994

An AddressON THE

FUNCTION OF HÆMOGLOBIN INTHE BODY.

Delivered at St. Mary’s Hospital on May 8th in the course oflectures on Pathological Research in its Relation to Medicine.

BY A. V. HILL, F.R.S.,PROFESSOR OF PHYSIOLOGY, UNIVERSITY COLLEGE. LONDON.

HEMOGLOBIN has two main functions in the body-the carriage of oxygen and the carriage of acids,including CO2; the modes of carriage are different,but equally important. That of oxygen has long beenknown, and its characteristics have been the subjectof many investigations ; that of acids and CO2, thoughoften discussed and doubted, at last seems certainbeyond dispute. A possible third function of hæmo-globin has recently appeared from the work of MissRobinson in Hopkins’s laboratorv at Cambridge.She found that haemoglobin and its derivatives,methaemoglobin and hsemin. are able to acceleratevery considerably the oxidation of unsaturated fattyacids-e.g., of linseed oil. Its non-iron containingderivative, hæmochromogen, produces, however, nosuch acceleration. The importance of this observationlies in the fact that a certain amount of haemoglobincertainly lies inside the tissues themselves ; it does notoccur merely in the blood. Possiblv its function in thetissues is that of an oxidative catalyst. This propertyof haemoglobin appears not to be a consequence of itseasily reversible combination with oxygen : it occursequally in heemin and methæmoglobin, or in haemo-globin fully saturated with carbon monoxide. Thefact that the katalytic action does not occur withhæmochromogen suggests that the accelerating powerof haemoglobin lies in its content of iron. Iron, as0. Warburg has found, is an important agent inkatalysing many oxidations analogous to thoseoccurring in the body. The action of haemoglobin onthe oxidation of unsaturated fats appears to be evenmore effective than that of an amount of inorganiciron equal to that contained in the hæmoglobin. Itwould seem possible that the faulty iron metabolismoccurring in such conditions as anaemia may resultin slow or defective oxidation in the tissues. Thismight be one cause of the dyspnoea associated withthose conditions. I have discussed this possiblefunction of haemoglobin first since little is known aboutit, and we may now turn to the other functions whichare better understood.

REASONS FOR EXISTENCE OF HÆMOGLOBIN INSIDECORPUSCLES.

In nearly all animals haemoglobin exists insideformed corpuscles in the circulating fluid ; it does notusually occur simply dissolved in the plasma. Therehas been much speculation as to the cause, or function,of the packing away of the haemoglobin into theseformed elements. It may be partly due to the neces-sity of retaining the haemoglobin molecule, which iscomparatively small, inside the blood-vessels ; were

it not inside the corpuscle, it rnight diffuse freely fromthe plasma into the lymph and from the lymph intothe tissues, and so its usual respiratory function wouldbe wasted. It may be to save it from excretion bythe kidneys, though, indeed, this end might havebeen achieved by making the kidneys impermeable tohaemoglobin, as thev are in animals in which thatsubstance is free in the plasma. The most importantreason, however, is probably to enable this sensitive erespiratory substance to be subjected to one set ofconditions, while the plasma-and the tissues which iit bathes-are subjected to another and more appro-priate set. The physico-chemical conditions obtainingin the plasma may be appreciably, possibly widely,different from those inside the red corpuscle. Bysuch means the body is provided, so to speak, with

another degree of freedom. It may be able, as anadjustment or an acclimatisation, to change theconditions affecting the haemoglobin without alteringthose obtaining in the plasma and in the lymphthroughout the body.

THE COMBINATIONS OF HÆMOGLOBIN.The haemoglobin molecule contains an atom of

iron. In its combination with oxygen the maximumamount of that gas with which it can combine is thechemical equivalent of the iron in the molecule. Theultimate unit of haemoglobin appears to have a

molecular weight of about 16,700 ; it is comparativelysmall therefore, the colloid molecule sharing many ofthe properties of crystalloids, being capable ofcrystallisation and obeying the laws of chemical massaction. It may combine also with carbon monoxide,a matter of great practical importance, since nearlyall the deaths which occur in coal-mining are due topoisoning caused by the high affinity of CO forhaemoglobin. The CO-combination appears to beexactly of the same nature as that of oxygen.Haemoglobin combines also with nitric oxide (NO), butthat reaction has been little investigated owing totechnical difficulties. The combinations of oxygenand carbon monoxide with haemoglobin are reversibleones. At a low pressure only a small amount is takenup, at very low pressures practically none, at higherpressures, however, the hæmoglobin becomes fullysaturated with the gas. If the pressure be reduced thegas originally combined will be again released. Theimportant characteristics of these combinations can beshown in a curve which is known as the " dissociationcurve.’’ The degree of saturation of haemoglobin isplotted vertically against the pressure of the gashorizontally (Fig. 1). The idea is simple but thetechnique difficult; it requires experience beforeaccuracy can be attained. It is necessary first to

