly
the proven functional dilatation of the vessels when adrenalin is
injected raises the question whether the same explanation does not
fully account for the chorda dilatation.

I say " fully " because I think it will scarcely be challenged that,
even if there were no vaso-dilatator fibres in the chorda tympani,
there would be some degree of dilatation of a functional nature when
it is stimulated. The question then really is this : Is the dilatation
due to stimulation of the chorda tympani the result of two mechanisms
or of one ? If the dilatator fibre in the chorda is put on its trial, can
it prove its title ?

Were I to act as its counsel and attempt this proof I would be
well advised to seek for some instances of dilatation, produced by
chorda tympani stimulation, which do not appear to involve the
functional activity of the gland. Two such may be cited.

(1) The most obvious case for consideration is that of the atropin-
ised gland. The experiment is one of the best attested in physiology,
namely, that if atropin is administered intravenously in certain doses
saliva will not flow on stimulation of the chorda tympani, but dilata-
tion takes place. With smaller doses of atropin saliva is obtained;
with larger ones dilatation is not evoked, but between these extremes
you may get the dilatation without the secretion. If the flow of
saliva is the true index of the functional activity of the gland, the
title is proved. But it is not a foregone conclusion that some sub-
liminal degree of functional activity may not take place unless it
finds its expression in a flow of saliva. Therefore it is necessary
to measure the metabolism of the gland. The results of numerous
experiments of this character are summed up in the table on p. 147.
In every case there is some degree of increased metabolism when
the chorda tympani is stimulated. Certain criticisms of the experi-
ments which we have quoted may rise in the mind of the reader.
The first is that on comparing atropinised with unatropinised glands
the increased metabolism elicited by the stimulation is considerably
less in the former than in the latter. That is true, but it is also true
that the dilatation is not maintained in the atropinised glands in the

The regulation of the supply of oxygen to the tissues 147

Effect of stimulation of chorda tympani on the
atropinised submaxillary gland.

Unstimulated

Stimulated

Exp.

Animal

Oxygen

Blood

Oxygen

Blood

Increase in

Increase in

used c.c.

flow c.c.

used c.c.

flow c.c.

oxygen %

blood flow °/0

1

Cat

0-024

0-40

0-027

0-36

11

40

2

Do-

0-026

0-94

0-034

32

31

240

3

Cat

0-018
0-016

0-25
0'34

0-026
0-023

0-157)
0-56}

44

106

4

> >

0-018

0-41

0-023

0-82

27

100

5

J )

0-020

0-60

0-031

2-6

55

333

6

» »

0-020

0-43

0-031

0-87

55

102

7

Dog

0-011

0-49

0-021

0-81

91

65

8

Cat

0-024

0-41

0-050

2-4

109

488

9

»1

1 0-026
j 0-022

0-25
0-25

0-046
0-049

2-2|
1-7|

50

812

same degree as it would be in the normal gland. The second assurance
that must be given is that this increased metabolism is not a fictitious
one caused by the increased rapidity of the flow. This was a matter
which greatly exercised me as it proved very difficult to devise satis-
factory control experiments ; however after testing numerous methods
of producing vaso-dilatation artificially, Prof. Franz Miiller(7) drew
my attention to the possibilities of yohimbin for this purpose, and in
collaboration we tested the matter and found that even the very
great amount of dilatation which can be induced by injecting small
quantities of this drug directly into the submaxillary artery does not
cause increased oxygen to be used by the gland.

(2) Bayliss(8) and Asher'9' pointed out another direction in which
it might be possible to find the necessary evidence for the proof of
the " title " of the vaso-dilatator fibre. They showed that when the
central end of the depressor nerve is stimulated on the one side, say
the right, the vessels of the left submaxillary gland dilate, whether
or no the sympathetic is cut, provided that the chorda tympani is
intact. The deduction is that a reflex stimulation of the chorda
tympani takes place which produces vaso-dilatation, and here it should
be said that there is no flow of saliva ; on the other hand it must also
be said that the phenomenon is only obtainable in a fraction of the
experiments.

Here again an appeal may be made directly to measurements of
the oxygen used by the gland to give an answer as to whether or no

10—2

148 Chapter IX

there is increased chemical action taking place in the gland. My
experience has been that in those cases in which I have been able
to obtain the reflex there has been increased oxidation, whilst in
the cases in which I have obtained no increased blood flow the
oxidation has remained normal. But the experiments which I per-
formed were few.

The answer then to the question, " Is it possible to demonstrate
that the vaso- dilatation which follows upon stimulation of the
chorda tympani involves a definite neuro-muscular vaso-dilatator
mechanism?" is "It is not possible on the evidence at hand either
to prove it or disprove it." The functional dilatation involved may
be held to account for all known cases of dilatation in the sub-
maxillary, but it is not proved to do so. The dilatation which takes
place on stimulation of the atropinised gland is of relatively short
duration. It is not impossible that under normal circumstances
dilatation may be instituted by dilator fibres and maintained by
metabolic products.

The pancreas. The transition from the submaxillary gland to
the pancreas is a natural one. Adrenalin does not cause any flow
of juice however from the pancreas; we must therefore give the
appropriate stimulus, namely secretin. The effect of secretin on the
vessels of the pancreas formed the subject of a research by Otto
May (10), who showed by plethysmographic tracings that there was an
increase in the volume of the pancreas when it was secreting and
also that there was an increase in the " pulse volume " on the tracing.
May attributes the dilatation to metabolic products acting on the
vessels, and it is certain that there is ample evidence of increased
chemical activity taking place in the pancreas itself inasmuch as
the amount of oxygen which the pancreas uses is increased about
four-fold when secretin is eliciting a flow of juice. The proof in the
case of the pancreas is however less satisfactory than in the case of
adrenalin in the submaxillary gland, because solutions of secretin
usually, and in the case of May's experiments admittedly, contained
depressor substance. The "depressor" substance is depressor in
virtue of the fact that it dilates the vessels all over the body.

Therefore the question really is this : Does the solution con-
taining secretin and depressor substance produce a greater degree
of dilatation in the pancreas than the same solution would do if the
secretin were absent ? The answer to this question is given by May
in the following words: "...there was an expansion of the small

The regulation of the supply of oxygen to the tissues 149

intestine and of the pancreas. But whereas that of the former
soon reached its maximum, and assumed its normal volume in less
than a minute, the vascularity of the pancreas continued to increase
gradually for some minutes ; indeed it was maintained as long as
the secretion of the gland (which began about two minutes after
injection) continued — usually a period of 10 — 16 minutes." Direct
experiments upon the outflow of the blood from the pancreatic vein
have given very variable results.

Before I leave the subject of secretin I must make some allusion
to the suggestion of Bayliss and Starling (11) that specific dilatation
materials existed for specific organs, for instance extract of jejunum
causes dilatation of the intestine but has no effect on the kidney.
Their statement of results however was rather of the character of a
preliminary communication.

FIG. 80. — Connexions of heart perfused from another animal. A, tube leading from
carotid artery of perfusing animal, into this drugs may be injected. B, tube leading
to jugular vein of perfusing animal. Samples from the heart are taken at E. The
clip is then removed from D and placed at C.

The heart. The heart presents a very close analogy in some
respects to the stibmaxillary gland ; this analogy has in fact been one
of the current themes of physiological discussion, its tissue like the
secreting tissue of the submaxillary gland has a double nerve supply ;
by playing upon this it is possible to increase or to decrease the
activity of the heart : we are face to face with the question, Do these
changes in activity alter the blood flow through the coronary system
and, if so, by what mechanism are the alterations brought about ?

In our research on the gaseous exchange of the heart, Dixon and
I (12) arrived at some data with regard to the functional regulation of
its blood flow, The arrangement of the experiment was as follows :

150

Chapter IX

Two animals were used, a large and a small one of the same species,
a cat and a kitten or a dog and a puppy.

The smaller animal's heart was perfused. The perfusion took
place after the method of Haymans and Kochmann (13) ; the blood
went from the carotid artery of the large animal to the aorta of
the small one, its course was thence through the coronary arteries,
capillaries of the heart-muscle and coronary veins back into the

FIG. 81. — Shows the rate of flow through the coronary vessels (dotted line) ; and the
output of carbonic acid (continuous line). The figures along the ordinates are in
each case c.c. per minute, those on the left-hand side of each curve refer to the
continuous line, those on the right-hand side to the dotted line. The periods are
arranged along the abscissae.

right side of the heart, out by the pulmonary artery and back into
the jugular vein of the perfusing animal. Between the pulmonary
artery of the perfused heart and the jugular vein of the perfusing
animal there was a T-tube. When one wishes to collect blood for
analysis one attaches a graduated pipette to the T-tube, opens the
clip and closes the tube at (7. The blood runs along the horizontal
pipette and its rate of flow can be measured directly.

The regulation of the supply of oxygen to the tissues 151

The activity of the heart was altered very much in degree in
different experiments, and in various ways, by the administration
of adrenalin, pilocarpine, atropine, chloroform, vagus stimulation,
etc. ; one general fact which emerged from the experiments was the
close relationship between the output of carbonic acid and the rate
of flow of blood. This is shown for the series of experiments in
Fig. 81. The correspondence between the two is extraordinarily
complete. It is incredible that such a coincidence should happen
merely by chance, but whilst it must be admitted, I think, that the

I II

FIG. 82. — Upper tracing represents record of puppy's heart. Lower tracing = blood
pressure of the perfusing animal. Period I = normal. Period II shows the effect of
injecting 1'2 c.c. of CHC13 water. The signal mark represents the time during which
the sample of blood was taken.

blood flow and the CO, output are in some way related, an analysis of
this relationship is necessary before the assertion that the change in
the cross-section of the coronary vessels was caused by the changes
in the metabolic activity of the perfused heart is warranted.

The other possibilities are as follows :

Drugs which tend to increase the activity of the perfused heart,
adrenalin for instance, would also tend to increase the activity of
the perfusing heart ; this would tend to raise the general arterial
pressure in the perfusing animal and so alter the blood flow through

152 Chapter IX

the perfused heart. This line of argument does not account for the
changes in blood flow. The drugs were always injected in the tube
leading straight to the perfused heart and in such small quantities
that whilst they affected it profoundly they had little or no effect
upon the general circulation of the perfusing animal. For example
take Exp. 6 (Fig. 81), in it there are two periods one before and one
after the activity of the perfused heart has been reduced by the in-
jection of 20 minims of chloroform water into the tube leading to the
perfused heart. In Fig. 82 both the tracing of the perfused heart
and that of the general arterial pressure are shown in the two periods
of the experiment. The general arterial pressure is scarcely altered,
the metabolic activity of the heart is much reduced. This reduction
is shown by the reduction in amplitude and frequency of the heart's
beats as seen in the tracing ; it is shown also by the change in the
metabolism.

Period I Oxygen taken in C02 given out

Before chloroform injection ... 3 c.c. per min. 8-8 c.c. per min.

Period II

After chloroform injection ... 0 '37 „ 1'9 ,,

As the actual effect of the chloroform in the coronary vessels them-
selves would be dilatation, we might therefore have expected that,
other things being equal, there would have been a more rapid flow of
blood through the coronary system of the perfused heart. Never-
theless this was not the case, the blood flow in the first period was
30 c.c. per minute, in the second 9 c.c.

Now that the effects of changes in the general arterial pressure
have been excluded and in some experiments the local effect of the
drug upon the vessels, what possibilities remain ? The direct action of
drugs on the vessels of the perfused heart may be still further elimin-
ated, for in some of our experiments the rate of flow was not altered as
the result of drugs but as the result of stimulation of the vagus of the
perfused heart ; this clearly does not affect the general arterial pres-
sure of the perfusing animal, nor is it claimed that the vagus carries
constrictor (fibres to the coronary vessels ; therefore the changes in
blood flow which we get as the result of vagus stimulation cannot be
regarded as due to either of the causes which we have considered.
The rate of blood flow however follows the C02 output very closely.

In the experiment, a tracing of which is given, there were three
periods : (1) before vagus stimulation, (2) during vagus stimulation,
and (3) after vagus stimulation.

The regulation of the supply of oxygen to the tissues 153

Period

Condition of
heart

Rate of flow through
coronary vessels
c. c. per min.

Oxygen consumed
per gram per min.

CO2 given out
per gram per min.

I
II
III

normal
vagus
after vagus

6-7
4-6
6-0

•014
•009
•022

•038
•005
•015

Another possible cause of the changes in the rate of flow through
the coronary system is the rhythmic contraction of the cardiac
muscles, which may be held to promote a more rapid flow of blood
along the coronary system.

in

FIG. 83. — Record of the movements of the heart of a small cat perfused from the circulation
of a large cat. Upstroke = systole. Period I = normal. In period II the signal mark
represents the time of vagus stimulation. The third period corresponds to the after
effect and in this period the third sample of blood was taken.

At first sight it seemed difficult to analyse the complex of
processes into the elements necessary for a decision whether the
variations in the activity of the heart produce their effect on the
blood flow by the mechanical method of squeezing the blood inter-
mittently along the vessels, or by the chemical method for producing
dilator substances. Fortunately nature has performed the analysis
for us. The output of C02, as I have already indicated in Chapter VI,
lags to some extent behind the actual variations of the frequency and
power of the contractions. The time relations of the activity and the
C0.2 output are shown in Fig. 84, also those of the changes in blood

154 Chapter IX

flow. The alterations in the rate of flow keep company with the
C02 output and like it lag after the changes in pulse rate. The
simultaneous "hysteresis" of the vascular dilatation and of the CO2
production shows that the increased blood flow caused by increased
activity is not due to a mechanical propulsion but to a chemical event.
In order to test the effect of metabolites on the coronary circula-
tion more completely Dixon and I determined to induce partial
asphyxia by restriction of the respiratory orifice of the perfusing
animal. The carbonic acid accumulates in the blood and as the
arterial blood gets darker the flow through the coronary vessels
of the supplied heart increases, at the same time the heart beat
slows it becomes less- efficient (14) (the call for oxygen being unabated
and tending if anything to increase).

Heart of puppy perfused from Dog.

The occlusion of tracheal tube commenced between periods I and II and ended

after period III.

CO2 used per g.
per min. c.c.

•043
•035
•008
•012

Kitten's heart perfused from Cat.

I 12-1 26-6 1-75 -010 -005

II 5-4 41-1 5-7 -Oil 002

Whilst the dilatation took place at the height of asphyxia this
experiment differed from those shown in Fig. 81 in that the C02
output in this period was particularly small. A possible explanation
of this (though not the only one) was that the asphyxiated blood
contained products of incomplete oxidation which likely enough
were more active than C02. Indeed Verzar has since shown that the
blood coming from the titanised gastrocnemius contains measurable
quantities of acid products*151. We performed further experiments
in which the effect of C02 was eliminated whilst that of other products
remained. The asphyxia was produced by making the perfusing
animal breathe a mixture of nitrogen and oxygen which became
increasingly poor in the latter. When the oxygen sank to about
4 % the heart beats began to fail. Period II corresponds to this.

Oxygen in CO2 in Blood flow 02 used CO2 prod.

Period art. blood °/0 art. blood °/0 per min. per g. per min. per g. per miu.

I 15-4 35 8 -036 '025

II 2-1 36-2 48 -033 —

Period

Oxygen in

art. blood °/0

CO0 in art.
blood %

Blood flow
c.c. per min.

02 used per g.
per min. c.c.

I

16-0

28-3

11-4

•027

II

13-6

34-9

11-4

•027

III

8-2

44-9

23

•036

IV

15-3

21-7

1

•018

The regulation of the supply of oxygen to the tissues 155

From this experiment which we have repeated on several occa-
sions it would appear that CO2 is not the only, nor is it the most
powerful dilating metabolite.

Before leaving the subject of the automatic supply of blood to
the tissues there is one aspect of it which must not be overlooked.
The metabolic products continue to exercise their influence for some
little time, amounting perhaps to minutes, after the functional
activity of the organ has passed off. Whether we call this a happy
accident or a beautiful mechanism it matters not, so long as we
understand the true inwardness of the facts which we are studying.
For the demand of the organ for oxygen is essentially something

10

FIG. 84. — Bates of pulse and CO., excretion and blood How from the coronary system.

which follows upon the activity of the organ, not something which
causes the activity nor yet something which absolutely synchronises
with it. The time relations of the activity of an organ and the
oxygen used up in it make a story which I have already told. So
far as can be gleaned from muscle and from the submaxillary gland,
organs which by means of nervous stimuli can be thrown instan-
taneously into violent activity, activity which can be suspended
almost as rapidly by a cessation of the stimuli, it would appear
that contraction of the muscle or the secretion in the gland is not
itself a manifestation of oxidation in the sense that the work of an
internal combustion engine is a direct manifestation of the oxidative

156 Chapter IX

explosion in the cylinder. The order of the processes is reversed ;
the contraction of a muscle is more like the running down of an
alarum clock. The clock is already wound, at a given moment the
potential energy of the spring is released and the alarm sounds.
It has then to be laboriously rewound. In some such way the energy
of muscle must be reinstated. During the period of reinstatement
oxidation is increased and blood is required. It is then automatically
supplied.

The very orderliness of the mechanism which I have described
has sometimes made me picture what would happen if it became
disordered ; if some cataclasm should take place in the cell as the
result of which it shed its products broadcast. The immediate effects
would be redness and dilatation but these words would but feebly
describe the condition of affairs. The increased vascular dilatation
might lead to swelling, but I would refer the reader to the work of
Martin Fischer <16), who supposes that acid can produce swelling
directly. Fisher's conclusions await further experimental support.

If the wholesale production of metabolic products can produce
rubor, color, and tumor, it must produce dolor in their trttin, and so
we get the picture of the cardinal symptoms of inflammation produced
not merely by a cataclasm in the cell, but by a perversion of a beautiful
physiological mechanism, a perversion which has been seized upon by
" nature " for the ultimate removal of the cause of the calamity.

REFERENCES

(1) Barcroft and Shore, Journal of Physiol. XLV, p. 296, 1912.

(2) Burton-Opitz, Quarterly Journal ofExp. PA?/m>/.ni,p.297, 1910; iv,p. 113, 1911.

(3) Severini, quoted from Gaskell (4).

(4) Gaskell, Journal of Physiol. in, p. 48, 1880.

(5) Dale and Laidlavv, Ibid. XLIII, p. 182, 1911.

(6) Barcroft, Ibid, xxxv, p. 19, 1907 and xxxvi, p. 53, 1908 ; Carlson, Greer and

Becht, Amer. Journ. of Physiol. xx, p. 180, 1907 ; McLean, Ibid, xxn,
p. 279, 1908 ; Barcroft and Piper, Journal of Physiol. XLIV, p. 359, 1912.

(7) Barcroft and Muller, Ibid. XLIV, p. 259, 1912.

(8) Bayliss, Ibid, xxxvn, p. 256, 1908.

(9) A slier, Zeitschr.f. Biologie, LII, p. 325, 1909.

(10) May, Otto, Journal nf Physiol. xxx, p. 400, 1904.

(11) Bayliss and Starling, Ibid, xxvm, p. 357, 1902.

(12) Barcroft and Dixon, Ibid, xxxv, p. 182, 1907.

(13) Haymans and Kochmann, Arch. Pharm. et de Tlierap. xm, p. 379, 1904.

(14) Report of the Brit. Assoc.for the Advancement of Science, 1907, pp. 401 — 2.

(15) Verzar, Journal of Physiol. XLIV, p. 256, 1912.

(16) Martin Fischer, Oedema.

CHAPTER X

THE UNLOADING OF OXYGEN FROM THE BLOOD

THE transference of oxygen from the red corpuscles to the tissue
cells involves at least two quite separate processes, firstly the chemical
breakdown of the oxyhaemoglobin and secondly the diffusion of gas
from the blood to the cells, through the capillary wall and through
the lymph. It is not very easy to separate these two processes so
completely that each can be considered apart from the other. Never-
theless it will be useful to start from a definite point of view with
regard to the principles involved in the maintenance of an efficient
diffusion ; by this we mean a flux of oxygen which will provide
for the maximal needs of the organ, not merely for its metabolism
during quiescent periods.

The point of view is simple enough — for the maximal flux the
maximal pressure -head is required. Herein lies physiological signi-
ficance of the factors which have been set out in our consideration
of the physical chemistry of haemoglobin. Let me place before the
reader the following picture, which when he has considered he may
vary to his liking. A certain quantity of haemoglobin is passing
through the vessels of an organ — it must needs impart a given
quantity of oxygen to the organ — in so doing it becomes reduced
to the extent of 50 per cent, of its oxygen, the haemoglobin is in a
pure salt-free solution and the temperature is 16° C. Can we dis-
cover the available pressure-head of oxygen in the blood leaving the
capillary ? It will be found from the dissociation curve of dialysed
haemoglobin at 16° C. that the pressure corresponding to 50 °/0
saturation is O'Smm. As the pressure in the tissues cannot be less
than zero, the figure 0*3 mni. stands for the pressure-head which is
available for the maintenance of the diffusion current through the
capillary wall just at the point where the haemoglobin is parting
with the last traces of oxygen which it loses. In the following
paragraphs I will focus the consideration of the reader on the

158

Chapter X

difference of pressure which for the sake of simplicity of expres-
sion I will call the "final capillary pressure-head." To what extent
have the effects of (1) rise of body temperature, (2) the presence
of salts in the blood, and (3) the presence of CO2, contributed so as
to admit of the reduction of the blood at a higher pressure? We
may take these in any order. Each of these factors — temperature,
salts and CO2 — has a general and a local effect, a general effect
because the whole blood of the body is at 37° C. and is impregnated
with salts and C02, and a local effect because the blood in the capil-
laries rises in temperature, acquires carbonic acid, and in some cases
at all events alters in its saline content as it traverses the tissue.

100

0 10 20 30 40 50 60 70 80 90 100

It will be best to consider the general effects of temperature,
salts and acids first. Let us take them in reference to the case
described above. Rise of temperature does not alter the mathe-
matical form of the curve, it merely alters the scale on which it is
drawn. From Fig. 85 which we have just been considering we can
see the effect of temperature : haemoglobin at 38° C. when 50 %
reduced is in equilibrium with oxygen at a pressure of about 7'5 mm.
of mercury ; this then becomes the " final capillary pressure-head," it
is twenty-five times higher than it would have been at 16° C., hence
oxygen can diffuse at 7*5 mm. pressure at least out of the vessels into

The unloading of oxygen from the blood 159

the tissue and therefore at twenty -five times its former velocity. Let
us take the influence of salts next. The effect of salts is to aggregate
the molecules of the haemoglobin ; this aggregation introduces the
double contour into the dissociation curves, causing the blood to give
out its oxygen much more readily at low oxygen pressures whilst it
takes it up more readily at high ones. Compare, for instance, the
thick lines in Fig. 86 A and B. The former is the dissociation curve
of dialysed haemoglobin at 37° C., the latter, that of human blood in
the absence of C02 at the same temperature, is very close to that of
haemoglobin in a solution of potassium chloride isotonic with blood.

The salts not only produce a clumping of the haemoglobin
molecules so that each clump has on the average 2'5 molecules,
but they have a very important effect in maintaining the state of
aggregation at this level in spite of physiological changes in the
reaction of the blood. The effect of C02 in moderate quantities in
the presence of the salts is therefore to alter the value of K in the

equation

y Kxn

100 ~ 1 + Kxn '

without more than a negligible alteration in n ; the effect of 40 mm.
CO2 pressure on the dissociation curve of blood is to replace the
thick line in Fig. 86 B by the dotted line, thereby raising the " final
capillary pressure-head" to 25 mm.; indeed in the blood of many
persons the oxygen pressure corresponding to 50 °/0 saturation is
as high as 30 mm. of mercury. This marvellous thing has there-
fore happened as the result of the combined effect of temperature,
salts and carbonic acid; the "final capillary pressure-head" has
become elevated about 100 times, making provision for oxygen to
diffuse into the tissues at 100 times the speed — and this without
sensibly reducing the percentage saturation of the blood in the
pulmonary alveoli. For the sake of simplicity we have only dis-
cussed the capillary pressure of oxygen at one point in the capillary,
namely the point at which the blood leaves it. The reader must
recollect that diffusion of oxygen is taking place all along the capillary
and he may work out for himself the effects of temperature, salts
and carbonic acid at any particular percentage saturation of oxygen
which he selects.

The effect of clumping produced by electrolytes is best grasped by
a comparison of the effect of a decrease in the equilibrium constant
(K} in clumped and unclumped solutions. A decrease in K alone

160

Chapter X

would simply spread out the curve. Compare the curves given in
Fig. 86. In both cases A and B, the dotted line shows a large pro-

20 30 40 50 60 70 80 90 100

10

0 10 20 30 40 50 60 70 80 90 100

FIG. 86. — A, dissociation curves of haemoglobin solution.

= dissociation curve of

haemoglobin at 38° C. - =the effect which C0.2 would produce if there were

no aggregation of the molecules. B, dissociation curves of blood. = without

C02. - = 40 mm. CO., pressure. Ordinate = percentage saturation with oxygen.

Abscissa = oxygen pressure.

portional increase in the pressure at any given percentage saturation
as compared with the thick line. At low percentage saturations,

The unloading of oxygen from the blood 161

however, the absolute pressure in the case of haemoglobin solutions
containing unaggregated molecules is so small that a considerable
proportional increase does not produce much absolute rise ; with
blood, however, the opposite is the case. This is most apparent near
the bottom of the curves. In the absence of C02 at 20 % saturation
the corresponding pressure of oxygen would be 2 mm. in the case of
the haemoglobin solution and 6 in the case of the blood ; suppose
enough C02 to be present to double each of these, the pressure of
oxygen in the case of the unaggregated haemoglobin solution be-
comes 4 and of the blood 12, the blood gains greatly in the matter
of millimetres pressure. The effect of the salts in steadying the degree
of aggregation of colloid molecules is probably very great, for in the
few observed cases where the change of reaction has been sufficient
to produce a measurable change in aggregation, in spite of the salts
present, the person observed has been in a very abnormal condition,
either very much distressed as the result of exercise at high altitudes,
or practically moribund as some of the cases referred to in the last
chapter of this book.

We have discussed the physiological significance of temperature,
acids and salts in their general aspect, but in regard to any particular
organ they have a special as well as a general aspect. The blood is
not the same when it leaves the organ as when it enters it ; in the
first place it is warmer ; in the second case it has acquired carbonic
acid and perhaps other acids ; in the third its saline content may have
been altered ; each of these will have its effect upon the pressure at
which the blood holds its oxygen. Of the salts we know nothing in
this connection, nor have we data which enable us to make any
allowance for the rise of temperature in the blood. It may be an
important factor in the case of actively contracting muscle. We
must therefore confine ourselves to the consideration of the effect
of the acids thrown into the blood in raising the head of oxygen
pressure and thus promoting rapid diffusion.

It remains for some physiologist in the future systematically to
determine the influence of acid and other factors which we have
enumerated in the various tissues of the body. He would then be
able to present a statement of the pressure of oxygen in the vein
leading from each of the organs which he had studied.

At present a single example must suffice to illustrate at once the
principles which are involved, and the order of oxygen pressure
which exists in the capillaries. The actual example forthcoming is

B. R. P. 11

162

Chapter X

furnished by the submaxillary gland. We start with (1) the arterial
blood, (2) the venous blood from the resting gland, (3) the venous
blood from the secreting gland. Can we tell the oxygen pressure in
all three of these?

Analysis gave the following data :

(1) The arterial blood was 94°/0 saturated with oxygen.

(2) The venous (resting) blood was 59 % saturated with oxygen.

(3) The venous (active) blood was 66 % saturated with oxygen.
From these we could infer the pressures of oxygen if we knew the

dissociation curves. The curves are probably all different since the
carbonic acid pressures may be supposed to be different, and possibly
60.

25 30 35 40 " 45 50 55 60
FIG. 87. — Ordinate = quantity of C02 taken up by blood in c.c. per 100 c.c. of blood.
Abscissa = pressure of C02. The points O correspond to the actual quantities of C(X
found in the samples of arterial venous and resting bloods. From these points the
C02 pressures are inferred.

other acids are introduced in varying quantities. We do not know
the carbonic acid pressures nor the concentrations of other acids
present — we have available data however concerning the quantity
of CO2 in each sample of blood which is as follows :
(1) Arterial 100 c.c. blood contains 36 c.c. CO2.

Venous (resting) „ „ „ 44 c.c. C02.

Venous (active) „ „ „ 29 c.c. C02*.

"* The venous blood coming from the active submaxillary gland is often anomalous
in that it yields less C02 when treated with acid than the arterial blood. The large
quantity of C02 and carbonates which are being discharged from the gland in the
saliva account for this phenomenon.

The unloading of oxygen from the blood 163

Our procedure is to expose each sample of blood to suitable known
pressures of C02 in a tonometer at 37° C. and to analyse the blood in
order to ascertain the amounts of C02 which it holds at these pressures.
The crosses in Fig. 87 represent such analyses. A graph is made for
each blood, relating the quantity of CO2 to known pressures of the
gas ; from this graph we can read the CO2 pressure corresponding to
the quantity held by each sample of blood. Were the samples alike
in the matter of other acids one graph would serve for the three, but
no assumption of that kind can be made, indeed it was known that
the reverse was the case.

These three graphs are shown in Fig. 87. From them it appears
that the carbonic acid pressures in the three samples of blood were
approximately :

(1) Pressure of C02 in arterial blood 36 mm.

(2) Pressure of C02 in venous (resting) blood 46 mm.

(3) Pressure of CO2 in venous (active) blood 34 mm.

We now proceed to determine the oxygen-dissociation curves for
the three samples of blood at or near these C02 pressures. For this
purpose the following data were obtained :

(1) Oxygen pressure 36mm. percentage saturation 60, 62 %>

mean 61 % • K = '000200.

(2) Oxygen pressure 43mm. percentage saturation 62, 65 %>

mean 63'5 u/0. K = '000147.

(3) Oxygen pressure 40 mm. percentage saturation 54, 54 °/0

mean 54 %. K = '000116.

The curves corresponding to these values of K are given in Fig. 88.
The arterial blood has been assumed to be the same throughout the
experiment, the curves given for it in the two portions of Fig. 88 are
identical. On the curve may be read off the pressures corresponding
to the observed percentage saturations of the blood with oxygen,
we thus find that the oxygen pressure is :

(1) In the arterial blood 93 mm.

(2) In the venous (resting) blood 39 mm.

(3) In the venous (chorda) blood 49 mm.

The two latter form the "final capillary pressure-head" for the diffusion
of oxygen from the capillary into the tissue under the circumstances
of rest and activity respectively.

This completes our survey of the general and the local mechanisms
for securing an efficient oxygen pressure in the capillary circulation.
In the examples which we have taken from the submaxillary gland,

11—2

164

Chapter X

general considerations applied to the blood as a whole have determined
a rise of pressure from about 0*3 mm. to 34 mm. in the blood leaving the
resting gland and 29 mm. in the blood leaving the active gland (these

100

90

80

70

60

50

40

30

20

10

10 20

30

40 50 60 7.0 80

90 100

100

90

80

60

50

40

30

20

10

JO

0 10 20 30 40 50 GO 70 SO 90 100

FIG. 88. — Dissociation curves of blood supplying and leaving the submaxillary gland,
determined at the pressure of C02 which was found in the blood in each case. Upper
series = resting gland. Lower series = gland during stimulation of chorda tympani.
— = Arterial blood =; Venous blood.

figures being read off from the arterial curves), whilst the local effects
of secretion of carbonic and other acids have secured an extra 5 mm.
pressure to the blood from the resting gland and 10 mm. to the blood

The unloading of oxygen from the blood 165

from the active gland. So much then for the pressure of oxygen in
the capillaries.

The amount of oxygen which will diffuse through the wall
depends, other things being equal, on the gradient of pressure
across the capillary wall ; let us therefore turn to the consideration
of the pressure of oxygen in the tissues.

The quantitative measurement of the extra-vascular oxygen
pressure in certain organs expressed in mm. of mercury has been
attempted by Verzar (1) during the portion of his Wissenschaftliche
Reise which he spent at Cambridge. The principle on which these
determinations were based was the following. The quantity of
oxygen which diffuses through the wall will depend, other things
being equal, upon the difference of oxygen pressure within and
without the capillary, and inversely on the distance of diffusion, i.e.
altogether on the pressure gradient.

If therefore Q be the quantity of oxygen which passes out per
minute, p the intra-capillary pressure and p' the extra-capillary or
intra -cellular pressure

Q cc p-p'.

Naturally Verzar's first endeavour was to test the current view
of intra-cellular oxygen pressure, which is that the pressure in the
tissues is nil. Though no quantitative experiments of this character
are simple or easy, this fortunately is one of the least difficult, for if

p' = 0, Q oc p.

On this hypothesis it is possible to institute a direct experimental
test as to whether the amount of oxygen which left the capillary
varied directly with the pressure of oxygen in the capillary. The
amount of oxygen can of course be directly measured from a know-
ledge of the oxygen in the blood going to and leaving the organ in
question, and of the quantity of blood which goes through it in a
given time. The measurement of p offers a much more difficult
problem. Theoretically Verzar should in each experiment have gone
through the whole gamut of determinations we have set forth in the
example given above. In practice we had to approximate. A first
approximation to the capillary oxygen pressure will be arrived at by
an application of the percentage saturation of the arterial and venous
blood to the dissociation curve of the arterial blood of an animal of
the same species. The measurement of Q involves the amount of
oxygen in the blood : if the oxygen capacity be also known the

166 Chapter X

percentage saturation is arrived at. We can, then, determine the
percentage saturation of the arterial and venous bloods and, by
laying them off on the dissociation curve, we can determine the
pressure of oxygen in the artery and the vein, and in that way at least
ascertain within limits the pressure in the capillaiy at a given time.
We need scarcely interrupt the course of our discussion to point out
to the reader the extreme fallibility of the methods in which we are
engaged. We are in the position of one navigating a difficult channel
in foggy weather. Nevertheless, it may be that the points which we
have to observe are sufficiently obvious to stand out even in the fog,
that in short there are fixed laws determining the pressure in the
tissues which can be appreciated by methods as fallible as the best at
present within our reach.

Let us assume then that we can determine Q and p. Can we
now vary one or other of these and see whether the relation holds
Q cc p ? By the following means Verzar varied p. The cat which was
anaesthetised with urethane was made to breathe gas from a gaso-
meter by means of water valves attached to the trachea tube. The
gasometer contained air, but as the animal drew upon the supply
of air, the gasometer in turn drew upon the contents of another much
larger gasometer which was filled with nitrogen, this in turn being
displaced by water. Thus the cat was getting gas which constantly
and gradually became less and less rich in oxygen, and in time the
oxygen pressure in its arterial blood began to fall visibly as indicated
by the colour of the blood. When this process had gone far enough,
samples of blood were again taken for analysis.

For the following reasons it will be best to consider the experiments
made on the salivary gland first. They gave very uniform results
which were easy to interpret. They were free from the complica-
tions of alteration in the blood flow since it has been shown by Miiller
and myself12' that the amount of oxygen acquired from the blood by
the resting submaxillary is not appreciably affected even by great
changes in the rate of flow. In the third place experiments on the
submaxillary follow naturally upon the investigations described
above relating to intra-capillary pressure of oxygen in the same
organ. What a satisfactory organ the submaxillary gland is — easy
of access, easy of control, relatively unaffected by the usual
anaesthetics !

The experiments can best be understood by following a single
one in detail. Let it be Experiment I.

The unloading of oxygen from the blood 167

The data of this experiment are given in graphic form in Fig. 89.
There are two periods, the first in which the animal was breathing air,
the second in which the quantity of oxygen in the inspired mixture
swells to 8'4 per cent, or about 63 mm. The pressure of oxygen in the
arterial blood is indicated by the line Al — Az, that in the venous
blood by Vl — V2. The arithmetic mean of these pressures is shown as
Ci — C2 and is taken to be the capillary pressure. The question is,
Can we draw a line T± — T.2 representing the extra-capillary or tissue
pressure. Let us consider what the inferior limit of the tissue pres-
sure must be. At no time can it be less than zero. Let us suppose

100

Exp.I Exp.II Exp.III

FIG. 89.— Al — A2 pressure of oxygen in artery, F, — V2 in vein, Ci — (72 in capillary and
Tj — T2 in tissue, during the periods of each of three experiments on the submaxillary
gland. In period I the animal was breathing air, in period II a mixture of oxygen and
nitrogen poor in oxygen. Ordinate = mm. pressure. The numbering of the experi-
ments follow that in Verzar's paper.

therefore that when the intra-capillary pressure p is at its lowest
the extra-capillary pressure p' is zero. Then the point C» would repre-
sent p and T» would represent p'. But C2 = 16 mm., T2 = 0. Therefore
p -p' = 16, or rather p -p' cannot be greater than 16. To find the
inferior limit of the extra-capillary pressure in the first period, we
need the following additional data. The quantity of oxygen used by
the gland Q1 was '025 c.c. per g. per min. and in the second period
Q2 = "022 c.c. Q therefore, so far from varying directly with p',
remained approximately constant in spite of the fact that p sank
from 61 mm. to 16 mm.

168 Chapter X

We wish to find p' in the first period, in which p is represented by
by T,.

Q\ '• Qz '•'• GI~ T! : G2 — T o,

•025

but d = 61, .'. ^ = 43.

On the assumption that T7. = 0, T± = 43, the extra-capillary pressure
cannot be less than 43 mm. But also 7\ cannot be more than 43 for
this is the venous pressure, and by a reversal of the above calculation
To cannot then be more than zero. It appears therefore that within the
limits of experimental error the line T^ — T2 does represent the extra-
capillary or tissue oxygen pressure, and that this almost coincides
with the oxygen pressure in the vein, so that the blood has nearly
got into equilibrium with the tissue before it leaves the gland.

