THE ROLE OF THE EXTRACELLULARFLUID IN THEMAINTENANCEOF A NORMALPLASMAVOLUME

By JAMESV. WARREN,ARTHURJ. MERRILL, AND EUGENEA. STEAD, JR.(From the Medical Service of the Grady Hospital and the Department of Medicine,

Emory University School of Medicine, Atlanta)(Received for publication January 8, 1943)

The importance of the blood volume in main-taining a normal circulation is now well recog-nized. The circulation will fail if the volume ofblood becomes too small to fill the vascular systemadequately. In general, the quantity of blood inthe vascular bed is of more significance in main-taining a normal cardiac output than is the pro-portion of red cells to plasma. Because the volumeof plasma can be increased or decreased with rela-tive rapidity, the body uses this means to maintaina comparatively constant blood volume. A de-crease in the volume of the circulating red bloodcells is normally compensated by an increase inplasma volume; and conversely, an increase in thevolume of the circulating red cells is compensatedby a decrease in plasma volume. B.ecause a con-stant blood volume is dependent upon these com-pensatory adjustments in the plasma volume, aknowledge of the factors controlling the plasmavolume is of both theoretical and practical im-portance.

If the arterial pressure, the venous pressure, andthe volume of the circulating red cells remain rela-tively constant, the plasma volume increases anddecreases as the amount of circulating plasma pro-tein is increased or decreased. Thus, the injectionof a concentrated protein solution causes a tem-porary rise in plasma volume (1); and the re-moval of plasma protein causes a temporary de-crease in plasma volume, even though the fluidlost in the removal of the protein is replaced byan equal quantity of physiological saline solution(2). The changes in plasma volume produced byvarying the amount of circulating plasma proteinare not permanent in normal subjects because thebody is able to add protein to the blood stream orremove it. When the blood volume is raised bythe addition of protein, the body withdraws pro-tein from the circulation. When the volume isdecreased by lowering the quantity of circulatingprotein, the body adds protein to the blood stream(3).

Ebert, Stead, and Gibson (2) studied thechanges in the plasma volume of normal subjectsafter venesection. Immediately after the removalof approximately one-fifth of the blood volume,hemodilution occurred, presumably because of thelowering of venous and capillary pressures. Ithas since been demonstrated (4) that the poolingof comparable quantities of blood in the extremi-ties by tourniquets does cause a significant de-crease in venous pressure. That this initial hemo-dilution was caused by adding to the blood streamfluid which contained little protein was shown bythe parallel fall in plasma protein concentrationand hematocrit reading. After one to two hours,no further fall in protein concentration occurred,though the hematocrit reading continued to de-crease. This indicated that protein was beingadded to the blood stream. The return of theblood volume to normal was not appreciably has-tened by giving physiological saline solution toreplace the blood removed. Most of the salineleft the blood stream and did not return to it untilthe quantity of circulating protein had increased.In the normal subject, there seemed to be enoughavailable fluid to accomplish the initial phase ofdilution, and after that time, a further increase inplasma volume was dependent on the addition ofprotein to the blood stream. Later experiments(5) demonstrated that after venesection, the bloodvolume could be immediately brought back to nor-mal by replacing the removed blood by an equalvolume of fluid containing albumin. These datademonstrated that in a normal subject, with arelatively constant arterial pressure, the amount ofcirculating plasma protein determined the size ofthe plasma volume and that a lowering of thequantity of circulating protein caused a decreasein plasma volume and served as a stimulus to thebody for the production of plasma protein.

The above data suggested that patients withnephrosis and nutritional edema would have a verysmall plasma volume because of the striking de-

635

JAMES V. WARREN,ARTHURJ. MERRILL, AND EUGENEA. STEAD, JR.

crease in the quantity of circulating protein. In-deed, from the studies made in normal subjects,it would appear that the reduction in plasma vol-ume would be so striking that the blood could notadequately fill the vascular system and circulatoryfailure would result. Observations on patientswith nephrosis (6) and nutritional edema (7) in-dicated that their plasma volumes are somewhatbelow normal but the decrease is not marked. Asthe plasma volume is adequate in nephrosis tomaintain a normal circulation and arterial pres-sure, it appears obvious that the plasma volume ismaintained by some other factor in addition to theosmotic pressure of the proteins, and that thisfactor is not operative to any great extent innormal subjects.