expose the solution of haemoglobin in a so-called" tonometer" or "saturator" to any requiredpressure of the gas, which must be analysed afterwardsto ensure that it is nnally of the expected composition.It must be revolved for some time under definedconditions of temperature in order to allow it fullyto attain an equilibrium. The solution containing thegas in combination must then be removed and placedin a blood-gas apparatus of the type designed byBarcroft, Haldane, or Van Slyke. In this way thedegree of saturation, or rather the degree of unsatura-tion, of haemoglobin may be measured. Allowancemust then be made for the gas dissolved in physicalsolution. So investigated, a dilute solution of haemo-globin, or a strong solution very carefully dialysed tofree it from salts, and adjusted to the right hydrogen-ion concentration, gives a dissociation curve whichappears to be a rectangular hyperbola governed bya simple chemical equation. The presence, however,of salts in a strong solution, such as occurs inside thered blood corpuscle, was shown by Barcroft and hiscolleagues in a series of very fundamental observationssome 15 years ago, very considerably to affect the

shape of the curve, turning it from the simple hyper-bolic type to one possessing an S-shape. It is a veryfortunate thing for the body that the actual dissocia-tion curve of blood is of the latter form and not of theformer. Respiration would have been difficult had theblood possessed the properties of a dialysed solutionof hæmoglobin. Owing to the S-shape, it is possiblefor the blood to take in its oxygen and to be fullysaturated at not too great a pressure, and yet to giveit off again at a pressure not so much less. The supplyof oxygen to the tissues, from the blood in the

capillaries, depends upon diffusion and requires,therefore, a comparatively large pressure-i.e., a.

concentration of the gas in ordel to drive it along.If the gas could be released from the haemoglobin onlyat a very low pressure, as is the case with dialysedhaemoglobin, the removal by the tissues of a largefraction of the oxygen contained in the blood would bedifficult or impossible. As it is, a high degree ofutilisation of the oxygen contained in blood is renderedpossible by the shape of the curve.

995

MOLECULAR CONSTITUTION OF HÆMOGLOBIN.

From the physico-chemical standpoint the changein shape produced in a strong solution by the presenceof salts appears to be due to an aggregation, or apolvmerisation of the haemoglobin molecule, or tosome other cause reducing what may be calledits thermodynamic

" activity.’’ In a concentrateddialysed solution, or, as Hartridge and Roughtonhave recently shown, in a dilute solution undialysed,the reaction appears to follow the chemical equation,Hb + O2 HbO2 (unaggregated molecules).In concentrated solution, however, in the presence

of salts, the molecules appear to be aggregated, andthe chemical equation to be,Hbn + n02 -’ (Hb02)n (aggregated molecules).The fundamental unit, therefore, of haemoglobin,

as it exists inside a red blood corpuscle, seems to beHbn containing n atoms of iron and combining withn molecules of oxygen. The quantity n seems to bean average number representing the average degreeof aggregation of the molecule. Most of the moleculesare present as Hb2, a minority as Hb3.

FACTORS AFFECTING DISSOCIATION CURVE.

The dissociation curve is affected by various factors. iA rise of temperature, both in the case of dialysedhaemoglobin and in that of whole blood, produces alarge fall in the affinity of the pigment for the gas.This is a thermodynamic consequence of the fact thathaemoglobin combines with oxygen with an evolutionof heat. It is of importance in the body, since thehighest temperature of the blood occurs in the mostactive tissues which need the oxygen most, and thehigh temperature helps to drive the oxygen out.The dissociation curve is affected also by a changein the hydrogen-ion concentration, whether pro-duced by the presence of C02 or by the additionof acid-e.g., of lactic acid from the muscles-to

FIG. 1.