The other two experiments tell the same story. The oxygen
pressure in the tissue is, within the limits of experimental error, equal
to that in the venous blood. So far from the intra-cellular oxygen
pressure being nil, as it is usually stated to be in elementary books on
physiology, it is rather considerable, over 40 mm. in each of the three
experiments cited.

Herein lies the importance of this fact : if the intra-cellular pres-
sure is 40 mm. with a certain value for Q, there is room for it to
fall down to zero, in which case the pressure gradient would increase
from 18 mm. to 61 mm., providing for a corresponding increase in the
value of Q should the cells demand it.

We have no corresponding data unfortunately with respect to the
active submaxillary gland ; so the submaxillary story must end here.

We pass to skeletal muscle ; the experiments are not so uniform,
nevertheless they seem to justify certain positive conclusions. I will
take the two extreme (Exps. V and VII in Verzar's paper) cases as
examples because they are satisfactory inasmuch as the blood flow
in each remained fairly constant. The first one for consideration is
Exp. V(Fig. 90). In this the values of d and (72 were 43 and 19 mm.
respectively. Let us endeavour again to ascertain the value of 7\ by
assuming that T2 is nothing.

Q! = '043, Qo = '016 c.c. per min., <72 - T. = 19 mm.,

•A4.Q

/. d-7^19 x i = 50mm.
DID

The unloading of oxygen from the blood 169

But Ci is 43 mm., therefore 7\ would work out at — 7 mm. As the
experimental error was not much more than 7 mm. the true value for
T.2 cannot have been much above zero.

This is not the point I wish especially to emphasise. The real
contrast between this experiment on muscle and those on the
salivary gland lies in the fact that in Period I the intra-cellular
pressure was 24 mm. below the venous oxygen pressure.

100

Exp. V Exp. VII

FIG. 9Q.—Al} A2, A3, oxygen pressure in arterial blood. Vlt F2, F3, oxygen pressure in
venous blood. Cj, C2, C3, mean oxygen pressure in capillary blood. Tlt T2, T3,
oxygen pressure in tissue, in two experiments on the gastrocnemius muscles. In the
first period the animal is breathing air, subsequently a mixture of oxygen and nitrogen
poor in oxygen. Ordinate = mm. pressure. The numbering of the experiments
follows that of Verzar's paper.

To consider Exp. VII next. There are three periods in the
experiment, the following are the data tabulated on the basis of

T, = 0 :—

i ii in

74 mm. 30 mm. 32 mm.

•0061 c.c. -0039 c.c. -0035 c.c.

26 mm. 0 mm. 4 mm.

Period
C

Q
T

The figure four in the last column has, of course, no significance,
not differing materially from zero. If Ts be assumed to be zero instead
of T2, T! would work out at about 20 mm., a figure which agrees
with that given by the independent method of calculation employed
by Verzar(1). The highest computation for muscle is 27 mm.*

* Value used on p. 178.

170 Chapter X

It is clear then that in this experiment there was a positive
pressure oxygen in the tissues during Period I. I picture the course
of the experiment thus : as the value of C fell along the line Cl — (72,
that of T fell along a parallel line Tl — Tx until it reached zero, after
which the intra-cellular pressure would remain at zero and the
quantity of oxygen diffusing out of the vessels would vary with the
intra-capillary pressure. The other two experiments quoted by
Verzar are intermediate between these two.

It seems to me to be certain (1) that there is a greater difference
between the venous oxygen pressure and the intra-cellular oxygen
pressure in muscle than in the submaxillary, (2) that the blood
therefore does not leave the muscle in equilibrium with the tissue,
(3) that there is at times a definite though small pressure of oxygen
in resting muscles.

Two other organs have been investigated, namely the kidney and
the heart. Of the latter we have nothing certain to say. It proved
impossible with varying capillary pressure to keep the quantity of
oxygen used by the heart constant. In that respect it resembled
muscle. On the other hand it also proved impossible to maintain
the beats at a constant rate under conditions of oxygen want. With
regard to the kidney very interesting results were obtained, but they
were also less simple than in the case of muscle and the submaxillary
gland. When the capillary pressure of oxygen falls, the amount used
by the kidney (Q) rises provided that the amount of oxygen brought
to the kidney is at least sufficient to allow it to rise.

For the moment I will pass over the interest of this as a problem
in metabolism and consider it solely as a problem in physics. It is
clear that whatever be the need of the cell for oxygen there cannot
be less than no oxygen pressure in it. Therefore let us consider the
limiting case where p = 0.

At this point Q is observed experimentally to have, and therefore
(p— £>') must have, its maximum value. With higher values of the
capillary pressure p, Q is observed experimentally to be less, and
therefore (p — p) must be less. Thus p', the tissue pressure, can no
longer be zero, but must be even nearer to p the average capillary
pressure than it was before. But p' can in any case not be greater
than the venous pressure of oxygen, for otherwise the blood could
never by diffusion have got reduced to its venous condition. But in

The unloading of oxygen from the blood 171

point of fact the venous Go-pressure was very small, the venous blood
was almost reduced, in the case when the animal was given only 4°/0 02
to breathe ; and assuming the tissue O3-pressure p' in this case to have
had its minimum value of zero, it was even so not so very far from the
venous Go-pressure. Thus when p is greater, and p — p less, p' must
be even closer to the venous O2-pressure than it was when p was
small. We see therefore that in the kidney the Go-pressure in the
tissue is determined very largely by the Go-pressure of the venous
blood, and is only definitely less than it when the venous pressure is
small owing to incomplete saturation of the arterial blood.

It therefore seems that the kidney like the submaxillary is an
organ in which the oxygen pressure in the tissue approaches that in
the venous blood, which, in the renal veins of animals breathing
normally, is remarkably high. Further than this we cannot make any
statement about the quantitative values of oxygen pressure in the
tissues except that in such organs as the parotid and the sublingual
the conditions are likely to be similar to those prevailing in the sub-
maxillary.

We have then two distinct types : in the first class the difference
between the final capillary pressure and the tissue pressure is so
small that the tissue oxygen pressure approximates to the pressure
of oxygen in the venous blood; in this class are the glands which
have been studied. In the other class, that of skeletal muscle, the
difference between the capillary pressure and tissue pressure is so
large that the latter is 25 mm. or less in the cases measured. In the
former organs the blood flow can be considerably diminished without
diminution of the oxygen taken in by the organ, in the latter it
cannot.

REFERENCES

(1) Verzar, Journal of Physiol. XLV, p. 39, 1912.

(2) Barcroft and Miiller, Ibid. XLIV, p. 259, 1912.

CHAPTER XI

THE RATE OF EXCHANGE OF OXYGEN BETWEEN THE
BLOOD AND THE TISSUES

AT the risk of being thought to have taken somewhat of a jump
I shall discuss a very different aspect of the capillary circulation.

At the end of Chapter V, I gave a short summary of the effect of
salts, temperature and carbonic acid upon the form of the dissociation
curve especially in relation to its physiological significance. I showed
that the form of the curve was such as to give it a double aspect ; to
favour both oxidation of the blood in the lungs and reduction of
the blood in the tissues.

Whilst this is true it must be borne in mind that the dissociation
curve is essentially an equilibrium curve. One cannot tell from it,
how long will suffice for the oxidation of blood under any one set
of circumstances as compared with the time necessary to reduce it
under any other.

But the conditions of respiration are not statical ones ; in no
tissue is there equilibrium between the tissue and the blood run-
ning through it ; therefore useful as is the knowledge which has been
acquired from a study of the effects of salts, temperature and acids
on the final equilibrium between haemoglobin and oxygen, the reader
cannot have any complete view of the significance of these factors
without some knowledge of the way in which they affect the rates of
oxidation and of reduction of the haemoglobin.

I will therefore give some account of the experiments which have
been performed for the purpose of gaining such knowledge.

The theory of the experimental procedure was very simple and was
an expansion of that which Hill and I used in studying the rate of
reduction of haemoglobin and which Mathison improved. It was to
pass a uniform stream of nitrogen through blood and to observe after
successive intervals of time the degree of reduction which had taken
place.

Rates of oxidation and reduction of blood 173

The practical difficulties about a method which sounds so simple
are considerable. As compared with a haemoglobin solution great
trouble is caused in the case of blood by frothing, which may easily
become sufficiently serious to vitiate the experiment.

On the whole the best method was arrived at by taking advantage
of the tendency to froth. It consisted in allowing nitrogen to bubble
at a uniform rate through a tube 9 mm. in bore of the shape
shown in the figure. The tube was placed in a bath at a known
temperature. The gas entered at A. Each bubble, as it arrived
at the surface of the blood B, formed a film which was pushed up
the tube until at last it was broken by a spiral of wire greased with
vaseline.

FIG. 91. — A, rubber tube leading from gasometer attached to glass tube of capillary bore,
the volume of which up to the tap should be '1 — '15 c.c. B, bubble of gas about
to form a film. C, mark on glass bath for purpose of levelling the apparatus. I),
surface of water in bath at 37° C. E, spiral of greased copper wire.

The films were formed and driven up the tube with quite
remarkable regularity. It seemed that in spite of a certain degree of
empiricism the method did give reliable comparisons between one
sample of blood and the next. It is a matter for regret that the
method, unlike the dissociation curve, is not one which has any
absolute physical significance. The validity of the comparisons made
depends upon their all being done in the same tube under the same
conditions of inclination, &c. If experiments of this character could
be carried out under conditions which could be reproduced at will,
they would be the source of the most illuminating information.

At suitable intervals of time the bubbling was stopped, a little
blood was abstracted for analysis by taking off the tubing at A and
using the portion of the tube below the tap as an automatic pipette.

174

Chapter XI

Thus a curve was obtained of the degree of reduction which took
place at any given time.

The first experiment which I shall discuss was one in which
Nikiferowski helped me; it was on a dialysed solution of haemo-
globin, and at the temperature of the room 14 — 15° C. We
endeavoured to reduce the oxyhaemoglobin by passing commercial
nitrogen, which contained perhaps a millimetre of oxygen pressure,
through the tube at the rate of 30 bubbles in 15 seconds.

This proved to be a hopeless affair; at the end of 150 minutes the
haemoglobin was still 90 per cent, saturated with oxygen, nor had it
become appreciably reduced in the last hour and three quarters of this

40

60

80

TOO 120 140 160

FIG. 92. — Relative rates of oxidation and reduction of a dialysed haemoglobin solution at
16° C. Ordinate = °/0 saturation of the solution with oxygen. Abscissa = minutes of
time reducing gas nitrogen. Oxidising gas nitrogen + 100 mm. oxygen.

time. Another experiment was performed on the same haemoglobin
solution. The haemoglobin was completely reduced at a higher
temperature, cooled down to 15° C. in the tube and then nitrogen
containing oxygen to the extent of about 100 mm. pressure was
bubbled through the haemoglobin at the standard rate. The haemo-
globin oxidised rapidly, in 10 minutes it was 63 per cent, saturated
with oxygen and in 15 more it was 93 per cent, saturated.

The course of the experiment may be followed from Fig. 92.

It may occur to the reader to ask why we chose this particular
mixture of oxygen and nitrogen for the oxidation process. The
reason is that we wished to imitate as nearly as might be the oxygen
pressure which is to be found in the lungs during normal respiration.

Rates of oxidation and reduction of blood 175

Through the experiment which I shall describe we have adhered to
this pressure of oxygen for the oxidation.

The above experiment lives in my mind as being one of the most
instructive with which I have had to do. Consider the task which
is set to Nature in providing a medium for respiration — a medium
which is at one moment in the lung acquiring oxygen, at another in
the tissue imparting the amount it has acquired. Think, too, that the
blood must be prepared to yield up its oxygen in a space of time of
the same order as that in which it acquires the oxygen. Then turn
to Fig. 92 and think of the rapidity with which the haemoglobin
acquires its oxygen in the experiment and the tenacity with which it
holds to the gas. I believe the haemoglobin would have gone bad
before it had become reduced under the conditions of this experi-
ment. Truly Nature has been set a wellnigh hopeless task.

But with the advent of the salts the process of reduction becomes
much more easy, as is shown by the work of Oinuma(1), who com-
menced his research by performing the experiment on blood instead
of haemoglobin solution. Here also (Fig. 93 A) the rate of reduction
of the blood is slow out of all proportion to the rate of oxidation ;
still the disparity is not of that apparently hopeless character which
it assumes in the case of the dialysed haemoglobin solution.

The results of Oinuma's experiments turned out to be intensely
interesting. They were as follows :—

The rate of reduction was increased

(1) by the addition of C02 to the reducing gas,

(2) by the elevation of temperature.
The rate of oxidation was retarded

(1) considerably by the addition of C02 to the oxidising gas,

(2) to a very slight extent by elevation of temperature. [It
should be stated explicitly that this statement does not actually imply
a diminution in the velocity constant of the reaction Hb + O2 — >• Hb02,
for reasons given in Oinuma's paper.]

The effect of these factors will be seen at once from Fig.
93 A, B, c. In the first there is no symmetry in the relation
between the rates of oxidation and reduction. The oxidation takes
place with much greater rapidity than the reduction. But as the
conditions of the experiment approximate more and more closely
to those of the body, the rates of oxidation and reduction become
more and more nearly equal until at last, when the conditions of the
body are imitated as closely as is possible, the curves which represent

176

Chapter XI

the two phases of respiration are almost like an object and its image.
These curves are shown in Fig. 93 c.

10 2Q 32 _ «o - so - 60

so - go - TOO Oxidation

17-5° C. no CO,

Eeduction

100

80

60

40

20

w

Oxidation

37-5° C. no C02

Reduction

100

80

60

40

20

7

\

\

Oxidation

37-5° C.

+ 40 mm. pressure
of CO.,

Reduction

FIG. 93. — Relative rates of oxidation and reduction of blood. Ordiuate = percentage
saturation. Abscissa = time in minutes. A, temp. 17'5° C., reducing gas hydrogen,
oxidising gas hydrogen + 100 mm. oxygen. B, temp. 37'5° C., gases as in A.
C, 37'5° C., reducing gas hydrogen + 40 mm. C02. Oxidising gas hydrogen + 100 mm.
oxygen + 40 mm. CO.,.

The contemplation of this figure gave me a great deal of enjoy-
ment till it occurred to me that the process of respiration only
involved the upper portion of the curve in each case. Now the

Rates of oxidation and reduction of blood 177

upper portion of the curve of reduction is steep whereas the upper
portion of the oxidation curve is not steep but gradual. If these
curves give a true picture of the time relations of oxidation and
reduction in the body, it would seem that the blood is much more
rapidly reduced in the tissues than it is oxidised in the lung.

This check made me anxious to pursue the matter and find out
just what portion of each curve was involved in the ebb and flow
of respiration. Data were required, firstly as regards the curve of
oxidation. It is probably true that the blood in the lung of the
resting person is exposed to an oxygen pressure of about 100 mm.,
and therefore that I was justified in using the curve of oxidation
as given. But does the blood become saturated up to the point of
being in equilibrium with this pressure ? This is a point on which
the most diverse views are held by physiologists, who have estimated
the percentage saturation of the arterial blood at various figures from
88 per cent, upwards.

The variety of the estimates is no doubt a measure of the difficulty
of obtaining an accurate experimental procedure for the determina-
tion of the facts. They are based upon various experiments on
animals in which the respiration was probably upset to some extent,
either in one direction or the other.

The only obvious way of getting at the percentage saturation in
the arterial blood was to get some person who would submit to
having a cannula put in an artery as he sat in a chair in his usual
health. Some of the arterial blood could then be withdrawn without
coming in contact with air, and analysed. It chanced that shortly
after (Minima's experiments my path crossed that of such a person.
Arthur Cooke (2) decided on the extreme measure of performing the old
operation of transfusion. The patient was transfused from the radial
artery of his sister and I am indebted both to the surgeon and the
lady for allowing me to collect a sample of arterial blood at the end of
the operation, which proved to be 94 per cent, saturated with oxygen.
In my attempt to discover how much of the oxidation curve was in-
volved in the process of respiration, I took Fig. 93 c. as my starting
point. I marked the curve at the point of 94 per cent, saturation and
blackened from here downwards. The next question was where to stop.
To this question there is no definite answer. In different organs the
blood becomes reduced to different degrees. I should like to draw a
diagram for each organ in the body. But for the present let me
suppose that on the average half of all the oxygen which blood can

B. R. F. 12

178

Chapter XI

hold is taken, i.e that the blood is reduced from 94 per cent, to
44 per cent, saturation. The blackened portion of the curve of
oxidation was to represent that portion of the curve which is involved
in the actual oxidation of the venous blood.

Fresh difficulties assailed me when I endeavoured to apply the
same sort of process to Oinuma's curve of reduction Fig. 93 c. In the
first place was it justifiable to make use of this curve ? It involves
the assumption that the oxygen pressure in the tissue is nil. Verzar's
investigations were undertaken for the purpose of testing this as-
sumption. It proved to be incorrect for resting organs generally.
The oxygen pressure in the tissue, like the percentage saturation
of the blood leaving it, might be anything below a certain maximum.

80
60
40
20
n

\

\

\

5

\

\

\

>^Z

-—

\

\

\

\

*• —

100

80

60

40

20

\

L

\

\

10 20 30 40 50 60
FIG. 94.

10

20 30
FIG. 95.

40

50

FIGS. 94 and 95. — Bates of oxidation and reduction of blood. Temp. 37° C. Ordinate =
percentage saturation with oxygen. Abscissa = time in minutes. Fig. 94. — Upper
curve, reducing gas nitrogen + 27 mm. oxygen + 40 mm. C02. Lower curve,
nitrogen + 40 mm. C02. Fig. 95. — Keduction curves as in Fig. 94, oxidising gas
nitrogen + 100 mm. oxygen + 40 mm. C02 .

We imagined a case in which it is 27 mm.* Fresh determinations
were made by Nikiferowski and I for the purpose of relating the
curves of reduction of blood in the presence of 0 mm. O2 and of
27 mm. O2 respectively. The result is given in Fig. 94. The 27 mm.
curve of course came to a different base line from the 0 mm. curve,
inasmuch as blood in presence of 27 mm. oxygen comes to an
equilibrium at about 40 per cent, saturation.

Here then was my base line for the 27 mm. curve. It is the
upper reduction curve in Figs. 94 and 95. It now remained to be
blackened between the points of 94 and 44 per cent, saturation to
arrive at the portion involved in the reduction.

* See p. 169.

Rates of oxidation and reduction of blood 179

As I had carried my speculation so far, I indulged in the luxury
of patching the curves of oxidation and reduction together, or rather

too

10 20 30

50 60 70 80 90 100

100

1 00

FIG. 96. — Curves of oxidation and reduction of blood corresponding to conditions
approximating to those of A, oxidation in the lung and reduction in a resting muscle.
B, oxidation in the lung and reduction in a small active muscle. C, oxidation in the
lung and reduction in muscle during general muscular exercise. Ordinate = percentage
saturation. Abscissa — minutes of time.

the "^working " portions of the curves, in order to form a picture of
the relative rates at which the tide of oxygen rose and fell in the

12—2

180 Chapter XI

blood. Fig. 96 A was the result of this process. Now the symmetry
has returned.

I was troubled by the apparent waste of time which these curves
revealed, the blood acquired and lost most of its oxygen in so
small a fraction of the whole time spent taking up or losing the
whole. The acquisition of the last portion of oxygen (that acquired
between 85 and 94 per cent, saturation) was such a slow affair, and
similarly the reduction from 55 per cent, to 44 per cent.

Then I applied the principle that the mechanism of respiration
must allow of a margin for activity. The probable significance of
these horizontal portions of the curve is that they can so easily be
dispensed with. The first effect of increased oxygen consumption
of the organ would be to lower the oxygen pressure in the organ.
The sharp bend in the reduction curve would then be far below
the blackened part which goes down to 44 per cent. The tail of the
curve would therefore be eliminated from the "working" portion
which would become like that shown in Fig. 96 B.

This drop in oxygen pressure would have secured a rearrangement
of the time relations which would go far to meet the requirements of
some smallish muscle were it thrown into activity. For blood would
traverse this active muscle at a greatly accelerated speed, it would
then be thrown into the general circulation which would scarcely be
quickened. The blood corpuscle would traverse the muscle quickly
and the lung slowly. The comparative time relations of the reaction

Hb02 ^± Hb + O2

taking place under the existing circumstances appear in Fig. 96 B
and are admirably adapted to the velocity of the corpuscle in the
lung and the tissue respectively.

But my speculation carried me a stage further. Suppose that
the exercise was so extensive as to cause great acceleration of the
general circulation as well as a fall in the oxygen pressure of the
active tissues. Suppose the available time for the acquisition of
oxygen by the corpuscle is cut down to a quarter of its former
value. It would become about 85 per cent, saturated, with a little
time to spare. Of course enormously more oxygen would be leaving
the lung under these circumstances. The blood then reaches the
tissue 85 per cent, oxidised, the tissue extorts its 50 per cent,
of oxygen, and therefore reduces the blood to 35 per cent, satura-
tion. This it can do in approximately the same time as that in

Rates of oxidation and reduction of blood 181

which it had previously reduced the blood from 95 to 45 per cent,
saturation. Now the corpuscle goes back to the lung where it has
to be oxidised from 35 to 85 per cent, to complete the cycle.
This process can take place within the time allotted, namely one
quarter of the period which the blood formerly spent in the lung in
which the leisurely oxidation from 45 to 95 per cent, had previously
taken place. This cycle is represented in Fig. 96 c. The difference
then between Figs. 96 A and 96 c are as follows : firstly as regards
the assumptions made, in the first case the " high and low water "
marks for oxygen are 95 and 45 per cent, and the pressure in the
oxygen tissue is 27 mm. In the second case the marks are 85 and
35 per cent, saturation and the oxygen pressure in the tissue is nil.
These alterations in the assumed conditions provide for a four-fold
increase in the rate at which the haemoglobin can acquire and expel
a given quantity of oxygen.

There is a certain amount of evidence in favour of the view that
the arterial blood is slightly less oxidised during activity than during
rest. The experiments most to the point are those of Hill and
Nabarro<3). The average of their normal animals gave 17'44 as being
the oxygen in the arterial blood (15 animals). During the tonic
phase of fits induced by absinthe the average was 16'11, whilst in
the clonic phase 17 '21. The subject is one which might suitably be
reinvestigated with due precautions to ascertain the percentage satura-
tion in each case and to ensure that there was efficient oxygenation
in all phases of the experiment.

REFERENCES

(1) Oinuma, Journal of Physiol. XLIII, p. 364, 1911.

(2) Barcroft and Cooke, Ibid. XLVII; Proc. of Physiol. Soc. Dec. 1913.

(3) Hill and Nabarro, Journal of Physiol. xvm, p. 218, 1895.

CHAPTER XII

THE ACQUISITION OF OXYGEN BY THE BLOOD IN THE LUNG

WE have now arrived at the point at which it would be de-
sirable to apply our knowledge of the dissociation curve of blood
to the solution of a number of physiological problems, and to
give some complete history of the transportation of oxygen from
the lung to the tissue. Before doing this however we find that there
is a difficulty confronting us in the nature of the transition of gases
through the epithelium of the lung itself. The older physiologists
would freely have assumed that because the oxygen in the alveolar
air of the lung exercises a pressure of about 100 mm. of mercury, the
blood would be saturated with oxygen up to the point corresponding
with that pressure ; in other words, that the pressures of oxygen in
the lung and in the arterial blood would be approximately equal.
This assumption was based upon the work of the Bonn School of
physiologists.

If I am to give a coherent account of the controversy which has
taken place between those who believe that oxygen is secreted by the
pulmonary epithelium and those who believe that it diffuses from the
alveoli into the blood, I must go a little further into the history of
the subject than I have been wont to do heretofore.

The aerotonometer methods of the Bonn physiologists seemed to
have established the theory of diffusion in the lung on a satis-
factory basis when Bohr, using similar but not precisely the same
methods, found that there was

(1) A greater pressure of oxygen in the arterial blood than in
the alveolar air.

(2) A greater pressure of carbonic acid in the alveolar air than
in the arterial blood :

neither of which phenomena could be accounted for by simple

The acquisition of oxygen by the blood in the lung 183

diffusion ; though Fredericq performed yet other aerotonometric
experiments which supported the diffusion theory, the fact remained
that those who studied tensions with aerotonometers were divided.

Whilst Bohr's theory of pulmonary secretion was more compli-
cated than its rival, it had a certain charm about it which has
appealed to many minds. Its reasonableness was based upon morpho-
logical considerations which the diffusion theory entirely ignored.
It started with the swim-bladder of the fishes ; this organ in the
cod if emptied gradually refills, and the gas on analysis yields the
remarkable result that some 80 % °f its volume is oxygen. What-
ever secretion may be, here is a case of it. Is it strange that the
swim-bladder of the cod should have this power ? It is an offshoot
of the alimentary canal; if the cell of the stomach can secrete
hydrochloric acid; if the cells of diverticula, such as the pancreas
and the liver, secrete sodium carbonate, is there anything strange
about the idea that another diverticulum of the great secre ting-
mass — the alimentary tract — should secrete oxygen ? Rather would
it be strange if it could not secrete. Moreover the unity in
function between the swim-bladder and other diverticula, such as the
glands of the stomach, is made more evident when we discern, as
Bohr did, that the secretory function of the swim-bladder is governed
by the vagns just as is the secretory function of the stomach. Cut
the vagi — the swim-bladder will secrete oxygen as little as the stomach,
similarly crippled, will secrete pepsin.

Passing from the fishes to the other vertebrate types, the alleged
secretion in the lung was studied very carefully by Maar in the
tortoise. It proved to be possible, and in fact simple, to study the
gaseous exchange in each lung of the tortoise separately. To quote
Krogh on the subject: "While the necessary operation is rather
difficult in the mammals, and may vitiate the results by injuring the
animal, the tortoise appears to be specially adapted for this kind of
experiment. The trachea divides just behind the head, and the main
bronchi run parallel to one another and loosely connected along the
whole length of the neck."

Maar discovered that when the vagus was divided the gaseous
exchange of the lung was so altered as to make it clear that the
vagus has a very definite action on the pulmonary respiration of the
tortoise, but the action is not easy of interpretation. As explained
by Maar, however, there was a sort of tonic action, the vagus pro-
viding inhibitory nerves. The figures were regarded as being lifted

184 Chapter XII

out of the domain of vasomotor effects in the two lungs by control
experiments on atropinised lungs.

The experiments carried out by Maar on the tortoise were, to
some extent, supported by those of Spallitta, an Italian physiologist,
on the turtle. This author found that whatever mixture of gas was
put into the turtle's lung, and enclosed therein, the oxygen was
practically all absorbed, whilst the C02 was always found to be
about 6-5%.

Xor was the action of the vagus on the respiration of warm-
blooded animals overlooked in the Copenhagen laboratory. Henriques
studied the effect of brief stimulation of the vagus on the gas exchange
of dogs. Practically it resembled that of the tortoise, namely, it
inhibited the secretion of oxygen relative to CO2, and in this way
caused the respiratory quotient to approach unity. Here again
circulatory changes were supposed to have been excluded since the
change in the respiratory quotient was produced in cases in which
the circulation was both quickened and slowed.

We are not now criticising the case for secretion in the lung but
merely stating it, and from the statement it appears that right up
the animal kingdom there is evidence, which sometimes appears good
and sometimes bad, for the possibility of active secretion of gas, the
activity being influenced, sometimes in one way, sometimes in another,
by the nervous system.

The diffusion and the secretory theories were then rivals. The
one is simple, inasmuch as it demanded of the lung no greater
powers than are possessed by a piece of parchment, but it involved
no great biological generalisation ; it is a theory which attracts the
physicist because it presents a simple and intelligible relation of facts
as they are, but which repels the biologist because his training has
taught him to regard the different organs of the body as specialised
groups of cells which confine their activity to carrying out individual
fractions of the functional complex appertaining to living protoplasm.
The other theory may be said to be complicated, but at least it
demands no greater mystery than is conceded to almost every other
cell of the alimentary tract — the mystery of metabolic activity on the
part of the cell itself. This may be among the greatest of all mysteries,
but it may fairly be claimed that an even greater would be a cell
whose purpose was in no way a reflection of its own metabolic
activity.

Nor must the reader in judging the theory of pulmonary secretion

The acquisition of oxygen by the blood in the lung 185

forget its close analogy to the glomerular function of the kidney.
The epithelium covering the latter organ is, in a general way, similar
to that covering the lung ; how vast is the mass of physiological
literature which tacitly assumes that if secretion can be proved in
the amphibian capsule of Bowman, it may be taken for granted in
the case of the primates. We do not wish to insist unduly upon the
inherent reasonableness of Bohr's point of view, but we regard his
position as at least more reasonable than that of the lecturer who in
one lecture denounces Bohr's view as extravagant, and in the next
teaches a doctrine of universal glomerular secretion on the basis of a
long series of well-established experiments on frogs. Neither Bohr,
nor any of his supporters, have ever gone so far as this ; their position
has always been that even granting secretion of oxygen in the cod
and other low forms, its existence in higher ones remains a matter to
be decided by experiment.

About twenty years ago, it was therefore of the utmost importance
that some totally new method should be sought, which would decide
between the rival aerotonometricians. Such a method was devised
by Haldane(1), and in his hands and those of his collaborators, Lorrain
Smith and Douglas, has played a very important part in the physio-
logical thought of the last two decades.

The principle of the method may be summed up in a few words.
If haemoglobin be exposed to a mixture of oxygen and carbon
monoxide, the resulting quantities of oxy- and carboxy-haemoglobin
depend upon the relative pressures of oxygen and carbon monoxide.
In short there is a balanced action

CO + O2Hb ^± COHb + O2.

If therefore any three of the quantities involved are known, the
fourth can be calculated. Haldane administered CO in known
quantities, i.e. at a known pressure (for the pressure of CO in the
air was taken as the index of its concentration in the plasma) ; he
estimated the relative quantities of 02Hb and COHb present, and
from these data calculated the pressure of oxygen in the blood.

Before this could actually be done, however, a great deal of
ground had to be cleared ; in the first place a method of estimating
the relative proportions of oxy- and carboxy-haemoglobin had to be
devised ; in the second place the necessary data had to be accumu-
lated for determining what was the precise relationship of the four
substances in question in all possible cases in which equilibrium existed.

186

Chapter XII

Haldane and Lorrain Smith exposed haemoglobin to mixtures
containing varying known concentrations of oxygen and carbon
monoxide, and determined the relative amounts of CO and 02 haemo-
globin. In practice the process is somewhat simplified by the fact that
if as much as '1 °/0 CO — and that is more than would be desirable for
the experiment — be mixed with air, there is but a negligible diminution
in the partial pressure of oxygen. We may conveniently make up
mixtures of CO in air — assuming that the oxygen is 21 °/0 in each

100
90
80
70
60
50
40
30

10

7

•05

-15

•20

•25

•30

35

•40

•45

• 50

. 97. — Curve of Haldane and Lorrain Smith indicating the partition of haemoglobin
between oxygen and CO. Ordinate = percentage of the total haemoglobin present as
CO haemoglobin, the remainder being oxyhaemoglobin. Abscissa = percentage of CO

in air.

case — shake them with haemoglobin solution, and when equilibrium
is established analyse the air for CO and the blood for the relative
amounts of the two pigments. Plotting such a result, Haldane and
Lorrain Smith obtained the above curve, which relates the relative
quantities of COHb and 02Hb to the relative pressures of CO and
oxygen gas.

We have now at our disposal, in theory at all events, the data
for carrying out one of Haldane and Lorrain Smith's experiments
(we quote their own figures).

The acquisition of oxygen by the blood in the lung 187

Suppose -08% of CO was in the air inspired by the subject; when
he had been breathing the mixture long enough for equilibrium to
have been established, some of his blood was drawn, and the relative
quantities of oxyhaemoglobin and CO haemoglobin estimated, the
result being that there was 46 % of the latter and 54 % °f tne former.
If we consult their curve (Fig. 97) it informs us that the oxygen and
CO must have exerted in the fluid, with which the haemoglobin was
in contact, pressures related in the proportion of '06 of CO to 20*9 of
oxygen. But the actual pressure of CO was '08 % of an atmosphere,

20'9
therefore the oxygen pressure in the plasma was '08 x ^~= 27'9°/0

of an atmosphere For the sake of simplicity we have omitted a
slight correction for aqueous vapour, which does not affect the main
point, namely that the oxygen pressure in the plasma exceeds that
in the atmospheric air, and still more that in the alveolar air, being
in this experiment almost double the latter.

Perhaps there was nothing more convincing about the researches
which we have described than the experiments with which they were
controlled. Their authors seemed to show not only that there was
evidence of secretion in the lung, but that under conditions of im-
paired vitality that secretion disappeared. Moreover, this aspect of
the question transferred the problem from one of merely theoretical
importance into one of great practical interest to the clinician,
for if the body depends upon the secretory activity of the lung, and
this is easily impaired, then the primary object of treatment in the
case of lung complaints should be a care for this precious function.
It should aim at the restoration of the lung from a condition analogous
to that of dyspepsia.

Of mice for instance the authors say " After exposure to a cold
atmosphere for a short time their body temperature begins to fall,
particularly if they are exposed to a somewhat high percentage of
carbon monoxide, and they become torpid." It was easy to cool
down the surroundings of the mouse by placing the bottle containing
it in cold water. The result invariably was that the oxygen tension
went down to about 15 % of an atmosphere.

In the light of more recent work I am not now greatly concerned
to quote the results of Haldane and Lorrain Smith at great length.
I am concerned to state the problem in such a way that the reader
shall be in a position to criticise it intelligently.

Indeed for fifteen years it divided the physiological world in a

188

Chapter XII

very interesting way; practically the critics were on one side and
the experimenters on the other ; those who taught and wrote for the
most part refused to accept the secretory theory, whilst with the
exception of a few senior physiologists, such as Fredericq and Zuntz,
the workers at respiration inclined more and more to the secretory
theory. Indeed matters seemed to have reached a deadlock. The
diffusion theory was supported entirely by aerotonometer experi-
ments, which were held by the supporters of the secretory theory
to be fundamentally vicious, on the ground that the blood reduces
itself when removed from the vessels; this charge seemed to be
unanswerable — at all events it was unanswered. Its weight lay in
the fact that, at high oxygen pressures, a very slight amount of
reduction would produce a very big drop in oxygen pressure in the
blood ; therefore, if any appreciable reduction took place in the tono-
meter, the blood in the vessels might easily have had twice the
oxygen pressure which was observed in the tonometer. This criticism
remained unmet; moreover those who refused to accept Haldane's
results proved quite unequal to the task of showing where he had
gone wrong.

Experiments on the effects of fall of temperature on the
oxygen tension of arterial blood.

Animal

°/0 of CO in
air

Temp, of
bath

Duration of
exp.

Saturation of Hb
with CO %

0-2 tension in arterial
blood
% of an atmosphere

Mouse

•200

47-5°

45 min.

79-5

15

)5

•078

5-0°

30 „

61-6

14-2

))

•103

5-5°

45 „

65-5

15-7

))

•101

6-0°

52 „

65-8

15-6

Indeed at the time of which I am writing (six or seven years
ago) the supporters of the physical theory had, in their endeavour to
throw the burden of proof upon their opponents, frankly taken refuge
themselves in metaphysics. They began their arguments with the
general statement that we must always believe the most simple of
two rival theories unless it can be disproved, because it is the most
simple. This statement, like that freely used by the opposite faction,
namely, that we must believe the theory which offers the greatest
possibility of benefit to the organism because evolution demands it,
contains an element of truth. To my mind they break down at

The acquisition of oxygen by the blood in the lung 189

the same point ; they assume a knowledge on the one hand of what
is simple, on the other of what is advantageous, whilst as yet the facts
and principles at stake are admittedly unknown.

Let me illustrate my meaning by an analogy. It is difficult to
conceive of a more simple form of column than a cylinder. Yet
I have before my eye a hexagonal column — a much more complicated
figure — in order to be sure what it is I have to go round it, to measure
its angles and count its sides. This column has come from that
magnificent natural pile, the Giant's Causeway, in the north of
Ireland, where sea over sea of molten rock has shrunk and fractured
into a structure composed entirely of columns just similar to that at
which I am looking. The man who would urge the rightness of the
simple view may retort that his point of view is unshaken, that one
would be right to hold that Nature was more likely to produce a
cylinder than a hexagonal prism until evidence had been adduced to
the contrary, and that only now that I have the hexagonal prism in
view am I warranted in changing my mind.

My meaning has escaped such an one. It is that the simplicity
of the two stones has been considered apart from their setting, apart
from the forces which brought them into being, and especially with
regard to the simplicity of those forces. Behind the complicated
hexagon there is a simple law of contraction, behind the simple
cylinder there is no simple natural process, in fact there is no natural
process at all — it has been produced like the dialysing membrane by
the hand of man — it might conceivably have been produced from the
prism by some natural process of wear and tear, as a dead membrane
might have been produced from a living one, which would probably
like a diphtheritic membrane be sloughed off" by the body. Taken
in its setting, the column of basalt from the Giant's Causeway is
the simple figure. The truth of the proposition that the simple
process was the more reasonable remains, but whilst the processes
are themselves unascertained it is impossible to be sure that the
process which seems the simpler does not involve consequences far
more complex.

There were not wanting those who said that they could not get
results at all by the method pursued by Haldane and Lorrain Smith.
A confession of one's own weakness is not a convincing proof of some
one else's.

But the cup of the diffusion theory was not yet full. Quite
another line of argument was adduced which, if correct, showed that

190 Chapter XII

even on their own ground the supporters of the diffusion theory had
gone far to prepare their own downfall.