Studies on patients and animals in shock havedemonstrated that the plasma volume decreases asthe quantity of circulating protein is diminished,and that death occurs from circulatory failure be-fore the quantity of circulating protein has fallento a level comparable to that which occurs innephrosis. In patients in shock, as in normal sub-jects, the size of the plasma volume is affected bythe amount of protein in the circulation. The moststriking difference between patients with nephrosisand normal subjects or patients in shock is thatthe former show a marked increase in the volumeof the extracellular fluid. It seems possible thatthe patient with nephrosis is able to maintain anadequate plasma volume because of the markedincrease in extracellular fluid, while the normalsubject and the patient in shock cannot maintainan adequate plasma volume because they do nothave this excess extracellular fluid. The purposeof this paper is to determine the effect of the ex-tracellular fluid volume on the size of the plasmavolume.

The general plan of the experiments was toreduce the quantity of circulating protein in theblood stream by causing it to leak out through in-jured capillaries into the tissues, and then to in-crease markedly the volume of extracellular fluidin order to determine whether the plasma volumecould be maintained at a normal level by a com-pensatory increase in the amount of extracellularfluid.

METHOD

Street dogs weighing between 5 and 16 kgm. were anes-thetized by the intravenous administration of sodium pen-tobarbital. Rubber tubing was wrapped tightly aroundthe proximal portions of the extremities, as close to thetrunk as possible. In the first 2 experiments, the circula-tion was occluded to both hind extremities; in the re-mainder, tourniquets were placed on all 4 extremities.The tourniquets usually prevented blood from entering theleg in appreciable quantities, although, at times, sufficientblood leaked past to cause swelling of the leg. After 5to 6 hours, the tourniquets were released. The intra-venous administration of physiological saline solution wasbegun either immediately after the release of the tourni-quets or within the first hour, and it was continued untilthe animal was very edematous. Infection was preventedby giving from 0.5 to 1.0 gram of sulfadiazine, intrave-nously, before the tourniquets were applied. At 4- to 6-hour intervals after the release of the tourniquets, 0.3gram of sulfadiazine was given intravenously.

The plasma volume, hemotocrit reading, and the plasmaprotein concentration were determined before the tourni-quets were applied. These measurements were repeatedafter the plasma volume had stabilized, as shown by aconstant hematocrit reading. The plasma volume wasmeasured by the dye method of Gibson and Evans (8),adopted to the Klett-Summerson photoelectric colorimeter.Heparin was used as the anti-coagulant. The plasmaprotein concentration in grams per 100 cc. was calculatedfrom the protein nitrogen, as determined by a standardmicro-Kjeldahl technique (9). The venous pressure wasmeasured directly by threading a No. 8 ureteral cath-eter through the external jugular vein into the superiorvena cava. The mean arterial pressure was measuredin mm. of Hg by direct arterial puncture.

To differentiate between the effects of the trauma tothe extremities and the effect of administering largequantities of physiological saline solution, experimentswere performed in which the extremities were not in-jured. Three dogs were anesthetized with sodium pento-barbital for 5 hours and at the end of that time, anamount of physiological saline solution equal to their bodyweight was administered intravenously. It required from6 to 8 hours to complete the infusion and during thattime the anesthesia was maintained. The venous pres-sure, the plasma protein concentrations, and the plasmavolume were measured before and 1 hour after the com-pletion of the saline infusion.

RESULTS

Experiments with tourniquetsFourteen experiments with tourniquets were

performed on 13 dogs. Five dogs died, apparentlyfrom heart failure. Four of these received lessthan 3200 cc., while the fifth received 5500 cc. of

636

RELATION OF EXTRACELLULARFLUID TO PLASMAVOLUME

TABLE I

Summary of observations on animals recovering after prolonged arterial occlusion of legs

Control ,~Stabilization

kgm.~~~~~~~9per cc.gas hcsatrc. o g. pr c rm

10 c 100 cc.1 .~

01.846.2 6 60 60 Comment

>v a

a. a>

1 11.8 34 6.25 2 6 60 6000 7 40 3.0 580 17.4 Recovery

1 * 10.9 35 6.3 720 45.3 4 5 20 10,500 8* 6.5 24 2.6 700 18.2 Recovery3 37 5.3 570 30.2 4 5 0 11,000 l0i 44 2.8 530 14.8 Recovery4 39 6.6 350 23.2 4 5 20 6,000 4* 38 2.1 320 6.7 Recovery5 15.5 45 6.9 780 53.8 4 5 10 16,500 8* 51 3.2 580 18.6 Recovery. Venouspressurenor-

mal throughout6 50 5.6 4 46 45 6,000 57 58 1.8 Recovery. Venous pressure ele-

vated, fell when saline stopped.Arterial pressure 112 mm. Hg

7 8.3 47 7.0 4 5 0 7,000 6 3.4 30 3.7 Recovery. Venous pressure ele-vated, fell when saline stopped.Arterial presure 120 mm. Hg