20 40 60 80 100 120

Dissociation curves of human blood, at different temperatures,and exposed to a CO! concentration which would have given apressure of 40 iiim. at 38°C. Calculated for n = 2’2, K38 =000060, q = 19000 cals. (blotted in tenns of oxygen pressuresat the actual temperatures involved, instead of in concentra-tion.- of oxygen.)

the blood. The effect of CO, in diminishing theaffinity of haemoglobin is a large one and exactlysimilar to that produced by a rise of temperature :it is of importance both in the tissues and in thelungs, since the increased pressure of CO2 in the Icapillaries of an active muscle helps to drive the Ioxygen out to the cells where it is wanted, whereasthe diminished CO, in the lunps helps to allow oxygento come in. The effect of fixed acid on the dissociation curve is important in connexion, both with asphyxiaas was shown by Barcroft, and also with muscular ,

exercise. The addition of 0-08 per cent. of lactic acidto blood, or even of 004 per cent., causes an extensivechange in the form of the curve, a shift to the right t(Fig. 1) similar to the shift produced by a rise of 5° C.Barcroft has shown that moderate exercise, such aswalking up a mountain, may produce similar effects,which he attributed to the acid liberated by themuscles appearing in the blood. When it is remembered

FIG. 2.

5 GAS PRESSURE.10 - MM. 2b 30 40 50 60 70 60 90 100 120 140 160

Diagram for plotting dissociation curves. Vertically log y/( 1—y)horizontally log x. If the observations obey the equationy/(1—y) = Kxn, the points will lie on a straight line ; theslope gives the value-of n, which may be read directly on thediagram. A change in K only moves the line parallel to itself.

that severe muscular exercise may cause the appear-ance in the blood of as much as 170 mgm. of lacticacid in every 100 c.cm.—i.e., of 0-17 per cent.-it will berea.lised how extensive are the changes in the respira-tory function of haemoglobin which may be caused inthis way. During severe muscular fatigue the blood,as a consequence of this effect of acid, may becomemore unsaturated as it passes through the capillaries,may have considerably more oxygen taken out of it ata given pressure than under normal circumstances--a fact for which the active muscles may well begrateful.The equation of the dissociation curve, which may

be deduced from the chemical formulae given above,is ly y = Kxn, with only two constants, K and n.

Fortunately for the physico-chemical theory of thecombination, the effect both of temperature and of achange in hydrogen-ion concentration appears to bedirected at K only and not at n. K is largelydiminished by a rise of temperature or of hydrogen-ionconcentration, while n remains unaltered. This makesit possible to adopt a very simple form of dissociation.

I curve, in w-hich log lyy is plotted vertically, and log gcurve, in which log is plotted vertically, and log x

horizontally. The resulting dissociation curve thusbecomes a straight line, which may be drawn with aruler if two points upon it be known (Fig. 2) ; a changeof temperature, or of hydrogen-ion concentration,merely moves this line backwards or forwards,parallel to itself, without changing its slope. Thismethod of plotting dissociation curves, though ittells us nothing physically about the nature of thereactions, is a great help in the rapid drawing andinterpolation of dissociation curves, and in clelnon-

strating the effects of the two chief factors whichchange them. From the slope of the line the valueof n may be read off.

This effect of acid on the dissociation curve is notcontmed to the case of oxygen. As Haldane and hiscolleagues have shown, CO has an exactly similar effectupon the dissociation curves for carbon monoxide ;on the distribution curves, however, of hæmoglobinbetween CO and O2, acids and CO have no effect.This may seem at first sight paradoxical, but whenwe remember that acid has the same effect on theCO dissociation curve as it has on the O2 dissociation

996

curve the result becomes obvious. There is no reasonwhy it should work one way more than the otherwhen both gases are present.An explanation of why acid has this most important

effect on the affinity of haemoglobin for oxygen, or CO,has long been lacking. Recent work, however, hasprovided a tolerably certain one. Haemoglobin, as itoccurs in blood, is an electrolyte, a weak polyvalentacid. We may regard it therefore as dissociatingaccording to the scheme :

H(Hb)n ⇄ H’ + (Hb)’nin which the ion (Hb)’u has a very much greateraffinity for the gas than has the molecule H(Hb)n.Raising the hydrogen-ion concentration, therefore,tends, according to the laws of mass action, to depressthe dissociation, to force the reaction to the left, tocause the formation of the undissociated moleculefrom the ions, and hence to produce a body having asmaller affinity for the gas. Raising the hydrogen-ion concentration, therefore, causes the haemoglobinto give up its oxygen, raises the oxygen pressure.This process occurs in the capillaries; acid and CO2pass into the blood and help to drive the oxygen tothe active cells which want it. Conversely, it isinevitable that raising the oxygen pressure shouldraise the hydrogen-ion concentration of blood. Thiseffect has been shown directly by Parsons and morerecently by Brown, indirectly-but equally certainly-by some observations of Haldane, of which I shallspeak later.