Those who opposed the secretory had thrown upon their antagon-
ists the whole onus of proving that there was a difference of oxygen
pressure between the arterial blood and the alveolar air. They said
in effect, " Unless you can show that the oxygen pressure in the
arterial blood is greater than that in the alveolar air, diffusion will
suffice to explain the passage of a gas from one to the other." Their
own aerotonometer results showed no measurable margin between the
two tensions ; they were therefore complacent enough. It was at
this point that Bohr replied, " Diffusion from one side of the alveolar
epithelium to the other must involve some difference of pressure ;
this may be immeasurably small, or it may not : the physical theory
at least presents the advantage that it offers the opportunity of
calculating the difference of pressure necessary to produce the flow
of gas."

He attacked the matter as follows. The first operation which
confronts a molecule of oxygen, which would pass from the alveolar
air through the epithelium, is the physical one of passing from the
gas into the fluid surface covering the lung. He therefore proceeded
to investigate the laws which governed the passage of gases into
fluid surfaces with the following result.

The argument was as follows :

(«) To calculate #, the number of cubic centimetres of gas which enters a surface
in a minute, let s be the area of the surface, p the pressure of the gas, and y the
invasion coefficient which is defined as the amount of gas which enters 1 sq. cm. of
the surface in one minute at the atmospheric pressure.

syp

Q= -*-=-

760

(6) To calculate &, the quantity of gas which leaves the surface of a fluid charged
with the gas in a minute, where s the area of the surface, £ the quantity of gas dissolved
in 1 c.c. of the fluid, and /3 the evasion coefficient which is denned as the quantity
of gas that leaves 1 sq. cm. of the surface in one minute when 1 c.c. of the fluid holds
1 c.c. of the as.

(c) Consider the special case of a fluid which is in equilibrium with a gas, the
pressure of the gas being 760 mm. Since the equilibrium exists, the number of
molecules of the gas which enter and leave the surface must be equal, and there-
fore

</ = &,
in other words

The acquisition of oxygen by the blood in the lung 191

but in this particular case p = 760 mm. and since the fluid is saturated with the gas
at 760 mm. the amount of gas which is dissolved in 1 c.c. of the fluid is a, the
coefficient of solubility, or the absorption coefficient of the gas in the fluid. Substi-
tuting 760 for p, and a for £, our equation therefore becomes

y = aj3.

Now since a, /3 and y are all coefficients, that is to say constant quantities, we
may, in any further consideration of the subject, replace /3 by yja, thus getting rid of
the evasion coefficient in favour of a which is easily determined experimentally.

(d) To calculate the volume M of a stream of gas passing per minute into a
fluid in which the concentration of the gas is £, the pressure of the gas in the fluid is

p', and without the fluid p.

M=g-b,

i.e. the difference between the amount of gas which passes into the surface and that
which passes out.

760

760 a
in the present case £ = «p'/760 ; substituting this value we get

(<?) For the human lung, M for a person not actively employed may be taken as
400 c.c. a minute, s as 90 x 104 sq. c. y was calculated by Bohr as being '012, p as 105 mm.,
and therefore the only unknown factor in the equation is p'. Now taking these
figures, the difference between p and p' is about 25 mm., and therefore p' is about
80 mm.

According to this calculation which contains nothing but mechanical reasoning,
the tension of oxygen in the fluid with which the epithelial cells of the lungs have to
deal cannot be more than 80 mm. or about 1 1 % of an atmosphere, i.e. is lower than
all computations of the tension of oxygen in the arterial blood as made by the
aerotonometer ; and the calculation was therefore regarded by its author as the final
and convincing proof of the secretion of gas by the epithelial cells from the place of
lower to the place of higher tension.

We may summarise the results obtained by the use of the invasion
and evasion coefficients, as follows :

(1) According to Bohr's statement of the diffusion theory, the
pressure of oxygen in the capillary would, apart from secretion,
be less than that in the alveolar air by at least the amount which
the invasion coefficient demands and which is measurable.

(2) The difference between the two becomes greater in direct
proportion to the oxygen used by the organism.

(3) The evasion coefficient of CO2 is such that the difference
between C02 pressure in the lung, and in the blood of the pulmonary
capillaries, is immeasurably small. Figure 98, given by Krogh,

192 Chapter XII

represents the actual measurement of the tensions of oxygen and
carbonic acid which he obtained, and which are not inconsistent
with the facts as stated above. So far, therefore, as our knowledge
of the difference of pressure necessary to maintain an adequate
current of oxygen through the lung is concerned, we may assume
that the oxygen pressure in the blood cannot exceed about 1 mm.
less than that in the alveolar air for every 100 c.c. of oxygen
absorbed by the organism, but we are at present somewhat in the
dark as to how much it is less than that amount.

The invasion coefficient applies only to the one portion of the
physical process of the passage of a gas through the lung, namely the
entrance of the oxygen into the moist surface. Bohr(3) made some
experiments shortly before the close of his life upon the physical
constants involved in the whole process — a subject which had pre-
viously been investigated by Hiifner'4', and Zuntz and Loewy(5).
What pressure head is necessary to force the amount of oxygen
which can be taken up by the human lung through the epithelium
in the time? The method of experiment is as follows: Carbon
monoxide is assumed to pass through the lung by diffusion, it is
then taken up by the haemoglobin. The amount that is taken up in
a given time can be measured. Within certain limits the pressure of
CO produced in the blood may be neglected. The amount which
diffuses through will then depend simply upon the external pressure
of CO. Now the pressure of CO in the alveolar air and the volume
of CO which passes through the epithelium in a given time are
both measurable and may be related directly to one another. This
relation is called the diffusion coefficient for the lung. When it is
once determined for CO it is assumed to have been determined for
oxygen from the known relation of the diffusion constants of the two
gases in physical experiments outside the body.

According to Bohr the diffusion coefficient for the lungs Avas of
such an order as to allow of the actual quantity of oxygen used at rest
to pass through the epithelium at the usual alveolar oxygen pressure,
but if his determination of the diffusion coefficient were correct it
would be impossible for the amount of gas which traverses the epi-
thelium during exercise to pass by diffusion, as also the quantity of
oxygen which goes through at rest, when the person is exposed to
low pressures of oxygen.

To sum up the situation as it was three or four years ago. The
diffusion theory rested upon aerotonometer experiments, which for

The acquisition of oxygen by the blood in the lung 193

the most part satisfied the teachers of the subject, but which more
and more failed to satisfy those who were actually at work upon it,
and there were not wanting signs that even in the " Landwirtschaft-
liche Hochschule" in Berlin some could be found who were prepared
to coquet with the views which emanated from Copenhagen and were
reflected from Oxford.

These views were supported by each piece of experimental evidence
that appeared. On the one hand they rested upon a biological
conception which could be illustrated from many parts of the animal
kingdom, on the other hand they leant upon data which were derived
from experiments upon man, not under operation, not under anaes-
thetics but in whom the normal functions appeared to have scope to
exercise their normal activity. Finally the physical theory seemed
to be disproved by the pushing of its own requirements to their
conclusion.

REFERENCES

(1) Nagel's Handbuch, i, pp. 54 et seq.

(2) Haldane and Lorrain Smith, xxn, p. 231, 1907.

(3) Bohr, Skand. Archiv, xxn, p. 221, 1909.

(4) Hiifner, Anncd d. Physik. u. Chem. Bd. LX, p. 141, 1897.

(5) Zuntz and Loewy, Arch.f. Physiol. p. 166, 1904.

B. R. F. 13

CHAPTER XIII

THE ACQUISITION OF OXYGEN BY THE BLOOD IN THE LUNG

(CONTINUED)

No apology is put forward for the point at which I have drawn
the dividing line between the present and the last chapter. It has
been drawn at the position in which the subject stood when Hartridge
commenced his work. The purely personal point of view from which
this volume has been written demands the treatment of the subjects
in a personal wray. But for Hartridge's work and a trifling contribu-
tion made by Cooke and myself the subject of pulmonary secretion
would probably have found no place in the book. The last chapter
then forms a statement of the problem as it appeared when Hartridge
commenced to work at it.

Whilst he was working, two researches of great importance ap-
peared, which quite altered the aspect of the problem. The first was
the work of Dr and Mrs Krogh(1), the second the work of Haldane
and Douglas12'. The first clearly proved that in animals under opera-
tion all the facts involved in the acquisition of oxygen could be
explained by diffusion. The second consisted in a withdrawal of the
secretory theory except during conditions of stress, i.e. either of
considerable exercise in which the demand for oxygen wras greatly
increased or of reduced barometric pressure under wrhich the supply
is diminished.

It will be desirable to treat of these two researches in some
detail.

The change which has taken place within the last three years has
largely been caused by an altered point of view with regard to the
requirements and possibilities of gas analysis as adapted to physiologi-
cal purposes. The generation wrhose hairs are grey in general held to
a sound chemical principle, namely that the greater the quantity of the
substance which could be obtained for analysis the more certain the

The acquisition of oxygen by the blood in the lung 195

analysis. Yet there are many limitations to this principle, especially
on the biological side. Many are the fortunes which have been lost
in commerce because the reaction which will take place in the test
tube will not take place in the works. The reasons are in many cases
apparent. I have been told that in a brewery of great repute the
chemical staff found it impossible to reproduce in the laboratory the
reactions which took place in the vat because the vat unlike the
beaker is practically an adiabatic system, its ratio of cooling surface
to volume as compared with that of the beaker is relatively infinites-
imal— its substantial walls are poor conductors of heat. With the
advent of an adiabatic beaker in the shape of a Dewar's vessel, the
conditions of the vat can be reproduced, and with the conditions
the reactions.

In the case of physiological requirements the younger generation
has grasped the principle that everything depends upon working
with the minimal possible quantity of gas.

This is so from whatever view you take it. Biological conditions
change so rapidly that the technique of the biologist must be one
which can be operated with great rapidity. Moreover they become
changed and abnormal by the mere abstraction of the quantities of
fluid required for the older methods. In the viscid fluids of the body
the process of abstraction of the gases is a much more difficult one
than it is in the case of the limpid liquids with which the chemist
deals, and is greatly facilitated by increasing the ratio of surface to
volume ; i.e. in general by decreasing the volume used. Then again
gas analysis is a very special sort of analysis inasmuch as the
measure of the gas is usually its volume, and this is extremely
sensitive to temperature and pressure. How great would be the
anxiety of the analytical chemist to reduce the amount of matter
which he weighed to the smallest possible bulk, if before he weighed
it he had to convince himself that it was of uniform temperature
throughout and that its temperature was known to a hundredth of a
degree. Nor have I stated the case as urgently as I might. I have
said that the measure of a gas is its volume. In practice this is not
the case, the measure of the gas is usually a measurement of length,
it is the distance between two points on a cylindrical burette. If
there is a temperature error, that will be three times as apparent as
it would be if the gas were free to expand or contract along all three
of its linear dimensions. It has therefore been the object of the
younger school of physiological gas analysts to cut down the quantity

13—2

196 Chapter XIII

of gas measured to the minimum quantity consistent with the preser-
vation of the measurement being a simple measurement of length.
This of course is accomplished by narrowing the bore of the burette,
and as it appears to be as easy to draw a capillary tube of uniform
bore as a more capacious cylinder, the real limit of the process
is arrived at where serious errors are introduced by unavoidable
inequalities in the surface of the glass. Moreover, gas in capillary
tubes is in many ways easy to manipulate ; if enclosed between two
fluid surfaces there is not the same tendency for the gas to escape.
You can lay the tube down, set it on its end, expel the bubble of
gas, withdraw it, manipulate it in all sorts of simple ways, and still
you have it, and have it under conditions in which it is, so to speak,
all surface.

Whilst in this country Brodie and I were pushing these principles
to what we regarded as considerable lengths for the purpose of
determining the gaseous exchange of the frog's kidney and found to
our own surprise that we could analyse 1/10 c.c. of gas with sufficient
accuracy, Krogh was working out a technique along the same
general lines, by the side of which ours appears crude to the last
degree. Nothing could be more beautiful or fascinating than his
method of analysing almost microscopic bubbles of gas, that is to say,
bubbles of about 1/100 c.c. This technique reformed aerotonometric
methods. A bubble of such small dimensions rapidly attains equi-
librium with any liquid with which it is in contact. If placed in
a stream of circulating blood, it is tossed to and fro, its surface and
that of the blood in contact with it are constantly changing, and the
stream in which it is may be so small that each corpuscle is out of
the animal but a very short time. Therefore there is no opportunity
for self-reduction of the blood, and moreover there is no amount of
haemorrhage which signifies. Krogh's method at once received the
acclamation of physiologists, partly because of its own beauty and
partly on account of the consciousness of its limitations which its
inventor betrayed. In a series of papers which are at once a model
of certainty of touch and modesty in presentation, he has revised
the whole field of the relation of oxygen pressures in the blood and
the alveolar air, in so far as it can be revised by methods of this
kind.

Krogh investigated the criticism levelled against aerotonometric
experiments generally, that their results yield too low an oxygen
tension owing to self-reduction of the blood. He found that self-

The acquisition of oxygen by the blood in the lung 197

reduction does take place, corresponding at most to '0007 c.c. of
oxygen per c.c. of blood per minute (or about half a per cent, of the
oxygen contained in the blood) and at least '0002 c.c. It is in some
respects fortunate, in others unfortunate, that Krogh chose the rabbit
as the object of this research. The blood of the rabbit was found
by Douglas to reduce itself with much greater violence than that of
higher mammals. If, therefore, the self-reduction does not matter in
the case of the rabbit — and it would seem that it must be very slight
in the infinitesimal time taken by blood to traverse a tube joining
two ends of the carotid artery — it clearly does not matter in the
higher types. On the other hand it is probable that if Krogh had
used the cat he would have obtained figures probably in the next
decimal place and therefore placed self-reduction in that animal far
out of the field.

As regards the actual tensions observed and the comparison of
them with the alveolar air, Krogh was at the disadvantage under
which all workers in this subject who depend upon animal experiments
find themselves, ihat of getting an actual sample of alveolar air to
analyse. Calculating it from the air of the bifurcation of the bronchi,
he found (1) that the C02 pressure in the alveolar air and the blood
were the same and kept the same whatever changes took place in the
former, (2) that the oxygen pressure in the blood was always below
that of the alveolar air by a certain amount, 15 — 20 mm., and that
this margin remained even with considerable variations of oxygen
pressure in the alveolar air.

For the moment I wish to focus the attention of the reader upon
this margin of 10 — 20mm., and to recall the fact that according to
Bohr the diffusion theory demanded for the human lung an excess of
about 25 mm. of pressure in the alveolar air over that present in the
film of moisture on the exposed surface of the lung cells. As well as
this a further difference of pressure was required to drive the oxygen
through the cells. Owing to the extreme solubility and diffusibility of
carbonic acid no measurable margin is demanded between the carbonic
acid pressure in the lung and in the alveolar air. Krogh's result then
agrees with Bohr's expansion of the diffusion theory inasmuch as he
shows a margin in the case of the oxygen and no margin in the case
of the carbonic acid.

The margin in the case of the rabbit was less than what Bohr had
postulated in the case of man, but that might easily be so. The
question, however, could not be left at this point because it would

198

Chester XIII

have meant shirking a difficulty that would surely have asserted
itself sooner or later.

To recall the statement made on p. 191

10 20 30 W 50

•spjredair

220 3ff

10

20 30 40

FIG. 98. — Oxygen and C09 pressures _ in the air at the bifurcation of the bronchi,

in the blood. Ordinate pressure of oxygen expressed in percentages of an

atmosphere. Abscissa = time. (Krogh.)

The margin referred to above is of course (p —p'}, its magnitude
is not less than and it varies directly with the quantity of

gas going through the epithelium. In the case to which I have
referred M is taken in man as 400 c.c. of oxygen per minute. It is
known, however, that 2000 c.c. of oxygen per minute is not an
excessive amount in cases of hard exercise in which case (p—p'}
would have to be increased 5-fold, that is, to at least 125 mm., in
other words p' would disappear altogether, seeing that even with
forced respiration 125 mm. probably exceeds the oxygen pressure in
the alveolar air.

The acquisition of oxygen by the blood in the lung 199

Therefore it became necessary to reinvestigate the value of the
invasion coefficient.

The bubble proved an excellent medium for this purpose.

"The method adopted is easily understood by reference to the
figure. About 250 c.c. of boiled distilled water contained in the
vessel (1) were saturated at 37° C. and at atmospheric pressure with the
gas in question which bubbled through (2), and left the vessel through
(3). The saturation was in each case continued for at least 24 hours
in order to drive out every trace of any other gas which might be
present. The oxygen was prepared by electrolysis of water. When
the saturation was completed mercury was allowed to flow into the
vessel until the water was above the tap (4) which was then closed.

In the experiments a current of water was conducted past a
bubble of gas identical with that in the microtonometer. When the
gas pressure in the bubble is equal to the tension of the dissolved
gas, the volume of the bubble must remain unaltered, but when a
definite difference is set up by raising or lowering the mouth through
which the water flows out, diffusion will take place through the
surface of the bubble, and the volume will become altered. The
mean surface being known and the alteration of the volume during
one minute being measured, the invasion may be calculated by the
formula given by Bohr

J/.760
p-p = .

...A uniform current of water was produced by the pressure from
the raised mercury vessel (5) which was suspended by the spring (6)
adjusted to maintain a nearly constant pressure in spite of the dis-
placement of the mercury. The flow was regulated by means of the
screw clip (7). When the ordinary clip (8) was opened water flowed
on to the microtonometer (9) and set the air bubble (10) in revolving
motion. The water flowed off through (11) and was collected and
measured in a vessel (12) which could be raised or lowered. The
water pressure was indicated by the level in the tube (13), placed on
a scale graduated in mm.

Before each determination the bubble was measured in the
graduated tube of the microtonometer. It was then carried down
into the funnel, the clip (8) was opened, and the pressure-gauges (13)
read off. After exactly one minute the clip was again closed, the air
bubble drawn up into the tube and measured anew, and the volume

200

Chapter XIII

of water which passed through the tonometer read off on (12) and
noted."

fl

FIG. 99. — Apparatus for the measurement of the invasion coefficient. (Krogh.)

As the result of 38 experiments tabulated, Krogh concludes that
the invasion coefficient for oxygen is 0*0776 ± 0'0024 as a minimum

The acquisition of oxygen by the blood in the lung 201

value. This value is nearly seven times that given by Bohr. If
therefore the calculation of p — p' for the human lung was repeated
with the new coefficient, it would seem to be not 25 mm. but about
3 — 4 mm. As a very rough calculation of the invasion coefficient
shows that it demands a difference of pressure on the two sides of
the lung surface of 1 mm. for every hundred c.c. absorbed by the
lung per min., it may seem to the reader that I have gone at undue
length into the invasion coefficient and that a matter which has been
proved by Krogh to be so trifling does not demand the extensive
discussion which I have given to it. This is so, perhaps, under the
ordinary conditions of life, but let him remember that at high altitudes
millimetres of oxygen pressure become precious. They are at best
all too few. If they are whittled away in passing through the lung
and if this process is magnified by the exact proportion of the amount
of work which is expended in climbing, then the reader will agree
with me that we do well to keep account of them one by one.

When the oxygen has entered the surface of the lung a further
pressure gradient must exist in order to drive it through the cells of
the epithelium. About this we have few reliable theoretical data.

In rabbits we have seen that the whole pressure head, as measured
experimentally by Krogh, amounts to 15 — 20 mm. (Fig. 98). How far
this reflects the condition of affairs in man we do not know. Rabbits
like all small animals use up a proportionately large amount of oxygen ;
other things being equal, man would use up less, but we have no
certain knowledge what the relative surfaces of the lung in man
may be as compared with the rabbit.

If we imagine a man situated at an altitude of 15,000 feet, his
alveolar oxygen pressure will be about 50 mm. of mercury. If now
we suppose him to be at rest and deduct 10 mm. from this for the
pressure head necessary to maintain the diffusion, the pressure of
oxygen in his blood will be 40 mm., and if further we suppose that
he does an increased amount of work, requiring twice the amount
of oxygen, and turn this 10 mm. into twenty we leave only a
possible 30 mm. as the oxygen pressure in the arterial blood. This
seems an impossibly low figure. However, as we have said, these
values which we have been using are speculative and 40 mm. oxygen
pressure would place the organism in a very different position from
30 mm. But the calculation illustrates the desirability of at once
doing experiments upon man himself and finding out under extreme
circumstances which, though abnormal, can exist, what the oxygen

202 Chapter XIII

pressure in the blood really is — after all, our theory must extend to
these facts. Beautiful, therefore, as Krogh's work is, its application
to man is beset Avith many difficulties.

Whilst Krogh emerged complete master of the field which he had
so successfully circumvented, his victory was to some extent a
Pyrrhic one. For whilst he had been laying his plans the enemy had
left the position which he was about to capture and they had retired
to another, which they disclosed almost simultaneously with the news
of Krogh's conquest. Their position as put forward by Haldane and
Douglas was this : Oxygen secretion is a function of oxygen want.
Under normal circumstances there is abundant oxygen and therefore
there is no oxygen secretion — diffusion is good enough — but when
the body demands more oxygen than it can readily obtain, whether
by reason of diminished oxygen supply or increased oxygen demand,
oxygen secretion in the lung sets in.

Krogh's investigations do not touch this new. situation, yet it is
one which cannot be neglected, if it is true that it affords one of the
most important and most beautiful forms of physiological adaptation.
That an essential consequence of exercise is an automatic means of
obtaining a forced draught for the necessary combustion is a physio-
logical conception which cannot be treated with anything but respect.
Moreover the conception is an extremely reasonable one, for were
there any possibility of secretion in the epithelium from which the
lung is derived, the process of natural selection would tend to
preserve it.

We must give some account of the reasons which led the Oxford
school to abandon their former position in favour of their present
one. The validity of the experiments of Haldane and Lorrain Smith
depends upon the correctness of the following assumptions.

1. That the curve given in Fig. 97 which was determined with ox
blood is correct for the blood of all the animals to which they sub-
sequently applied it.

2. That this curve which was determined at the laboratory
temperature, correctly represents the facts at the body temperature.

3. That the relative affinities of the blood of different species for
oxygen and carbon monoxide are the same.

4. That the relation of oxygen and carbon monoxide to blood is
unaffected by dilution.

5. That the same relation is unaffected by the presence of
carbonic or other acids.

The acquisition of oxygen by the blood in the lung 203

6. That all the haemoglobin is either oxyhaemoglobin or CO
haemoglobin. They really measured the ratio of CO haemoglobin to
total haemoglobin. In so far as there was reduced haemoglobin in
the blood on which they worked, this would vitiate their result.

We must add one consideration, which, though dealt with by
Haldane in another way, was not investigated quantitatively by him—
the influence of light.

These points had either been assumed or had been incompletely
investigated in the research of Haldane and Lorrain Smith to which
I have alluded. Such assumptions if not warranted were at least not
in serious conflict with the scientific knowledge of the time (1895).
But in the light of the additions to knowledge recounted in the first
five chapters of the present volume it will be seen that each of the
six propositions was eminently likely to be erroneous.

Probably the fourth and fifth are approximately correct. The
rest are certainly wrong in a greater or less degree. In the light of
these facts the work of Haldane and Lorrain Smith was done over
again by Haldane and Douglas (5) for the purpose of ascertaining
whether these errors were serious or negligible.

The Oxford observers concluded that carbonic and non-volatile
organic acids did not affect the partition of the haemoglobin between
the oxygen and the carbon monoxide, but that the balance was
affected to some extent by temperature and that there were individual
differences between different persons and different species which
could not be overlooked. But although these complications had
arisen Haldane and Douglas were able to balance their errors in
a way which had not been open to Haldane and Lorrain Smith, for
actual samples of alveolar air had become available by the discovery
of Haldane and Priestley. Therefore the blood of a person was
exposed outside his body to his alveolar air at 37° C., and within
it to alveolar air, modified by the pulmonary epithelium, each con-
taining the same concentration of carbon monoxide, and the partition
of the haemoglobin between the oxygen and the CO was compared
in the two cases. The titrations were both carried out in the same
light. If the proportion of CO haemoglobin in the blood from the
body exceeded that in the blood exposed to the alveolar air, the
pressure of CO being the same in each case, then it would follow that
the blood was exposed to a less pressure of oxygen in the lung than
in the alveolar air : and if the proportion of CO haemoglobin in the
blood from the body was less than in the blood exposed extra vitam

204 Chapter XIII

to the alveolar air then the blood must have been exposed to a
greater pressure of oxygen in the lung than in the alveolar air. The
results of some of their experiments are given in the table which
faces this page.

In discussing the work of the Oxford School I have intentionally
withheld all reference to their method of estimating the relative
quantities of oxygen and carbon monoxide associated with the
haemoglobin. Clearly the whole value of the results which they
obtained depends upon the accuracy of their estimation and no
critical account or investigation of the subject can omit to take
cognisance of it.

The essence of their method is as follows. Some blood is taken
from the object of the research before the CO is administered : this is
diluted and set on one side : let us call it A. The carbon monoxide
is administered : some more blood B is taken from the patient. B is
a different colour from A since it contains CO haemoglobin, but by
adding carmine to A the colours of the two may be matched. The
amount of carmine solution added is noted down, call this x. B is
now saturated with CO, its colour again changes, but again A may be
made to match it by the addition of more carmine to the extent of
y c.c. From the relation of x to y (which is not a simple linear one),
the percentage saturation of B with CO may be calculated.

Apart from the statement made above, that the real estimation is
that of the ratio of CO haemoglobin to total haemoglobin, whilst it
professes to be an estimation of CO haemoglobin to oxyhaemoglobin,
i.e. it assumes that there is no reduced haemoglobin in the blood—
an assumption which is probably justifiable when there is no oxygen
want and but little CO haemoglobin — the possible sources of error in
the above method fall into two general categories which have to do
with (1) the possibility in practice of matching the tints, (2) the
theoretical soundness of the method. Haldane and Douglas obtained
results which closely agree for any particular sample of blood, though
they have told me that the quantities of carmine they use in the
operation differ. The method is claimed to be accurate to 2 °/0. This
statement must be treated with great consideration. I have no
sympathy whatever with persons who attack a method of this kind in
a light-hearted way, and having failed to get consistent results after
a few titrations or even a few days spent in the endeavour, condemn
the method. The method is essentially a subjective one. The accuracy
which can be obtained by subjective methods with long and incessant

The acquisition of oxygen by the blood in the lung 205

a

H

o>

m

1

a

1

S
•_

0
(23

t

a «

<UQf

I

^
tn
0

S

1

..

g

-

.

am jTt|o3A|« jo

ajnjsioui

joj pajoajjoo K5

jon :ajai[dsoraji: <N rH op IN <N -H< gs

ui poo'iq' leuajji; I-H I-H rH rH I-H rH I-H
jo aanssMd f-n

•=

I IN o co ao I-H co to

3 2 ™ ' i . .»

$o>'? iJrHOlplNlNaOt—

o 2 '> CO CO CO CO IN CO CO

o~w Vc"NiN(NC<llNcvieN

O X«

L ( <:\ J^1 CO rH 1-H CO , ^-rtjH

i 00 co CO <N i — i

5

V r4

OS

O5

/

j
3,

- ^j

;"> «^ a.

*

<•*

0

f ,? O

1M

«S

. O

O O O

(N C- 00 CO
OS

IN Oq

o

o

C5

0

O5
<M

o

OS

a

03

OJ

(N

~f o

U7> cb

00
00

IN

CO
O

CO

6

OS
rH

O
CO

to

rH
O

O

rH
CO

CO
rH

O

«O

OlO

CO CO rH CO C>J CO
IN <N Ol IN C^ (N

O)

1

<N

O
OCO

CO 00 00 Tt<

«5 CO 00 OS o CO
CO •** OS OS -^ «5

00 rH QO <N

rH rH rH rH

O O O O

OS "H< CO ^

rH CO O rH

rHO

OO

C<l CO •"** rH
O O O O

O O O O

OOOO

oOOO

O)(N(NINrHOl
r>q CO i— 1 i— 1 rH i— 1

00

OS

o

IN

CO

O5
CO

IN

00

CO CO «5 <N

OO

Ip

00

xi

O

rHOSCO

•rfl

-*JI

O]COrH(Nr-l

<N

O

to

<N

CO
C

o

00
0

co

?•
rH
IN

2
IN

IN

O

OS

rH

i-H
IN

OOOCO

IN i— 1 <N >-H

O i-l t- OS

COrH-rXCO
<N<NiN<N

OSINCOO

CONOCO

lOrHOsO

rH

i-HCOOO
rHOO

OOO-*
(NrHC^

<N
O

OJD COCpO

O rHOrH

Cp
O

COOOOit~

rH 00 O

rH CO OS

OOO

1-HCOCOlN

IHCOOSO

OOOO

O017>(NO

OOOSCO>Cl

T-Ht-'NIN

(NrHCMrH

O

-

3

•43

•T3

d

in

_3
o "08

?~ >

O 3
-u o

Bro

'c3

13 «
0) ^3

O S

? I

0)

2

ffl

OOOOOrHrH OrH

rH_ rH^ r- (^ r- (^ rH^ rH rH i— I rH

•g(N(NlN(N'N»O«5 cJ" «5~

i-HrHrH(N<NlNlN IN<N

td ffi

CN"

o

R tfl R
6ccOO
o r,' do

rH rH rH i— 1
i-HrHrHrH

cT^T^S"

C^'cD't^'oO"
ININlN

TJ

Q

•206 Chapter XIII

practice is very great. A farmer by mere inspection will tell you the
weight of a bullock with an accuracy which probably exceeds that
claimed by Haldane and Douglas for the carmine method of titra-
tion.

I have never myself made a carmine titration, and had I made a
few I have little doubt that I should have failed to obtain concordant
results : but such a failure would have formed no just criticism of the
work of those accomplished in the method and would therefore have
been a mere waste of my own time.

Anyone who reads the recent work of Haldane and Douglas must
be struck by the way in which the method has served them in the
determination of the physical relation of oxygen and carbon monoxide
to shed blood. Take for instance the curves given in Fig. 35 — and
these are a sample of many determined since — they are mathemati-
cally fair curves with equations, and not only so, but there is inde-
pendent reason to believe that they are not only fair curves, but that
the properties of haemoglobin are such that they could not be of any
other shapes than the ones shown. Such curves do not result from
unsound experiments.

In spite of the last consideration enough has been said to form
a sufficient apology — if apology were necessary — for Hartridge to
take the ground that an independent* method of determining the
CO haemoglobin present in blood was a desirable preliminary to an
independent examination of Haldane's results.

Apology is not necessary however. It is of the essence of sound
experiment that the experimental methods should be varied in every
possible way. This, at all events, is the school in which I have been
brought up. It is the school in which my foretime lecturer, now my
friend and colleague Mr C. T. Heycock, F.R.S., untiringly pointed the
moral from the work of Stas on the atomic weight of silver.

I should never concede on the one hand that a blood gas worker
in the Cambridge laboratory owed an apology to his predecessors for

* Hartridge commenced in 1909 with the idea of revising the method as used by
Haldane and Lorrain Smith. Between the summer of that year and the spring of the
following, my results on salts, temperature, &c. were published. They materially
altered the problem for him and also caused the work to be taken up again by Haldane
and Douglas who published their work to some extent as they went. Apart from this
I believe neither they nor Hartridge were cognisant of the work of the other until
a short time before their final publication. The findings of the two sets of observers,
which agree very closely in so far as the physical properties of haemoglobin are con-
cerned, have therefore the merit of being independent.

The acquisition of oxygen by the blood in the lung 207

devising a fresh method of experiment, or on the other hand that his
doing so was of the nature of a challenge.

Hartridge's method is based on the fact that the substitution of
CO for O2 in haemoglobin is associated with a movement of the bands
towards the violet end of the spectrum.

Hartridge claims that in mixtures of the two the degree of
movement of the band is a measure of the relative amount of CO
haemoglobin present. Hartridge's spectroscope has two slits ; when
you look into it you see two spectra, one immediately above the
other — but these spectra present this difference, that the blue end of
the uppermost one is to the right of the field, the blue end of the

A

\

, \
1 \

I \

, 3.

'/,

,< I.

FIG. 100. — The method of formation and adjustment of the reversed spectra. A shows
the ordinary arrangement with single slit. B shows the arrangement with double slit.
C, the bands of O2Hb as they would appear in the spectroscope. D, the a bands have
been adjusted into line. E, the oxygen has been replaced by CO, and the a bands
have shifted away from one another. F, the lower spectrum has been displaced by
the micrometer, so that the a bands are again in line.

lower one to the left. One of the spectra is " reversed." Suppose C
in the figure to be the two spectra. Let us fix our attention on the
sharper and narrower of the two haemoglobin bands — the one nearer
the red end. By a movement of a lever attached to the instrument
the upper spectrum may be shifted in the field of vision till the
two a bands are exactly one above the other (D). If now the
oxyhaemoglobin be replaced by a CO haemoglobin the a bands
will no longer appear to be one above the other, but as in E. The
final operation is to make them coincide once more as in F. This is
accomplished as I have said by moving a lever on the side of the
apparatus — the arc through which it is moved is read off on the

208

Chapter XIII

scale, and is the measure of the degree of shifting of the band, which
in its turn is a measure of the percentage of CO haemoglobin present.

I do not propose to enter here into a description of the manipula-
tive details, such as screening fluids for blocking out parts of the
spectrum, not germane to the inquiry. For these I must refer the
reader to the original papers of the author, nor do I propose to
discuss the abstract foundation of the method from the point of
theoretical optics.

I will confine myself to some criticisms of its practical sufficiency
as a piece of laboratory technique.

©

FIG. 101. — The apparatus as employed for the calibration of Hartridge's spectroscope

in plan and in elevation.

Like the carmine method it is a subjective method, it depends
upon the matching of two things as seen by the eye. Other things
being equal it is an easier proposition to put two dark lines into
exact juxtaposition than to match two tints to within 2°/0. Much
depends upon the total range of difference in each case and the errors
involved in the two operations are not easy to compare.

Hartridge has been at great pains to define the subjective element.
He has spared no labour to satisfy himself that the method gives

The acquisition of oxygen by the blood in the lung 209

readings which are consistent with those given by the blood gas
pump. This comparison proved very illuminating. It was of course
necessary to calibrate the instrument in some way ; there was no
antecedent probability that the movement of the band would be a
direct linear function of the percentage of CO-haemoglobin present.

100

50

o so 100

FIG. 102. — The calibration curve. Ordinate = percentage of haemoglobin which is
CO haemoglobin. Abscissa =r micrometer readings of instrument.

Hartridge began by calibrating the instrument with mixtures of CO-
and oxyhaemoglobin in known amounts or amounts at least presumed
to be known — i.e. it was presumed that the amounts bore the same
relation in the mixture that they did in the aliquot parts. So long as
Hartridge continued to use this method his comparisons with the

B. R. P. 14

210

Chapter XIII

pump always broke down, especially so at low percentages of CO-
haemoglobin. He therefore discarded it in favour of another method
of calibration. Oxyhaemoglobin and carboxyhaemoglobin of equal
concentrations were placed in wedge-shaped troughs in apposition
(Fig. 101): by turning a screw these could be shifted perpendicularly

100

50

eo.

0 5O 100

FIG. 103. — The results given by spectroscope and gas pump compared.
Actual °/0 CO vertical. Observed readings °/0 CO horizontal. Pump o. Spectroscope «.

to the beam of light so that the light was always shining through oxy-
and carboxyhaemoglobin in known proportions to one another before
reaching the grating of the spectroscope. After the adoption of this
method complete harmony with the pump was obtained by Hartridge
;and the subjective element to this extent removed. In the case of

The acquisition of oxygen by the blood in the lung 211

the pump the oxygen and CO were pumped from blood boiled in
vacuo in the presence of an alkaline solution of potassium ferri-
cyanide. The CO in the gas obtained was estimated by combustion
with a glowing platinum wire.

The accompanying figures (102 and 103) will give the reader an
idea of the accuracy obtained by Hartridge. The first shows the
relation between the linear movement of the band and the percentage
of CO-haemoglobin, the latter being the ordinate and the former the
abscissa. The important point at the moment is the coincidence of
the individual observations with the line representing their mean.
In Fig. 103 is shown the correspondence (1) with the pump, and
(2) with the spectrometer (calibrated with the wedges) of samples of
blood containing known proportions of O.,- and CO-haemoglobin.
There are no published comparisons between the carmine method
and the pump and so far as I know none have ever been made. In
this way Hartridge checked his subjective method by an objective
one.

By another method Hartridge has endeavoured, and with success,
to define the subjective element in his method. Two persons in-
specting the same solution of O2- and CO-haemoglobin will obtain
different readings with the instrument, i.e. they will place the index
at different places on the scale because their eyes see differently.
This of course is the demonstration of the subjective element.

The carmine method and the spectroscopic one are alike in this
respect : there is no presumption that any two people see the same
tint when they look at a mixture of blood and carmine. In the case
of Hartridge's (5) method the subjective element is demonstrable in
units, and tests can therefore be instituted for the purpose of as-
certaining the degree to which the subjective element vitiates the
accuracy of the method. Winfield, whose eye differed considerably
from that of Hartridge, carried out control observations. He calibrated
the instrument which Hartridge had in use. The calibration was
made as in Hartridge's case with the wedge-troughs of oxy- and
CO-haemoglobin and gave the curve which is shown in Fig. 104.
This curve was roughly speaking parallel to Hartridge's curve. They
then independently made observations upon various samples of blood
containing O2- and CO-haemoglobin in unknown proportions, and each
interpreting his readings by his own curve arrived at identical
results.

Lastly the method is susceptible of treatment in the way in which

14—2

212

Chapter XIII

subjective methods should be treated — statistically. Each observation
takes but a few seconds and does not affect the composition of the
fluid observed. With a given sample of haemoglobin it is but the
work of a few minutes — perhaps as long as would be required for
a single carmine titration — to take 50 observations from which an
average in the strict meaning of the term could be determined.