8 5.3 38 4.7 320 15.0 4 51 5 3,000 3 1.7 37 2.0 330 6.6 Recovery. Venous pressure ele-vated, fell when saline stopped.Arterial pressure 110 mm. Hg

* Following recovery from the first experiment, this dog was used on a second occasion.

physiological saline solution. The heart failureappeared to result from the effects of either theanesthetic or the tourniquets, and not from theadministration of saline. These data will be re-ported later. The animals recovered in 9 instances,and these experiments form the basis of this report(Table I).

As noted by other observers (10, 11), the arte-rial pressure fell when the tourniquets were re-leased, the legs became swollen, and the hematocritreading rose. There was at times a moderate in-crease in plasma protein concentration. Themarked rise in hematocrit reading without a cor-responding increase in the protein concentrationindicated that protein, as well as fluid, was leavingthe blood stream. The marked swelling of theextremities showed that the leak was occurring incapillaries which had been damaged by the pro-longed arterial occlusion. In the dogs that recov-ered, the circulation improved as soon as salinewas given and remained adequate as long as thesaline administration was continued. The infu-sion was given at a rate of 1 to 2 liters per hour,so that each dog received from 3 to 16.5 liters ofsaline in a period of from 3 to 12 hours. After1 to 2 hours, the dogs voided copiously; therefore,a variable portion of the administered fluid waslost through the kidneys. The extremities became

warm and very swollen. Later, generalized edemadeveloped, which was most apparent around theeyes. It was never as marked as in the injuredlegs. If the administration of saline was stoppedbefore the edema became generalized, the plasmavolume decreased rapidly, as shown by the steadyrise in the hematocrit reading to a level muchhigher than that present before the tourniquetswere applied. When the eyelids were swollen,the infusion was stopped. After 30 minutes, bloodwas drawn for an hematocrit reading. This wasdone 3 times at 30-minute intervals. The plasmavolume was determined when the hematocrit read-ing became stabilized.

The plasma volume was determined before theapplication of the tourniquets and after the ad-ministration of saline in 6 experiments. In 5 ofthese, there was no significant change. In 1 ex-periment, sufficient saline was not given to bringthe hematocrit reading back to the control leveland the plasma volume was smaller at the end ofthe infusion. In 3 experiments, plasma volumestudies were not done, but the hematocrit readingsindicated that a marked degree of hemoconcen-tration was not present.

After the plasma volume had become stabilized,the plasma protein concentration was strikinglylowered. In the 6 experiments in which measure-

637

JAMES V. WARREN,ARTHURJ. MERRILL, AND EUGENEA. STEAD, JR.

ments were made, the amount of total circulatingprotein at the end of the experiments ranged from29 to 51 per cent of the quantity present before theapplication of the tourniquets. All of the dogshad hemoglobinemia and hemoglobinuria. In 2experiments, the hematocrit readings were consid-erably lower after the infusion than before theapplication of the tourniquets. In 1 of these 2dogs in which it was measured, the plasma vol-ume was not increased. This indicated a measure-

able destruction of red cells.None of the dogs showed any signs of circu-

latory insufficiency after the plasma volume was

stabilized, and at no time was there evidence ofpulmonary edema. In 3 of the 4 experiments inwhich the venous pressure was measured, it rose

while the saline was being given but fell to normalshortly after the infusion was stopped. In thefourth, the venous pressure remained normalthroughout. After the plasma volume had becomestabilized, the mean arterial pressure was measuredin 3 dogs and found to be normal. In the 3 dogswhich were weighed, the body weight had in-creased 32, 41, and 60 per cent.

In the 12 to 24 hours following the infusion,the amount of plasma protein increased rapidly.The dogs diuresed and after a few days becameedema-free. The destruction of red cells con-

tinued, however, and the hematocrit reading de-creased steadily for the next 72 hours. The legswere completely paralyzed for several days butfunction gradually returned, and at the end of a

few weeks, the dogs were able to walk.