METHOD OF CONVEYANCE OF ACID AND CO2BY BLOOD.

We will turn now to the question of how acid andCO2 are carried by blood. It has long been disputedwhether, on the one hand, CO2 combines directly withhaemoglobin, either (1) as oxygen does, or (2) to formhaemoglobin carbonate or bicarbonate,

(1) Hb + CO2 ⇄ HbC02, or

(2) HbOH + H2COa HbHCO3 + H2O;or whether, on the other hand, it is carried as bicar-bonate by alkali released by the hæmoglobin or otherweak acids in the blood,

NaHb + H2CO3 -> NaHCO3 + HHb.It has been shown by Barcroft that the effect of CO 2on the dissociation curve is precisely equal to thatof the change in hydrogen-ion concentration whichit produces. There is nothing left, therefore, forany specific effect : CO produces its change on thedissociation curve merely by its action as an acid and inno other way. Conversely, therefore, haemoglobinmust deal with CO2 as it would with an acid, not byany specific reaction such as that with oxygen andcarbon monoxide. It is known, moreover, that inthe ordinary range of C02 pressures there are no Ispecific absorption bands corresponding to a directcompound of carbon dioxide and haemoglobin. Therewould seem, therefore, to be only two possible waysin which the CO could be dealt with: either theformation of sodium bicarbonate by means of sodiumreleased by haemoglobin, or the direct formationof haemoglobin bicarbonate. The fact that CO2under all ordinary conditions is present simply as

bicarbonate is shown, moreover, by the accuracyof the method of calculating the hydrogen-ionconcentration by means of L. J. Henderson’sequation,

cH = k.pC02/VC02,If the term vCO contained a large proportion presentin some form other than bicarbonate, then themethod would necessarily go astray ; it is widelyemployed, however, and no serious discrepanciesare found. Now an easy way of deciding whether thehaemoglobin combines with CO2 directly, or liberatessodium to combine with it, is provided by the deter-mination of the sign of the electric charge on thehmmoglobin ion. A recent, hitherto unpublished,research by H. Taylor has shown beyond question

by electrometric methods that the hæmoglobin isthere as an anion, carrying a total negative chargeadequate to account for all the CO which blood iscapable of taking up. There is no doubt, therefore,that under ordinary circumstances haemoglobin ispresent as the sodium or potassium salt of an acid-e.g., as Na3Hb, an alkaline salt capable, like thesalts of other weak acids, of acting as a

" buffer," thatis, of giving up its alkali to neutralise other andstronger acids. Such reactions as the following wouldoccur :--

Na3Hb + HN → Na2HHb + NaXNa2HHb + HX > NaH2Hb + NaX.

Had there been haemoglobin bicarbonate present,then the haemoglobin would necessarily have had apositive electric charge ; this is quite definitely notthe case. As the CO2 pressure is increased, the totalcharge on the haemoglobin diminishes-as it shouldif the latter be present as an anion-and indeed byadding acid may be made to disappear altogether andbecome positive, though small, on the opposite sideof the iso-electric point. This point, however, occurswell outside the range possessing any practicalphysiological significance. Such considerations showthat CO2 not only may be carried but must be carriedin the manner indicated. One could wish that thecertainty of these facts would induce people, who stillhold the contrary view, to spend their time in thediscussion of something more profitable.

IMPORTANCE OF CO2 DISSOCIATION CURVE.The CO2 dissociation curve is of considerable

importance (Fig. 3). It differs from the oxygendissociation curve ; it has no S-shape at the beginning; .