In comparing the two methods there is one other point on which
I must touch. A few pages back I expressed admiration of the

-20

+20

FIG. 104. — Curves showing Winfield's calibration of Hartridge's instrument. Ordinate =
percentage saturation of haemoglobin with CO. Abscissa = micrometer readings.
• Observations from which curve was drawn. J/ = mean of 4 or 5 observations.
o Observations made at a later date to test whether the eye changed from time to
time.

carmine method, in that it appeared to yield reliable data concerning
the physical laws governing the reactions of haemoglobin ; my admira-
tion of the spectroscopic method — and my early bent filled me with
nothing but suspicion of spectroscopic methods in general — is not
more qualified. It is difficult to suppose the method, were it an
essentially bad one, could yield the figures which Hartridge put

100

The acquisition of oxygen by the blood in the lung 213

before the reader about the effect of temperature on the CO-02-
haemoglobin equilibrium.

Hartridge (6), like Haldane and Douglas, reinvestigated the whole
field of accessory conditions which influence the reaction between
haemoglobin and carbon monoxide — the eifects* of C02, acids, tem-
perature, &c., on the equilibrium. One observation has reference to
the effect of temperature on micef. I give it in his own words:
" There are however, oddly enough, experiments on mice in the same
paper quoted, which do agree, the experiments being as follows: the
mice were placed in the cold so that their rectal temperatures fell in
certain cases below 19°, and it was found that then the CO saturation
reached was higher than that obtained at normal temperature ; that
is to say the same effect as similar experiments performed in vitro.
The mere effect of temperature may therefore explain them." Per-
haps the most striking data which he obtained were those of the
influence of light on the reaction.

This had not been overlooked by Haldane, nevertheless the
following figures show the need of special precautions for excluding
the effect of light in a very striking way.

Percentage saturation of a haemoglobin solution with CO
under various circumstances of illumination.

Dark 96 °/0 Sunlight 40 °/0 Dark 91 °/0

Sunlight 42 °/0 Electric light 88 °/0 Daylight 80 °/0

Sunlight 35 °/0 Dark 92 °/0 6" Magnesium 64 °/0

Of these the most interesting is, I think, the daylight. When one
considers that the method of carmine titration deals with solutions of
only one-tenth of the concentration of that used by Hartridge and
for that reason much more accessible throughout its whole mass to
the active (ultra violet) rays, one will see that there is at least the
possibility of grave errors entering into the carmine method unless
special precautions are taken which are defined by quantitative
estimations, and if it is necessary to define such a source of error in
units in the notoriously gloomy atmosphere of these islands, it is
much more necessary to do so in the intensely actinic atmospheric
conditions which obtain at higher altitudes and lower latitudes than
our own.

* The same results were arrived at independently by Haldane, Haldane and
Douglas, Journal of Physiol. XLIV, p. 303, 1912.

t These experiments are quoted on p. 187 and tabulated on p. 188 of this volume.

214 Chapter XIII

The actual results which Hartridge (7) obtained may now be given.
In the first place he agrees with the now universally accepted opinion
that under normal circumstances the oxygen pressure in the alveolar
air is very close to that in the blood.

The following figures show the percentage saturations with CO
(A) of the blood of the body and (B) of blood of the same person
previously withdrawn, but exposed at body temperature to the alveolar
air, a sample of which was taken from the individual.

Percentage saturation with GO in blood exposed to alveolar air.

Series 1. Rest.

(1) (2) (3) (4) (.5) (6) (7) (8) (9) (10) Mean

A In vivo 12 18 21 29 20 36 35 25 43 44

B In vitro 10 18'5 225 26'5 15 33-5 33%5 24 42-5 40

A/B 1-2 -97 -93 1-09 1-3 1'07 1-05 1-Oi 1-01 1-1 1-08

These figures show that, if anything, there is a trifling excess of
CO-haemoglobin in vivo, i.e. a slightly lower oxygen pressure. In two
of these experiments Hartridge's blood became 40 °/0 saturated with
CO ; up to this point therefore there is no sign of secretion.

More interesting than these results, however, are those which he
obtained whilst establishing a greater or less degree of oxygen want.

The first method he adopted for this purpose was to breathe
mixtures of oxygen, CO and nitrogen which were much poorer in
oxygen than is atmospheric air. The following were the results:

Percentage saturation with CO in Mood exposed to alveolar air.
Series 2. Diminished oxygen supply.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

Mean

A

In

vivo 31

30

32

30

29

27

27

29

30

B

In

vitro 30

22

25

28-5

28

18-4

23-5

28-5

26-5

A/B 1-03

1-4

1-3

1-06

1-04

1-45

1-14

1-02

1-14

1-17

Oxygen tension

in

alveolar

air mm.

... 55

42

36

49

40

If a comparison be instituted between these oxygen tensions and
those which were actually observed by us at the top of Monte Rosa,
at an altitude of 15,000 ft., it will be seen that for the most part they
correspond to a considerably higher altitude.

In a third series of experiments Hartridge, whilst breathing from
the respiration apparatus, performed a considerable amount of actual
physical exercise though not so much as causes dyspnoea to pass into
distress ; the following were the figures which Hartridge obtained :

The acquisition of oxygen by the blood in the lung 215

Percentage saturation with CO in blood exposed to alveolar air.

Series 3. Worl:

(1) (2) (3) (4) (5) (6) Mean

A In vivo 30 24 24 26 27 25

B In vitro 28 23 20 24 23 25

A/B 1-Ofi 1-04 1-2 1-08 1-24 1-00 1-10

If then we regard the work of Hartridge as that in which the
possible errors have been most clearly defined quantitatively, and
accept it as being the most mature expression of opinion at the time
of our writing, our verdict must be that there is no evidence of
oxygen secretion in man as represented by a single individual—
Hartridge — either (a) at rest, or (b) under such restricted conditions
of oxygen want as may be induced by diminished oxygen pressure for
short periods of time, or (c) during moderate exercise of 4 minutes
duration.

For what it is worth my own view on the question of oxygen
secretion at the present time is this : It is agreed on all hands
that there is no oxygen secretion in the resting individual. The
issue then may be stated thus. Is the lung capable of oxygen
secretion? It is alleged to be so during oxygen want, i.e. during
exercise, or at high altitudes or with insufficient functional haemo-
globin (as during partial CO-poisoning). The real point to my mind
is that of exercise, for it is from its advantage during exercise that
such a function in the lung would naturally have been evolved and
maintained. Exercise then is the prima facie test case — here the ex-
periments of Hartridge and of the Oxford School disagree. Hartridge
finds a pressure of oxygen in the blood which is below that in the
alveolar air, Haldane and Douglas find the reverse. Yet their methods
are similar, the differences consisting in (1) the method of estimation
of the CO-haemoglobin in the blood and (2) the nature and duration
of the exercise. The issue refines itself largely into the question of
which of these methods is the more correct. Reading the accounts
of the two methods, I never fail to be greatly influenced by the follow-
ing facts: (1) Hartridge has tested his so exhaustively against a
purely objective one — that of the pump. (2) He found, when
testing his apparatus against the pump, the method of actual
mixtures unreliable and had to replace it by that of optical
mixtures. (3) The carmine method has never been tested in any
other way than by that of actual mixtures. (4) Of the rival
workers Hartridge has tabulated his errors the more exhaustively.

Chapter Xlll

Therefore, of the two methods I have the greater reliance on the
spectroscopic one. So I am forced at present to think that during
the exercise taken by Hartridge, with or without partial CO-poisoning,
there was no secretion. But perhaps I may take the reader into my
confidence so far as to say that I wish it were otherwise.

But while it seems to me fairly certain that there was no secretion
in Hartridge's case, there is one possible line along which Hartridge's
results might perhaps be brought into harmony with those of Haldane
and Douglas. It has been shown that the call for oxygen follows
upon the activity. The exercise taken by Hartridge was of a violent
but spasmodic type, that of Haldaue and Douglas was less violent and
more sustained. It is not an impossible supposition that the secre-
tion might show some lag after the exercise, and therefore be most
evident in the case of sustained work.

Lastly, apart from questions of the detail of individual experi-
ments there are certain general questions that cannot be overlooked
by the holder of the diffusion theory. Before this theory can be
regarded as proved on the positive side it must be tested under the
circumstances which are likely to strain it most. Two questions it
must answer. The first is, can it account for the passage of 3500 c.c. of
oxygen per minute through the pulmonary epithelium ? Perhaps it will
be able to do so one day, at present it cannot. At the end of the last
chapter I spoke of the diffusion coefficient, here I will only say that
the more recent determinations of the diffusion coefficient must be
materially changed before they will admit of 3500 c.c. of oxygen
traversing the lung along a pressure gradient dropping from 100 mm.
of mercury in the lung to 50 mm. in the blood. But if the test
of exercise is exacting that of exercise at high altitudes is more
exacting. I have granted a 50 mm. gradient and incidentally assumed
that the blood in the pulmonary artery is 80 °/0 saturated with
oxygen. But what if the whole alveolar pressure is but 30 mm.
(approximately that which must have obtained in the case of the
Duke of Abruzzi18' and his party at 24,000 feet altitude). The
work of recent observers on the dissociation curve shows that if the
blood were in equilibrium with such a gas it would only be 50 °/0
saturated* with oxygen. What then are we to allow for a gradient?

* The fact that the dissociation curve of man at rest scarcely changes at high
altitudes is treated in Chapter XVII. During activity it becomes meionectic, and
that this renders the position even more difficult on the diffusion theory is shown in
Chapter XVIII. When Roberts and I arrived at the Capanna Margherita our bloods
would have been about 30 °/0 saturated at 25 mm. O2 pressure.

The acquisition of oxygen by the Hood in the lung 217

If 5 mm. be allowed, the arterial blood would be saturated to the
extent of about 40 °/0. Now if 50 mm. pressure head in the lung will
not suffice for the passage of 3500 c.c. of oxygen per minute, 5 mm.
would not suffice for the passage of 350 and it is clear that each of
the party must have been using three or four times this amount. If
the arterial blood during work at high altitudes contained less oxygen
than the mean value for normal venous blood at the sea-level, what
would the venous blood contain and what possibility would there be
of obtaining a pressure gradient sufficient to transfer the oxygen from
the blood to the tissues? At present the diffusion theory fails to
answer these questions though it may do so one day.

On the secretory theory these considerations are explained, for the
members of the Pikes Peak expedition (8) found evidence of oxygen
secretion living at an altitude of 14,000 feet. They consider that the
secretion increased as their residence continued. Further experi-
ments are necessary in two directions. Hartridge should extend his
series to sustained work, both at low and at high altitudes, and should
he find no greater pressure gradient than he has already done, fresh
determinations of the diffusion coefficient are required in order to
prove that the actual quantities of oxygen used can pass by diffusion
through the pulmonary epithelium when driven by a difference of
pressure of only a few millimetres between the oxygen in the alveolar
air and that in the arterial blood.

REFERENCES

(1) Krogh, Skand. Arch.f. Physiol. xxm, pp. 179, 193, 200, 217, 224, 236, 248, 1910.

(2) Douglas, Journal of Pht/siol. xxxix, p. 453, 1910.

(3) Douglas and Haldaue, Skand. Arch.f. Physiol. xxv, p. 169, 1910.

(4) Hartridge, Journal of Physiol. XLIV, p. 1, 1912.

(5) Hartridge, Royal Soc. Proc. Ser. B, Vol. LXXXVI, p. 128.

(6) Hartridge, Journal of Physiol. XLIV, p. 22, 1912.

(7) Hartridge, Ibid, XLV, p. 170, 1912.

(8) See Douglas, Haldane, Henderson and Schneider, Phil. Trans. Ser. B, ccin, p. 310.

PAET III

THE DISSOCIATION CURVE CONSIDERED AS AN
"INDICATOR" OF THE "REACTION" OF THE BLOOD

CHAPTER XIV

THE DISSOCIATION CURVE IN MAN

WE are now in a position to discuss the dissociation curve in man
from what I may call the practical rather than the theoretical side.
Firstly, let us consider the constancy of its position in the same person
at different times. Secondly, the degree of coincidence in the curves
obtained from different persons. Here let me say at the outset that
I am only dealing with the individual in his normal condition and
not under special conditions such as violent exercise.

The first question which confronts the worker is whether he is to
study the dissociation curve in the presence or absence of carbonic
acid. Three possible courses are open :

(1) To study the dissociation curve in the absence of C02.

(2) To study it at a standard CO2 pressure.

(3) To study it at the C02 pressure of the alveolar air of

the person.

Each of these courses has something to be said for it, but it is
clear that the first two give slightly different information from the
third. The third tells of the blood as it actually circulates in the
arteries and is the most interesting and most valuable measurement
of the three : it tells most about the condition of the individual. But
the first and second methods have a value of their own if you wish
to inquire whether the blood of all persons is the same ; it is best to
study this question under the simplest conditions possible, hence I
shall start without the complication of the alveolar air. Nor need the
whole curve be studied, let me rather take some convenient pressure of

The dissociation curve in man

219

oxygen, say 17 mm. of mercury, and ask whether the blood of every
individual in the absence of carbonic acid is equally saturated with
oxygen when exposed at 37° C. to 17 mm. oxygen pressure. Fig. 10.5
shows that my own blood at this pressure is 75*5 °/0 saturated with
oxygen. This figure was arrived at as the result of some determina-
tions made by Dr Boothby of Boston, U.S.A., and myself. These data
will give an idea of the standard of accuracy of the methods employed.

too

17 MM.

FIG. 105. — Dissociation curve of Barcrof t's blood with the addition of 0, '02, '065 and
•2 °/0 lactic acid. The last is only an approximation. Temperature 37° C.

Percentage saturation of blood with oxygen at 17 mm.
O2 pressure and 37° C.

Barcrof t's blood ... 78 77 73
Boothby's blood ... 79 77 76
Mean values .

74 75 76 difference between extremes 4 °/0

75 77 „ ,, „ 5°/0
Barcrof t's blood 75 -5 %

Boothby's blood 76-8 °/0

Corresponding figures of a number of workers are :

Boothby 76 -8%
Barcroft 75 -5
Winfield 78
Burn 76
Higgins 74-3

Hill 74%
Miss Chick 73
Miss Pillmann 76
Miss Martin 79
Dr Klein 68

Were it not for the single case of Dr Klein it might almost be
supposed that the whole of the cases given above were sufficiently

220 Chapter XIV

close to fall within the six per cent, limit, and the deduction might be
made that there was no difference between the blood of healthy
persons.

Yet another series of figures obtained by Mathison, at Pisa, on
the members of our Monte Rosa party, seems to indicate that the
blood of different persons differs.

Ryffel 77 °/0
Eoberts 75*5
Camis 74
Barcroft 72
Mathison 63, 66

These figures run slightly lower than the previous series,
probably because the same care was not taken to exclude C02
completely. In the former series the gas in the tonometers stood
over alkali before it was used, in the latter it was used as it came
from the cylinder. The gas contained about 1 mm. pressure of CO2.
The point is that the figures in each series are comparable with one
another. Clearly the differences are greater than can be accounted
for by experimental error.

Having established the fundamental fact that there is a difference
between the blood of normal persons, apart entirely from any question
of the quantity of carbonic acid present, let us turn to the dis-
sociation curve of the blood as it leaves the lung.

In order to gain reliable information concerning the blood of a
given individual it is necessary that his dissociation curve should be
studied over a considerable period of time, running if possible into
years. There are now several persons for whom this has been done,
Douglas, Roberts, Ryffel, Higgins and myself. The question arises with
regard to each of these — Can any difference be made out as between
one time and another? This point seems to me to be of such great
importance that I propose to set the evidence on the subject in some
detail before the reader.

Firstly then as regards Douglas whose normal alveolar carbonic
acid pressure is found to be 40 — 41 mm. of mercury. His dissocia-
tion curve is given in Fig. 106. Particular interest attaches to it not
only because it has been determined more often probably than that
of any other person, but because it has been determined inde-
pendently by two sets of observers, working by slightly different
methods. The points determined by me were each done on 0*1 c.c. of
blood in the small differential apparatus, those determined by Haldane

The dissociation curve in man

221

and Douglas were each determined on 1 c.c. of blood in the old
Barcroft-Haldane apparatus. Moreover in some of their determina-
tions the Oxford observers used sodium carbonate and in others

100

20

40

120

100

100

100

100

100

100

80

100

120

FIG. 106. — Dissociation curve of Douglas' blood as determined on various occasions,
extending over about two years. The curve is the same in each case.

ammonia for the purpose of laking their blood. An examination of
the figure will show that no difference can be observed in the curve
between one time and another. In the figure the curve is the same

222

Chapter XIV

in each case. It is repeated six times corresponding to six occasions
on which the determinations were made. In all it represents about
forty different determinations no one of which is seriously off the line.

20 40 60 80 100.

100

100

100

100

100

20 40 60 80 100

FIG. 107. — Dissociation curve of Barcroft's blood determined on various occasions.

The curve is the same in each case.

My own curve has been investigated from time to time — over a
period extending from the summer of 1909 to that of 1912. The

The dissociation curve in man

223

following points have been recorded at different times and in different
places. In this case also the points have not all been determined by

100

20

40

100

100

100

100

80

100

FIG. 108. — Dissociation curve of Higgins' blood, top curve in Jan. 1911 and Sept. 1911,
second curve Nov. 1912. I and II show the two upper curves side by side.

the same observer, most were determined by me, but some by
Higgins. The places and dates of their determinations were as follows :

224

Chapter XIV

Cambridge, Feb. 1910, C02 40 mm.
Teneriffe, March 1910, C02 40 mm.
Cambridge, May 1910, COS 40 mm.
Pisa, August 1911, CO2 40 mm.

Data for my curve similar to those given for Douglas' in Fig. 106
are shown in Fig. 107. Similar curves might be given for other
observers, but I will only burden the reader with one more, namely
that of Higgins. His has a special interest inasmuch as it is the one
recorded case of the curve in a normal person changing to an extent
that is appreciable. The determinations were made (l)in Cambridge
in January 1911 by myself, (2) in Boston (Mass.), in September of
the same year by Higgins himself, and (3) in November 1912, also by
Higgins, at Boston. The first two sets of observations fall upon the
same curve, the last falls upon a slightly different one. These two
curves are shown with the corresponding observations as 1 and 2 in
Fig. 108, whilst the two are shown together in the lower part of the
figure in order that the reader may judge of the disparity between
the two. It will be seen that the difference between the two curves
is very small. This is the greatest difference discovered in any one
person under normal conditions.

Let me turn to the differences which may be observed between
the curves of different individuals. The reader will remember that

ii Kxn

any one of these curves is represented by the equation -^ = yri? ~n »

where y is the percentage saturation with oxygen, x the oxygen
pressure, K the equilibrium constant, and n the average number of
molecules of haemoglobin aggregated together. The simplest way of
presenting these curves will be to give the values of K in each case,
n being in all cases 2*5. The following is the result, the curves being
taken in order from the greatest to the smallest value of K.

Name

Value of A"

Alveolar CO2 pressure at which
K was determined

Kyffel

0-000363

44—45

Roberts

0-00033

39

Graham

0-000317

41

Camis

0-000315

40

Higgins (1) •

0-000301

38

Barcrof't

0-000292

40

Higgins (2)
Douglas

0-000264

) (

38
41

Haldane

0-000212

40

Mathison

39

The dissociation curve in man 225

These curves all lie within the black area marked in Fig. 109, the
upper boundary of this area being Ryffel's curve and the lower
boundary that of Douglas, Haldane and Mathison. This area, there-
fore, represents the limits within which the curves of human beings
may be regarded as normal at the present time.

Having determined the limits between which any single individual
may be regarded as possessing a normal dissociation curve and the
degree of variability which may be regarded as normal in any single
person, we may now proceed to discuss some special cases of variation
when the body is subjected to abnormal conditions. Before doing so,
however, it is necessary to explain the meaning of certain words used
to denote the changes which we are about to describe.

If the curve of a particular person moves from its normal position
it may move in either one of two directions, i.e. it may be either
above or below the normal curve of that person. In other words,
at a given oxygen pressure the blood may take up either more or
less oxygen than it is wont to do. To express these facts Mr Harrison,
Fellow of Trinity College, Cambridge, at the instance of Dr Fletcher,
suggested the following nomenclature ; when the dissociation curve
is above its normal situation, i.e. when at a given pressure of oxygen
the blood takes up an abnormally great percentage of its total
possible load of oxygen, the curve is called "pleonectic*." When on
the other hand it takes up less than the usual percentage of oxygen
it is "meionectic" ; when it becomes saturated to the normal extent
under any specified conditions it is "mesectic." If the curve becomes
meionectic the value of K will diminish, if pleonectic it will increase,
as compared with the normal or mesectic value. There is one point
which must be made perfectly clear to the reader ; normal blood in
any individual must yield a mesectic curve — but blood which yields
a mesectic curve need not necessarily be normal. Let me give a case
in point. In Chapter V it was shown that if carbonic acid were
withdrawn from blood, the value of K would increase, i.e. the curve
would rise ; if this happened in a person and nothing else, he would
acquire a pleonectic curve, but if at the same time that the CO.2 was
withdrawn lactic acid was added to the blood the value of K would
diminish, and the curve would tend to return to its former position,
and if enough lactic acid were added the curve would become as it
was to start with. Again, if this process took place in a human being

os-, disposed to take more than one's share. From 7rAeoi>e£t'a, a dis-
position to take more than one's share. (Liddell and Scott.)

B. R. F. 15

226

Chapter XIV

his curve would at the end be mesectic, because the value for K
would be what it normally was, but the blood would be abnormal
because it would contain more lactic acid and less carbonic acid than
usual. The position of the curve, therefore, whether pleonectic,

100

10 20 30 40 50 60 70 80 90 100

FIG. 109. — Limits within which the dissociation curves of normal persons fall.

mesectic, or meionectic, depends not upon the actual quantity of any
one electrolyte in the blood, but upon the balance which is main-
tained between them.

CHAPTER XV

THE EFFECT OF DIET ON THE DISSOCIATION CURVE OF BLOOD

IN company with my colleagues I have made efforts, in several
directions, to move the dissociation curve in individuals away from
its mesectic position. Naturally such attempts would take the line
of subjecting the patients to some form of treatment which is known
to alter the acid radicles in the blood either in quantity or in kind.
The first of these attempts which I will chronicle was made by
Higgins and myself. It is perhaps the simplest, and as it illustrates
a good many of the principles involved in the more complicated
researches which follow, I will describe it in some detail. My col-
league on this occasion was a specialist not only in gas analysis,
but in the physiology of dietaries, coming as he did on a visit to
Cambridge from the Nutrition Laboratory at Boston, U.S.A., presided
over by Professor Benedict. Moreover, since Higgins visited this
country he has continued the work at Boston. The motif of the
research was as follows : all carbohydrate was strictly eliminated
from the diet, and this of course led to an acidosis the most obvious
evidence of which was the appearance of /3-oxybutyric acid, diacetic
acid, &c. in the urine. Here let me explain my use of the word
acidosis in reference to the blood. In the following pages it will
signify the appearance of acids (exclusive of C02), abnormal in kind
or perhaps only in quantity in the blood or even a decrease in the
bases present. By acidosis then I mean an increase of acid relative
to basic radicles in the blood, C02 not being considered. But by the
term acidosis I will signify nothing concerning the final "reaction" of
the blood which is largely regulated by the amount of C02 present.
The question for decision was, Would the dissociation curve depart
from its mesectic position owing to these bodies being thrust into the
blood by the tissues, i.e. owing to the acidosis? And here we are at
once face to face with the distinction which we drew at the close of

15—2

228

Chapter XV

the last chapter, between mesexy and normality. It is easy to ascer-
tain that the blood is abnormal, for it is known that such diet causes a
lowering of the carbonic acid pressure in the alveolar air which must
correspond to a reduced pressure of CO2 in the blood, but it is also
known that organic acids appear in the urine which have come from
the blood. To what extent is the loss of carbonic acid in the blood
balanced by the gain in other acids so far as the total reaction of the
blood is concerned?

Table of diets, &c.

Date

Sept. 5
Sept. 6

Sept. 7

Sept. 8

Sept. 9

Time

12.30

8.15
12.30

7.30
10.00

3.30
7.30

8.00

10.30
3.30
8.00

Diet

Beef stew and vegetables

2 slices bread and butter

No supper

4 eggs

1 cup beef tea

1 tin sardines
6 ozs. cheese

2 pork chops

2 pieces pepsin gum

3 ozs. butter

4 eggs

3 ozs. butter

4 soft boiled eggs

1 piece fried ham and a
little butter

A few sips of coffee

1 piece pepsin gum
Lemon juice

2 lamb chops
1 box sardines
Beef steak

No food taken till 12.00 when
ordinary diet was resumed

Table of abnormal constituents in urine for

24 hours, collected about 8 a.m. on the

following morning

NH3 1-296 g.

NH3/N2 °/0 6

Acetone 1-348 g.

Oxybutyric acid 4-0845 g.

Vol. 1120 c.c.

N-, 18-084 g.

NH3 2-015 g.

NH3/N2 °/0 11-2

Acetone 1-725 g.

^-oxybutyric acid 1-577 g.*

Vol. 1300 c.c.

N, 18-933 g.

NH3 3-350 g.

NH3/N2 °/0 17-6

Acetone 2-791 g.

/3-oxybutyric acid 8-025 g.

* Later work shows that these are minimal values.

Here is an account of a complete experiment, with all the data of
urine, respiration, &c. set forth (pp. 228 and 230), and from it we see
that the alveolar carbonic acid pressures fell from 37 to 29'5 mm. of
mercury, the latter value being that observed on the fourth day of the
diet. On this day blood was taken for analysis and compared with that
found on the day before the special diet was commenced. The following
curve is Higgins' (Fig. 110 A) mesectic curve. The points marked by

The effect of diet on the dissociation curve of blood 229

circles are those determined on the fourth day of the diet with an
alveolar pressure of 29 '5 mm. So far as can be seen from inspection
of Fig. 110 A the curve remains mesectic in spite of the changed

100

90

80

10 20 30 40 50 60

100

10

50

an 70

FIG. 110. — Dissociation curves of Higgins' normal blood (normal alveolar CO._, pressure).
A, exp. Sept. 8, 1911, o = points determined at existing alveolar C02 pressure when on
diet; «=the same blood exposed to his usual C02 pressure. B, exp. Jan. 1911,
o = points determined during diet at existing C02 pressure; x = points determined
previous to, and + subsequent to, diet at normal C02 pressure.

condition of the blood caused by the decline in the alveolar pressure
of carbonic acid. The most obvious question which will occur to the

•230

Chapter XV

mind of the reader is — Were the experimental methods at our com-
mand equal to the task of discerning any shifting of the curve, should
it exist? The answer to this question is given in the figure. The dark
spots (Fig. 110 A) are those obtained by exposing Higgins' blood,
taken during the diet, to the alveolar pressure of carbonic acid which
existed before the diet 37 — 38 mm. of mercury. It is clear that all
these points are below the curve.

The absence of any carbohydrate metabolism is shown by the
figures which were obtained for the respiratory quotient.

C02
c.c. per min.

oa

c.c. per min.

B.Q.

Pulse

C02 in alveolar air

Normal

198

237

0-84

_

36-5, 37 mm.

Sept. 7

8.50 a.m.

206

305

0-68

85

—

9.14

197

292

0-67

83

—

9.40

198

301

0-66

84

—

Sept. 8

9.25 a.m.
9.51
10.15

209
196
189

281

279
280

0-75
0-70
• 0-68

76
75

74

I 29'5 mm. when blood was
taken for analysis

Sept. 9

9.53 a.m.

198

276

0-72

70

—

10.16

191

268

0-71

74

—

Sept. 11

210

243

0-86

69

38 mm.

Higgins' dissociation curve had been very carefully determined
before the experiment began. In each case blood was exposed to
the carbonic acid pressure of the alveolar air at the time the blood
was taken.

The comparison may be put before the reader in another way.

It has been pointed out in Chapter IV that all human dissociation
curves fit the general equation

Kxn

100 ~ 1 + Kxn '

in which n may be taken as 2'5. The only variable then is K. If
the values of K be calculated for the observed points (1) of the
normal blood exposed to normal alveolar C02 pressure of 39 mm.,

(2) blood drawn on the fourth day of the carbohydrate-free diet
exposed to the alveolar C02 pressure of the person at that time,

(3) the blood drawn on the fourth day of the carbohydrate-free diet

The effect of diet on the dissociation curve of blood 231

exposed to 39 mm. C02 pressure, the following values for K are
obtained :

(1) (2) (3)

•000319 -000373 -000227

•000361 -000216 -000280

•000307 -000307 '000201

•000299 -000352 -000236

•000311 -000261 -000221

•000242 -000350

•000257 -000305

•000258
•000400
•000283
•000273
•000281

Mean -000299 -OU0309 '000233

Of the twelve determinations given in column (1) five are above
the mean value, six below it, while one corresponds with it, all the
values in column (3) however are below the mean in column (1), and
all but one are below any individual figure in column (1). Since the
blood referred to in columns ( 1 ) and (3) was analysed at the same car-
bonic acid pressure there can be no doubt that the blood changed.
But a comparison of columns (1) and (2) shows that the values of
K as calculated in the two sets of figures are not really to be
distinguished from one another, five determinations in column (1)
being above the mean of column (2) and seven below it. Indeed the
mean value of K given in columns (1) and (2) would not yield curves
distinguishable from one another in diagrams of the scale and style
of those in this volume.

The conclusion then is that Higgins' blood in this experiment was
mesectic on the fourth day, but that whilst mesectic it was abnormal
inasmuch as it contained less carbonic acid and an equivalent excess
of other acids — presumably those which appeared in the urine.

A point of some subtlety arises here — let the following points be
accepted :

(1) That the alveolar carbonic acid falls.

(2) That the dissociation curve remains constant in position.

(3) That the respiratory centre is regulated chemically by the
same factors as affect the dissociation curve, of which the most
important is probably the hydrogen ion concentration of the blood.

The immediate deduction from considerations (2) and (3) would
be that the respiratory centre worked at the same rate on the days

232

Chapter XV

of the diet as on normal days since the curve remained mesectic.
How then was the carbonic acid got rid of? If the same quantity of
C02 was produced on the days of diet as on the days of rest, and if
also the C02 in the alveolar air was less than normal, it would appear
that the CO.2 produced was diluted with a greater quantity of atmo-
spheric air than usual, which would be another way of saying that
the ventilation was increased, a deduction which seems to be incon-
sistent with our original statement, for it would involve stimulation
of the respiratory centre and meionexy. The other possibility would
be that the actual C02 production went down.

In a later experiment which yielded similar data as regards the
dissociation curve, the rate of respiration and the total ventilation
were measured. There was no very obvious increase on the 8th and
9th of November, the days of diet on which the dissociation curve
data were obtained, as compared with the 4th of November.

Date

Nov.

C02 per min.
c.c.

02 per min.
c.c.

R.Q.

Average rate of
respiration

Ventilation per min.
litres

f

4

229

254

•90

15-0

6-31

Normal \

5 204

242

•84

14-1

5-61

1

6

191

244

•78

13-4

?

f

7

210

302

•70

16-2

6-93

Diet...

8
9

198
190

250
272

•79
•70

14-2
14-9

6-33
6-44

I

10

180

264

•68

14-9

6-28

At the end of the period of diet there is an evident tendency for
the total C02 output to fall. The point is not one of great import-
ance in the present connection, but with a greater degree of acidosis
it is clear that dyspnoea, such as takes place in diabetes, must
supervene — a definite dyspnoea prompted by the effects of oxybutyric
acid &c. on the respiratory centre. Here, however, I wish merely
to indicate to the reader the nature of the problems in which he
is likely to involve himself if he uses the dissociation curve as an
" indicator."

The last point which must be considered is the degree of acidosis
which took place. Here let me repeat that I mean by this phrase
the degree to which acids other than C02 were added to the blood,
assuming the bases to remain constant. This point was tested in
another experiment ; in it also the blood remained mesectic through-
out the dietetic period, i.e. it gave the same dissociation curve when

The effect of diet on the dissociation curve of blood 233

exposed to the C02 pressure of the alveolar air at the time the blood
was drawn. Before the diet the alveolar C02 was 37 mm., on the diet
31 mm. What degree of acidosis compensated for these 6 mm. of CO2?
This could be directly found by using the dissociation curve of the
blood without any CO2 present as an indicator. The curve was the
same as that of normal blood to which was added 0*03 % lactic acid,
but no doubt the actual acids present were yS-oxybutyric and others
found in the urine.

In getting a clear understanding of the distinction between blood-
acidosis and meionexy, as I use the phrases, the reader may find the
following figure useful, though it pretends to no quantitative accuracy

CO2mm
30

40

10 -

I II HI W IT
FIG. 111. — Scheme illustrating the relation between C02 pressure in alveolar air and
blood, acidosis and A'. The figures in this are purely hypothetical, they do not
correspond to the blood of any actual person.

and it takes many things for granted, such as the constancy of the
bases present. He may get a conception of the acids which rise and
fall in blood as the "floating" acids — they consist of C02 and organic
acids, lactic, /3-oxybutyric, diacetic, &c.

Consider samples of blood I, III, V and VII. The alveolar CO,
decreases progressively from I — VII. The added organic acids, repre-
sented by the black area, increase progressively from I — VII, but the

234 Chapter XV

value of K remains unchanged at '00025. This represents the reaction
of the blood. Such were the samples of Higgins' blood which have
been considered so far.

Now take two specific cases V and VI. Let me compare them
with (I), the normal blood.

The normal blood I, exposed to no C02 pressure, gives a value
for K of '0025 and exposed to the CO2 of the normal alveolar air of
the patient (40 mm.) gives a value of '00025 for K. Either of these
may be found by a single determination of the percentage saturation
y at an oxygen pressure of oc. Then, taking n, the number of
molecules y aggregated together, as 2 '5,

y Kxn

100 ==l

xn(lOO-y)'

With this compare blood V. The alveolar C02 is found to be
27 mm. A point on the dissociation curve is found in the presence
of this quantity of C02. It also gives the value of '00025 for K,
therefore the blood is mesectic. Now some of blood V is taken again
for the purpose of determining the acidosis. Free from CO2, it
gives a value of K = '00065. From this it is clear that the amount of
the acidosis is equivalent to about '07 °/0 lactic acid. Now pass to
blood VI. The alveolar C02 is also found to be 27 mm. Exposed to
27 mm. C02 the value of K works out to be '00024. The blood is
therefore off the line, it is meionectic. The value of K in the absence
of CO2 is '00075. This corresponds to an acidosis equivalent to '085 °/0
lactic acid. In this case therefore the acidosis and the CO2 together
amount to more than the original C02. Blood VI being meionectic
the respiratory centre would be stimulated ; blood V being mesectic
the respiratory centre would not be stimulated.

In all the cases I, III, V and VII, not only would the blood be
mesectic, since K remained normal, but the rhythm of the respi-
ratory centre and no doubt a hundred other things would remain
normal also.

Now consider another series, I, II, IV, VI. In this as in the first
the alveolar CO2 decreases progressively, but the value of K decreases
also, since the aggregate of floating acids increases. In this case the
blood would become increasingly more meionectic and the respiratory
centre would become increasingly active.

The effect of diet on the dissociation curve of blood 235

In all three experiments made upon Higgins the curve remained
mesectic ; one experiment of the same character was performed upon
Graham.

Two other experiments have been performed with a different
result. In each the blood became pleonectic. One of these was
performed upon Higgins, another upon myself, and as the contrast in
Higgins' case between this experiment and the one given above is
striking I will give the results in the same form, i.e. (1) I will give the
curves of the experiment, and (2) I will tabulate the values of K.
(They are plotted in Fig. 110 B.)

In the column (1) are the values of K before the experiment
began, in column (2) the values a week after it finished ; between these
there is no clear difference. In column (3) will be found the values
of K on the third day of the diet; in every case the curves were
taken at the existing alveolar CO2 pressures.

(1) (2) (3)
K before diet K a week after diet K on third day of diet

•000243 -000256 -000477

•000303 -000260 -000444

•000218 -000326 -000434

•OOU252 -000280 -000538

•000302 -000362
•000374

Mean -00026 -00030 -00047

The difference which was noticed between the two experiments
in which the subject became pleonectic on the carbohydrate-free diet
and some subsequent ones, in which he remained mesectic, was that
in the pleonectic experiments the subject became upset, whilst in
the mesectic experiments he remained in what appeared to be his
normal health. In the most recent experiment which was carried out,
Higgins, however, was as much upset as in those just cited and his
blood remained mesectic. It is not possible, therefore, to accept this
explanation. In the one experiment in which Higgins became pleo-
nectic, he also became anaemic ; his haemoglobin value dropped to
80 °/0- Whether the two things are connected or not I do not know.

CHAPTER XVI

THE EFFECT OF EXERCISE ON THE DISSOCIATION CURVE

OF BLOOD

So far the effect of exercise upon the dissociation curve of blood
has been but slightly studied. No doubt there is a great field for
work in this direction. What information there is, however, bears
directly upon the relation of the reaction of the blood to the activity
of the respiratory centre.