Experiments without tourniquetsAfter 5 hours' anesthesia, 3 dogs were

intravenously an amount of physiologicalsolution equivalent to their body weight.

Summary of observations

given

salineFrom

6 to 8 hours were required to complete the infu-sions. After the first 2 liters, the dogs voidedprofusely. The eyelids became somewhat edema-tous and the head and body appeared puffy. Therewas no pitting edema. The massive localizededema of the extremities, which appeared quicklyin the dogs whose legs were injured by the tourni-quets, did not occur. One hour after the infusionwas completed, the dogs had gained from 1.8 to4.2 kgm. in weight. The venous pressure was un-

changed in 2 dogs and 8 cm. higher than thecontrol level in 1. The protein concentration wasmoderately lowered and the plasma volume eitherunchanged or slightly larger than before the salinesolution was given. The amount of total circu-lating protein ranged from 74 to 91 per cent ofthat present before the saline was given. Theedema rapidly disappeared and after 24 hours,the animals appeared to be in a normal state.

COMMENT

The prolonged arterial occlusion caused damageto the capillaries in the extremities. When thetourniquets were released, protein leaked out intothe legs. In a preparation of this type, the fasterthe circulation, the greater the protein loss in theinjured areas. In these experiments, there was no

evidence of circulatory failure after the infusionwas started, and protein and fluid continued topass into the legs until they became tight withedema. As the skin of the extremities becametight, the rate of swelling decreased. Though thefluid and protein loss into injured areas was nowgreatly decreased, an adequate plasma volumecould not be maintained if the infusion was stoppedat this time, because the plasma protein concentra-tion was now so greatly lowered that fluid waspassing from the blood stream into the tissues of

TABLE II

on uninjured anesthetized animals receiving saline solution

Control One hour after infusion

Dog Plasma ~~~~Total Saline Time Plama TotalDogWeight|Hemato- protein Plasma circ.given require

Weight |Hemato protein Plasma circ.Weightreading concen- volume plasma increase reading concen- volume plasmaraig tration protein raig tration protein

kgm. gramsoc cc. grams cc. hours kgm.

grams Pr cc. grams100 cc. 100 cc.1 13.6 42 5.9 670 39.5 14,500 6 4.2 42 4.2 695 29.22 10.2 42 5.0 530 26.5 10,000 8 2.7 43 4.7 520 24.23 10.5 37 5.5 550 30.2 10,500 7j 1.8 41 4.3 600 25.8

638

RELATION OF EXTRACELLULARFLUID TO PLASMAVOLUME

the entire body. The osmotic pressure of the pro-teins was no longer sufficient to support an ade-quate blood volume. In time, generalized edemadeveloped, and eventually, the plasma volume re-mained adequate to maintain a normal circulation,in spite of the great decrease in the quantity ofcirculating protein. The amount of generalizededema that had to be formed before an adequateplasma volume was maintained seemed to varywith the dogs. In the animals with the tight skins,the plasma volumes became stabilized with muchless visible swelling than in those with loose skins.Though the pressure of the extracellular fluid wasnot measured, it probably rose as the volume ofthe fluid increased, because, in these acute experi-ments, there was not sufficient time to have de-stroyed an appreciable amount of tissue. In pro-longed experiments, a loss of tissue substancewould allow the accumulation of large quantitiesof extracellular fluid without increasing the pres-sure of the extracellular fluid. In addition to theincrease in the volume and pressure of the extra-cellular fluid, production of plasma protein by thebody and the return of the protein from the tis-sues of the injured extremities by way of thelymphatics probably helped to stabilize the plasmavolume.

The question arose as to whether a fall in arte-rial pressure, with the resultant fall in capillarypressure, occurred during these experiments andaccounted for the fact that the plasma volume waswell maintained, in spite of the marked decreasein the amount of circulating protein. Measure-ments in 3 dogs indicated that a significant dropin arterial pressure had not occurred.

In the uninjured animals, the intravenous ad-ministration of an amount of physiological salinesolution equivalent to their body weight producedgeneralized edema and some loss of protein fromthe blood stream. It is not possible to estimate theprotein concentration of the edema fluid from theamount of protein apparently lost from the bloodstream and the quantity of fluid retained becausenew protein may have been added to the bloodstream during the experiment. Assuming that noprotein is added to the plasma, the edema fluidwould have contained from 0.1 to 0.25 gram ofprotein per 100 cc. of fluid.