CO2 dissociation curves of oxygenated and reduced humanblood (J. S. H.). Christiansen, Douglas, and Haldane, 1914.

it does not approach a maximum, but continues to riseas the CO2 pressure is increased. One of the mostimportant-the most astonishing, the most intriguing--observations in connexion with the C02 dissociationcurve was that made by Christiansen, Douglas, andHaldane on the effect of oxygen upon it. Thecombination of haemoglobin with oxygen diminishesthe amount of CO with which the blood can combine !The explanation of this appears to be as follows.The haemoglobin molecule is a very large one, with anumber of groups OD it capable of releasing a hydrogenion and acting as an acid. At one particular groupin the immediate neighbourhood of the iron atomvery special conditions obtain, and when the ironcombines with oxygen, as it probably does when thatgas is taken up by haemoglobin, the acid dissociationconstant of the local hydrogen ion is very considerablyincreased. What was previously such a weak acidthat it remained almost completely undissociated, hasnow become such a strong one that it becomes almost

997

completely dissociated and requires sodium toneutralise it. Consequently less sodium is availablefor the neutralisation of acids or C02, hence less C02is capable of combining with the blood when oxygenhas been taken up. This effect, which Haldane andhis colleagues observed in 1914, is, as a matter of fact,a necessary consequence, from the physical standpoint,of the converse effect of CO2 on the dissociation curvesof blood. It is amusing to reflect that a phenomenonwhich could have been predicted from otherwell-knownexper:mental facts had to wait for its discovery byactual experiment ; it is strange how often one

begins to predict only after the event ! This effectof oxygen in changing the affinity of blood for CO2 isof importance in the capillaries and lungs, in bothof which it aids the transport of C02. It can

scarcely be called an adaptation, for it deals witha definite inadaptable chemical property ; it isa striking example, however, of the way in whichthe body is apt to find and utilise chemicalreactions which seem so admirably to suit itspurposes.

METHOD OF RETENTION OF HAEMOGLOBIN INRED CELLS.

Haemoglobin is retained inside the walls of thecorpuscle by the presence of an impermeablemembrane; so also are Na., K, and phosphate ions,but not H* or Cl’. The wall of the corpuscle is verythin and elastic. Probably the total osmotic pressureinside it is equal to that outside. When the plasmahas water or any chemical body added to it, thecorpuscle swells or shrinks, so as to maintain an equalityof osmotic pressure. The osmotic pressure of haemo-globin is not inconsiderable and is probably balancedby the difference in the pressures of other indiffusiblebodies, such as Na., K, and P04. By special methodsit is possible to destroy the impermeability to Na*and K’ while leaving that to Hb. If the boundary ofthe corpuscle be permeable to all salts, but not tohaemoglobin, the osmotic pressure of the latternecessarily causes the corpuscle to swell up. Duringthe dialysis of blood corpuscles, without previouslaking, this phenomenon was discovered by Barcroft,who found that the corpuscles swelled up into contactwith one another, appearing to be a homogeneoussolution, but becoming cloudy at once, as Brownshowed later, by the addition of salts. The effect ofadding salts is only temporary, and disappears as thesalts distribute themselves evenly inside and outsidethe now permeable corpuscles ; the latter then swellagain under the osmotic pressure of the haemoglobininside. This experiment may be repeated again andagain, the cloudiness appearing and then disappearing each time the salt is added. Another explanation ofthe phenomenon has been given by Brinkman, anexplanation, however, requiring some credulity. Itis supposed that the haemoglobin in such cases isreally outside in the solution, but that on the additionof salts it returns, by some process of adsorption or otherwise, to the inside of the corpuscle. A roughcalculation, however, shows that if the haemoglobinwere distributed only on the surface of the corpuscle,it would form, not a monomolecular layer, but oneabout 100 molecules thick. It is difficult to see howit could hold on ! Other explanations of thephenomenon than that advanced by Brinkman wouldseem more plausible.Burker has described an interesting relation between

the linear dimensions of a blood corpuscle and theamount of haemoglobin which it contains. Over a widerange of animals the amount of haemoglobin percorpuscle is apparently proportional to the surface ofthe latter. This also suggests, at first, that the hæmo-globin is connected with, or carried by, the surfaceof the cell. The amount, however, of haemoglobinwould seem to prohibit this explanation, since itis difficult to conceive adsorption, or any otherphysical process, to go on through a layer 100 moleculesdeep. It is far more likely, as L. E. Bayliss hassuggested, that Burker’s relation depends uponthe fact that the thickness of all the corpuscles

is about the same, in which case their content ofhaemoglobin would naturally be proportional to theirarea.