Investigations have been made upon the dissociation curve of
blood before and after a climb of 1000 feet from the sea level, or
nearly so. For this purpose Carlingford Mountain, in the North of
Ireland, was chosen. On that mountain a climb of 1000 feet was
marked out. The first experiments were made by myself. I chose
a rate of climbing which did not entail any sort of " distress."
Roughly speaking this meant going up at the fastest rate at which
respiration could comfortably be performed through the nose. As
the results of these tests it seemed that 30 minutes for the climb was
a suitable speed ; 25 minutes was to me a definite effort whilst
anything slower than 30 minutes gave the impression of loitering.
I am of course fully conscious that any such subjective index is of
the roughest possible character, that probably one changes from day
to day and that what appears to be an effort one day is not so on
another. It is difficult, however, to arrive at any standard for moun-
tain climbing ; the staircases of some of the tube stations in London
would form a very good standard course for ascents of varying
degrees of rapidity. We however desired something which could be
compared more definitely with an Alpine climb, to which I shall
refer presently. The equipment which I took on the actual ascent
consisted in a couple of Haldane's vacuous tubes for the collection of
alveolar air, a long rubber tube for the same, some clean needles with
which to prick my finger, a crucible in which to whip the blood and

Effect of exercise on the dissociation curve of blood 237

some feathers for the whipping, and a small stoppered test tube in
which to put the whipped blood. At the top of the ascent I obtained
a sample of alveolar air which proved to have a partial pressure of
35 mm. of CO2 as opposed to 40 mm. at the sea level, also a sample of
my blood for analysis. The analysis was two-fold — firstly the blood
was exposed to 17mm. oxygen pressure in the absence of CO2 for the
purpose of determining by Mathison's method* the degree of acidosis,
if any. The blood at 17 mm. was, as the result of two determinations,
54 % and 57 % saturated with oxygen, whilst before the climb it was

.-.„,.

' - ~

FIG. 112. — Carlingt'ord, showing the climb of 1000 feet.

75 °/0. The difference corresponds to an added amount of acid which
is equivalent to '023 °/0 lactic acid.

Against this however the amount of CO2 in the alveolar air
and presumably in arterial blood went down from 40 mm. to 35 mm.
The question at once arose, did this fall in CO2 compensate for
the increase in other acids? In other words, was the actual blood
in the body pleonectic, mesectic, or meionectic? The answer of
course could only be found out by experiment. This experiment
formed the second part of the analysis. The blood taken at the end
of the ascent was exposed to oxygen pressures in the presence of CO2

* See p. 257.

238

Chapter XVI

at 35 mm. pressure, its dissociation curve was thus determined and
compared with that of my normal blood at 40 mm. C02 pressure.

The data for my normal curve have already been given. The
data at the end of the ascent compared with those of my normal
blood were

C02 pressure

O2 pressure

Percentage
saturation

K*

After ascent ...
Normal .

35 mm.

40 mm.

27-5 mm.
27 '5 mm.

43) An 01

37 j /0
53

•000168
•000292

The blood proved to be meionectic.

This experiment was controlled by one in which the ascent was
made much more slowly — in three-quarters of an hour. This slower
speed is much more nearly the mountaineer's rate of climbing.
There was scarcely any departure from my normal condition when
I reached the summit of the 1000 feet, my respirations were 18 per
minute (they are 17 as I sit writing), though no doubt they were
deeper than usual ; the carbonic acid in my alveolar air proved to be
38mm. and while exposed to that CO.2 pressure and to 31 mm. oxygen
pressure my blood became 56 and 55 °/0 saturated in a couple of
determinations. As compared with my normal blood, the following
were the data :

C02 pressure in
alveolar air

Oo pressure of
analysis

Percentage
saturation

K*

After ascent ...
Normal

38
40

31
31

56
61

•00023
•000292

The slower ascent showed all the features of the faster one but to
a less marked degree, i.e. the drop in alveolar CO2 pressure and the
meionexy : the departure from the normal was just, but only just,
appreciable. The first of the two experiments was also controlled by
another of which Roberts was the subject. He made the ascent more
rapidly than I did, going up the 1000 feet in 20 minutes. The data
which he yielded were as follows :

* See Appendix III.

Effect of exercise on the dissociation curve of blood 239

CO2 pressure in
alveolar air

O2 pressure of
analysis

Percentage
saturation

K

After ascent ...
Normal

35—36 mm.
41—42 mm.

33 mm.
33 mm.

53

71

•00018
•00033

Moreover, as in the first of the two experiments upon myself which
I cited, Roberts' blood was tested by Mathison's method for the
degree of acidosis. At 17 mm.* oxygen pressure his blood when freed
from C(X was 55 °/0 saturated, which, according to his curve, corre-
sponds to an added amount of acid equal to '029 % °f lactic acid.
Now in this experiment the actual amount of lactic acid in Roberts'
blood was analysed by Ryff'el before and after the ascent. The blood
used was a portion of the same sample used for the gas analysis. The
method depends upon the conversion of the lactic acid into acet-
aldehyde, which is distilled off and estimated colorimetrically with
Schiff 's reagent. Before the climb commenced Roberts' blood con-
tained '014 % °f lactic acid, at the end of the ascent it contained
*046 % °f lactic, the difference '032 °/0 corresponded within the limits
of experimental error to the acid added to the blood as indicated by
Mathison's method.

Here then is a complete story. During the ascent, at the rate at
which Roberts made it, lactic and carbonic acids, and these only, are
added to the blood. The blood becomes meionectic, therefore the
respiratory centre is stimulated, the increased respirations cause the
excessive carbonic acid to be expired and not only the excess of
carbonic acid but somewhat more than this. The carbonic acid
pressure in the alveolar air therefore falls. The lactic acid, however,
is not got rid of so quickly as the carbonic acid, and is retained in
sufficient quantities to make the blood meionectic so long as the
exercise is taking place. We have no data which enable us to judge
of the subsequent duration of the meionexy

So much for the facts : let us turn to their physiological significance,
that is to say to the consideration of the extent to which the changes
are beneficial to the organism or otherwise.

First as regards the blood itself. The change in the curve suggests
a decrease in the value of K. (1) In the tissues each corpuscle will
lose its oxygen more quickly, other things being equal. (2) For a
given degree of reduction the pressure of oxygen in the capillaries

* See Figs. 105 and 114.

240

Chapter XVI

will be higher. These factors form the necessary chemical and
physical complement to vascular dilatation and more rapid flow of
blood through the capillaries. Our picture of the capillary circula-
tion through the active tissues is that each corpuscle spends less

100

80

7C

60

50

30

20

10

IP 20 3O 40 50 60 70 80 90 100

0 0-

•to-

70

100

20

60

7O 60 90 100

BARCROFT

Mesectic

Meionectic

CARLINGFORD
lOOO'in 45 mins

100

10 20 30 40 50 60 70 80 8O 100

FIG. 113. — Showing changes in the dissociation curve as the result of climbing. The
dotted curve is in each case the one which results from exercise.

time in the capillaries ; each corpuscle is reduced to about the same
extent as during rest ; this is made possible by the diminished value
of K. More oxygen leaves the blood, and this again can only diffuse
out by a rise in the intra-capillary oxygen pressure combined with
a fall in the extra-capillary oxygen pressure. It must be borne in

Effect of exercise on the dissociation curve of blood 241

mind that the extra-capillary oxygen pressure is, in the case of muscle,
never high, being perhaps 20 mm. or less and at most 27 mm.

In the capillaries we have then a most beautiful mechanism for
rapid unloading of the red corpuscles during their speedy transit
through the active tissues.

It might at first sight appear that what is gained in the tissue is
lost in the lung. The decreased value of K tends to make each
corpuscle take up oxygen less quickly in the lung, other things being
equal ; but so far as the lung is concerned other things are not equal.

80(
7n

) -01 -02 -03 -04 -05 -06 -07 -08%

\

L

actic &

cid ad

Jed

60
50
40
30
20
10
n

N

CM

\3

6**t

o

E

P

^
J

^

>/m

r^
j_i

<0

E
E
N

^x

_o

E

<^

e^^
^\^

f>-
'A

c

0

Jj
«J

o

JLJ
tJ

0)

O

-Q

i.
£

^3.

^^

3

A3

id
CO

gj

X

V

"o->

C
~

(d

HI

o->
id

_U

c

^

o
_o

-O
</>

U

L.
1)
-u
Sfc

o

c£

"t!

V

_D
O

or

(0
CM

O

FIG. 114. — Showing relation of percentage saturation of Roberts' blood at
17 mm. oxygen pressure to acid added.

For the meionectic condition of the blood stimulates the respiratory
centre and causes increased ventilation of the lung. We have seen
that there is a fall in the carbonic acid in the alveolar air. The
counterpart of this is that there is a rise in the pressure of oxygen in
the alveoli. The rate at which the oxygen is taken up, other things
again being equal, depends upon the pressure of oxygen.

Apart from the fact that nature can more than compensate for
the diminished value of K by increasing the respirations, it must be

B. R. P. 16

242 Chapter XVI

borne in mind that it is more important that the blood should become
more rapidly reduced in the active tissue than more rapidly oxidised
in the lung of the active person. This is most evident if a very small
muscle be considered. Let it be in full activity: the blood will go
through it with ten-fold rapidity, hence the necessity for rapid reduc-
tion of the haemoglobin : but the muscle being small the quantity of
blood involved would not be so large as greatly to quicken the circu-
lation through the lung — the blood would have its usual time for the
acquisition of oxygen.

It seems impossible to close this chapter without some reference
to the controversy as to the exact nature of the chemical stimulus to
the respiratory centre.

Haldane and Priestley (1), some ten years ago, made a great advance
in the physiology of respiration by demonstrating in the most con-
vincing manner that the respiratory centre was normally stimulated
by carbonic acid. Since the issue of their paper there have been two
views held by physiologists as to the status of carbonic acid as
a stimulus. One school, including Laqueur and Verzar (2), hold that
C02 is a specific stimulant and that other acids would not have the
same effect : others, of whom Winterstein has been one of the most
outstanding, have held the view put forward by Haldane and
Boycott that C02 merely acts because it is an acid and that any
other acid would do as well. The climbs at Carlingford seem decisive
on this point; in them we have the C02 concentration decreasing
while, as Haldane and Boycott (3) predicted, the total hydrogen ion (4)
concentration increases. The stimulus to the respiratory centre
cannot be simply the C02, for were it so the breathing would be
slower rather than faster.

REFERENCES

(1) Journal of Physiology, xxxn, p. 225, 1905.

(2) Laqueur and Verzar, Arch.f. d. ges. Physiol. CXLIII, p. 395, 1911.

(3) Haldaue and Boycott, Journal of Physiology, xxxvn, p. 355, 1908.

(4) Appendix IV.

CHAPTER XVII

THE EFFECT OF HIGH ALTITUDES UPON THE DISSOCIATION
CUEA7E OF AN INDIVIDUAL

THIS problem has now formed the subject-matter of a couple
of expeditions, the first of which went to Teneriffe in 1910 under the
auspices of the International Commission for the Study of High
Altitudes and Solar Radiation. The President of the Commission,
Professor Pannwitz, is much to be thanked for the completeness of
the arrangements, which made it possible to carry through a great
amount of scientific work in a very short time.

The island of Teneriffe is in some ways especially suitable for
such work, owing to the fact of its very temperate climate, temperate
in the sense of being neither too hot, too cold, nor too windy. Our
object in working there was the study of the human subject when at
rest. In Teneriffe this object is particularly easy of attainment. No
one in the island, so far as my experience goes, either takes or wants
to take violent exercise. In the Alps no one has any other object in
view than exercise in some form or other, but to walk up the Peak of
Teneriffe would be only less peculiar than to ascend to Col d'Olen on
a mule.

Our sea-level station in Teneriffe was at Puerto Orotava, where
we stayed at the Grand Hotel Humbert, and every facility was given
us for carrying out our work. The hotel was perhaps 300 feet above
the sea, the climate when we were there, which was in March, was
much like that of the English summer and was cool compared to
that which we subsequently encountered at the sea-level station of
our Italian expedition — Pisa.

The work performed at Orotava consisted in getting a base line,
so to speak, for our subsequent observations. The two persons
studied most completely were Mr C. G. Douglas, Fellow of St John's
College, Oxford, and myself.

16—2

244

Chester XVII

Our dissociation curves had been mapped out in England before
we left, they have been re-determined on several subsequent occa-
sions in this country. We attached great importance to the accurate
determination of them at the time and still greater importance
now. At that time they were important merely as objects of com-
parison for our other curves. Now they are more important, for it is
known that if one curve is accurately ascertained, any other curves
for the same individual can be found by the accurate determination
of a single point.

FIG. 115. — The station in the Caiiadas (7000 feet). The peak in the background. (Douglas.)

We gave in Chapter XIV the dissociation curve of Douglas,
showing the points on it which have been determined by us, together
with those which have been determined by Haldane and Douglas for
the same carbonic acid pressure, namely 40 mm.

The second station in Teneriffe was at Las Canadas. From the
point of view of the meteorologist this station is of especial interest.
It is about 7000 feet above the sea-level and therefore not much
higher than many centres of population. Johannesburg, for instance,
is close upon 6000 feet — putting aside therefore such places as the
mining towns in the Andes, the Canadas are a fair example of the

The effect of altitude

245

higher altitudes at which the working life of man is carried out on
a normal scale. As a station for the study of climate the Canadas
offer, among other advantages, a relative immunity from wind. This
in view of the investigations published by Lyth (1) is a factor worthy of
consideration. The island of Teneriffe consists roughly of a huge
crater about 8000 — 9000 feet in height. The diameter from lip to

FIG. 116. — The Canadas showing the living house and the laboratory. Espigone, one
of the summits, on the lip of the " old crater" in the background. (Douglas.)

lip is eight miles. On the south side of the island the lip is incom-
plete. The inside of the tip is a steep, not quite precipitous, cliff,
down which you must climb for a thousand or two feet, unless you
enter the crater as we did by a gap in the cliffs we called the Portillio.
We were then inside the old crater ; our back was to the cliff", which
in places rises in named summits, Guajara and Espigone for instance.
Our faces were towards a level plateau of sand.

246

Chapter XVII

From the description I have given so far it might be supposed
that a sandy plain now stretched before our eyes, and that in the
distance six or seven miles away we saw the opposite lip of the old
crater before our eyes. But this is not so, for the new crater arose
out of this plateau ; this was in front of us ; it appeared as a majestic
peak, bursting suddenly upon us as we emerged from the Portillio,
and rising to a height of about 12,000 feet. All that is left of the
plateau is a ring of level sand, in close proximity to the almost vertical
lip of the old crater. On the outside of this ring rise the cliffs to
about 1000 feet in height, on the inside the gradual ascent of the

FIG. 117.— View from the Alta Vista hut, showing the Caiiadas and the Portillio. (Douglas.)

peak. It was on this sand that our station was placed. No place at
this altitude could have been more sheltered by natural barriers.

It was quite unlike any place to be seen in Europe. Compare it
with much higher altitudes in the Alps, and the comparison is a very
remarkable one. The complete dryness of the atmosphere at the
Canadas spells the lack of the beautiful vegetation which makes the
Alpine snow line so attractive. Go out of the laboratory at Col
d'Olen, everything is moist underneath your feet, the cracks in the
rock are filled with saxifrages and gentians. Not so at the Canadas
although the vegetation at lower altitudes in Teneriffe is no less

The effect of altitude

247

beautiful than in the Alps. To get to the Portillio you must ride
through woods of heath which rises seven or eight feet in the air
and bursts into blossom above your head. But this has all been
left behind. The moisture condenses into clouds, which hang over
the island in a sheet at an altitude of 4000 — 5000 feet ; through these
clouds you penetrate. Once you arrive within the old crater you
have reached a new climate. Between you and the clouds there is
an impassable rampart. No vestige of mist was seen either above
us or around us during our sojourn in Las Canadas, the occasional
appearance of a cloud top above the lip of the crater was the only
reminder that such a thing as a cloud existed.

FIG. 118. — Transport of apparatus to the Alta Vista. (Douglas.)

With the absence of moisture there was a corresponding absence
of vegetation — scarcely anything green was to be seen. There is a
broom, called retama, that grows in considerable quantities, there
is a viola which seems to survive, but, beyond these, I noticed no
plant. Everything was arid. Imagine a mountain of coke from the
gas-works rising 5000 feet above you and you have pictured to your-
self the peak of Teneriffe as seen from the Canadas.

In some places, the " coke " is replaced by pummice, in others the

248

Chapter XVII

pummice is reduced to sand, as in the Montana Blanca or the
extreme summit of the Peak ; otherwise all is black, shading into
brown or perhaps red, all is crumbling and broken. How different is
the ascent of the peak from the Canadas to the ascent of Monte Rosa
from the Col d'Olen Laboratory. There is no element of exhilaration
about the former ; you start in the afternoon, you sit on a mule, you
wonder at its skill in putting its fore feet on the appropriate blocks
of broken lava, you think perhaps it sees wiiere it puts them ; but
when it comes to finding an explanation of how your mule places its
hind feet with equal certainty you "give it up."

??2*M& ^2&&

f*f»tt^ c- k- -\--fWWfi

-i*

FIG. 119.— The Alta Vista Hut (11,000 feet). Standing near door dressed in black
is Geheimrat. Professor Zuntz. (Douglas.)

How different from the Alps, from the bustle of guides and
porters and ropes before sunrise, from the hope and the beauty
of new things that comes with the rising of the sun as you stop for
the party to be roped before it ventures among the crevasses, from
the sense of 2000 more feet behind you as you rest and have some
refreshment at the Capanna Gnifetti. Then, as you tramp across the
Lysjock Glacier, you make up your mind for the last 500 metres which
is to bring you to the Margherita hut on the summit. How white
and exhilarating it all is and how far from the mind is any thought

The effect of altitude 249

of " giving it up " on Monte Rosa. Surely if Excelsior is the motto
of the Alps, Mariana (to-morrow) is that of Teneriife.

What fields of research lie in front of the physiologist before he
can explain how the subtleties of climatic conditions affect the
human mind, that entity of which all human activity is a mani-
festation.

How gross seem the avenues at present open to such investi-
gations. You are one person in one place, another in another. At
the Alta Vista, I became as one incapable of arithmetic — I vow that I
would have been at the bottom of the class with the dunce's cap on
and that I could not have helped it. At Col d'Olen I have heard two
clever and distinguished physiologists pause to discuss whether or
no four times eight made thirty-two. At Johannesburg I have been
told that a cricket team representing England so lost their nerve
that they laughed like children with quite trivial turns in the
course of the game and fell an absurdly easy prey to their South
African opponents. At the Margherita hut I have seen one of the
pleasantest and most considerate of companions behave as though he
were suffering from alcoholic excess in a mild degree. What of the
surprise that comes to us when we hear of cautious and skilful
climbers losing their lives doing extravagantly reckless things. Such
incidents are caused by the little recked of cerebral changes which
appear from time to time as the incidents of life at high altitudes.
They are doubtless the effects of acid intoxication — but of this later.

The climber depends for the most part on his cerebellum, his
cerebrum takes its chance and is little considered. One day these
psychological changes, which, in my opinion, appear much earlier
than cerebellar ones, such as defective coordination and giddiness, or
medullary ones, such as vomiting, will be studied for their own sake.

In the meantime we have got the pioneer work to do, we have got
to make a road into this forest wherever we can, we have got to find
out the changes which take place in the blood at such altitudes, and
in truth this is enough to tax our powers.

It may seem that I have depicted Teneriffe, as compared to Monte
Rosa, in a light unfavourable to the former. Let me disabuse the
reader's mind of this idea. Teneriffe has many advantages as a
place for the study of climate. Truly it takes longer to get there,
more truly it takes longer to get back, and most truly it costs more
money. If, however, you can secure the time and the money, there is
much to be said in favour of Teneriffe. From England, at all events,

250 Chapter XVII

the problem of getting your equipment there is much more simple.
You put it on the steamer, with due leisure, in Liverpool, London or
Southampton, and you take it off at Santa Cruz. If you do not mind
travelling in a steamer of 2000 tons, you may disembark under the very
windows of the Hotel Humbert and have your things carried up by
the hotel porter. We were fortunate enough at Teneriffe to be passed
through the customs — the same consideration, indeed, was shown to the
Monte Rosa expedition, for which I should like, here and now, to record
my thanks to the Ambassadors of France and Italy : yet, even taking
these acts of encouragement to scientific workers into consideration,
the difficulty of getting your apparatus intact to Col d'Olen is very
great. Think of the embarking and the disembarking on the channel
steamer, think of the terrors of the custom house, even if the luggage
is unopened, of the justifiable resentment of your fellow-passengers if
you take it in bulk in the railway carriage and of the impossibility of
putting it in the van ; delicate as my apparatus was, I brought all the
important pieces back intact from Teneriffe : little but broken glass
arrived in London from Col d'Olen.

But, perhaps, the greatest advantage of Teneriffe is that you can
start your work at the sea-level. On our Monte Rosa expedition we
made Pisa our base of operations. This, of course, is a far cry from
Col d'Olen as compared with the mule ride from Orotava to the
Canadas — that perhaps is a minor consideration and would not be a
consideration at all if Turin were made the base — but there seemed
to me a much greater difference between the climate of the plains of
Tuscany and that of the high Alps than exists between Orotava and
the Canadas. In the latter case there was the difference in the
moisture, upon which we have already touched, but the difference
in the temperature nothing near so great.

In spite, however, of these differences, altitude clearly had an
effect which was the same in both cases. There were two obvious
changes, both of them well known from the work of previous authors,
which must, in some way, affect the amount of oxygen in the
arterial blood — both of these are evident from a study of the alveolar
air. The first is the diminished oxygen pressure in the inspired air,
the second the diminished carbonic acid pressure in the expired air.
Of these two, the latter shall claim our interest first.

Great stress was laid by Mosso upon the diminished CO, in the
breath, not because its diminution is of any importance in the breath,
but because this is but the reflection of lowered CO2 pressure in the

The effect of altitude

251

body generally. To this absence Mosso attributed the symptoms of
mountain sickness, as Henderson (2) has more recently attributed the
effects of surgical shock ; this is, in fact, the "Acapnia theory."

That the want of carbonic acid would, other things being equal,
affect the affinity of the blood for oxygen is, of course, clear from what
has been said in the earlier chapters of this book. This had not escaped
Bohr, who pointed out that the increased affinity of the blood for
oxygen in the presence of a diminished carbonic acid pressure in the
blood, would form a very beautiful adaptation on the part of the
organism to the rarefaction of the air.

Here then was an obvious line of work.

The results were so different from the predictions made by Bohr
that we have been at great pains to verify them. Inasmuch, however,
as we have obtained them with a number of different individuals
both in Teneriffe and on Monte Rosa, and as these results have been
confirmed by Haldane on Pikes Peak, it seems to be established
that so far from the blood gaining in affinity for oxygen by its loss of
CO.,, the affinity of the blood for oxygen remains as a first approxi-
mation unaltered in spite of the lowered CO2 tension.

The following are the carbonic acid tensions in the alveolar air of
various workers at various altitudes expressed in mm.

Sea-level

7000 feet

10,000 feet

12,000 feet

15,000 feet

Douglas

40—41

36

32

Barcroft

40

40

33—32

38

29

Camis

39

37

29

Roberts

39

33—34

26

Math i son

39

32

28

Ryffel

45

35

30

Zuntz

35—36

29

27

The accompanying Figures 120 and 121 show the affinity of the blood
for oxygen at these different places and at different carbonic acid
tensions. The result is sufficiently astonishing ; whatever be the place,
the blood exposed to the carbonic acid pressure in the blood at that
place always possesses the same affinity for oxygen. It is different in
different people ; in Douglas's and Mathison's blood, for instance, it is
less than in that of Roberts, Camis and myself ; this does not matter,
the blood of each person has a certain affinity for oxygen, that affinity
remains almost unaltered. It was the same in Douglas's case at Orotava
with 41 mm. C02 pressure, at the Canadas with 35 mm. and at Alta

252

Chapter XVII

Vista with 33. My own blood had the same affinity for oxygen at
Orotava with 40, at the Canadas with 40 rnm., at Alta Vista with 38,
at Col d'Olen with 30 mm. and so on. This is the principle of the

100
90
8

70
60
50
40
30
20
1O

10

20 30 40 50 60 70

* —

i

^>^r

i

__ — —

ZUNTZ

• Berlin
+ Canadaf
OAlta Vist

i

x

^

f.

i

a

,

/o

/'

10

20

30

4

0

50

60

70 80 90 100

/

<r~

100

^

^^

. — '

/

Ofl

^^

DOUGLAS

1

r

°/

5 ""

— Normal Curve

-t- Canadas
O Alta Vista

/

GO

7

/

/

_^_

7

-40
30
20
10

7

X

^

/

7

p

/

/

r~

/*

7

/

^y

'

/

X,

9j

)

BARCROFT

+ Canadas 41
OAItaVista38
K Col.d'OlenCo230min
after return from Marg:

7

y

1

10 20 30 40 50 60

100
90
80
- 70
60
50
40
30
20
10

To

FIG. 120. — Dissociation curves of Zuntz, Douglas and Barcroft. The points in each case
were determined with the blood exposed to the existing alveolar CO., pressure of the
place.

The effect of altitude

253

individual constancy of the dissociation curve, which seems to be true
as a first approximation.

100

10 2.

0 3O 4O 50 6O

?o 8u eo

100-

^

_—

_-

|0 8O

90

100

/-

X

^

MATHISON

A

— Mesectic
0 Col. d'Olen

60
50

/

' 1

o 2

0 3

0

4

0 5

0 6

0

7

•

/Qfl

100

_

100'
90
80
7O
60
60
4O
30
20

30
20

x

/

• —

C. A M 1 <5

/

x

/

• Mesectic Pisa
° Margherita

/

2

0

y

^

/

^~-'

^ —

40
30
20
10
0

/

/

^

i

•

/

o

0

/

i

/

ROBERTS

•Mesectic Cambridge
° Margherita

/

/

J.

/

I

1

1

^/

3 10 20

30 40

50

6O 70 SO

90 100

FIG. 121. — Dissociation curves of Mathison, Camis and Eoberts, with points determined
at high altitudes, the blood being exposed to the existing alveolar C02 pressure of
the place.

The figures to which we have just referred the reader will
convince him, at all events they have convinced us, that the principle
is true, within the limits of the method of which we have been
treating. But may it not be that these limits are too wide for us to
be able to discern the changes due to the decreased carbonic acid
pressure? Have we a right to expect that we can observe the

254 Chapter XVII

differences which the diminished CO2 pressure would produce ? Could
we, in short, by the same methods detect any difference in the dis-
sociation curves of our blood when exposed at sea- level to the two
carbonic acid pressures which obtain in the body at, say, Berlin and
Alta Vista ? This is a simple experiment to try. I have here three
cubic centimetres of blood withdrawn by means of a needle from one
of the large subcutaneous veins in Prof. Zuntz's arm. I divide the
blood into two portions, the dissociation curve of each is determined,
one in the presence of 35 mm. CO2, the other in the presence of
23 mm. C02, the alveolar C02 pressures at Berlin or Orotava and
Alta Vista respectively. The difference in the dissociation curves
is evident enough.

Moreover, the same is clear in the case of Douglas where the
disparity of tensions was not so obvious. It is shown in Fig. 122 in
which the dotted curve corresponds to Douglas's normal blood
exposed to 34 mm. CO2. In Zuntz's case the tension at Alta Vista
was only two-thirds of what it was at Orotava — in Douglas's case
it was about three-quarters ; but it is perfectly clear that the same
blood exposed to these different tensions gives dissociation curves
which are distinguishable from one another with ease and certainty.
What then are we to say ? In Berlin the same blood exposed to CO2
tensions of 23 and 35 mm. respectively gives quite different curves,
yet the blood of the same individual exposed to 35 and 23 mm.
C02 tension in Orotava and Alta Vista respectively gives quite iden-
tical dissociation curves.

There is only one conclusion : as we ascend, the C02 has been dis-
placed in the blood by something else which produces an equal effect
on the affinity of the haemoglobin for oxygen. Probably this something
else is another, but not a volatile, acid ; it is something which does not
go away in the breath. It follows from what we have said above that
we should be able to discern its presence. When we were in Teneriffe
we did not fail to try. With Orbeli (3) I repeated the experiment when
we arrived home by subjecting cats to a diminished oxygen pressure.
In the interval between the Teneriffe expedition and the Monte Rosa
expeditions Mathison worked out the method for estimating the
amount of acid so thrown in the blood. We used this method and
made estimations by it and by other methods at Monte Rosa. Up to
a certain point the matter is perfectly clear. Let us say what there
is to be said about it. We will take up the points in the order
in which we have enumerated them :

The effect of altitude

255

(1) Firstly then, what evidence have we that the blood changed,
apart from the C03, as we ascended the Peak ? The simplest way of
attacking this problem is to expose the blood to equal CO2 pressures
in the various places and determine its dissociation curves. In
Fig. 122 are given the dissociation curves of Douglas's blood, with-
drawn respectively at Orotava, Canadas and Alta Vista and, in each
case, exposed to 41 mm. C02 tension.

10 20

30 40 50 60 70 80 90 100

100

90

10 20 30 40

80 90 100

FIG. 122.

FIG. 123.

FIG. 122. — Dissociation Curves (I — III) of Douglas's blood exposed to 40 — 41 mm. CO.,
pressure: III at sea level, II at 7000 feet G, I at 11,000 feet <g). • Blood taken at
7000 feet exposed to 36 mm. C02, X at 11,000 exposed to 33 mm. CO2, O at sea level
exposed to 34 mm. C02.

FIG. 123. Lines as in Fig. 122. Douglas's blood at Oxford. • +no lactic acid, O + 0-03—

0-04 °/Q, and <g> +0-07—0-08 °/0 lactic acid.

The difference between them is clearly to be discerned ; with each
rise of altitude the curve is displaced somewhat to the right — enough
to be seen. This suggests an increase in the acid radicles, or a
decrease in the bases of the blood. It becomes a matter of interest,
therefore, to see whether these curves could be imitated by the
addition of small quantities of acid to blood. Fig. 123 shows that

256

Chapter XVII

this can be done. To some of Douglas's blood lactic acid was added
to the extent, in one case of about 0*037 per cent, in another 0*075,
the dissociation curves were then determined ; their similarity to
those shown in Fig. 122 is very striking.

That these changes in the blood were due to simple oxygen want
was tested by trying to induce them in animals. In two cases a cat

100

90

80

70

60

50

40

30

20

10

yp-

7

/

-$7

7

10

50

60

70

80

90

100

20 30 40

FIG. 124. — O Defibrinated cats' blood. ® Points from Cat I after partial occlusion of
trachea and 15 minutes breathing of gas of increasing poverty in oxygen. • Cat II
at beginning of exp. O after 15 minutes gas-respiration, x after 21 minutes ditto.
The gas was becoming continually poorer in oxygen, at the end it was about 4°/0 oxygen.

was subjected by Orbeli and myself to an atmosphere in which the
oxygen gradually got rarer and rarer as time went on. The curves
obtained are shown in Fig. 124.

(2) We have said enough to indicate the possibility of a method
for the purpose of estimating the effective strength of the acid or
acids thrown into the blood apart from the carbonic acid.

The effect of altitude

257

If the blood be thoroughly shaken the CO2 may be shaken out.
With the small quantities of blood which we used (1 — 2 c.c.) about
ten minutes shaking was usually employed. Suppose now, we have
some blood which has been so treated. Let it be some of my own for
instance, it will give a certain dissociation curve : with another
portion of the same blood to which '01% lactic acid has been added
another dissociation curve will be obtained, and if '02°/0 lactic
acid be added we get yet a third curve. Now let us take a certain
oxygen pressure, say 17 mm. of mercury, it will be seen from an
inspection of Fig. 125 that at 17 mm. oxygen pressure we get
percentage saturations corresponding to the stated quantities of

100

I7MM.

10 20 30 40 50 60 70 80 90 100

FIG. 125.

acid added to the blood of a particular person. We may express such
results graphically : if we do we get a curve such as that shown in
Fig. 126. This curve may then be used as the basis of determinations
of unknown quantities of acid added to the blood in the body of the
person for whom it has been determined. The blood is withdrawn from
the finger or the arm, the CO2 is shaken out, it is exposed to 17 mm.
pressure of oxygen at 37° C. and the percentage saturation with oxygen
is determined ; the amount of acid is then read off from the curve.

It is necessary, of course, to make a curve for the blood of each
person ; moreover, the method is only approximate, but it has the

17

B. R. P.

258

Chapter XVII

advantage as compared with methods of hydrogen ion determination
by means of a gas chain battery, that the apparatus is very easily
carried and set up. I imagine direct determinations of hydrogen ion
concentrations at the Alta Vista hut would be attended with con-
siderable difficult}'.

At present I will only deal with one of the party, namely, myself,
as I was in most respects a typical case. The changes which took
place in my blood may be traced in the various stages of the journey.

80

60

50

40

30

20

10

•01

—i —

•02

-i —

•03

— i —

•04

•05

•06

•07

•08

FIG. 126. — The curve shows the percentage saturation of Barcroft's normal blood, with
oxygen, when exposed to 17 mm. oxygen pressure with various quantities of lactic
acid added as determined at Pisa. Ordinate = °/0 saturation. Abscissa = °/0 of lactic
acid added. The dots correspond to blood taken under the circumstances indicated :
the percentage saturations at 17 mm. were determined by analysis, the points were
then referred to the curve and thus the degree of acidosis was found.

We walked from Alagna to Col d'Olen, an ascent of 5500 feet (1700 m.),
in about four hours ; immediately on arrival some of my blood was drawn
and a determination made; the determination showed that, leaving
out of account the carbonic acid, the acidosis in the blood amounted
to the equivalent of '042 °/0 of lactic acid ; this however was not a
permanent co'ndition, it was a mixed effect due to exercise and
altitude together ; at rest at Col d'Olen the acidosis amounted to the

The effect of altitude 259

equivalent of 0'02 °/0 lactic acid. After we had been at Col d'Olen some
ten days we went up to the Margherita. The ascent is about 5000 feet,
we made it in seven hours, of which we rested for perhaps an hour at the
Capanna Gnifetti ; on reaching the summit there was again a further
large accession of acid to the blood. It reached 0'54 °/0. This again
had ebbed to some extent by the following morning. At present we
are only discussing the question of the condition of the blood when
the person under observation is at rest or at all events not exerting
himself. There is the most distinct and definite evidence that the
blood changes at each different altitude, in that apart from the C02
present the blood gains in acid, or loses in alkali, at all events that
its reaction alters in the direction of decreased alkalinity at the
higher altitudes.

(3) Though the method we have described, namely, that of
measuring the reaction by the affinity of the blood for oxygen, proved
most satisfactory and was operated with great skill and rapidity by
Mathison and Roberts, we naturally wished to have some other methods
at our disposal. One of these was the method of Boycott and Chis-
holm (1) for determining the reaction of blood. It depends upon the
fact that the haemoglobin precipitates at a definite acidity. The
blood is therefore titrated with acetic acid until a precipitate appears,
the original reaction being calculated from the amount of acid
necessary to bring the blood to the point of precipitation. This
method was not used so systematically as that previously employed.
It gave qualitatively the same result (2).

Lastly, as it seemed to us probable that the acidosis which we had
observed in Teneriffe might have been due to a lactic acidosis, we
determined to make actual determinations of the lactic acid present
by Ryffel's (3) method (see Chapter XV). Let us state precisely what it
is that this method estimates. It is the total quantity of the radicles
of the «-OH organic acids. It therefore has to do with something
essentially different from the others ; the others are a measure of the
hydrogen ion concentration, for it has been shown by Mathison that
the change in affinity of the haemoglobin for oxygen when acids are
added to the blood depends upon the concentration of hydrogen ions,
and the same has been shown by Boycott and Chisholm for their
method. But Ryffel's method is different ; it measures the lactic acid
radicle irrespective of the degree of ionisation of the acid or of the
bases with which it is united. Moreover it includes an}7 other a-OH
acids which may be present in the solution.

17—2

260

Chapter XVII

Our conjecture that the change in the reaction was due to the
increase of lactic acid turned out to be entirely erroneous. Ryffel
obtained the following figures for the "lactic acid" in the blood of
various members of the party at Pisa, Col d'Olen and the Margherita
hut respectively.

Subject

Lactic acid
Pisa

Increase of
lactic acid
Col d'Olen

Increase of
acidity* by
Mathison's
method
Col d'Olen

Increase of
lactic acid
Cap. Margherita

Increase in
acidity* by
Mathison's
method

Eoberts

0-018

+ 0-006

0-023

+ 0-027

+ 0-035

Camis

0-017 +0-004

0-026

—

—

Ryffel
Mathison

0-019
0-013

+ 0-004
-0-005

0-025
0-016

+ 0-021

+ 0-048

* The acidity is expressed in terms of lactic acid.

From these figures it will appear that the lactic acid at Pisa and
Col d'Olen was practically the same, whilst at the Capanna Margherita
there was a sensible lactic acidosis.

Our Monte Rosa expedition therefore left us in the following-
position. We found:

(1) In common with previous workers that the higher the altitude
the less CO2 in the alveolar air and presumably less in the blood.

(2) The higher the altitude the more marked the acidosis in the
blood when the CO2 is shaken out(4).

(3) At any altitude the acidosis and the diminution of C02 so
nearly balanced one another that the reaction of the blood remains
practically constant and the dissociation curve is therefore mesectic.
This is true at all events on a first approximation. It is only by
a statistical treatment of a great number of determinations'5' that
a degree of meionexy, corresponding to a fall of about 7% in
K, may be discovered — sufficient to give the respiratory centre the
slight stimulation (6) which would account for the increased ventilation
observed.

Every man at rest has therefore a dissociation curve which remains
approximately constant in spite of changes in those individual factors
whose balance preserves the constancy of the curve, its form is a cal-
culable quantity involving almost fixed values of n and K. You might
label the man with these letters as a mark of his identity. What acid
is responsible for the acidosis in the blood is yet to be ascertained, it is
clearly not lactic acid or any of its close relations ; on the other hand

The effect of altitude 261

it may be that there is no increase of acid at all but rather a diminu-
tion in the amount of alkali present. From the point of view of the
dissociation curve the matter is of little interest, but from a wider
point of view it is of great interest ; so far we are left without a know-
ledge of mechanism of the adaptation which we have discovered. The
idea of oxygen want producing lactic acid is familiar enough ; were the
acid lactic, we should at once say that it had been produced in the
tissues as the result of oxygen want — but we should be in another
difficulty for lactic acid is secreted by the kidney. If we found
it continuously in the blood we should expect to find it in largely
increased quantities in the urine. Ryffel tested this point and did
not find any great excess.