The plasma volume remained unchanged orslightly increased in spite of a moderate fall in

the quantity of circulating plasma protein. Thedecrease in plasma protein was not within therange seen in the traumatized animals, but theobservations are consistent with the hypothesisthat an increase in volume and pressure of theextracellular fluid is one mechanism by which anormal plasma volume can be maintained in thepresence of a decrease in the osmotic pressureof the plasma proteins.

The data obtained from the experiments onboth injured and uninjured animals indicate thatthe amount of extracellular fluid is an importantfactor in determining the size of the plasma vol-ume. Other observations support this view. Indehydration, the plasma volume is decreased, inspite of the fact that the protein concentration ofthe plasma is elevated. Thus, in the presence of asmall volume of extracellular fluid, the greaterthan normal osmotic pressure of the proteins isnot able to maintain even the normal plasma vol-ume. In the normal subject, a rise in the salt andwater content of the body causes the plasma vol-ume to increase. Thus, desoxycorticosterone,which increases the volume of the extracellularfluid, also increases the plasma volume (12). Thesalt and water retention from cardiac failure in-creases both the quantity of extracellular fluidand the plasma volume (13).

There are several ways in which an increase inthe volume and pressure of extracellular fluidmight enable the body to maintain a normal plasmavolume, despite a marked lowering of the quantityof plasma protein. An increase in extracellularfluid pressure might cause (1) fluid to pass moreslowly from the arterial end of the capillary intothe tissues; (2) fluid to flow more rapidly fromthe tissues into the venous end of the capillary;(3) an increase in lymph flow. It is probablethat all three of these mechanisms are operative.

Patients who have a marked lowering of thequantity of circulating protein, as in nephrosis andnutritional edema, are able to have a plasma vol-ume large enough to maintain an adequate circula-tion because of the marked increase in the volumeand pressure of the extracellular fluid. In normalsubjects after extensive injury to the capillariesof the extremities, the plasma volume quickly be-comes too small to support a normal circulation,because the quantity of protein in the blood rap-idly diminishes as protein leaks through the in-

639

JAMES V. WARREN,ARTHURJ. MERRILL, AND EUGENEA. STEAD, JR.

jured capillaries into the tissues. It is probablethat if physiological saline solution was given in-travenously in large enough quantities, the increasein the volume and pressure of the extracellularfluid would be sufficient to keep the plasma volumeat a level adequate to maintain a normal circula-tion. This is not practical in the treatment ofpatients because of the enormous quantity ofsaline solution which would be required. It isobvious that in shock caused by the loss of plasmaat the site of injury, it is much more efficient togive plasma than physiological saline solution.When plasma is given intravenously, it leaks outthrough the injured capillaries until the capillariesrecover. If sufficient plasma is given, however,to keep the quantity of plasma protein at a normallevel, fluid is lost from the blood stream only inthe area of injury, and an increase in the extra-cellular fluid content of all the tissues is notnecessary to keep the plasma volume at a normallevel.

The plasma volume is really a part of the extra-cellular fluid volume. A portion of this fluid iskept in the vascular bed because the sum of theforces causing fluid to leave the blood stream isbalanced by the sum of the forces causing fluid toenter it. The osmotic pressure of the plasma pro-teins and the extracellular fluid pressure are thetwo forces which balance the capillary pressure.If the osmotic pressure of the protein is decreased,the plasma volume will become smaller unless thecapillary pressure falls or the extracellular fluidpressure rises. An increase in extracellular fluidpressure can compensate for a marked decrease inthe osmotic pressure of the plasma proteins. Thisis the reason why patients with nephrosis maintainan adequate plasma volume. A rise in capillarypressure normally causes a decrease in the plasmavolume. This can be prevented if the extracellularfluid pressure is elevated. Thus, an increase inextracellular fluid pressure is one of the mecha-nisms by which patients with cardiac failure main-tain a large plasma volume in the presence of highvenous and capillary pressures.

SUMMARYANDCONCLUSIONS

1. The quantity of plasma protein was reducedin dogs by injuring the capillaries of 2 to 4 ex-tremities by a 5- to 6-hour period of arterial oc-clusion, produced with tourniquets. After release

of the tourniquets, the dogs were given physiologi-cal saline solution, intravenously, in large quanti-ties. The infusion maintained the circulation at anormal level, and there was a continued loss ofplasma into the extremities. The plasma proteinconcentration, therefore, fell to a low level.