At the surface of the corpuscle there exists what isknown as a Donnan membrane equilibrium. The

presence in the corpuscle of charged ions, incapable ofpassing through its walls, produces a difference ofelectrical potential across its surface and a differenceof concentration on its two sides of ions otherwiseable freely to diffuse. This is not the occasion todiscuss this membrane equilibrium, but it would seemcertain that the differences which are known to exist,and the transformations which are known to occur,between the inside and the outside of the blood

corpuscle, are due largely to the existence of thisDonnan equilibrium.

"BUFFER" PROPERTY OF HÆMOCHLOBIN.

The property possessed by haemoglobin of actingas a weak acid which, in combination with alkali, is.a very potent " buffer," explains the capacity of blood

FIG. 4.

Normal human blood cH varied by varying CO2 pressure’Relation between vCO2 and cH. The heavy line, which isexactly straight, denotes the relation for the " average *’

I normal man, the broken lines enclose the relations for eightnormal persons. The steepness of the straight line is a measureof the efiricieiicy of " buR’ering," and is determined mainlyby the amount of hæmoglobin in the blood. Barcroft,Bock, &c., 1922.

for holding-for neutralising--relatively large quan-tities of acid, without a serious rise of the hydrogen-ionconcentration. Recent work by Barcroft and hiscolleagues has defined the relations which exist, inthe blood of the average normal man, between thehydrogen-ion concentration and the C02 pressure onthe one hand, and between the volume of CO carriedby the blood and the hydrogen-ion concentration onthe other (Fig. 4). The steepness of the line, in thelatter case, relating the volume of CO carried in theblood to the hydrogen-ion concentration, is a measureof the efficiency with which the blood is buffere; theposition of the line depends upon the amount ofbicarbonate in the blood and may be altered byartificial means. Its slope, however, depends mainlyupon the amount of haemoglobin present in the blood ;it will be less in the case of anaemic persons, morein that of persons acclimatised to high altitudes, witha high haemoglobin content.

TYPES OF HÆMOGLOBIN.

It has long been discussed whether all types of

haemoglobin are the same, or whether different animalseach possess their own characteridic pigment. It was

998

known, of course, that crystallographically the haemo-globins of different animals were different: this,however, might be due merely to the specific globinportions of the molecules, and not be associated withany very striking physical or chemical differences.Recent work by Barcroft on the blood of a marineworm, Arenicola, has shown quite definitely that, bothin its physical and its chemical properties, that animal,at any rate, possesses its own specific haemoglobin.Barcroft’s work was carried out with a wonderfullittle instrument, the Hartridge reversion spectroscope,which is becoming of more and more value in thestudy of the behaviour of haemoglobin and blood.The principle of this instrument is not unlike that ofthe Barr and Stroud range-finder, in which twoimages are placed, one above the other, and a

coincidence is obtained by the observer betweencorresponding lines or points. In the Hartridge instru-ment two spectra are placed, one above the other, oneof them being reversed, and coincidence is obtainedby placing two absorption bands in line. This methodenables far greater accuracy to be attained than bythe usual method, in the measurement of the wave-length of an absorption band, and has been adaptedfor determining the dissociation curves of dilutesolutions of haemoglobin, both with oxygen and CO.Employing Hartridge’s instrument Barcroft found that the oc-absorption-band of haemoglobin from Arenicola Iis in quite a different place from that of humanhaemoglobin, and-even more striking-that thedistance between the CO x-band and the oxygenx-band is also very different. Correlated with theseshifts in the oc-band is a large change in affinity. Theblood of Arenicola has an enormously greater affinityfor oxygen, and a greater affinity for carbon monoxide,the latter, however, being increased nothing like somuch as the former. Furthermore, Barcroft foundthat if we vary the animal, or the temperature, or thegas with which the haemoglobin is combined, thenunder all conditions there is a linear relation betweenthe affinity and the wave-length of the resulting a-band.Clearly there is something of a very fundamentalphysical character behind these observations. Inany case, Barcroft has answered without question thelong-debated problem of whether all types of haemo-globin are the same, and has shown, moreover, theextraordinarily low pressures at which, in Areaaicola,oxygen dissociation can take place. An animal whichcan utilise oxygen, which can reduce its haemoglobinat these very low pressures, must have an amazingly Ieffective oxidative mechanism.

EFFECT OF CO ON CAPACITY OF HÆMOGLOBINFOR OXYGEN.