Normal urine, sleep O'OOl °/0 lactic acid

Col d'Olen 9.15 p.m— 7 a.rn 0-0012

There seem to be two alternatives before us ; one is that oxygen
want under these circumstances produces acids in the tissues which
are not readily excreted : the other is that oxygen want so affects
the kidney that it excretes alkali more freely. Certainly Verzar's ex-
periments seemed to show that oxygen want did increase the activity
of the kidney. We are now entirely in the region of speculation,
however, so let us retrace our steps a little.

A few lines back we spoke of the altered condition of the blood as
an "adaptation" to the altered conditions of barometric pressure.
In doing so we introduced a fresh idea into our narrative, namely,
that the change in the blood was beneficial to the organism. This is
true; so far we have treated the matter merely as an interesting
observation that the dissociation curve of the man remains constant ;
but the alteration of the individual factors while the balance is
preserved leads to a very important result. Over any considerable
interval of time there is always a certain relation between the CO2
in the alveolar air and the oxygen in the same.

The proportion of the one to the other depends upon the respi-
ratory quotient and ultimately upon the food that is eaten. If the
CO2 changes the oxygen will change, so that if Cv be the pressure of
CO2 in the alveolar air and CA that in the atmospheric air, while Ov
be the oxygen in the alveolar air and OA that in the atmospheric air,

C — C

-7^- —T~ will remain approximately* constant.

* The determinations of C02 pressure in the alveolar air were made at the end of
expiration. To obtain the respiratory quotient more exactly a slight correction has

262 Chapter XVII

For the purpose of the present calculation CA may be neglected,
it is under half a millimetre and therefore we may say as an approxi-

Q

mation that ^- ^~- is constant.
VA - Uv

Let us see by a specific example the bearing of this constancy on
the question. In Teneriffe at the sea-level Douglas's alveolar CO2
(CV) pressure was 41 mm., his respiratory quotient was 75, therefore

41

(0A — OV) was ._- =51. The value of 0A was 160mm., or corrected
'/5

for the pressure of aqueous vapour in the lung 150 mm., therefore
OA = 99 mm. Now at the Alta Vista hut, where the barometer was 514,
the value of OA corrected for aqueous vapour in the lung is 104.
Had his C02 pressure been as before 41, OA — Ov would have been 51
and his alveolar oxygen pressure 53 ; but owing to the acclimatisation
which we have described his C02 pressure dropped to 32, and therefore

32
OA — OK was ^rr = 43, instead of 53, therefore his alveolar oxygen

pressure would work out at 104 — 43 = 61 mm. * He had gained
10 mm. at this altitude, or one-sixth of his whole oxygen pressure by
the change of the distribution of acids in his blood.

My own case afforded a control experiment. My carbonic acid
was scarcely altered as between the sea-level and 11,000 feet in
Teneriffe, with the result that my oxygen pressure at that altitude
was only about 49 — 50 mm., while that of Douglas was 61, a fact
which made a great difference to our comparative comforts at the
Alta Vista hut. This again was controlled by another observation.
For reasons of which I shall speak elsewhere, at Col d'Olen my
alveolar C02 did drop to 33 mm., my oxygen did reach 64 mm., whilst
later at Col d'Olen, after I had been up to the Margherita hut, where
my blood had acquired another dose of acid, displacing more C02,
and had retained it after my return, my alveolar C02 pressure was
only 30 mm. and my oxygen up to 70 mm. Therefore, even allowing
a mm. or two for experimental error, my oxygen pressure was half as
much again as it had been at the Alta Vista. On my dissociation
curve the difference between 70 and 50 mm. is the difference between

to be made. This may be done from a table found in the appendix copied from
Haldane's " Methods of Air Analysis."

* These calculated values agreed very closely with the observed values, which
were 58'5 — 62'5 mm.

The effect of altitude

263

FIG. 127 a.— Col d'Olen showing track. (Durig.)

FIG. 127 &.— Hut at Col d'Olen. (Aggazzotti.)

264

Chapter XVII

85 °/0 and 93 % saturation, an important difference in the satu-
ration attainable. But probably the rise in the limiting percentage
saturation is less important than the effect on the rate at which
oxygen is taken up. Since this rate is directly proportional to the
pressure of the oxygen, the oxygen is taken up nearly half as fast
again in the lung of the acclimatised person.

There are two points in the above argument which have been
passed over rather lightly because the argument itself only demanded
the mere mention of them. Yet they are both of considerable interest
on their own account. They are the following : firstly, why did I be-
come acclimatised at once on Monte Rosa when I did not change at
Teneriffe? and secondly, why did Douglas acclimatise at Teneriffe
when I did not do so? The fundamental principle is that the
acclimatisation is due to oxygen want. In order that you should
acclimatise satisfactorily you must do so gradually, and therefore
court oxygen want from the beginning of the ascent, then the acid
gradually finds its way into the blood and the CO., gradually finds its
way out. The most satisfactory way of doing this is to walk up the
mountain. The difference between my ascent of Monte Rosa and
that of the Peak was that the former was made on foot, the latter on
a mule. As regards the acidosis when I reached Col d'Olen there
was no doubt. Some of my blood was taken directly I arrived there ;
it was analysed by Mathison and contained the equivalent of 0*42 °/0
lactic acid. The following figures, obtained with Roberts' blood and
my own, illustrate this point.

Equivalents of lactic acid in blood.

Barcroft

Roberts

Mathison

Reaching Col d' Olen

0-042%

0-08%

Living at ,,

0-02

0-02

0-00 °L

Reaching Margherita Hut

0-054

0-09

Living at

0-05

Living at Col d'Olen after return from Margherita ...

0-038

0-02

0-016°/0

At each effort of climbing then there was an acidosis which
only partially disappeared in the ordinary life of the place. An
interesting point was the fact which we discovered on coming down
to Col d'Olen from the Margherita hut, that the acidosis which
had been induced by going up there remained some time and
materially strengthened our position owing to the lowering of the

The effect of altitude 265

C02 pressure and the corresponding rise in oxygen pressure which we
have already discussed. The beneficial effect of exercise at high
altitudes is of course commonplace amongst the persons who frequent
mountains, and especially the benefit derived from going up a little
higher than the point at which one is living, and then coming down
again. Our results show that this benefit is no intangible affair which
may be vaguely included under the general term "training," but that
it is a very definite change in the blood of the individual which may
be detected by the chemical analysis of that fluid.

There is no difficulty then in stating the reason why I should
have become acclimatised when I walked up the mountain and not
when I went up on the mule — in the former case the effort of the
climb induced oxygen want and consequently acid production, in
the latter case this element was absent.

To answer the other of the two questions set above, why did I not
become accommodated at Teneriffe when Douglas did so ? is at first
sight more difficult, as the consideration of mountain sickness is bound
up with that of exercise at high altitudes. I will postpone the answer
till the next chapter.

REFERENCES

(1) Boycott and Chisholm, Biochemical Journal, v, p. 23, 1911.

(2) This research is published in abstract in the Physiol. Proc. ; Journal of Physiol.

XLV, p. xl, et seq. and will be sent for publication in extenso to the Philo-
sophical Transactions of the Royal Society.

(3) Ryffel, Physiol. Proc. ; Journal of Physiol. xxxix, p. v, 1909.

(4) See also Aggazzotti, Arch. ital. de Mol. XLVII, pp. 54, 66 ; Galleotti, Arch. ital.

de biol. XLI, p. 80.

(5) Barcroft, Physiol. Proc. ; Journal of Physiol. XLVI, p. xxx, 1913.

(6) Campbell, Douglas, Haldane and Hobson, Journal of Physiol. XLVI, p. 301, 1913.

CHAPTER XVIII

THE EFFECT OF ALTITUDE ON THE DISSOCIATION CURVE
OF BLOOD CONSIDERED IN RELATION TO EXERCISE

IN each of the two preceding chapters I have tried to draw
a picture, in one of the man who is taking exercise at ordinary
altitudes, in the other of the man who is at a high altitude but
not taking exercise.

I have depicted the man who is taking exercise as meionectic ; his
blood is unusually acid, it takes up oxygen less readily than usual, it
parts with it more readily, while at the same time the change in the
reaction of the blood quickens the respiration and the heart beat.

I have pictured the man living at a high altitude as possessing
blood of usual or almost usual reaction. Mesectic, or nearly so, his
blood contains an unusually small quantity of carbonic acid and an
unusually large quantity of other acid radicles. The diminution of
carbonic acid in the blood is reflected in the alveolar air with the
result that CO2 pressure in the pulmonary alveoli is unusually low
and consequently the oxygen pressure is higher than it would other-
wise be.

In the present chapter I wish to consider the effect of altitude
upon exercise. In a couple of words it is this : a given degree of
meionexy would be produced by a lesser amount of exercise at a high
altitude than at a low one, or, to put the matter in another way,
a given amount of exercise would produce a greater degree of
meionexy at a higher altitude than it would at a low one.

We shall now give some account of the experiments on results of
which these statements are based.

Our party had two mountain tracks as similar to one another as
we could make them, each was a climb of 1000 feet. The low level
course was at Carlingford, co. Louth. It has already been described. It
started from practically the level of the sea. Our high level course

Exercise at high altitudes

267

commenced at a point 1000 feet below Col d'Olen and ended
at the Laboratory. It extended from an altitude of 9000 feet to
one of 10,000. Our scheme was to go down 1000 feet from Col

10 20 30 40 50 60 70 80 90 100

90

80

70

60

50

40

30

20

10

ROBERTS

Mesect/c

— Meionectic

CARLINGFORD

IOOO' in 2O mins

ROBERTS

Mesect/c

— Meionectic

)'OLEN

IOOO' in 33 mins

100

90

80

70

60

50

40

30

20

10

0 10 20 30 40 50 60 70 80 90 100

FIG. 128. — Showing approximately equal change in the dissociation curve as the results of
1000 feet climb near sea-level in 20 minutes and in the high Alps in 33 minutes.

268 Chapter XVIII

d'Olen and ascend at the required speed. The ascent is a steep
mountain path all the way from a small pond, right up to the Albergo :
thence a few minutes' walk over some rocks to the right brought us
to the Laboratory. Everything was read}r when the subject of the
experiment came in ; we took his alveolar air before his respirations
slackened, then he was at once bled ; the blood was divided into
two portions, one given to Mathison and Roberts for the acid deter-
minations, the other retained by Camis and myself for the determina-
tion of the dissociation curve at the subject's alveolar C02 pressure,
i.e. the dissociation curve of the man as he stood.

Parallel climbs at Carlingford and at Col d'Olen were made by
Roberts and myself. Let me first discuss those in which the indi-
vidual endeavoured in each case to do his climb with the same
amount of effort at the two places.

Roberts in each case climbed as fast as he could walk. He did
not run. The time occupied at Col d'Olen was 33 minutes and at
Carlingford 20 minutes ; that is, he was able to walk about half as
fast again with the same degree of effort at Carlingford as at Col
d'Olen. The degree of meionexy induced at each place can be
judged from the change in the value of K in the equation

y Kxn

100 ~ 1 + Kxn '

Values of K (Roberts}.

Carlingford
1000 feet in 20 mins.

Col d'Olen
1000 feet in 33 mins.

Before start

0-00033

0-00033

At finish

0-00018

0-00016

Change in K

0-00015

0-00017

The changes in K which were induced by the two ascents were
almost identical. The curves are given in Fig. 128 and are scarcely
to be distinguished. Climbing at a much slower rate the experiment
made upon me yielded much the same result as that just quoted. In
each case I climbed at such a rate that I could just respire efficiently
without departing from nasal breathing; had I gone faster I should
have had to breathe through my mouth.

Exercise at high altitudes

269

The dissociation curves are given in Fig. 129 and were, as in
Roberts' case, almost identical at Col d'Olen and at Carlingford.

100

90

80

70

60

50

40

30

20

10

10 20 30 40 50 60 70 80 90 100

BARCROFT

Mesectic

Meionectic

CARLINGFORD

1000' in 30 mins,

BARCROFT
-Mesecfic

Meionectic

COL D'OlEN
1OOO in 45 mins

+

100

90

80

70

60

50

40

30

20

10

10 20 30 40 50 60 70 80 90 100

FIG. 129. — Showing approximately equal change in the dissociation curve as the result of
1000 feet climb near sea-level in 30 minutes and in the high Alps in 45 minutes.

270

Chapter XVIII
Values of K* (Barer oft).

Carliugford

Col d'Olen

Before start

0-00029

0-00029

At finish

0-00017

0-00019

Change in K

0-00012

0-00010

It is clear therefore that as a rough and ready index of effort in
climbing, the maintenance of a rate which just fell short of making
one breathe through the mouth served its purpose, in that it caused
changes in the chemistry of the blood which were identical within
the limits of our experiments.

I have dwelt upon the identity of the results obtained climbing
1000 feet, from the sea level to 1000 feet altitude in one case, and
from 9000 to 10,000 feet in the other. There remains to be stated
the important fact that the former climb was accomplished in 30
minutes, whilst 45 minutes was necessary for the latter. In my case
as in Roberts' the rate at Carlingford was half as fast again as at
Col d'Olen.

Now let me consider an instance of climbs at different altitudes
which occupied the same time in each case. This consists of a com-
parison between an ascent made by me at Carlingford and one made
at Col d'Olen at the same speed. The result in this case is clear : whilst
a very obvious degree of meionexy was produced at Col d'Olen that
at Carlingford is scarcely to be appreciated by these methods. The
following are the data as regards K :

Values of K* (Barcroft).

Carlingford

Col d'Olen

Before start

0-00029

0-00029

At finish

0-00024

0-00017

Change in A'

0-00005

0-00012

The slow ascents are perhaps more instructive than the rapid
ones, for if the climbing be a little slower than that of the last experi-
ment which I have mentioned, the degree of meionexy produced at

* For hydrogen ion values see Appendix III.

Exercise at high altitudes

271

low altitudes will become inappreciable. Not so at high altitudes,
however. This was most strikingly shown on our ascent from Col

10 20 30 40 50 60 70 80 90 100

100

90

80

70

60

50

30

20

10

BARCROFT

• Mesectic

Meioneotic

CARLINGFORD
IOOO' in 45 mins

BARCROFT

Mesectic

Meioneotic

COL D'OLEN

lOOO7 in 4 5 mins

100

90

80

70
60
50
40
30
20

0 10 20 30 40 50 60 70 80 90

FIG. 130. — Showing the changes in the dissociation curve which result from a
climb of 1000 feet near the sea-level and in the high Alps.

d'Olen to the Margherita hut. It is an ascent from an altitude of
10,000 to one of 15,000 feet. For the ascent we took about eight

272

Chapter XVIII

hours. This time included a rest of an hour at the Capanna
Gnifetti. The gradient does not become steep until the last 1500
feet, which are up steps cut in the ice. The ascent up this final stair-
case, when made by a party roped together, is the most leisurely affair.
Apart from this staircase there is nothing that could even be called
climbing : a path along the summit of a ridge soon brings the party

FIG. 131. — The Punta Gnifetti on the summit of which may be seen the
Capanna Margherita. (Durig.)

to a couple of glaciers. These are crossed diagonally with no par-
ticular effort, then bending round the corner of a rock and making
an ascent of a few metres the party found itself at the Capanna
Gnifetti, the point at which the route to Zermatt diverges from that
which leads to the summit. After this point it is all snow and ice.
Immediately in front of the party is the summit of Lyskamm, that

Exercise at high altitudes

273

conical peak of pure white which commands the Gressoney Valley.
To the right of this was our way, up a snow slope and still up, till
we reached the Lysjoch glacier, and then we were on a vast plateau
from which rise the peaks which form the crown of the Monte Rosa
group. Lyskamm was on the left, the punta Parrot, the punta Gnifetti,
the Ludwigshohe, and the Pyramid Vincent on the right ; for about
half an hour's walk it was level, though the snow entailed a certain
amount of effort. Looking down to the left, past Lyskamm, was the
Gorner Grat and Monte Cervino (Matterhorn) beyond. Our height
may be grasped from the fact that we were then about the level of

FIQ. 132. — The Capanna Margherita. (Aggazzotti.)

the top of the Matterhorn. Suddenly we turned to the right and as
I have said crept slowly up the punta Gnifetti to the summit.

Whilst I cannot hope that my description will really bring the
scene before the eye of the reader, it will suffice to show that such
a walk would be an intolerably tedious stroll at ordinary levels. Yet
it taxed our powers to the full. I cannot for a moment suppose that
any appreciable degree of meionexy would be induced by a walk
from say Fort William to the top of Ben Nevis in seven or eight or
nine hours. The only symptoms from which the pedestrian would
suffer would probably be those of cold. Not so when we arrived at

B. R. F. 18

274

Chapter XVIII

the Margherita hut. At these rates of climbing the effect of altitude
to all intents and purposes is quantitative, the effect is clear at
15,000 feet, it is not to be found at low levels. It is in fact an acid
intoxication ; its extent may be gathered from the two sets of curves
given in Fig. 134. They are those of Roberts and myself. At the
present time they are the record cases of meionexy.

To pass from the actual degree of meionexy produced, let me
treat of the degree of acidosis, confining this term to the appearance
of unusual acids in the blood. We must ask two questions :

(i) To what extent did acid radicles appear in the blood ?

(ii) What were the acid radicles which appeared ?

PIG. 133. — View of Matterhoru from the Capanna Margherita at sunset. (Durig.)

Here again we must recapitulate what has been said about
acidosis in the last two chapters. The acidosis due to exercise
appears to be a lactic acidosis, that due to altitude does not appear
to be a lactic acidosis at altitudes of 10,000 feet, though at altitudes
of 15,000 feet there is some degree of lactic acidosis. As regards the
degree of acidosis the best determinations are those performed on
Mathison and on Camis, but especially the former, and this for
the following reasons : firstly Mathison started his climb free from
acidosis both at Col d'Olen and at the commencement of his low level
station, which was the Sugar Loaf at Abergavenny ; secondly his
climb was in each case a very strenuous one, 1000 feet in 20 — 21 mins.
The data are as follows :

Exercise at high altitudes

275

Mathison, acidosis expressed in equivalents of lactic acid in blood.

Before climb After climb Due to climb

Abergavenny O'OOO 0-021 0-021

Col d'Olen . O'OOO 0-040 0-040

10

20 30 40 50 60 70 80 90 100

ROBERTS

Mesectlc

Meionectlc

Climb to
MARGHERITA HUT

X

x

BARCROFT

Mesectic

Meionectic

Climb to
MARCHERITA HUT

100

90

80

70

60

50

40

30

20

10

0 10 20 30 40 5O 60 70 80 90 100

FIG. 134. — Changes in dissociation curve caused by walk from Col d'Olen

to the Margherita hut.

18—2

276

Chapter XVIII

Camis, acidosis expressed in equivalents of lactic acid in blood.

Before climb After climb Due to climb

Cold'Olen 0'023 O'Ool 0'028

It is clear from the figures which have been obtained from
Mathison's blood that the effect of altitude has been to increase the
acidosis to a very marked extent.

And now as to the nature of the acidosis ; so far as a single experi-
ment can settle that point, the following figures show that the
acidosis, produced by exercise at high altitudes as at low ones, is a
lactic acidosis.

Acidosis in Camis s blood.

Pisa

Col d'Olen

Before climb

After climb

Due to climb

Acidosis in equivalent of 1
lactic acid... . j

o-ooo

0-023

0-051

0-028 °/0

Lactic acid in blood

0-014

0-017

0-043

0-0215 /

^ "•"«

As to the length of time taken for the effects of his exercise to
pass off we have but scanty data. Here again we must distinguish
between the acidosis and the meionexy.

Changes observed as the result of ascent of WOO feet by

11 August

13 August

Before
starting

Immediately
after arrival
6 p.m.

6.20 p.m.

8 p.m.

Alveolar air

36mm.
•00

37 mm.
•040

27 mm.
•040

37 mm.

•018

•on

Acid in blood in equivalent)
of lactic acid j

Before the climb began Mathison was mesectic. Immediately after
the exercise there was considerable excess of acid in the blood and no
diminution of CO2 in the alveolar air. This of course meant a total
reduction in the alkalinity of the blood, and consequently a meionectic

Exercise at high altitudes

277

condition, the physical signs of which were increased frequency of
the pulse and violent panting. The effect on the dissociation curve
is typical. The curve was pushed to the right. Mathison's normal
dissociation curve as determined before going to Col d'Olen is shown

in Fig. 135. It has the equation y = 100 — „ n, the constants being

1 T J\.3C

K = '000212, n = 2'5. The curve obtained at Col d'Olen immediately
after returning from the climb is also shown, together with the
points upon it which were actually determined. It has the following
constants: w=2'5, ^='000140.

IUU

90
80
70
60
50
40
30
10
10
0

^

^rr-

S s-

^

-"""

/<

X

/

/

//

J /

/

MATHISON

c

H

y '•

/Wesee/Yc
Meionectlc
COL D'OLEN
lOOo'in 1 9 mins.

/

I t

^

'/

10 20 30 40 50

60 70 80

90 IOC

FIG. 135. — Changes in Mathison's dissociation curve caused by climbing 1000 feet
in 19 minutes, and their disappearance. • 5 ruins. x 20 mins. o 2 hours
after exercise.

In twenty minutes Mathison's alveolar carbonic acid pressure had
dropped from 37 mm. to 27 mm., and a couple of determinations of
his dissociation curve under the new conditions indicated that it was
mesectic within the limits of error of the determinations, and by
8 p.m., two hours and a quarter after the experiments, he had settled
down to his permanent condition. There was an acidosis equivalent
to '018 °/0 °f lactic acid, but from this he scarcely, as far as is known,
receded while he was at Col d'Olen.

278 Chapter XVIII

And now to turn to the physiological significance of the meionexy
and of the acidosis.

The essence of any mechanism of adaptation to high altitudes
must be a " speeding up " of the whole process of respiration, physical
and chemical. The blood in the lung is exposed to a lower oxygen
pressure than usual ; it is saturated to a less extent with oxygen
than formerly. The tissues begin to suffer from oxygen want : a very
trifling change in the blood is enough to produce a great change in
the circulatory conditions ; the pulse quickens, respiration becomes
deeper and more rapid, the amount of blood which leaves the lung
increases perhaps twofold. Let the reader picture to himself each
corpuscle as a ship with its little cargo of oxygen and twice as many
of these are leaving the lung as before ; they go to the tissue and go
through the capillary at perhaps twice their former velocity ; but stay
—how are they to unload their cargo in the reduced time at their
disposal? How futile would be the whole scheme if the corpuscle bolted
through the capillary carrying its oxygen into the vein with it. Here
is the advantage of meionexy. The meionectic blood parts with its
oxygen with much greater rapidity than does the normal blood under
given circumstances. Therefore when the corpuscle gets to the capil-
lary it can discharge its cargo with unusual facility.

It is true that the meionectic blood is at a disadvantage in the
lung. It probably has not time in the lung to become saturated to
quite the degree that normal blood would. But blood saturates
itself up to 80 or 85 % very rapidly, and the organism must take the
risk of doing without the rest, it meets the small deficiency in the
degree of oxidation by a large increase in the quantity of blood
leaving the lung. The thing of course is a compromise, you must
lose something somewhere. You cannot pretend that the organism
is not working at a disadvantage at an altitude of 15,000 feet. It is
making the best of a bad business. The best it can make is meionexy.
Meionexy involves stimulation of the neuro-muscular mechanism of
respiration and quickening of the proportionate rate at which the
blood loses oxygen in the tissues.

But at slow rates of climbing the degree of acidosis is much
greater than the degree of meionexy. Including the carbonic acid
the blood becomes to a small extent more acid than formerly, but
the blood becomes much richer in other acids and poorer in C02 than
before. The carbonic acid is displaced in the blood and after a pre-
liminary rise it ultimately decreases in the alveolar air. Therefore the

Acid intoxication 279

oxygen pressure in the alveolar air is higher than it would otherwise be.
This relative increase in the oxygen pressure has a twofold effect — it
increases the rate at which the blood can acquire oxygen in the lung
and it increases the limiting percentage saturation. In effect it brings
you thousands of feet down the mountain again. So even from the
chemical point of view the blood is not under such bad conditions in
the lung after all. The meionexy reduces affinity of the blood for
oxygen, but indirectly betters the conditions for its acquisition. What
it takes away with one hand it gives back with the other. It is not
all loss in the lung and in the tissue it is all gain — to the organism.

The adaptation to altitude, being as I have said a compromise,
must break down if the conditions become too severe, that is to say,
if the altitude becomes too high for the necessary exercise, or the
exercise too great for the altitude ; the result is mountain sickness.
I have sometimes been asked the following question : If the effect of
altitude is merely to produce a given degree of meionexy with a less
degree of exercise, why are the final effects of exercise at low and
high altitudes different ? No one suffers from mountain sickness in
say the University Sports, though the effects of meionexy are evident
enough. Mountain sickness is doubtless caused by want of oxygen
in the brain itself. It is not merely that the vomiting centre is
stimulated by a meionectic condition of the blood in the general
circulation. At high altitudes even tissues which are comparatively
inert are suffering to some extent from oxygen want ; or at least
to prevent their so suffering there is some degree of dilatation. The
result of this dilatation, combined with faulty oxygenation of the
arterial blood, is to reduce the supply of oxygen of the medulla
to a minimum. Then some trifling change takes place ; the wind
meets you in the face and you hold your breath ; the digestion
of food, or perhaps even the contemplation of it, sends an extra
supply of blood to the abdomen ; sleeps comes and the blood
tends to leave his brain. Any of these will make you feel sick. The
actual symptoms of mountain sickness resemble those of haemorrhage
rather than of exercise.

Is it possible to tell those who are likely to suffer from mountain
sickness from those who are not? It is difficult to say without
studying a much larger number of cases than those which have already
been subjects of research. In Teneriffe(1) the individuals whose
bloods had the smallest affinity for oxygen at a constant OCX pressure
suffered least from mountain sickness. Zuntz' and Douglas' blood,

280 Chapter XVIII

for instance, both had a value for K at 40 mm. CO2, which was about
'000212, mine was '000292. Neither of these suffered in the least
from mountain sickness ; I did, though it never actually amounted to
vomiting. In the case of Douglas' and my blood, these values for K
represented our actual curves. It appeared that Douglas' more
gradual dissociation curve meant a more gradual degree of adapta-
tion. I did not begin to adapt till my alveolar oxygen pressure was
10 mm. below that of Douglas, mine being about 50 mm. of oxygen
and his being 60 mm.

At that pressure the percentage saturation in each of our bloods
was roughly the same, but I was in a very much worse position than
Douglas

(1) Because a trifling further drop in the oxygen pressure in the
blood meant a large drop in percentage saturation in my case and a
small one in Douglas' case.

(2) Because 60 mm. oxygen as compared to 50 meant that, other
things being equal, oxygen could be taken up more rapidly in the
ratio of 6/5.

(3) Such a trifling drop in the oxygen pressure to which the
actual blood was exposed might be due simply to an increase in the
pressure gradient between the outside and the inside of the lung
epithelium. An increase of this sort would be the direct result of
a little exercise.

Zuntz' case was a little different from that of Douglas. It is true
that with the same CO2 pressure their bloods gave the same dissocia-
tion curves, but in point of fact Zuntz had a less alveolar C02 pressure
and his dissociation curve at this pressure was more like mine than it
was like Douglas'. But the lower CO2 pressure gave him the advantage
over me, for it meant of course a higher oxygen pressure.

Between the dates of the Teneriffe and Monte Rosa expeditions
two mountaineers of eminence were kind enough to give me blood
for analysis, these were Oscar Eckenstein, who claims to have lived
at an altitude of 20,000 feet longer than any other man, and
Longstaffe, the Himalayan explorer. The fact that the blood of these
two persons, when exposed to the standard C02 pressure, gave the
lower values for K than any other human blood that I have analysed,
was an interesting confirmation of my theory.

It became a matter of some little interest to me at Pisa to specu-
late as to which of the members of the Monte Rosa party would
stand the high altitude best. The five persons fell into three groups.

Pr oneness to mountain sickness 281

In the first was Mathison, whose blood at 40 mm. gave a value
for £"='000212. In the next were Camis, Roberts and myself,
£"'00029 — '00033, and lastly Ryttel, whose blood had a value
> '00037. According to the theory then we predicted that Mathison
should suffer least, Ryffel most, and the rest of us should occupy
an intermediate position, and whether by coincidence or otherwise
our classification proved to be correct.

REFERENCES

The equations of the dissociation curves discussed in the above chapter were
published in the Physiol. Soc. Proc., January 1913 ; the data of acidosis in the
Report of the Committee on High Altitudes, British Association Proc. 1911.

(1) Journal of Physiology, XLII, p. 43, 1911.

CHAPTER XIX

SOME CLINICAL ASPECTS

MY concluding chapter will tell of a couple of researches, one at
University College Hospital (1), the other at Guy's Hospital (2), in which
the dissociation curve " used as an indicator " has revealed new facts
with regard to disease and done something to shed a new light on
certain pathological conditions.

The first of these researches with which I will deal has been
carried out by my two former colleagues, Ryftel and Poulton. Their
concern was to explain the dyspnoea which takes place during
uraemia. Naturally they pursued the methods with which the reader
is now already familiar, the methods which had already given positive
results at Carlingford, on Monte Rosa and in the Alta Vista hut.
These methods showed clearly enough that the dissociation curve of
the patient had shifted in the direction of meionexy, a result which
forms the counterpart of the discovery by Schlayer and Straub — the
latter an old member of " the firm " -that the alveolar carbonic acid
sank during uraemia.

Expressing the degree of meionexy by the depression in the value
of K, the reader will see from the following table that the uraemic
patients were very meionectic. He must bear in mind that the normal
limits of K x 104 for active persons are 3'6 - 2'1, whilst for middle-
aged patients, who may be regarded as control cases, K x 104 varies
within even narrower limits, 3'4 — 2'6.

Alveolar air

Blood

Case

Age

Dissociation

Saturation

Urea

Lactic acid

C02 mm.

02 mm.

curve

per cent.

K x 104

per cent.

per cent.

I

51

25

119

40

61

1-57

0-36

0-018

II

21

14

132 29-4

44

1-67

0-34

0-018

III

29

24-5

119

27-4

32-4

1-22

0-30

0-023

IV

61

24

121

26-3

34

1-46

0-21

0-030

Normal

— •

35—45

—

30

51—63

3-6—2-1

0-02*

0-015—0-02

* On hospital diet.

Some clinical aspects 283

Schlayer and Straub's paper, in which they described the low
alveolar pressure of CO2 in uraemic cases, bore the title " Is uraemia
an acidosis? " The answer to that at the present time, in so far as it
can be given, is as follows. It is not a lactic acidosis, yet the change
in the value of K in the equation

y Kxn

Too ~= i

is such as would be produced by the addition of acid to the blood,
and as far as is known this change can be produced in no other way.
The addition of urea to the blood does not reduce the value of K.
Uraemia therefore is accompanied by a change in the reaction of the
blood in the acid direction. This change is not due to excess of C02
but to excess of other acids. In this sense it is an acidosis.

Now to turn from uraemia to the class of cases on which I myself
have worked. I came upon them in this way. Dr Thomas Lewis,
Physician to University College Hospital, London, told me of certain
cases of dyspnoea in which the distress was unaccompanied by
cyanosis or other physical signs of a sufficiently grave character
to account for the degree of dyspnoea. In the absence of sufficient
physical signs he inquired if there were chemical signs to which the
dyspnoea could be attributed. After a few preliminary experiments
we determined to make an investigation on the following lines.

Lewis undertook the selection of cases and their treatment ; Ryftel,
the estimation of lactic acid in the blood and of abnormal organic
acids in the urine ; Wolf, the analysis of the blood for urea, ammonia
and " rest " nitrogen ; Cotton, the analysis of the blood for urea
(hypobromite) and the determination of the constant of Ambard* ;
whilst the tests for meionexy and acidosis and the alveolar air deter-
minations were undertaken by me.

The reader will get the most clear account of these cases of
dyspnoea by reading the reports on a typical case. From these he
will see that there is :

(1) Considerable acidosis (in the sense in which I have defined
the word).

(2) Low alveolar C02 and respiratory quotient.

(3) Meionexy.

It is clear then that the dyspnoea is explained : there is no further
mystery about it. The meionexy sufficiently accounts for it.

* See Appendix IV.

-284 Chapter XIX

(4) There is no increase in the amount of lactic acid in the
blood. It is about equivalent to the lactic acid found in control
cases.

(5) There is no increase in other organic acids above those
present in control cases. But after the C02 has been shaken out
there is a change in the balance of acids to bases in the blood in the
acid direction.

(6) The reports on the urea and other forms of nitrogen in the
blood and urine have turned out negative.

The following is a report of a typical case.

CLINICAL REPORT

H. L. (Dr Lewis.)

A maiTied woman of 66 years, admitted to hospital on April 22nd, 1913, com-
plaining of great shortness of breath, vomiting and pains in the sternal region.

History. Several members of the family have died of tuberculosis. There is no
history of past illness. She had one child (now dead) ; there have been no mis-
carriages. The illness began two years before with shortness of breath and this has
increased ; it has been continuous and is increased by exercise. For several weeks
frequent vomiting and pain in the sternal region have been present. She has had
a cough for years and there has been a good deal of expectoration ; attacks of
palpitation and giddiness are common. Her appetite is poor ; she has been wasting
a good deal. Sleep is very disturbed ; she says she wakes repeatedly in the night with
a feeling of suffocation. She has had to get up to micturate four or five times each
night for two years.

Condition on May 5th, 1913. A frail woman (weight 8 st. 4 Ibs.), who sits propped
up in bed with pillows. Modified Cheyne Stokes breathing is present ; the hyper-
pnoeic periods are very long (85 seconds, approximate rate 45 to 60), the breaths
gasping. During the hyperpnoeic period and especially towards the end of it the lips
and tongue are pink or but very slightly cyanosed, during the intermediate periods the
breathing is slower (rate 33), shallow and very irregular (from curves taken 1st May).
During the apnoeic period, the lips and tongue are but very slightly or only moderately
cyanosed. The hands are cold and moderately blue ; the finger tips are clubbed.
There is a slight yellowish tinge of the conjunctivae and skin ; the vessels of the
cheek are injected. Dropsy is present in the legs. The right apex of the lungs is
dull, there are no crepitations. No retinitis.

The heart's limits of dullness are increased (if and 7 inches to right and left of
the mid line) ; the apex beat is in the 7th left interspace in the anterior axillary line.
A systolic murmur is heard at the apex ; the second sound is reduplicated and
intensified at the aortic cartilage. The arteries seem a little thickened, the blood
pressure is 1 50 mm. Hg ; the pulse rate is 90, extra systoles are present and a trace
of alternation follows them.

Some clinical aspects 285

The liver is enlarged, the edge is hard and feels nodular, and there is tenderness
to pressure over it ; there is a suspicion of some ascites.

The urine flow is not free ; sp. gr. 1017, dark, acid with faint smell of (?) acetone.
Albumen and granular casts are present. She has leucorrhoea.

May 5th, 1913. (Dr Cotton's report.)

11-97 xN/36-06

T (duration of observation, minutes) 65 mm.

v (vol. of urine in time T) 15

V (calculated vol. of urine in 24 hours) 332 c.c.

C (concen. of urea in grs. per litre) 36-06 (grs' per htre) '33

D (urea in 24 hours) 11-97

D25 -nr- - = 14-36

K pTg-087*.

May 5th, 1913. (Dr Gloyn's report.)

Red blood cells 3,300,000

Haemoglobin 82 °/0

Colour Index -9 (nearly)

Eed cells showed marked poikelocytosis.

May 5th, 1913. (Dr Ryffel's report.)

Blood : 0'030 grm. lactic acid per 100 c.c.

Excess about O'Olo grm. per 100 c.c.
Urine: day's volume 385 c.c., sp. gr. 1022.
No acetone. Marked excess urobilin. Alb. 0-7 %.
Total N. 1-432 per cent., 5'51 grms. per diem.
Ammonia N. 0-27 grm. per diem.
Amm. N

Total N " /0'

N
Total acidity. 279 c.c. y^ per diem.

Lactic acid 0-0062 per cent., 0-024 grm. per diem.
Large deposit of urates and some uric acid crystals.

May 5th, 1913. (Mr Barcroft's report.)
Alveolar CG-2 ... ... ... 31mm. 31 mm. (abnormally low)

Respiratory quotient (corrected) -63 -53, '61 (abnormally low)

Blood acidosis ... ... ... Saturation at 17 mm. 44 °/0 , abnormally low. Evidence

of considerable acidosis.
Meionexy ... ... ... Gas in tonometer, 02 pressure 30 mm.

,, ,, C02 ,, 29 mm.

Saturation of blood exposed to this gas 38-5 % .
Normal limits of saturation at 30 mm. 51 — 64.
K= -000127.
Very meionectic — attributable to considerable acidosis.

May 23rd, 1913. (Dr Lewis' report.)

More specimens taken. In much the same condition as regards breathing. Getting
weaker. Cyanosis, as before, very slight.

June llth, 1913. Progressively weaker ; evidently dying, but very slowly. C. S.
breathing continues as before.

* See Appendix IV.

286 Chapter XIX

May 23rd, 1913. (Mr Barcroft's report.)

Alveolar air : 02 C0.2 mm. E. Q. correct

14-24 4-05 28-6 -56

Blood acidosis : Saturation at 17 mm. 43 °/0 = -04 lactic.
Meionexy : Saturation at 29'5 mm. 02 and 27 mm. C02 47'5 — 48 °/0 .
K= '00019.
Meionectic.