2. If the saline infusion was stopped in thefirst few hours after the release of the tourniquets,the plasma volume decreased rapidly in size, andsigns of circulatory insufficiency developed. Thequantity of protein in the circulation was notsufficient to maintain an adequate plasma volume.If the saline infusion was continued until the ani-mal developed sufficient generalized edema, no fur-ther tendency to hemoconcentration was noted, inspite of the fact that the quantity of circulatingprotein was still 50 per cent or more below thelevel present before the tourniquets were applied.The plasma volume was either normal or mod-erately reduced. At this time, the circulation andarterial pressure were normal.

3. Uninjured anesthetized animals who receivedlarge amounts of physiological saline solution in-travenously, developed generalized edema butshowed much less decrease in total circulatingprotein than did the injured animals. The mod-erate fall which occurred in circulating plasmaprotein was not accompanied by a decrease inplasma volume.

4. The quantity and pressure of the extracel-lular fluid is as important in determining the sizeof the plasma volume as is the quantity of circulat-ing plasma protein. A marked decrease in theosmotic pressure of the plasma protein can becompensated by a marked increase in the pressureof the extracellular. fluid, without affecting thecirculation or the size of the plasma volume.

The work described in this paper was done under acontract, recommended by the Committee on Medical Re-search, between the Office of Scientific Research andDevelopment and the Emory University School of Medi-cine. It was carried out with the technical assistance ofRosamond Piotti, S.B.

BIBLIOGRAPHY1. Freeman, N. E., and Wallace, W. M., Effect of con-

centrated serum on plasma volume and serum pro-tein concentration. Am. J. Physiol., 1938, 124,791.

2. Ebert, R. V., Stead, E. A., Jr., and Gibson, J. G., II,Response of normal subjects to acute blood loss.Arch. Int. Med., 1941, 68, 578.

640

RELATION OF EXTRACELLULARFLUID TO PLASMAVOLUME

3. Madden, S. C., and Whipple, G. H., Plasma proteins:Their source, production, and utilization. Physiol.Rev., 1940, 20, 194.

4. Warren, J. V., and Stead, E. A., Jr., The effect ofthe accumulation of blood in the extremities on thevenous pressure of normal subjects. Am. J. M. Sc.,1943, 205, 501.

5. Stead, E. A., Jr., and Ebert, R. V., Studies on HumanAlbumin. Section in Blood Substitutes and BloodTransfusion, Mudd, Stuart, and Thalheimer, Wil-liam. C. C. Thomas Pub. Co., Baltimore, 1942, p.

185.6. Harris, A. W., and Gibson, J. G., II, Clinical studies

of the blood volume. VII. Changes in blood vol-ume in Bright's disease with or without edema,renal insufficiency, or congestive heart failure, andin hypertension. J. Clin. Invest., 1939, 18, 527.

7. Chang, H. C., Plasma protein and blood volume.Proc. Soc. Exper. Biol. and Med., 1932, 29, 829.

8. Gibson, J. G., II, and Evans, W. A., Jr., Clinicalstudies of blood volume: Clinical application of

method employing azo dye "Evans blue," and thespectrophotometer. J. Clin. Invest., 1937, 16, 301.

9. Koch, F. C., and McMeekin, T. L., A new directNesslerization micro-Kjeldahl method and a modi-fication of the Nessler-Folin reagent for ammonia.J. Am. Chem. Soc., 1924, 46, 2066.

10. Bywaters, E. G. L., and Popjak, G., Experimentalcrushing injury. Surg., Gynec. and Obst., 1942,75, 612.

11. Wilson, H., and Roome, N. W., Effects of constric-tion and release of extremity: Experimental studyof tourniquet. Arch. Surg., 1936, 32, 334.

12. Thorn, G. W., and Emerson, K., Jr., The r6le ofgonadal and adrenal cortical hormones in the pro-

duction of edema. Ann. Int. Med., 1940, 14, 757.13. Seymour, W. B., Pritchard, W. H., Longley, L. P.,

and Hayman, J. M., Jr., Cardiac output, blood andinterstitial fluid volumes, total circulating serum

protein and kidney function during cardiac failureand after improvement. J. Clin. Invest., 1942, 21,229.

641