There is one curious and paradoxical observation, °

which is the last that I will discuss. If blood bebrought in contact with a mixture of CO and 02, andthe pressure of the CO be varied from a finite valuetowards zero, then the degree of saturation of thehaemoglobin with oxygen will change ; it will rise atfirst as the CO pressure is diminished, and will fallagain as the CO pressure becomes very low andultimately vanishes. It would have seemed at firstsight almost inevitable that, as the CO pressure fell,the amount of oxygen taken up would rise continu-ously ; CO turns out 02, its removal should allow moreO2 to combine. There is a physico-chemical explana-tion of why the apparently common-sense view is

wrong. We might-after the event-have predictedthat the oxygen concentration would fall again.

Douglas, Haldane, and Haldane, in any case, estab-lished the experimental fact, which explains an earlierobservation by Haldane and Lorrain Smith-viz.,that the condition of mice exposed to a very lowpressure of oxygen did not deteriorate, but actuallyimproved, on the admission of a small amount ofcarbon monoxide into the chamber in which theywere ! It would seem possible--though perhapsunlikely—that the breathing of a small amount ofcarbon monoxide might be of assistance in this w-ayin aiding mountain-climbers suffering from oxygen-want at extreme heights.

DIRECTIONS OF FUTURE RESEARCH.

Haemoglobin is one of the most interesting substancesin the world. Many views are possible about it, andsome of these views I have expressed here. Othersyou will find in books, particularly in Bayliss’s"Principles of General Physiology." In 1919 Mr.Barcroft and I persuaded Sir William Bayliss to cometo Cambridge to discuss with us, and with Dr.Hartridge, the problems of haemoglobin. After ourdiscussion I remember jestingly remarking that weought to found a " Haemoglobin Club " ; Sir WilliamBayliss took the suggestion seriously, approachedthe Medical Research Council, and the " HaemoglobinCommittee was formed. Much work has sincebeen done ; we have written, with great labour,and with many searchings of heart, a pamphleton the reaction of the blood. Some of the mostbeautiful work of recent times, of a physico-chemicalcharacter, has been achieved by Hartridge andRoughton at Cambridge, on the chemical dynamics ofhaemoglobin ; my colleague Brown has established thevalidity of thermodynamic reasoning and chemicalequations applied to the reactions of haemoglobin.There is still, however, much to be done ; one wouldlike to know why the reaction between haemoglobin,oxygen, and carbon monoxide is affected by light,the reason why the ion of haemoglobin has so muchgreater an affinity for oxygen than has the molecule,the exact nature of the corpuscular envelope, and theorigin of the equilibria which take place at its boundary.One would be glad to discover the chemical natureof haemoglobin, and its manner of synthesis in thebody. We know still very little, but we knowenough to be quite sure that we are on the rightlines in regarding the red blood corpuscle as a physico-chemical problem, and in investigating it by physico-chemical means.

THE POSOLOGY OF BEVERAGESAS DETERMINED BY THE FRACTIONAL-DIURESIS TEST.

BY DR. P. L. VIOLLE,CONSULTING PHYSICIAN AT VITTEL ; DIRECTOR OF THE

LABORATORY, INSTITUTE OF HYDROLOGY, PARIS.

IN certain patients it is necessary to know whatbecomes of liquids ingested by them and traversinga more or less abnormal organism. It is importantto learn how the water intake is absorbed, diffusedthroughout the tissues and body-fluids, and, finally,how it is eliminated by the kidneys. Such knowledgeof the course of water within the organism, especiallyimportant in cardiac, renal, and cardiorenal cases,is indispensable in view of the following objects:(1) exact regulation of the quantity of beveragesemployed; (2) logical distribution of their intakethroughout the 24 hours ; (3) indication of the condi-tions under which the absorption, diffusion, andelimination of water may be most perfect.

, The term " beverages " should include all liquidsreceived by the organism. In considering the totalwater intake, the water present in various foods mustbe included, since it constitutes an important partof the daily ration of water supplied to the organism.For present purposes, however, the following dis-cussion refers to the water present in beverages.Huchard and Fiessinger long ago pointed out the

necessity for restricting beverages in asystolic patients.Widal has insisted that physicians should continuallyinterest themselves in the quantity of liquids to beprescribed in cases of nephritis and in the regulationof food and drink according to given indications andparticular cases. No method for such regulationhas been hitherto available. In a case of Bright’sdisease, the permeability of the kidney must beknown not only for urea and chlorides, but also forwater. Whoever undertal-es the management ofBlight’s disease must be familiar with these threefactors. In ascertaining the general renal permeabilityit is a mistake to depend upon any single method,