May 23rd, 1913. (Dr Cotton's report.)

Urea in blood -78 grm. per litre.

Chlorides in blood 6'10 grms. per litre (normal quantity of chloride per litre is 5'62).

May 23rd, 1913. (Dr Gloyn's report.)
Haemoglobin =90%.

With the case of H. L. may be compared that of another patient
E. M. with distressed breathing which is of quite a different type.
Here also there is meionexy to a slight extent but there is no
acidosis, no fall in the C02 pressure in the alveoli, but rather a rise,
and deep cyanosis. The meionexy is due to the rise in C02 which in
turn is due to deficient circulation.

CARDIAC CONTROL

B. M. xiv/18. Age 29. (Dr Lewis' report.)

A married woman of 29 years, admitted to hospital for shortness of breath and
swelling of the legs and stomach.

There is a history of St Vitus dance at the age of 6 years.

Her illness commenced 11 years ago when she began to get short of breath on
exertion, about the same time the legs and abdomen began to swell ; she has suffered
from these symptoms off and on ever since, having been admitted on several occasions
to hospital for exacerbations of the symptoms. On the present occasion the breath-
lessness has been considerable for 14 days and the swellings have been present for
the same period. She has a cough with fairly copious, sometimes haemorrhagic
sputum for several months. For two weeks occasional vomiting has been ex-
perienced.

Condition June 4th, 1913. A thin subject who sits propped up in bed. Cyanosis is
deep, the lips being plum coloured and cheeks dusky and the ears tinged with blue.
The conjunctivae and skin generally are tinged with urobilin. The respirations are
hurried and irregular. The veins of the neck are very distended and pulsate freely.
Slight oedema of the skin is present. Temperature is normal.

The cardiac impulse is very diffuse in the 5th, 6th and 7th spaces. The shock of
the heart-beat is seen and felt over a wide area and there is epigastric thrust. An
early diastolic or full diastolic thrill is palpable at the apex. R. L. C. D., 7 inches,
L. L. C. D., 2 inches, U. L. C. D., 2nd rib. Early diastolic or full diastolic, and systolic
murmurs are heard at the apex. The aortic and pulmonary sounds are normal. The

Some clinical aspects 287

heart's action is grossly irregular. Rates 90 — 100. Auricular fibrillation is present.
The liver and spleen are both enormously enlarged, the former reaching the
umbilicus, is pulsatile, the latter occupies the greater part of the left flank.

A few crepitations are heard at the bases and in the interscapular region and
axilla ; sputum is still present.

The urine is dark of high sp. gr. (1026 — 35), blood was found in it on one
occasion ; it contains albumen, but no casts. Quantity reduced.

June 4th, 1913. (Dr Cotton's report.)

Chlorides in the blood, 6-90 grms. per litre (considerable retention, Normal 5-62).
Urea in the blood -38 grm. per litre.
Bed blood cells : 4,420,000.
Haemoglobin content : 92-5.

June 4th, 1913. (Mr Barcroft's report.)
Alveolar air :

02 C02 mm. K.Q.

14-05 5-5 39 -77

12-51 6-0 43 -66

12-84 5-7 41 -67

Acidosis : Saturation at 17 mm. 71 °/0 = 0 -006— -008% lactic.

Meionexy : Saturation at 31 mm. 02 and 40 mm. C02 49 and 55. ..52 °/0 average.

Value of K, -000207.

Slightly meionectic.

June 4th, 1913. (Dr Wolf's report.)

Total non-protein nitrogen per 100 c.c. of blood 32-4 mg.
Urea ,, ,, ,, ,, 17-2 mg.

Best ,, „ ,, ,, 15-2 mg.

It seems then that these cases of dyspnoea of cardiac or renal
origin may be split into two, and that from the obviously cyanotic
cases there may be detached a definite type of clinical case in which
dyspnoea is a prominent symptom but which differs from the majority
of cases of dyspnoea in the absence of what may be regarded as
equivalent cyanosis. Clinically and pathologically the cases show
cardiac and renal degeneration, though the actual lesions are not
destructive. Amongst the common symptoms are those already
enumerated, namely, dyspnoea without, or with but a slight grade of,
cyanosis, Cheyne-Stokes breathing, restiveness and a relatively high
pulse frequency (both of which are more conspicuous in the evening),
and also thirst. In its fully developed form it rapidly ends fatally.
On the chemical side the symptoms are acidosis (in the sense of an
increase of the ratio of acid to basic radicles), absence of abnormal
nitrogenous metabolism, and meionexy.

To Lewis, the clinician of our party, I freely leave what is his due
—the privilege of providing this particular " asthma " with a suitable

288 Chapter XIX

clinical adjective ; and I will write it down here in the terms in which
I think of it — " Lewis's dyspnoea."

Nevertheless there are certain broad aspects of these cases which
seem to justify some comment. The prevalence of acids in the blood
coupled with the absence of acids which are abnormal in kind suggests
that the condition is due rather to renal than to metabolic disturbance.

Had we found excess of lactic we might have supposed that there
was general oxygen want in the tissues, had we found /8-oxybutyric
we might have tried to link the condition with metabolic disorders,
but the evidence so far as it goes is that the kidneys, instead of
keeping the blood at a certain composition which we call normal keeps
it at another and more acid character which we call abnormal. Such
a change might be wrought by a functional renal disturbance so slight
as to be quite remote from the region of a visible lesion of the kidney.

The analogy between the condition which we have been discussing
and the condition of the body at altitudes of about 10,000 ft. is in-
evitable. Meionexy, fall in alveolar C02 pressure, increase in acid
character of the blood, are the obvious points of resemblance.

In each case the immediate cause of the change in the blood
seems to be renal, beyond this we cannot at present go.

In our study of the cases of Lewis's dyspnoea we compared them
with various other cases which served as controls. In addition to
cases of dyspnoea referable to evident cardiac trouble such as E. M.,
we compared cases of " Lewis's dyspnoea " with ones which had no
dyspnoea but were in hospital for quite other reasons such as con-
valescent appendix or gastric cases. In these we found the value of
K to be singularly constant, much more so than in the case of persons
living an everyday life. This no doubt was due to the fact that these
control cases were all of middle age or rather more and were living
the same sort of life and eating the same sort of diet. The value of K
varied from '00024 to '00034, being thus higher than in the cases of
dyspnoea from whatever cause they arose.

But the most interesting controls were those of the patients
suffering from "Lewis's dyspnoea" whose condition changed con-
siderably whilst they were in our hands.

For instance one case J. P. was seen by me first on February 14, he
was then suffering from the complaint, a month later he was sufficiently
well to be discharged; just before this I saw him on March 13th.
Within a month (April 7) he returned in a very grave condition and
died that evening. The day before his death his blood was tested again.

Some clinical aspects 289

The following changes in the value of K sh,ow the degree of
meionexy which obtained in each case :

K

Normal limits of A' ......... -00024— '00034

-00018

1. J. P. 7 March, a week after admission

"00018

2. ,, 13 March, shortly before discharge ... (1) '00023

3. ,, 7 April, a few hours before death .... •! ' | -00011 "

Another case which showed variations was that of a female
patient, M. P.

She was a mild case on May 23rd when first tested, by June 24th
the dyspnoea had disappeared and the case was discharged.

K

1. 23 May, dyspnoeic ......... (1) -00020

2. 24 June, breathing normal ...... (2) -000274

Not only had the value of K become normal but the alveolar C02
rose from 33 mm. to the normal value of 39 mm. and the blood in the
absence of C02 on May 23rd was 61 % saturated with oxygen at
17 mm. oxygen pressure, whilst on June 24th the percentage satu-
ration at 17 mm. had risen to the normal figure of 74 °/0; the former
corresponded to normal blood with '01 % °f lactic acid added.

One more point : in the account of the case of J. P., given above,
when tested just before death the values of K given do not appear
to be very concordant. The interpretation of this irregularity, which
appears in very meionectic curves, e.g. Figs. 30 and 134, where the
points observed are below the curve at low pressures and above
it at high ones is, in the light of more recent* work, that blood has
gained sufficient in acidity to cause a measurable change in the value
of n, the average number of molecules of haemoglobin clumped
together. The effect is of course in the direction of increased clumping
of haemoglobin and, as in other cases of the same phenomenon, its
significance lies in the fact that a similar clumping of other protein
molecules probably takes place and alters the properties of every cell
in the body by a reduction of the velocity of its molecular vibrations.

REFERENCES.

(1) Lewis, Ryffel, Wolf, Cotton and Barcroft, Heart, v, p. 45, 1913; Journal of

Physiol. XLVI, p. liii.

(2) Poulton and Ryffel, Physiol. Proc. ; Journal of Physiol. XLVI, p. xlviii.

* See Appendix II.
B. R. P. 19

APPENDIX I. ON METHODS

ESTIMATION OF THE OXYGEN CAPACITY OF BLOOD BY THE
FERRICYANIDE METHOD OF ESTIMATING OXYGEN IN BLOOD

THE simplest measurement which it is possible to make of the oxygen in blood is
that of the total oxygen capacity of the haemoglobin. For this measurement it is
necessary to be careful about three things quite apart from apparatus.

(1) The blood must be fresh so that it does not reduce itself as the result of
bacterial action. The length of time which may elapse from the time at which the
blood is drawn till the estimation is made depends upon the cleanliness with which
the fluid is collected and the temperature at which it is kept. In ordinary weather
blood taken by pricking the finger should be estimated within 24 hours. Blood may
be kept for some two or three days by placing it in a stoppered tube which is kept
in ice in a Dewar's flask. It cannot be kept indefinitely even at 0° C.

(2) The blood must be thoroughly shaken with air in order that it may be
fully saturated with oxygen. At room temperature the degree of saturation in air
is not measurably less than 100 %•

(3) The blood must be shaken just before it is used in order that no sedimenta-
tion may take place.

The theory of the ferri cyanide method is obscure, the blood in faintly alkaline
solution gives off accurately the quantity of oxygen which could be abstracted from
the haemoglobin with a blood gas pump. Nevertheless the substance formed by the
reaction is methaemoglobin, a body which is credited with containing the same
quantity of oxygen as oxyhaemoglobin, though in a stable form.

The following equation is given by Haldane for the reaction. The alkali in the
formula is sodium bicarbonate. The general nature of the formula is the same no
doubt if ammonia or other alkali be substituted.

| + 4Na3(FcCv6)+4NaHC03 = O, + Hbf + 4Na4(FcCy6
X) V)

Haldane's demonstration of the accuracy of the method was made with an
ordinary Dupre's apparatus for the estimation of urea. 50 c.c. of blood and 100 c.c.
of dilute alkali were mixed in the bottle of the apparatus ; into the small tube in the
interior of the bottle were placed 20 c.c. of a saturated solution of ferricyanide of
potassium. The blood and the alkali were thoroughly mixed in order to allow of
complete laking of the corpuscles. The rest of the operation was conducted just

Appendix I. On methods

291

like an estimation of urea in urine with sodium hypobromite. In theory at all events
this was so, in practice very much greater precautions had to be taken ( 1 ) to maintain
the temperature of the great mass of colloidal fluids constant, and (2) to get off all
the oxygen by an immense amount of shaking. These two processes were of course
wholly inimical to one another. Nevertheless the method sufficed in the hands of
Haldane to establish the theoretical accuracy of the ferricyanide reaction.

The differential method of Blood Gas analysis.

The apparatus consists of a manometer, with a blood gas bottle at the head of
each limit. The apparatus is shown in Fig. 136.

FIG. 136.

The bottles are similar to those of the apparatus previously described.

No control apparatus is necessary, the principle of the differential apparatus being
that changes in the temperature of the bath, &c. affect the two bottles alike and
therefore counterbalance one another.

The tubing of the manometer should be 1 mm. bore approximately ; the bottles
each about 25 c.c. in capacity. They should be of the same size within 1 tenth
of a c.c.

To Jill the apparatus with clove oil.

This fluid is in many ways ideal, it has however one drawback, namely that the
apparatus must be chemically dry before the oil is put in. It must also be chemically
clean. It is best, I think, to avoid alcohol and ether in the cleaning. I clean the
tubes by filling them with potassium bichromate and sulphuric acid and then plunging
them in a bath of the same, which I can warm up. The bichromate is then thoroughly
washed out with distilled water and the apparatus is dried by aspirating air (which
has passed through sulphuric acid) gently through the manometer in a warm air bath.

19—2

292 Appendix I. On methods

The clove oil is put in the following way. Starting with clean apparatus, the
taps of which are free from grease, a fine pipette is
drawn out which will go down the broad portion of the
tubing in the vicinity of the tap, the stopper of the tap
being removed. The stopper on the other side is turned
so that no air can get out of the manometer on that
side. The pipette is charged with clove oil, of which
some slight excess is put into the broad portion of the
tubing, the pipette is then removed, the tap on the
other side is opened gently and sufficiently to allow of
the oil descending the fine tubing and getting just
round the bend of the manometer ; the tap is then
closed again.

All excess of oil is now removed from the broad FIG_ 137.

portion of the tubing with a pipe cleaner. It is necessary

that this should be done very thoroughly, as any oil which remains here is apt to
creep down and enclose a column of air in the manometer at a later stage. When all
excess of oil has been removed the tap may be opened and the oil allowed to find its
own level.

Should bubbles form in the apparatus they may be expelled by forcing the fluid
up to the top on that side. This may be done by putting the bottle on the opposite
side and with the tap suitably turned, warming the bottle. When the fluid has been
forced up, or rather as it is being forced up, it may be necessary to break the bubble
by making the fluid meet the end of a pipe cleaner which has been pushed down as
far as it will go.

In no case should one blow into the manometer. This will inevitably cause
a dampness that will make the oil form beads and prevent it rising and falling
accurately

To calibrate the Differential Blood Gas apparatus*.

The calibration consists in a determination of the relation between the quantity
of gas x evolved in the apparatus and the difference of pressure p, measured in the
manometer, x=kp. The method depends upon the liberation of a known quantity
of gas inside the apparatus ; x then is known and p is observed, therefore if k is the
constant of the apparatus it is directly determined

*-*.

P

The reaction which has proved most serviceable so far has been the liberation of
oxygen from hydrogen peroxide in the presence of acid by the addition of potassium
permanganate. The equation is as follows :

Therefore 316 grams of potassium permanganate yield five molecular weights or
111,000 c.c. of oxygen, at normal temperature and pressure. A solution of hydrogen
peroxide must be made up and freshly titrated with approximately decinormal
potassium permanganate, of such a strength that 1 c.c. of the H202 gives off 0'2 c.c.

* See also Hofmann's method, Journal of Physiol. XLVII. p. 272.

Appendix I. On methods 293

of oxygen. It must of course when made be accurately titrated and the exact

quantity of oxygen which it will give off determined.

The following example may serve to help the experimenter :

Potass, permanganate solution used, strength being 1 c.c. = '00295 gr.

Pure 20- volume hydrogen peroxide diluted 100 times, then 100 c.c. H2O2 required

10'3 c.c. KMn04 to set the oxygen free,

From this equation it is seen that 316 grs. KMnO4 liberate 160 grs. oxygen.
But '0896 x 16 grs. oxygen occupy 1000 c.c. at N.T.P.

160 x 1000

Since 100 c.c. hydrogen peroxide contain oxygen liberated by 10'3 c.c. KMn04,

1 "5x1 0-3

*

1 O C.C.

,, „ „ „ ,, „

1-5 x 10-3 x -00295
which contain - - grs. KMn04.

„,, 160x1000

Then 316 grs. KMn04 liberate - c.c. 02.

' x lo

1-5 xl 0-3 x -00295 l'5x 10'3 X "00295 X 160 x 1000

100 100X '0896x16x316

= •1537 c.c. 02.

Temp, was 13'4° C., pressure 755 mm.

Now 1000 c.c. air, saturated with water vapour at this temp, and press., occupy

932 c.c. at N.T.P.

153'7
.'. vol. of O2 actually liberated was cub. mm.

Difference of press, observed was 46'2 mm.

But vol. in bottle was 3'5 c.c. which would add "35 to the constant

.'. Constant for empty bottle = 3'92.

Perhaps the greatest drawback in this method is due to the difficulty of obtain
ing pure hydrogen peroxide. The commercial reagent is made up with glycerine or
other substances which are added for the purpose of preventing the H202 from
deteriorating in strength. Such peroxide appears to give unreliable results.

The calibration of the apparatus is carried out as follows :

Into each of the bottles put 1 c.c. of H2O2 and 2 c.c. of H2S04 strength N/100.

If the left bottle is the one which is to have its constant determined, place "2 c.c.
of permanganate in the reservoir and the same quantity of water in the correspond-
ing reservoir of the other side. After the permanganate is in its place cut a piece
of filter paper 2 or 3 cm. in length and 1 mm. in breadth and place it so that it projects
a little from the reservoir but does not become wet with the permanganate, till it is

* This was volume of H202 put in blood gas bottle.

294 Appendix I. On methods

desired to spill the permanganate from the reservoir into the bottle. It will ensure
the proper spilling of the permanganate. Put on the bottles, the stoppers being
efficiently greased with lard or vaseline. Place the apparatus with its bottles in the
water bath.

After five mimites read the meniscus on each side :

Left Eight

120 119-5

Close the taps. See that the meniscus remains constant for two minutes, then
remove the apparatus from the bath ; upset the fluid into both bottles. Shake for
one minute. Replace the apparatus in the bath. After a minute has elapsed shake
again in the air for a minute, put back in the bath and after 60 seconds more read

Left Eight

90 159-5

Difference 69'5 mm. +'.3 = 70 mm. =p.

Now by calculation 1 c.c. of H202 gives '2 c.c. of O2 at 0° C. and 760.

."} o o ^ ("• f\

.'. it gives "2x r— x — - at 15° C. and 755= 212 c. or 212 c. mm

27 o / »)o

.'. &=~ = 3-03.
70

Comparison of this result with another which depends upon the actual measure-
ments of the physical constants (i.e. the volumes of the bottles and the bore of the
tubing) of the apparatus, shows a slightly higher constant by this method. As
a matter of fact very careful research carried out by Burn has shown that this method
gives results consistently higher by about 2 % than those given by a method which
depends upon the measurement of the sizes of the bottle and tubing. This is scarcely
surprising. The old method is of course theoretically accurate ; but for certain
small quantities which are negligible and have been neglected in the calculation,
the method was mathematically correct. Nevertheless in practice one was dealing
with fluids which leave films on whatever they touch, which exert vapour pressures
and so forth, and it is not surprising that a theoretical method when checked by
a practical one should depart from it to the extent of 2 °/0.

To determine the total oxygen capacity of a sample of
dejibrinated blood by the differential method.

Apparatus : a differential blood gas apparatus, a burette, 1 c.c. pipette (calibrated),
1 finely drawn out pipette, potassium ferricyanide, ammonia solution (4 c.c. of strong
ammonia per litre), vaseline.

In each blood gas bottle put 2 c.c. of ammonia from the biirette and 1 c.c. of
defibrinated blood. Shake up the two together so that the blood becomes thoroughly
laked.

Grease the stoppers with vaseline.

See that the taps are open.

With the fine pipette put 0'2 c.c. ferricyanide solution into the reservoir on one
side.

Appendix /. On methods 295

Put the apparatus in the water bath.

After five minutes read the meniscus on each side

Left Right Zero error iu p

120 119-5 -5

Close the taps ; after two minutes if the meniscus has not moved upset the ferri-
cyanide on one side with the precautions indicated above ; shake for one minute and
place in the bath for one minute ; read again

Left Eight p

93 146-5 53 -5 +-5 = 54

Shake again for one minute and allow to stand for one minute ; read

Left Right p

91 148-5 57-5+ -5 = 58

Shake again for one minute and allow to stand for one minute

Left Right p

91 148-5 57-5+ -5 = 58

The value of p is now constant, had it not been so the operation of shaking, &c.,
must have been continued.

The volume of gas given off is p+A; = 57x3'03 = 171 cubic mm. or '171 c.c.

This must be reduced for temperature and pressure and corrected according to
the calibration of the 1 c.c. pipette. Suppose the latter delivers "96 c.c., that the
temperature is 15°C. and the pressure 755 mm. The oxygen capacity is then

Repeat the operation on the other side of the apparatus.

Correction of Haldane's standard haemoglobinometer.

The importance of the above determination is great, owing to the fact that it is
the only practical way of ascertaining the correctness of a standard haemoglobino-
meter.

Haldane's standard instrument is constructed so that 100 on the scale corresponds
to an oxygen capacity of '185 c.c. of oxygen per c.c. of blood. The reading of any
blood on Haldane's scale should therefore be

The oxygen capacity as determined

^185"
That given above should be

If it is not so, as the result of careful determination the standard solution in the
haemoglobinometer is incorrect.

296 Appendix I. On methods

Determination of the oxygen in unsaturated blood with the

differential apparatus.

Apparatus as above, with the unsaturated blood in addition.

Place 2 c.c. of ammonia in each bottle, put 1 c.c. of the unsaturated blood in the
left bottle and 1 c.c. of saturated blood in the right bottle. The blood must in each
case be run from the pipette gently to the bottom of the bottle so that it lies in
a layer underneath the ammonia and is protected by it from oxidation by the air.

Grease the stoppers, place the bottles on the stoppers, place potassium ferri-
cyanide in the left reservoir. With the tap open place the apparatus in the water
bath ; after five minutes read the meniscus. Suppose it to be

Left Eight

120 119-5

Shut the taps, see that the meniscus does not move in the next two minutes.
Then shake thoroughly so as to lake both samples of blood. The level of the clove
oil will have changed, as the unsaturated blood will have taken up some oxygen.
This change may be disregarded, however, as this extra oxygen combined with the
haemoglobin will be turned out again by the ferricyanide producing a corresponding
change in the opposite direction. When the blood is completely laked turn the
bottle so as to let down the ferricyanide, shake as in the above example. Let the

final reading be

Left Eight p

110 129-5 19-5+ -5 = 20 mm.

The volume of gas as measured therefore is pxk: 20 x 3'03 = 60-6 cubic mm. or
0'06 c.c. This must be corrected in several ways before a correct result is obtained.
The considerations which must be borne in mind are as follows: (1) temperature,
(2) pressure, (3) the calibration of the pipette, (4) the fact that the plasma is not
saturated with oxygen, (5) the temperature at which the blood was in equilibrium
with oxygen and nitrogen. Taking the first three together and using the values
given in the determination of the total oxygen capacity the corrected reading would
be 0'059 c.c.

As unsaturated blood is usually taken directly from the body or from a tonometer
at body temperature we will suppose that it lias been exposed to gases at 37° C.

In the lung the partial pressure of nitrogen is about 560 mm., in the air in the
blood gas bottle about 590. The question then is how much nitrogen will blood
which has been exposed to 560 mm. nitrogen at 37° C. take up when exposed to
590 mm. at say 15° C.

Exposed to 760 mm. N2 pressure at 15° C. blood take&up 0-016 c.c.

760 ,, ,, 37° C. ,, 0-011 c.c.

„ 590 ,, „ 15° C. „ 0-012 c.c.

therefore ,, 560 ,, ,, 37° C. ,, 0-008 c.c.

The blood in the apparatus will therefore take up nitrogen from the air to the
extent of 0'012-0'OOS c.c. = 0'004 c.c. This must decrease the amount of oxygen
which appears to have been given out by this amount, therefore 0'004 c.c. must be
added to the answer given above, namely 0'059 c.c. ; it therefore becomes 0'063 c.c.
Blood which is unsaturated at room temperature may be regarded as having no

Appendix I, On methods 297

oxygen in the plasma. Such blood therefore when shaken with air at 1 5° C. will
take up '006 c.c. of oxygen per c.c. This also must be added to the reading obtained
above, which will bring it up to (0-063 + 0'006)=0'069 c.c. This therefore is the
corrected oxygen reading, and represents the oxygen in the blood from this animal,
which when cold will be practically entirely in the haemoglobin, when warm it will
have been to a trifling extent in the plasma.

Determination of Percentage Saturation of Blood with oxygen
by means of tJie differential method.

The percentage saturation is of course the quantity of oxygen which the blood
contains A, divided by the total oxygen capacity (7, multiplied by 100, i.e.

A

100X ^.

0

We have already described methods for the estimation of A and C ; but as they
cannot both be performed upon the same sample of blood it is convenient to make
use of the following very simple device.

If the blood in question be simply shaken with oxygen it will take up oxygen
until it becomes saturated. If the quantity which it originally held was A, and that
which it takes up on shaking was B, then

B = C-A or A = C-B.

The percentage saturation then is

We proceed to measure C and B. Of these two measurements B is made first.

Measurement of B.

Place 2 c.c. of ammonia in each bottle, underneath the ammonia run 1 c.c. of the
blood to be estimated into the bottle on the left hand side (L) and 1 c.c. of saturated
blood (not necessarily from the same animal) in the right hand bottle (R).

Grease the stoppers.

It is most important that the stoppers should be thoroughly clean inside and free
from all traces of ferricyanide.

See that the taps are open.

Place the bottles on the stoppers and the apparatus in the bath.

After five minutes read. Suppose the reading to be

Left Bight

120 119-5

Close the taps ; see that the surfaces do not move for two mimites, then shake for
one minute, or until the unsaturated blood ceases to become redder. Replace in the
bath, after one minute read, shake again for one minute. Read and shake for
alternate minutes until the reading becomes constant. Suppose it then is

Left Eight p

129-5 110 19-5-0-5 = 19 mm.

B then = 19.

298 Appendix I. On methods

Measurement of C.

Take the apparatus out of the bath, open the taps. Take the bottle L. Put
potassium ferricyanide into the reservoir, replace the bottle, put the apparatus in the
bath, and proceed as in the determination of the oxygen capacity.

Let the measurement of C give a pressure of 60 mm.

The percentage saturation = 100 i.e. 68 °/0 .

This determination gains considerably in accuracy owing to the fact that neither
the quantity of blood used nor the constant of the apparatus enters into the calcula-
tion. There is therefore no need to calibrate the pipette used or even to exercise
extreme caution in measuring the amount of blood used for the experiment.

It is of course unnecessary to correct for temperature and pressure, the correction
affecting B and C alike. On the other hand one must reckon with the fact that the
unsaturated blood will take up nitrogen and oxygen from the air of the apparatus in
which the blood is shaken in quantities which depend upon the composition and
temperature of the gas with which the blood has previously been in equilibrium.

The general principles upon which the calculation for this correction is made are
set out under the measurement of oxygen in unsaturated blood. For blood taken
from a tonometer at 13 % or fr°m tne body it is generally sufficiently accurate to add
4 °/0 to the observed saturation.

Determination of the difference between the quantities of oxygen
contained in 1 c.c. of arterial and 1 c.c. of venous blood tcith
the differential manometer.

If, as is usually the case, the arterial blood is nearly saturated with oxygen,
relatively to the venous blood, this determination is one of the easiest in blood gas
analysis.

Place 2 c.c. of ammonia in each of the bottles, in L place 1 c.c. of venous blood, in
R 1 c.c. of arterial blood.

See that the stoppers are greased and that they are free from all traces of ferri-
cyanide and that the taps are open.

Place the bottles on the apparatus, and the apparatus with its bottles in the
water bath.

After five minutes read. Suppose the readings to be

Left Eight Zero error

120 119-5 0-5

Close the taps ; see that the meniscus does not move for two minutes.
Shake for one minute and place in the bath for one minute alternately till a
constant reading is obtained. Let this be

Left Eight Diff. Zero error p

127 112-5 14-5 -5 14

If k be the constant of the apparatus, p x k is the difference between the oxygen
in the arterial and venous blood.

A few words may be said about the limitations of this method.

Appendix I. On methods 299

In the first place it measures the difference of the total oxygen in each sample of
blood and not merely the oxygen in the haemoglobin.

Secondly it assumes that the two samples of blood are of the total oxygen
capacity. The reasoning which underlies the method is as follows. Let the oxygen
in the arterial blood be AA and that in the venous blood Ay. It is required to
measure AA — Av.

Let the amount of oxygen necessary to saturate the arterial blood be BA, and the
venous blood Br, and the total oxygen capacity C.

A,= C-BA,
A,.=-- C-B,.,

AA-Ar= C-BA -
= B,.-BA.

The question then arises, Is the method applicable to cases where the oxygen
capacity of the arterial and venous bloods is different, such for instance as in the
case of blood flowing through an active gland ?

There is one special case in which the method may be applied ; fortunately it is
the most common, but care must be taken to see that it obtains. It is the case in
which AA and G are almost equal, that is to say in which so little oxygen is taken
up on the arterial side of the apparatus that a small error in its quantity may be
neglected as compared with the difference between AA and Ar.

Thirdly we must consider the accuracy of measurement of the blood. The guiding-
principle is the same. So long as BA is almost nothing it is not material that the
artei'ial blood should be very accurately measured. The practical point that has to
be considered is the accuracy of the measurement of the venous blood.

Moreover if any considerable quantity of oxygen is taken up by the arterial blood,
one has to allow for any difference which may exist between the constants of the
two sides of the apparatus.

For all these reasons it is best when working with arterial blood which is un-
saturated to place the arterial and venous bloods each in a different apparatus,
and analyse each separately with saturated defibrinated blood in the other bottle
of the apparatus, and, having obtained the value of either A or B for each blood, to
subtract the one from the other by arithmetic.

A useful check on the measurement of the blood is a determination of C for each
sample of blood after B has been determined,

These complications, as Verzar found, make the estimation of the oxygen used by
organs during oxygen want very difficult ; for the reasons given above he made
separate estimations of arterial and venous bloods.

Differential method with apparatus for 0*1 c.c. of blood.

There is no difference in theory between this apparatus and that which we have
just described. The smaller apparatus was designed for work on human blood.
With it all the work described in this book on the human dissociation curves has
been performed, except where otherwise stated.

300

Appendix I. On methods

The apparatus consists of a manometer of 0'5 mm. bore glass tubing. At the top
of each limb of the manometer is a three-way tap with
a T boring. Of the three ways of each tap one opens to
the open air, the other goes to the blood gas bottle. The
blood gas bottles are of the shape indicated in the figure ;
they are each about 3 c.c. in capacity up to the top, and
should of course be identical in size. The apparatus is
fixed to a wooden stand, fitted to a clip by which it hangs
on the edge of the water bath.

There are certain practical details about which the buyer
of one of these pieces of apparatus should satisfy himself :

(1) That the tubing of the manometer is not more
than O'o c.c. bore.

(2) That the point at which the fine tubing of the
manometer joins the coarser tubing of the tap should be
above, not below the bend of the tubing.

(3) That when the manometer hangs in the water bath

with the tubing vertical the bottles should be completely submerged in the water.

(4) That the clip be hung in such a position that the apparatus can stand in
the bath with the bottles vertical or nearly so.

FIG. 138.

FIG. 139. — Correct position of junction
between fine and coarse tubing.

FIG. 140. — Incorrect do.

FIG. 141.

FIG. 142.

Appendix I. On methods 301

(5) That the shape of the bottles be such that when hanging on the apparatus
with the manometer vertical 0'02 c.c. fluid placed in the pouch at -X" has no tendency
to find its way down the tubing to Y.

(6) That the diameter A — A of the bottle be such that if 0'2 c.c. of ammonia
and O'l c.c. of blood be placed at Y the fluid will assume the position shown in
Fig. 142 when the bottle is placed in that position. If A — A is too narrow the
fluid sticks in the bottom part of the bottle even when the latter is placed on
its back.

Calibration of the differential apparatus for analysis of
oxygen in 0*1 c.c. of blood.

Two methods are available for this calibration.

(i) A direct comparison of the oxygen given out by blood with that given out by
the same blood in the larger form of apparatus, the constant of which is known.

(ii) The hydrogen peroxide method. The strengths of solutions necessary and
their standardisation are carried out in the small as in the case of the larger apparatus
already described. The only point which needs amplification is the use of the solu-
tions in the small apparatus. This is very simple. Into each bottle is placed 0'2 c.c.
of N/100 sulphuric acid and O'l c.c. of standard hydrogen peroxide.

FIG. 143.

These should be run into the bottles from a pipette with a long narrow end, so
that the sides of the bottle are not wet. For putting the potassium permanganate
into the bottle a special pipette is necessary. It is drawn out and bent at the end.
As shown in the figure it may be used to place the permanganate in the pouch.

The danger, of course, is that the H2O2 and the permanganate may become
prematurely mixed. If the bottle is of the proper shape and reasonable care is taken,
this danger should not cause serious preoccupation. The other bottle is filled in the
same way except that water, not permanganate — a drop in each case — is placed in
the pouch.

The bottles are then placed on the apparatus which has, of course, been adequately
greased. Care must be taken to see that the taps are open when the bottles are put on.

The apparatus is placed in the bath and allowed to stand for five minutes the
meniscus is read, the taps are closed, if in another minute the meniscus has not changed
the bottles may be shaken. Care being taken that both the permanganate and the
water become mixed with the acid hydrogen peroxide.

The bottles are then shaken and left in the bath for alternate minutes until
constant readings are obtained.

The method of working out the constant is similar to that already given for the
larger apparatus. Suppose k to be '38.

30*2 Appendix /. On methods

To determine the total oxygen capacity of a sample of deftbrinated
blood by the differential method, using O'l c.c. for each determi-
nation.

Apparatus: a differential blood gas apparatus of the form shown in Fig. 138;
a burette holding about 1 — 5 c.c., which delivers ^ c.c. with fair accuracy ; a
pipette which delivers ^Q c.c. of blood ; a curved pipette similar to that shown in
Fig. 143 ; a saturated solution of potassium ferricyanide ; ammonia (4 c.c. of strong
ammonia per litre) ; vaseline.

Into each blood gas bottle put 0'2 c.c. of ammonia from the burette and O'l c.c. of
defibrinated blood. Shake up the two together so that the blood becomes thoroughly
laked.

Grease the stoppers with vaseline, see that the taps are open.

Put one drop of ferricyanide into the pouch of the left-hand bottle, place the
right-hand bottle with the pouch uppermost. Place the
apparatus in the water bath. After five minutes read
the meniscus on each side,

Left Right Zero error
120 119-5 -5

After one minute, if the meniscus has not changed,
shake up the blood and ferricyanide, being careful that
the pouch on the right side does not become wet or
soiled. Shake and place in the bath for alternate
minutes.

Final reading :

Left Eight p FlG 144

91 148-5 57-5 + -5- 58

The volume of gas given off is p x k = 58 x '38 or 22'0 cubic mm.

This must be reduced for temperature and pressure and corrected according to
the calibration of the pipette. Suppose the pipette delivers 0'096 c.c., that the
temperature is 15° and the barometer is 755 mm., the oxygen capacity is

273 755 0-1

Open the taps, take the bottle off the right side, the pouch of which is as yet
uncontaminated. Put a drop of ferricyanide in it, replace it on the apparatus, and
determine the oxygen capacity of the blood in the right bottle. Two concordant
results should thus be obtained.

Correction of Haldanes standard haemoglobinometer.

See page 295.

Determination of the oxygen in O'l c.c. of unsaturated blood
with the differential apparatus.

Apparatus as above, with the unsaturated blood in addition.
Place 0'2c.c. of ammonia in each bottle, put O'l c.c. of the unsaturated blood in the
left bottle and O'l c.c. of saturated blood in the right bottle, the blood must in each

Appendix I. On methods 303

case be run from the pipette gently to the bottom of the bottle so that it lies in a layer
underneath the ammonia and is protected by it from oxidation by the air.

Grease the stoppers, put a drop of ferricyanide in the pouch of the left bottle. See
that the taps are open, place the bottles on the apparatus and the apparatus in the
water bath. After five minutes read the meniscus on each side. Suppose it to be

Left Right Zero error

120 119-5 -5

Shut the taps, see that the meniscus does-not move in the next minute. Shake in
such a way that the blood becomes thoroughly laked but that the ferricyanide and
the blood do not mix. When satisfied that the laking is complete, mix the ferricyanide
and the laked blood and shake for one minute. Place in the bath and shake for
alternate minutes. Suppose the final reading to be

Left Bight

110 129-5 19-5+ -5 = 20 mrn.

If k is '38 the volume of gas which has come off is

p x k = 7 '6 cubic mm.
This must be corrected in accordance with the instructions on p. 296.

Determination of the percentage saturation of O'l c.c. of blood
with oxygen by means of the differential method.

The percentage saturation is, of course, the quantity of oxygen which the blood
contains (^4) divided by the total oxygen capacity (C) multiplied by 100, i.e.

100 x^.

L>

We have already described methods for the estimation of A and C ; but as they
cannot both be performed upon the same sample of blood it is convenient to make
use of the following very simple device.

If the blood in question be shaken with oxygen it will take up oxygen till it
becomes saturated. If the quantity which it originally held was A and that which it
takes up on shaking was B, then

B=C-A or A = C-B.
The percentage saturation then is

We proceed to measure C and /?, of these two measurements B is made first.

Measurement of B.

Place 0'2 c.c. of ammonia in each bottle, underneath the ammonia run O'l c.c.
of the blood to be estimated into the bottle on the left-hand side, and O'l c.c. of
saturated blood (not necessarily from the same individual) in the right-hand bottle.

Grease the stoppers.

It is important that the ends of the stoppers should be thoroughly clean inside
and free from all traces of ferricyanide.

304 Appendix I. On methods

See that the taps are open.

Place the bottles on the stoppers with the pouches upwards. Place the apparatus
in the bath.

After five minutes read, suppose the reading to be

Left Right Error

120 119-5 -5

Close the taps and see that the surfaces do not move in one minute, then shake
for one minute at least or until the unsaturated blood ceases to change colour ;
replace in the water bath and shake alternately for a minute until a constant reading
is obtained. Suppose this to be

Left Eight p

110 129-5 19-5-0-5 = 19 mm.

Measurement of C.

Take the apparatus out of the bath, open the taps. Take off the left bottle, the
pouch of which should be clean and dry — place a drop of ferricyanide in the pouch
and replace the bottle on the apparatus with the pouch downwards. Proceed to
measure the oxygen capacity.

Let the measurement of C give a pressure of 60 mm.

_ JQ\

The percentage saturation = 100 — — = 68 °/0.

The corrections which must be applied to this will be found on p. 298.

To expose blood to a gas of a given composition.

Apparatus :

A useful form of tonometer for the purpose consists of a cylindrical vessel of
250 c.c. capacity.

FIG. 145.

At one end the cylinder can be corked with a rubber stopper, at the other there
is a three-way tap. The bore of this tap is To mm. At the point A the tubing is
somewhat broader, so that there is a portion of the apparatus which will hold about
1 — 2 c.c. of blood when the apparatus is placed vertically. That portion of the
tubing between B and C requires to be made carefully if, as is sometimes the case,
the apparatus is expected to deliver a definite amount of blood into one of the bottles
of the small apparatus just described.

Appendix I. On methods

305

5cm

The following are the requirements which should be fulfilled for this purpose :

(1) The point of the tube must reach easily to the bottom
of the blood gas bottle.

(2) The tubing between B and C should hold about 0'12 c.c.
of blood.

(3) At no place must the bore be so great that bubbles of
gas become mixed with the blood as the blood passes down.

(4) At no place must it be so fine that a froth of mercury
and blood which sometimes has to be dealt with, will stick in
the tube.

(5) The stronger it is the better.

The tubing therefore should be 7 cm. in length, B to C,
2 cm. from B to Z>, and 5 cm. D to C. The bore should be
uniformly To mm. throughout the whole length of the tube.

The outside diameter of the fine part of the tube should
be 4'5 mm.

For many purposes the tonometer is neither required to
deliver a given quantity of blood nor to deliver it into a
narrow mouthed bottle. A tube 7 cm. long and 1'5 cm. bore
will do, without any constriction. FIG. 146.

Laboratory method of fillmg tonometers from stock

gas mixtures.

The following is a description of the plant which is in the Cambridge Laboratory
for this purpose.

For each gas mixture two Mariotte
bottles of a capacity of 15 litres each are
connected. It is better not to get the
bottles with ground glass bungs, these are
liable to break and the bore of the taps
is much too narrow to be serviceable.
Rubber bungs are quite satisfactory. They
should be bored and glass taps inserted
which have a bore of 1 cm. in the stopper.

Into the top of bottle B should be
securely fixed with wire a good rubber
cork bored to take the glass tube Y which
goes to the bottom of the bottle. The
cork in the bottle B has another hole ;
through this a short glass tube passes
which is connected by rubber to a brass
tap. This tap connects at right angles to
a brass tube W (see endwise in Fig. 147)
which we shall call the vacuum main. In
the installation at Cambridge there are
five such pairs of bottles, each forming a
gas holder, joined to the vacuum main.

B. R. F. 20

FIG. 147.

306

Appendix I. On methods

At one end the main connects with a glass tap (Fig. 148) A, and beyond this
there is a Geryk pump. At the other end there is a T-tube C, a vacuum gauge Z>,
and a glass tap.

A B

TG TT

FIG. 148.

To fill a tonometer with gas from one of the gas holders — say the one attached to
the tap G. Place the tonometer, which is of course securely corked, on the T-tube (.',
open the tap /' of the tonometer, close all the other taps except A. Turn the Geryk
pump until the gauge shows a good vacuum. Then close A and gently open G. If
G be opened suddenly the gauge may be broken. Close G, open A exhaust once
more, close A, open G and after a minute or so close G again. Close /'and take off
the tonometer.

The holders may be filled with nitrogen through the tap at the bottom of bottle
B, Fig. 146, from a nitrogen cylinder. This is done with the bottle full of water and
just enough water in A to cover the tap at the bottom. The water is then displaced
upwards into the bottle A by the nitrogen — the nitrogen must not be run in too
rapidly or some part of the apparatus may blow out.

,B

' Fio. 149.

To obtain a given oxygen pressure add 20 c.c. of oxygen for each mm. required.
The oxygen may be transferred from an oxygen cylinder — in an apparatus consisting
of a 250 c.c. graduated measuring cylinder— fitted with a rubber cork and with glass
tubing after the manner of a wash bottle. This is attached to a wash bottle of about
a litre capacity, the attachment being as shown in the figure. Start with the whole
measuring cylinder and tubing connected with it full of water, and the wash bottle
partly full. Place the glass tube A on a piece of rubber tubing along which oxygen

Appendix I. On methods

307

is gently coming from the cylinder, remove the clamps at B and C. Let the oxygen
displace the water backwards into the wash bottle till the required amount has been
obtained. Close the clips and disconnect.

Connect A by rubber tubing to the tap Z, Fig. 147. See that the tap X is
closed and that there is no considerable positive pressure in the gas holder by
opening \V for a moment. Then with W closed and Z open, transfer the oxygen by
blowing into the wash bottle.

It is now necessary to analyse the gas, by filling a tonometer in the manner
described above and analysing the contents with a Haldane's gas analysis apparatus.
By the addition of a little air or nitrogen in calculated quantities it is possible to get
the gas in the holder pretty exactly to a known composition. It is best however to
have the gas with a trifle less oxygen than what is actually likely to be required.
For instance if 20 mm. oxygen is being aimed at one would have the gas in the
gasometer 19'5 mm. oxygen. One has to analyse the gas in the tonometer before
using it ; having made the analysis it is very easy to add a small quantity of air from
the gas burette of the Haldane's apparatus. With a 250 c.c. tonometer air is added
to the extent of 0'15 c.c. for every tenth of a mm. oxygen pressure required ; of this
we shall speak later.

We now have a mixture of oxygen and nitrogen in the proportions required ; we
may want a known pressure of carbonic acid.

It is not possible to keep carbonic acid in the stock mixture on account of its
solubility in the water. It is best to add it from a burette. For this purpose I use
a burette of the form shown in figure 150. It is of 15 c.c.
capacity and is graduated in c.c. and Yotlis of a c.c., the gas is
manipulated over mercury. The tonometer is fixed at A by a
junction of rubber pressure tubing. Mercury is run up to the
tap of the tonometer, which is shut. The B tap is also shut.
The mercury reservoir is now placed so that the surface of the
mercury is at the level C. The tap B is opened so as to connect
C with the interior. Some excess of C02 is forced in by placing
the nozzle of a tube connected with the CO2 cylinder and from
which C02 is gently coming, in the rubber tube at C. Close the
tap. Take away the tube by which the carbonic acid entered.
Level the mercury within and without the burette with the tap
shut. Then open the tap towards 6', gently raise the mercury
reservoir and expel C02 to the air until the burette contains the
required amount. Shut the tap of the burette. Open the tap
of the tonometer, raise the mercury reservoir considerably so
that when the tap of the burette is opened towards the tono-
meter the CO2 may go up into the tonometer and not the mercury
and blood from the tonometer come down to the burette. Open
the tap of the burette towards A and send up the C02 into
the tonometer.

It is usual to analyse the gas in the tonometer before using
it. If this is to be done remember that some oxygen and car-
bonic acid are lost in the analysis and therefore to this extent
more of these gases must be put into the tonometer than are to be ultimately
required.

20—2

FIG. 150.

308

Appendix I. On methods

Alternative method for filling tonometers.

On our expeditions on Monte Rosa and in Teneriffe we had not at hand the
plant which has been described in the foregoing paragraph. The tonometers were
therefore filled over mercury. This method has the merit of simplicity, otherwise
I think it is not particularly desirable. The cork is taken out of the tonometer, which
is placed with the mouth downwards in a mercury bath. The pressure is reduced
inside the tonometer either by means of attaching a pump to the delivery end or by
sucking the air out with the mouth. When the tonometer is filled with mercury the tap
is shut. A tube from a nitrogen cylinder (or as was the case in Teneriffe, a phosphorus
pipette in which nitrogen was made from atmospheric air) wras then attached to the
end of the fine tubing and the tap so turned that the nitrogen ran to waste through
the tap. When the dead spaces have all been cleared of air and nitrogen is found to
be running- at a convenient rate the tonometer tap is turned through 180° and the
mercury replaced by nitrogen. A few bubbles are allowed to escape through the
mouth. The tap is then turned back to its former position, the nitrogen tube
removed and the gas shut off. The tonometer is corked up, and the oxygen and
carbonic acid are added by the method previously described for the addition of ('< >L..

To analyse the gas in a tonometer.

Make ready the gas analysis apparatus by clearing all oxygen out of it and getting
the various fluid surfaces to their marks.

Turn the tap C of the burette as in the figure. Connect the tonometer with the
gas burette in the manner indicated in the diagram. Turn the tap C into the vertical
position, raise the mercury reservoir till mercury reaches A. Then
turn the tap B through 180° and lowyer the mercury reservoir till
sufficient gas for analysis — 8 or 9 c.c. — has gone into the gas burette.
There is in general a positive pressure in the tonometer so that the
surface of the mercury in the gas burette and that in the reservoir
will not be the same. This pressure is an essential factor in the ex-
periment and must therefore be measured. The simple way of
doing this is to place the mercury reservoir as close to the stem of
the gas burette as possible and read off the pressure in terms of the
number of graduations on the stem. These graduations are, of
course, TJjj of a cubic centimetre, but if it be known how many of
them correspond to 1 mm. in height the measurement may readily
be expressed in terms of mm. of mercury. In finally calculating
the partial pressure of the gases from the percentage composition
of the contents of the tonometer, this positive pressure in the tono-
meter must of course be added to the height of the barometer.

When the gas has been transferred to the gas burette, the taps
are turned to the positions shown in the figure. The tonometer is
removed and the gas in the burette is analysed in the usual way
with Haldane's apparatus.

It may here be helpful to give an actual calculation of the partial
pressure of oxygen as obtained from the data at which we may be supposed to have
arrived.

Let the percentage of oxygen in the air of the tonometer be 0, the height of the

FIG. 151.

Appendix I. On methods 309

barometer P, the positive pressure in the tonometer 2>, the absolute temperature of
the bath in which the blood will be exposed to the gas 71, that of the room £, and the
required partial pressure of oxygen x,

T

x=Ox(P+p)x-.

There is a slight correction which is sometimes worth making at low pressures.
It has reference to the fact that the air in the tonometer acquires oxygen from the
blood. Suppose 1'5 c.c. of blood are placed in the tonometer and the partial pressure
is such that half of the oxygen is given off. This will amount to 0'15 c.c., which in
250 c.c. will exert a pressure of upwards of half a millimetre.

In analysing the gas after an experiment, as is frequently necessary, there are
some additional points to be considered.

(1) The tonometer contains blood and this must not be allowed to foul the gas
burette. It is therefore advisable to connect the two by a long rubber tube of 1 mm.
bore. The tonometer is placed with the cork downwards ; the air is cleared from the
dead space of the rubber tubing, etc., by running mercury out to the open air through
the tonometer tap ; this is now turned and a few drops of mercury run into the
tonometer ; the mercury reservoir is then lowered and the gas is drawn into the
burette.

(2) The positive pressure must be measured as before, but it must be borne in
mind that in taking the gas into the burette the pressure in the tonometer has become
reduced. If 8'3 c.c. are taken out of a 250 c.c. tonometer the value (P+p) will be
3 % below the pressure which existed in the tonometer.

Measurement of composition of alveolar air.

There are two methods, each of which has warm adherents, for measuring the
composition of alveolar air. (1) The method of Zuntz and Loewy ; (2) the method of
Haldane and Priestley. Before any detailed description of either of these is given
a few words may be said about the principles on which each rests.

The method of Zuntz and Loewy consists in estimating the composition of the
inspired and expired air, and computing that of alveolar air from a number of factors
of which the most important is the "dead space of the lung." The dead space
really is the volume of the nose, throat, trachea, bronchi and bronchioles. There is no
profession that the alveolar air is ever actually obtained, and the correctness of the
method really hinges upon the accurate measurement of the dead space.

In the method of Haldane and Priestley it is claimed that if the most sudden and
violent expiration possible be made, the volume of air expired is so great as to clear
the dead space entirely and to give a considerable volume of actual alveolar air at the
end of the expiration. In a sense this involves a knowledge of the size of the dead
space too, but in practice this may be evaded. The further one gets down the
passages which lead to the alveoli, the more does the composition of the air resemble
that of the alveolar air and the more does it depart from that of the atmospheric air.
If, then, as the respiration comes out, one gets to a point after which successive
portions of the respiration are of constant composition it may be assumed that one
has arrived at the alveolar air — that at least is the contention ; and further it is the
contention of Haldane and Priestley that such a point is reached, and that about the
last third of the respiration consists of alveolar air, which may be collected.

310 Appendix I. On methods

Within the last few years a good deal of energy has been expended by the
protagonists of each theory; the general motif of which has been to shake confidence
in the results obtained by the opposing theory.

In so far as I have had an opportunity of seeing the two methods at work on the
same individual (and I believe that this test has rarely been performed) the two
methods gave an identical result, that being so it seemed to me rather a work of
supererogation for the supporters of one method to be so suspicious as to the
validity of the other.

This however was at or near the sea-level, and it is at great altitudes that the two
methods are described as giving most divergent results, the method of Zuntz giving
lower readings for the oxygen alveolar air than the method of Haldane and Priestley
We may therefore make some criticisms upon the two methods.

(1) There is underlying the method of Zuntz and Loewy, or at all events under-
lying their application of it, the assumption that the dead space is a constant quantity.
This, as Haldane has pointed out, is probably very far from being the case. The
respiratory passages, at all events the smaller ones, all have a rich musculature in
their walls. They are, in fact, like arteries which admit air instead of blood. The
physiology of the bronchial muscles leaves much to be desired. It has mostly been
studied from the point of view of asthma, and therefore it is the contraction of the
muscles as compared with the normal tone which has claimed attention. It is, how-
ever, not improbable that our notions about the bronchial muscles are as distorted as
those of the arteries would be if we merely considered the changes from the normal
arterial tone of the resting organ to that of pathological arterial spasm. In the case
of the artery the functional significance of the musculature of the arterial wall is that
it can relax to let the organ have the amount of blood which it requires when it is in
full activity. There is no reason to doubt that the functional significance of the
bronchial muscles is to relax during conditions of exalted respiration and allow
perhaps ten times the ordinary quantity of air to pass through the tubes without let
or hindrance. When we recollect that the volume of a cylinder increases with the
square of its diameter it will be easily seen that the dead space of the lung may
increase without difficulty two or three-fold under conditions such as those obtaining
at high altitudes, in which the respiration is quicker than usual. Of the mechanism
of such dilatation we are in ignorance, it may be medullary ; it may be the effect of
changes in the blood*.

The neglect of this factor may be one reason why Zuntz's method diverges from
that of Haldane's at high altitudes. This neglect would tend to make Zuntz and
Loewy's results for the oxygen pressure in alveolar air too low.

If however one concedes that the dead space of the lung may be two or three times
as great during conditions of deep or rapid respiration, as under ordinary circumstances
one must ask oneself whether by a violent deep expiration such as is made for the
purpose of Haldane's alveolar air determinations, it is possible to clear the air space
as efficiently as one does under normal circumstances.

It would be most desirable that someone should make himself thoroughly con-
versant with both methods and make a more systematic comparison of the two than
has yet been done.

* Since this was written the work of Krogh has appeared (Journal of Pliysiol. Oct. 1913).

Appendix I. On method*

311

I have always used Haldane's method, and it is by that method that all the
results in the foregoing chapters have been obtained. Indeed the simplicity of
the apparatus required and the ease and rapidity with which the determinations
can be carried out by it, makes it much the more suitable of the two methods for
work outside the laboratory.

I will therefore proceed to describe the method. A rubber tube six feet in length
and | — 1 inch in bore, is fitted with a glass mouthpiece round which the lips fit
tightly. This mouthpiece must be at once large
enough to offer no appreciable resistance to the
respiration and small enough to be easily occluded
by the tongue. About one inch from the end of
the rubber tubing is a hole about 4 or 5 mm. in
diameter, just large enough to take an ordinary
piece of glass tubing.

In the laboratory this piece of tubing is put directly on to the burette of a
Haldane's gas analysis apparatus. The tap is turned diagonally as shown in Fig. 153,
after the whole burette has been filled with mercury up to the point A. The
mercury reservoir of the burette is then lowered almost
to the bottom of the burette. (The operator must be
careful that he has a tap good enough to stand this
procedure.) The apparatus is now ready for the deter-
mination. The subject of the determination applies
his mouth to the mouthpiece, taking care to breathe
normally through his nose all the while — this is the
most difficult part of the experiment — then suddenly
either at the beginning or at the end of one of the
normal respirations he shuts his nose and makes an
expiration so forcible through the tube that he expels
all the air he possibly can along the tube and at the
end closes the tube with his tongue. Either he or an
assistant then turns the tap to the vertical position for
a moment, the burette fills itself with alveolar air and
the stopper is then turned back to the position shown
in the figure. The analysis is carried out in the manner
described.

It is best to take one sample at the end of a normal
inspiration and one at the end of a normal expiration and average the two.

In many cases it is not convenient to carry a gas analysis apparatus to the place
where one wishes to make a determination of the alveolar air. One then takes the
sample into a Haldane's gas sampling tube instead of directly into the gas burette.

This tube is shown in Fig. 154 attached to the gas burette.

The taps of the tube are adequately greased, and if it is necessary to carry the
tube about it is best to be able to prevent all possibility of the stoppers either coming
out or accidentally turning by fixing them in position by means of stout rubber rings
(for this and similar purposes I use the rings sold for use on umbrellas). The tube is
rendered vacuous in the laboratory and the taps placed so that the vacuum is shut
oft' from the air. When it is necessary to take the sample of alveolar air this sampling

FIG. 153.

312

Appendix I. On methods

tube is put on to the large rubber tube. The operation is that already described
except that the tipper tap of the sampling tube, not that of the burette, is that
which is rapidly opened and closed.

The gas is transferred to the Haldane's apparatus for analysis in the following
way. A small piece of bent capillary glass tubing is joined by rubber junctions
to the sampling tube and the gas burette of the Haldane's apparatus. The dead
space is filled with mercury. The lower tap of the sampling tube is connected
with a mercury burette such as shown in Fig. 150 ; the dead space here is filled with
mercury also. A sample of the gas is then taken over into the gas burette, its place
being filled with mercury from the mercury burette.

FIG. 154.

To expose the blood in a tonometer at a known temperature.

This is done in a water bath. According to the circumstances and the scale on
which the experiment is being performed, one of two methods may be adopted.

Method I is a laboratory method which is best adapted for use where power can
be obtained and is suitable for the investigation of a number of different samples at
a time. The plant which I have is for five tonometers ; these might for instance
contain the samples of the same blood exposed to gas containing five different oxygen
pressures and thus five points on a dissociation curve might be obtained ; or on the
other hand the tonometers may all be filled with different samples of blood and
exposed to similar atmospheres as in experiments for the determination of the
alkalinity of the various samples of blood.

The bath consists of a brass cylinder or barrel 44 cm. in height and 16 cm. in
diameter. This cylinder is mounted horizontally on bearings from which it can be
taken by simply lifting it, it being kept in place by a weak spring. One end of the
cylinder screws off, the spindle attached to this and which rests on the bearing is
hollow ; through this a thermometer may be inserted. On the spindle at the other
end a V-pulley is attached by which the cylindrical bath is geared by a belt to some

Appendix I. On methods

313

source of power, the gearing being driven in the case of the bath at Cambridge by
^g horse-power motor.

Into the cylinder fits a rack, this rack can stand vertically on a metal base, from
the centre of which rises a hollow tube which is concentric with the bath itself.
Attached to this tube are five spring clips, into each of which a tonometer can
be fixed.

The bath is placed on its end and filled almost full of water at about 1 — 2° C.
higher than is desired for the experiment.

The rack with the tonometers is then put in the water bath. The lid containing
the thermometer is screwed on, the bath is then placed horizontally on its bearings
and the belt put on the driving wheel. It is then rotated at about 15 revolutions
per minute' by the motor. The thermometer is just put in so far that the mercury
surface can be read from without the apparatus. A spirit flame playing on the
bath is sufficient to keep the apparatus up to the required temperature, should the
tendency for the temperature to fall be too great.

After ten minutes the motor is stopped, the bath is taken off the bearings and
allowed to stand vertically on the driving wheel for two minutes, so that the blood
may drain down into the small receptacle near the tap of each tonometer. The lid is
then unscrewed and the tonometers taken out one by one. It is of course essential
that the tonometers be carried vertically and not shaken, otherwise the blood will
acquire oxygen from the atmosphere of the tonometer at the reduced temperature.

If the composition of the blood is required the tonometer is placed vertically with
its lower end touching a duster, the tap is gently turned and any air and mercury in
the tubing together with a drop of blood are allowed to find their way out ; the tap is
then closed and the end placed under the ammonia of a blood gas bottle, and if
a known volume of blood is not required, the tap is simply opened again cautiously
and a suitable amount of blood is allowed to flow gently out.

If the delivery tube of the apparatus is to act as a pipette of known volume, the
tap is turned in the opposite direction, in which case the amount of blood which is in
the pipette will run out by its own weight. Should it fail to
do so it may be assisted by placing a rubber tube of large
bore in the position shown in the figure and blowing down
it ; the air pressure resulting will probably be adequate. It
is a risky proceeding to attach a rubber tube to the glass
one and then blow.

Meth<><! 77, which is very useful for work outside the
laboratory as well as in it, consists of using a water bath
from the side of which project two arms, each of which
has a crook at the end which acts as a bearing. The
arms are placed at such a distance apart that the bearings
catch the tonometer, one close to the cork, the other close to
the tap. When the tonometer is immersed in the bath it of
course tends to rise, but finding itself caught by the bearings
which are above it, is unable to do so. A piece of string is FlG-

put round it, by pulling the ends of the string alternately the tonometer is rotated.
The bath is heated and kept warm from underneath. The tonometer is rotated for
ten minutes and then held for a short time in the bath with the cork at the surface

314 Appendix I. On methods

and the end of the delivery pipette on the bottom, the blood can then drain into the
cup for it before the tonometer is allowed to cool.

In emergencies the domestic bath fitted with hot and cold water serves excellently,
it is large enough to accommodate the tonometer vertically as well as horizontally.

The method of determination of the percentage saturation of the blood has already
been described. It is usually desirable to analyse the air in the tonometer after the
blood has been taken out. For this purpose the tonometer must be placed with the
cork downwards, the nozzle being connected to the gas analysis apparatus by a bent
tube (either of glass or rubber) of fine bore. The dead space is cleared of blood by
forcing some mercury into the tonometer, gas is then withdrawn. In calculating the
pressure allowance must be made for the expansion of the gas into the analysis
apparatus.

APPENDIX II. ON THE AGGREGATION
OF HAEMOGLOBIN

Whilst this book has been in the press some further information has been
acquired relating to the concentration of carbonic acid to which a dialysed solution
of haemoglobin is exposed to the degree of aggregation of the haemoglobin molecules.
It has been shown in Chapter V that alterations in the concentration of C02 have
very little effect on the value of n in the case of blood, H being the average in
number of molecules (each containing one atom of iron) in a clump, and having the
value 2'5. In the case of high concentrations of CO2 there seemed to be some
evidence of a slight increase in the value of n. The most natural interpretation of
this was that the salts present in the blood determined the value of n and the CO.,
the value of K in the equation

y Kxn

Too"

the two effects being therefore quite different in kind. From this supposition it
would follow that the dissociations of dialysed haemoglobin in the presence of in-
creasing quantities of C02 would be a series of rectangular hyperbolae presenting the
general appearance shown in Fig. 13 in which n in each case was unity and K
diminished as the C02 pressure increased.

Experiment shows that this conception is quite at variance with the facts. The
following are the values of K and n for the dissociation curves of dialysed haemo-
globin in the presence of the stated pressures of C02 at 40° C.

CO., pressure n K

0 mm. 1 'HI

8 ,, 1-78 -0062

33 ,, 2-5 '00049

67 2-7 -000192

Appendix II. Aggregation of haemoglobin 315

A better conception of the effect of C02 will be derived from the following figure :
3

2-5

FIG. 156.— Relation of ;i (ordmate) to CO2 pressure (abscissa) in dissociation curve

of dialysed haemoglobin.

It is clear that the reason why the value of n is almost unaffected by C02 in blood
is because 2'5 is nearly the maximal value. This has been effected by the salts
present.

It is clear also that extreme concentrations of C02 may raise the value of n
to a slight extent as it appears experimentally to do in the body. The effect of the
salts then is to render the system stable and prevent large changes in the clumping
of the molecules which would otherwise occur as the result of changes in reaction.
The relative stability of the two systems may be judged from the following data :

Haemoglobin

Blood

71

A'

no C02

I
•111

40 mm. C02

2-5
•00043

no C02

2-5
•00258

40 mm. C02

2-5
•00029

316 Appendix III. "Reaction' of Blood

APPENDIX III. ON THE HYDROGEN ION CON-
CENTRATIONS OF REDUCED BLOOD

Measurements have been made by Peters (Physiol. Soc. Proc., Jan. 1914) of the
hydrogen ion concentration of Barcroft's blood subjected to various pressures of CO2.
The blood was completely reduced, and the following curve was obtained. The result
was nearly but not quite the same as that obtained by Hasselbalch and his colleagues,
whose admirable work on the reaction of the blood may be read parallel with the
later chapters in this book.

C02 pressure mm. 0 10 20 30 40 60 80 100

PH 8-11 7-67 7-52 7'39 7'30 717 707 699

Till further experiments have been performed Peters would not insist on the
absolute accuracy of his measurements but there seems no doubt about their com-
parative accuracy.

We already have (Chapter V) a relation between the C< >2 pressure in Barcroft's
blood and A", the equilibrium constant in the equation

y Kxn
100 \ 1+A>- '

We may therefore directly compare A' and the hydrogen ion concentration :

Log A' 3-40 3-20 3-00 4-80 4-60 4-40 4-20 4-10

Kx, x=-2-60 2-80 3-00 3-20 3-40 3-60 3-80 3-90

PH -8-10 7-933 7-72 7'59 7'42 7-25 7-075 6-99

The comparison is instructive enough for it appears that the product of CH
is constant as shown by the fact that the figures in each column of the last 2 lines
in the above table add up to — 10"8±'l. Should this relation prove on further
examination to be substantiated, Hill suggests that it is evidence of actual chemical
combination between the hydrogen ion and the haemoglobin. The hydrogen ion
would then have two actions: (1) a tendency to cause the globin molecules to
aggregate and (2) an actual union with the haematin portion.

We are not so much concerned with the theoretical qiiestion as with the fact that
we can state a figure for P,, at the various places at which we have measured K in
Barcroft's blood.

K PH

Normal -00029 -7 '29

Carlingford after climb of 1000 feet in 45 miu. -00024 7 -22

30 „ -00017 7-09

Cold'Olen „ „ 45 „ 7'14

Margherita Hut arrive -00013 7 '02

Appendix IV. Constant of Ambard 317

APPENDIX IV. ON THE CONSTANT OF AMBAKD

The determination of the Constant of Ambard (Ambard and Weill, Journal de
physiol. et path. gen. xiv, p. 73, 1914) is for the purpose of discovering whether
there is retention of urea or not. When the concentration of urea in urine is
25 grams per litre of urine, the concentration of urea in the blood (ur) divided by
the square root of the output of urea per 24 hours (Z>) is nearly constant, varying

tJL¥

between 6'3 and S'O, i.e. —=- = ^(constant of Ambard).

When the concentration of the urea in the urine is not 25 grams per litre but
C grams per litre, then D must be corrected for the change in concentration. It is
multiplied by the root of C and divided by the root of 25 ; this gives a factor termed
Z>2r, which replaces D in the above equation, so that

- = A'(constant of Ambard)

and

An actual observation does not last for 24 hours, but for a time T which is about
| of an hour, the volume of urine secreted in this time is v c.c. and the volume of
urine as calculated for the 24 hours at this rate of secretion is V,

Jr rx24x60..

= Tooo^T Iltres'

the other measurements necessary are of course the concentration of urea in the
urine (7, and in the blood ur. If the final answer does not come between 8 and 6'03
the secretory action of the kidney is upset.

INDEX

Acetic acid, effect of, on rate of reduction
of blood 54

Acid, effect of, on dissociation curve of
haemoglobin 53 et seq.; from salivary
glands 162

Acidosis 227 ; due to altitude 255 et seq.,
276; due to exercise 238, 276; due to
disease 282 et seq.; due to diet 232; in
relation to respiratory centre 232 ; deter-
mination of, Mathison 259 ; Boycott and
Chisbolm 259

Adsorption, theory of union between oxygen
and haemoglobin 24. See also Aggrega-
tion and Colloids

Aggregation of haemoglobin molecules 32,
50, 51, 60 et seq.; in disease 289; with
CO2 314

Alps 249 et seq.

Alta Vista 246 et seq.

Altitude, effect of 243 et seq. ; on alveolar
C02 251, 266; acidosis 255; on blood
130; on dissociation curves 252, 267; on
lactic acid production 259; on oxygen
capacity 130; "pulmonary secretion"
217 ; on, respiratory centre 261

Alveolar air, estimation 310

Ambard, constant of 283, Appendix IV

Ammonia, effect of, on dissociation curve
of haemoglobin 43

Ammonium carbonate, do. 44

Anaemia, oxygen used by blood in 11, 124

Atropine on salivary glands 147

Blood, dissociation curve of, dog's 49 ;
effect of acid on 53 et seq. ; effect of
CO2 on 58, 65 ; effect of lactic acid on
54 ; effect of temperature on 158 ; human
47, 62, 63, 64 ; oxygen used by 120 et seq. ;
oxygen used by, during anaemia 124
- rate of reduction of 173 ; effect of CO-,
on 54, 175; acetic acid on 54; lactic acid
on 54 ; hydrochloric acid on 54 ; tem-
perature on 175

Blood-flow 134 ; effect of metabolites on
137; heart 149; liver 134; pancreas 148;
salivary glands 139 et seq.

Calibration of differential apparatus 293,
299

Cafiadas 244 et seq.

Capillaries, oxygen in, 163

Carbonic acid, aggregation of haemoglobin
with 314 ; effect of, on COHb 57 ; effect
of, on haemoglobin Appendix II ; effect
of, on dissociation curve of blood 53 et
seq., 58, 65, 67, 160; effect of, on rate
of reduction of blood 54 ; in blood from
salivary glands 139, 162; in alveolar air
251; do. in disease 283; ."hysteresis"
86

Carboxyhaemoglobin 202 et seq. ; dissocia-
tion curve of 57, 67, 186; effect of CO->
on 57, 67, 68, 69

Carlingford 236, 267

Chemical basis of variation 72

Col d'Olen 250 et seq.

Colloid nature of haemoglobin 50, 51

Constancy of n for human blood 65

Coronary system, blood-flow through 149

Corpuscles, salts in 47

Cyanosis 283

Diet on dissociation curve 228
Difference of arterial and venous blood 298
Differential blood-gas apparatus 291, 299
Dilatation, vascular, in relation to functional

activity 138

Dilution in relation to oxygen capacity 25
Dissociation curves, altitude 252, 267 ;
ammonia on 43 ; ammonium carbonate
on 44, 62; COHb 186; exercise 237,
267; electrolytes 41 et seq.; potassium
chloride 46, 61; Einger's solution 62;
salts on 41 et seq.; do. in corpuscles 47;
do. canine corpuscles 47 ; do. in human
corpuscles 47 ; sodium bicarbonate 61 ;
sodium chloride 46, 61 ; sodium phosphate
67
Dyspnoea 285

Exercise, acidosis in 238, 276; hydrogen
ion concentration in 238, 270; lactic acid
in 239, 276; meionexy in 238, 267

Index

319

Ferricyanide reaction with haemoglobin
290

Gases, storing of 305

Haemoglobin, aggregation of molecules 32,
50, 51, 60 et seq., Appendix II; do. in
disease 289 ; analysis of 2 ; colloidal
nature of 50, 51 ; dissociation curve of
18, 19, 20; do. a rectangular hyperbola
18, 23, 36, 60; effect of: acid on, 52,
53 et seq.; ammonia on 43; ammonium
carbonate on 44, 62; electrolytes on 41 et
seq.; potassium chloride on 46 ; Ringer's
solution on 46, 62; salts on 41 et xcq.,
159 ; salts in red corpuscles on 47 ; tem-
perature on 27 et seq . ; trivalent ions on
51 ; increase at high altitudes 130 ; iron
in 3; molecular weight of 31; molecular
heat of 31 ; specific oxygen capacity of
3 et seq. ; reaction with ferricyanide 290;
rate of oxidation 174; rate of reduction
174; effect of salts on 175

Haemoglobinometer, standardisation of 295,
302

Haldane and Douglas' experiments on lung
203

Hartridge's experiments on lung 207

Heart, blood-flow in 149 ; oxygen pressure
in 170; oxygen used by 80 et seq.

Heat of formation of oxyhaemoglobin 31 ;
liberation of heat in muscle 77 et seq.

Human blood corpuscles, effect of salts in
47

Hydrochloric acid, effect of, rate of re-
duction of blood, 54

Hydrogen ions 238, 270, Appendix III

Hysteresis of carbonic acid 86

Increase of haemoglobin at high altitudes

130

Individual differences in blood 64
Inflammation in relation to metabolites 156
Intestine, oxygen used by 116
Invasion coefficient 190
Ions, trivalent 51 ; hydrogen 238, 270,

Appendix III
Iron in haemoglobin 3
Isotonic and isometric cbntractions of the

heart 85, 89

Johannesburg 244

Kidney, oxygen pressure in 170; oxygen
used by 93, 107 et seq. ; work done by 93 ;
effect of clamping artery 115

Krogh's experiments on lung 196

Lactic acid in blood 241, 257; in disease
283 ; during exercise 239 ; at high alti-
tudes 259, 276; in urine 261; effect of,
on dissociation curve 54

Law of mass action 17, 60, 314

Lewis's dyspnoea 288

Liver, oxygen used by 104, 134

Lung, Krogh's experiments on 196 ; Haldane

and Douglas' do. 205, 217 ; Hartridge's

do. 207
Lyophil colloids 51

Margherita Hut 249 et seq.

Mass action, law of 17, 60, 314

Mathison's method of acidosis determina-
tion 259

Meionexy, definition 225 ; in exercise 238,
267 ; in disease 283 ; at high altitudes
on respiratory centre 262

Metabolites, effect of 137, 156

Molecular weight of haemoglobin 31

Monte Rosa 249 et seq.

Mountain sickness 279

Muscle, oxygen pressure in 169 ; oxygen
used by 75; heat liberated in 77

Nitrogenous substances in blood in disease

283
Nuttallia equi, oxygen capacity of blood in

13

Orotava 243

Oxygen capacity 130; determination of 294,

302 ; in relation to dilution 25 ; specific

3 et seq.
Oxygen exchange in capillaries 179 ; in lung

182 et seq.
Oxygen in blood, determination of 296,

302; in capillaries 163 et seq.
Oxygen pressure, in tissues 178 ; in salivary

glands 167 ; in muscle 169 ; in heart 170 ;

in kidney 170
Oxygen used by blood 120 et seq. ; do. during

anaemia 11, 124; heart 80; intestine

116 ; do. during absorption 116 ; liver

104; kidney 92, 110; muscle 75; pancreas

101 ; salivary glands 98, 146

Pancreas, oxygen used by 101

Percentage saturation, determination of 297,
303

Pressure of oxygen in, capillaries 163 et
seq. ; arterial blood, coronary system 170 ;
kidney 170 ; muscle 169 ; salivary glands
167 ; of C02 in blood 144, 162

Pleonexy defined 225

Potassium chloride, effect of, on dissocia-
tion curve 46, 61

Rectangular hyperbola, dissociation curve

of haemoglobin a 18, 19, 20, 36, 60
Reduction of haemoglobin, rate of 38 ; effect

of temperature on 38
Reflex dilatation in submaxillary gland

147

Relation of K to CO-> pressure 70
Renal artery, effect of clamping 115
Respiratory centre 233, 242; altitude 261

320

Index

Kinger's solution, effect of, on dissociation
curve of haemoglobin 62

Salivary glands, carbonic acid from 145,
162 ; blood-flow in 138 ; effect of atro-
pine on 147 ; oxygen used by 98 ; reflex
dilatation in 147

Salts, effect of, on dissociation curve 41 et
seq., 159; in human blood 47

Sodium bicarbonate, effect of, on dissocia-
tion curve 61

Sodium chloride, effect of, on dissociation
curve 46, 61

Sodium phosphate, effect of, on dissociation
curve 61

Specific oxygen capacity 3 et seq.

Spectroscope, Hartridge's 207

Storing gases 305

Submaxillary. See Salivary

Suspensoids 51

Swim bladder 183

Temperature, effect of, on dissociation curve
27 et seq., 160; on rate of reduction of
haemoglobin 38; do. of blood 176; on
"pulmonary secretion" 188

Teneriffe 243 et seq.

Tension. See Pressure

Titanium method of estimating iron 5

Tonometer 304 ; filling 305, 308 ; treatment
of blood in 314

Trivalent ions on aggregation of haemo-
globin 51

Urine, lactic acid in 261; secretion of 92,

110
Uraemia 282

Work done by kidney 93

CAMBRIDGE: PRINTED BY JOHN CLAY, M.A. AT THE UNIVERSITY PRESS

o

*o

•**"

BIOLOGICAL
LABORATORY

WOODS HOLE,
W. H. 0. 1.

, , *

M.B.L. LIBRARY -WOODS HOLE, MASS.